INTERNATIONAL LABOUR OFFICE STUDIES AND REPORTS Series C (Employment and Unemployment) No. 20 THREE SOURCES OF UNEMPLOYMENT The Combined Action of Population Changes, Technical Progress and Economic Development BY WLADIMIR WOYTINSKY GENEVA 1935 Published in the United Kingdom For t h e INTERNATIONAL, LABOUR OFFICE (LEAGUE O F NATIONS) By P . S. KING & SON, Ltd. Orchard House, 14 Great Smith Street, Westminster, London, S.W. 1 4164'(6 ( ^ No. 8561.—Printer :16,OFFICE DE PUBLICITÉ (Ane.giuÉtabl rueMaroq, Brussels (Bel m). iss. J. LEBÈGUE & O") PREFACE One of the major social problems t h a t has been exercising public opinion for some years back is t h a t of the influence of technical progress on employment. I n view of this fact t h e International Labour Office published in 1931, under the title The Social Aspects of Rationalisation \ a number of preliminary studies on such aspects of t h e problem as output, hours of work, wages, employment possibilities, industrial hygiene and accident prevention. I n course of subsequent discussions on the reduction of hours of work, very special attention has been paid t o t h e importance of technical progress as a factor in unemployment. I t was even suggested 2 t h a t Governments should be asked t o contribute to a sort of permanent enquiry on the subject b y supplying, a t regular intervals, direct or indirect information concerning the changes in t h e volume of employment resulting from technical advances in various industries or occupations or among certain groups of workers. One of the main obstacles t o progress in this direction seemed to be the difficulty of isolating technical improvements from t h e innumerable other factors affecting the total volume of unemployment, and determining their exact relative importance. Professor Wladimir Woytinsky's study, which the Office has pleasure in publishing in this volume, suggests an ingenious solution for this delicate problem of scientific method—a solution t h a t will carry all the more weight because the author has been able to apply it to the actual course of recent economic developments in several countries. The reader will see how the method enables him to distinguish what fractions of the fluctuations in employment and unemployment can be attributed to demographic factors, to the increase in individual output and to changes in 1 2 Studies and Reports, Series B, No. 18. Cf. The Reduction of Hours of Work. International Labour Conference, Geneva, 1934, Report I. — IV — the volume of production respectively. The value of such a distinction is obvious, since it would enable the various causes of disturbances on the labour market to be to some extent localised. The conclusions are, it should be noted, advanced with due reservations, for their positive value is inevitably affected by the degree of approximation of the available statistics. In this connection, the author has also certain suggestions to make for the improvement of employment and unemployment statistics. CONTENTS Page PREFACE INTRODUCTION in 1 CHAPTER I : Three Sources of Unemployment CHAPTER I I : The Mathematical Formula of Unemployment CHAPTER I I I : Unemployment before the War General Survey Great Britain Germany United States Summary CHAPTER IV : Unemployment after the War United States Great Britain Germany France Italy Czechoslovakia Belgium Sweden, Norway, Denmark Poland and Hungary Japan 5 13 19 19 21 28 32 37 40 40 62 78 100 114 126 133 139 148 150 CONCLUSIONS 157 APPENDIX : Application of the Mathematical. Formula for ployment to Empirical Series of Figures Unem165 LIST OF TABLES TABLE I.—Unemployment in Great Britain from 1831 to 1930. II.—Movement of Population in the United Kingdom, 1871-1911 III.—Movement of Population in Germany, 1882-1907 IV.—Development of Manufacturing Industry in the United States, 1899-1913 V.—Development of Manufacturing Industry in the United States, 1913-1927 VI.—Gainfully occupied Persons in the United States, 1920 to 1927 VII.—Classification of the Occupied Population of the United States by Economic Branches, 1910, 1920 and 1930 VIII.—Classification of Wage Earners in Employment in the United States by Economic Branches, 1920 to 1927 IX.—Estimated Average Minimum Volume of Unemployment in the United States, 1920 to 1927 X.—Industrial Production in the United States, 1923,1929 XI.—Development of Manufacturing Industry in the United States, 1923 to 1933 , 19 26 29 34 41 43 44 45 46 50 51 VI P a TABLE XII.—The Depression in the United States : Production and Employment in Manufacturing Industry, 1929 to 1934 XIII.—Estimated Total Numbers of Persons aged 16-64 insured against Unemployment in Great Britain, 1924 to 1934 XIV.—Estimated Number of Workpeople insured against Unemployment in Great Britain and Northern Ireland (Males and Females) : A. 1923-1927 B. 1927-1934 XV.—Number of Insured Persons recorded as Unemployed in Great Britain, 1924 to 1934 XVI.—Number of Insured Persons recorded as Unemployed in Great Britain and Northern Ireland (by Econo- Se 54 63 65 66 68 mln 13. n n n )i^\ inno 4.~ innj tyc» **xav u i t u i l / U V O ; , X ( / A U \J\J X C t f * é \J XVII.—Estimated Number of Insured Persons in Employment in Great Britain, 1924 to 1934 ".. 72 XVIII.—Estimated Number of Workers in Employment in Various Economic Branches in Great Britain and Northern Ireland, 1923 to 1934 73 XIX.—Unemployment in Germany, 1925 to 1934 80 XX.—Percentage Unemployed among Trade Union Members in Germany, 1924 to 1932 83 XXI.—The " Rationalisation " Depression in Germany 84 XXII.—Percentage of Workers on Short Time among Trade Union Members in Germany, 1925 to 1932 87 XXIII.—Number of Wage Earners in Employment in Germany, 1929 to 1934 1 91 XXIV.—Visible and Invisible Unemployment in Germany, 1929 to 1932 92 XXV.—The Decrease in Employment in Germany during the Depression, 1929 to 1932 93 XXVI.—Classification of the Occupied Population in France by Economic Branches, 1921 and 1926 101 XXVII.—Unemployment in France, 1926 to 1934 104 XXVIII.—Index Number of Wage Earners employed in large Establishments in France, 1931 to 1934 107 XXIX.—Corrected Index Number of Wage Earners employed in large Establishments in France, 1931 to 1934 . . 108 XXX.—Short Time in large Establishments in France, 1930 to 1934 110 XXXI.—Unemployment in Italy, 1920 to 1934 115 XXXII.—Unemployed in Italy by Occupational Groups, 1921 to 1934 117 XXXIII.—Distribution of the Occupied Population in Italy, 1921 and 1931 118 XXXIV.—Italian Emigration and Immigration, 1921 to 1934.. 119 XXXV.—Estimated Occupied Population of Italy, 1921 to 1931 120 XXXVI.—Estimated Number of Wage Earners in Employment in Italian Industry, 1921 to 1931 122 XXXVII.—Distribution of the Occupied Population of Czechoslovakia by Occupational Groups, 1921 and 1930 . 128 XXXVIII.—The Depression in Czechoslovakia, 1929 to 1934 132 XXXIX.—The Depression in Belgium, 1930 to 1934 138 VII 3 Page XL.—Index Numbers of Industrial Production in the Scandinavian Countries, 1920 to 1934 XLI.—Population of Working Age (15 to 60 years) in the Scandinavian Countries, 1910/11, 1920/21 and 1930/31 XLII.—Distribution of the Occupied Population in the Scandinavian Countries by Occupational Groups, 1910/11 and 1920/21 XLIII.—Unemployment in the Scandinavian Countries : Percentage of Trade Union Members Unemployed, 1913 and 1920 to 1934 XLIV.—Index Numbers of Wage Earners in Employment in Industry in the Scandinavian Countries, 1923 to 1933 XLV.—Index Numbers of Individual Output in Industry in the Scandinavian Countries, 1923 to 1933 XLVI.—Industrial Establishments in Japan, 1909 to 1928 . . . XLVII.—Manufacturing Industry in Japan, 1926 to 1934 139 141 141 144 145 146 152 154 LIST OF DIAGRAMS DIAGRAM I.—Index Numbers of Population and of Employed Persons in Germany from 1925 to 1940 II.—The Influence on the Number of Unemployed (Ch) of an Increase in the Number of Wage Earners (S) in Technical Progress (T) and in the Volume of Production (V) III.—Unemployment in Great Britain from 1831 to 1930 . rV.—Development of Unemployment in Great Britain, 1860 to 1910 V.—The German Labour Market, 1882, 1895 and 1 9 0 7 . . VI.—Industrial Production in the United States, 1899 to 1913 VII.—Number of Wage Earners in Manufacturing Etablishments in the United States, 1899 to 1913 VIII.—Industrial Production in the United States, 1913 to 1927 IX.—Development of Unemployment in the Manufactures in the United States, 1920 to 1927 X.—IndustrialProduction in the UnitedStates, 1923tol933 XI.—The Depression in the United States : Production and Employment in Manufacturing Industry, 1929 to 1934 XII.—Recent Phase of the Depression in the United States : Employment and Production in Manufacturing Industries, 1933 to 1934 XIII.—Number of insured Workers and Volume of Employment in Great Britain, 1924 to 1934 XTV.—Unemployment in Great Britain : Number of Insured Persons recorded as Unemployed, 1924 to 1 9 3 4 . . . XV.—Index Numbers of Employment in Great Britain, 1923 to 1934 8 16 20 24 31 35 36 42 49 52 53 60 64 68 74 — Vili — DIAGRAM XVI.—Development of Unemployment in British Industry, 1924 to 1933 XVII.—Distribution of Unemployment in Great Britain, 1924 to 1934 XVIII.—Unemployment in Germany, 1925 to 1934 XIX.—Unemployment among Trade Union Members in Germany, 1924 to 1933 XX.—The " Rationalisation " Depression in Germany XXI.—Complete Unemployment and Short Time among Trade Union Members in Germany, 1925 to 1930.. XXII.—The German Labour Market during the Depression, 1929 to 1932 XXIII.—Classification of Unemployed Persons in Germany, 193Ö and 1Ô32 XXrV.—The Depression in Germany : Employment and Industrial Production, 1929 to 1932 XXV.—Unemployment in France, 1926 to 1934 XXVI.—Index Numbers of Wage Earners Employed in large Establishments and of the Volume of Industrial Production in France, 1931 to 1934 XXVII.—Classification of Persons Employed in large Establishments in France by Hours of Work, 1930 to 1934 XXVIII.—Employment in large Establishments in France : Index Numbers of Persons Employed, Hours of Work and amount of Work performed, 1930 to 1934 XXLX.—Industrial Production in France : Index Numbers of Production and of the amount of Work performed in large Establishments, 1930 to 1934 XXX.—Unemployment in Italy : Unplaced Applicants for Employment, 1920 to 1934 XXXI.—Growth of the Occupied Population in Italy (excluding Agriculture), 1921 to 1931 XXXII.—Development of Unemployment in Italian Industry, 1921 to 1931 XXXIII.—Unemployment in Czechoslovakia : Unplaced Applicants for Employment, 1921 to 1934 XXXIV.—Number of Wage Earners and Employment in Czechoslovakia, 1921 to 1934 XXXV.—Development of Unemployment in Czechoslovak Industry, 1921 to 1929 XXXVI.—Unemployment in Belgium : Percentage of Insured Wage Earners Unemployed, 1921 to 1934 XXXVII.—Index Numbers of Industrial Production and of the Population in Industrial Occupations in the Scandinavian Countries, 1923 to 1934 XXXVIII.—Development of Unemployment in Industry in the Scandinavian Countries, 1923 to 1933 XXXIX.—Index Numbers of Employment in Industry in Poland and Hungary, 1927 to 1934 XL.—Industrial Production in Japan, 1913 and 1919 to 1930 XLI.—Industrial Production in Japan, 1926 to 1933 Page 76 77 80 82 85 87 94 95 96 105 108 109 111 112 116 121 123 127 130 132 136 143 147 150 153 155 INTRODUCTION Even during the period of prosperity from 1925 to 1929, unemployment in several industrial countries had reached alarming proportions; when the world depression set in, it grew to an unprecedented extent, and economists and politicians found themselves face to face with problems of such seriousness that the very fate of this generation may depend on their solution. A society that fails to provide normal opportunities of employment for a large proportion of its members has forfeited its right to exist, and even if it still retained that right it would not have the power to defend itself against disruptive forces within and without. Does this mean, then, that our modern civilisation is doomed? Or does it still possess the power to master the blind, destructive forces of economic anarchy? The events of the next few decades can alone provide the answer to these questions, and that answer will depend in no small measure on the extent to which the world succeeds in comprehending the ties that unite the economic and the social factors of modern life. The old theories in this field have proved inadequate to meet the new problems that have arisen, and they must therefore be supplemented by empirical research. This means that a wide expanse of unexplored territory now awaits statistical investigation ; it is with the tasks of statistical science in this field that the present study is more particularly concerned. The path of statistical analysis is long : it comprises three stages, which must be taken successively. First comes the selection of the best method of observation; then the actual compilation of the figures, during which the selected method is tested, checked, improved and perfected; lastly, when the data have all been collected, comes the task of utilising them to scientific and practical ends. The degree to which unemployment statistics have advanced along this path is far from being the same in every country. It is surprising, for example, to find that the United States, which have the finest statistics in the world in other domains, could not — 2 — tell, when the depression was at its worst, how many millions of their workers were unemployed. Again, it is significant that in Germany, where exact unemployment figures were published every month, it should suddenly be discovered that the statistics overlooked between 1.5 and 2 million " invisible " unemployed. The attempts of the International Labour Office to evolve international statistics of employment and unemployment reveal the differences in the methods of compiling labour statistics in the various countries and how difficult it is to co-ordinate the results. Yet unemployment is an international problem; just as its practical solution requires international collaboration, so its theoretical study presupposes research on an international scale. Unemployment exemplifies all the contradictions of the present economic system. Before the problem can be thoroughly grasped, various social facts must be carefully studied. Every separate aspect of the problem—movements of population, migration, technical progress, changes in the international distribution of labour and in world trade, etc.—must be systematically studied, always in the light of the experience of all the different countries, since the same phenomena occur in several of them, but in a variety of forms. But these isolated investigations, concentrating as they do on single factors in economic life, are not enough. The balance of the social system results from the interplay of numerous factors, and the fact that this balance has been upset, as is proved by the existence of such widespread unemployment, means that these factors are no longer in harmony. What is required, therefore, is an international investigation into the conditions of harmony and disharmony of the various elements that make up the economic and social life of peoples. Only thus can one hope to determine the degree of responsibility of the various factors (structural, cyclical or other) for the recent disastrous collapse of the labour market. And that is the immediate purpose of this study. It may be thought that the title " Three Sources of Unemployment " is too wide. The writer is making no claim to discover hitherto unknown sources of our present great distresses or to proclaim some new panacea for their removal. All that he is endeavouring to do is to reduce the existing statistics of population, production and the labour market in the various countries to a simple formula that will clearly bring to light the interplay of population changes, technical progress and economic development, — 3 — In other words, the task the author has undertaken is essentially a matter of statistics and scientific method. He has had, it is true, to go into some of the material causes of the exceptionally severe unemployment in some countries, but the method of investigation remains the important thing. If that method throws a fuller light on the mechanism of the labour market in times of ample employment and in times of growing unemployment, the author will have achieved his purpose. The practical proposals made at the end refer not to the best means of combating or abolishing unemployment but simply to the best way of studying the phenomenon and representing it statistically. It must be left to the reader to decide whether the author has simply been engaging in statistics for statistics' sake or whether he has been helping, by the observation of facts, to pave the way for a sounder and more far-reaching economic and social policy. The author is fully aware of several lacunae in the study, but these can also be explained by the fact that he is concerned primarily with a problem of scientific method. For instance, only passing reference is made to seasonal unemployment; the question of women's work and that of juvenile unemployment are not touched upon at all, nor is any study made of the incidence of unemployment by occupations. No account whatsoever has been taken of the economic, social and political consequences of unemployment. All these problems and many more should find a place in a sociological study of the recent depression. But the task which the writer had set himself did not call for an examination of these questions. This is merely a preliminary study, the plan of which was restricted from the outset by the special aim in view and also by considerations of space. CHAPTER I T H R E E SOURCES OF U N E M P L O Y M E N T The growing unemployment of recent years has given fresh point to a saying of T. R. Malthus, which he included in the second edition of his famous work, published in 1803, b u t deleted from the next edition three years later : " A man who is born into a world already possessed, if he cannot get subsistence from his parents on whom he has a just demand, and if the society do not want his labour, has no claim of right to the smallest portion of food, and, in fact, has no business to be where he is 1 ." Malthus proceeds, with purposeful savagery, to depict the fate of this unwanted one, for whose labour those who possess the world can find no employment : " At nature's mighty feast there is no vacant cover for him. She tells him to be gone, and will quickly execute her own orders. . . " Nowadays, with systems of social insurance, there can be no question of executing " Nature's orders " in the sense in which Malthus used the term. B u t the fact of receiving a weekly dole from society does nothing to alleviate the moral distress of those whose labour is unwanted. Since Malthus' day the population problem has undergone a radical change. Gone is the conception of N a t u r e preparing her mighty feast for a certain given number of guests a n d turning the unwelcome guest from the door. W h a t Malthus considered a law of nature is now thought of as the work of social institutions created b y man and therefore capable of being changed by man. Nor is the mentality of those who are not admitted to the feast the same as it was 130 years ago. 1 An Essay on the Principle of Population or a View of its Past and Present Effects on Human Happiness, p . 531. A new edition very much enlarged. London, 1803. The italics are the present writer's. Cf. Wl. WOYTINSKV : Bevölkerungsbewegung und Ökonomie der menschlichen Arbeitskraft; published by the " Comitato italiano per lo studio dei problemi della popolazione ", Rome, 1932. — 6 — Yet in one sense the picture drawn by Malthus is still true of our age : society can at any given moment employ only a certain quantity of labour, which is sometimes equal to the available supply, but is often far removed from it. Whenever society's demand for labour power falls short of the supply, there is largescale unemployment, and those who do not happen to be in the employed quota are superfluous, unwanted guests. It is this fact that gives such point and appositeness to Malthus' phrase in a world in which unemployment is so rife as in the last few years. Unemployment has always a dual source : (a) an inadequate supply of labour in view of the demand of the economic system; (bj an inadequate demand for labour in view of the available supply. In short, there is always a lack of harmony between two groups of conditions, and the first task in the investigation of unemployment must be to define those two groups. The supply of labour depends on a number of factors, the most important being the natural movement of population—not, be it noted, the excess of births over deaths, which determines the trend of the total population of a country, but the changes in the number of persons engaged in gainful activity (occupied population) 1. Experience shows—and I need quote no statistics in support of this assertion—that over a period of several decades the occupied population in any country maintains a more or less constant ratio to the number of persons of working age. After the war, the idea was current for some time that this ratio was no longer constant, having been upset by the recent rapid increase in the number of women in employment. But careful research failed to confirm this view : the change in the ratio of the occupied population to the total population corresponded very closely to the change in the age distribution of the population, which was marked chiefly by a decline in the number of children. In any case, the divergence between the increase in the occupied population and the increase in the population of working age is less than the probable error resulting from the difficulty of defining exactly the term " gainfully occupied population ". Generally speaking, the occupied population of a country comprises roughly three-quarters of the persons between the 1 L. HERSCH : LABOUR OFFICE : " Population and Unemployment " in : INTERNATIONAL Unemployment Problems in 1931. Studies and Reports, Series C (Unemployment), No. 16, pp. 173-217. Geneva, 1931. — 7 — ages of 15 (or 16) and about 65 years 1 . The proportion can easily be determined for any given country from the census figures, and thus the probable influx of persons to gainful activity can be estimated with sufficient accuracy in advance. Reference should be made here to the estimates made by A. L. Bowley for the League of Nations in 1926 2. On the basis of the censuses and the mortality tables he estimated the increase in the population of working age in 11 countries from 1921 to 1941. Using these figures, it would be a simple matter to reckon the probable increase in the occupied population of these countries over the same period z-\¿s But it is o'bvious that the whole occupied population does not come on the labour market of the country. To go back to Malthus' phrase, it is only those who " cannot get subsistence from their parents ". Independent work of all kinds and the assistance given by members of a family (especially in agriculture) do not affect the labour market; from its point of view all that counts is dependent work (wage-paid work in the widest sense of that term) 4. In determining what relationship exists between the influx of wage-paid labour into economic activity and the increase in the total occupied population, there are two problems to be considered : the one concerns the general trend of population movex In the age group mentioned, practically all the men and almost 50 per cent, of the women engage in some form of gainful activity. To these must be added a comparatively small number of persons above or below these age limits who also work for gain. The proportion is lower in agricultural countries, where the members of the family (e.g., peasants' wives) who assist the head of the household are not reckoned as forming part of the occupied population. 2 LEAGUE OF NATIONS, ECONOMIC AND FINANCIAL SECTION : Estimates of the Working Population of Certain Countries in 1931 and in 1941. Geneva, 1926. 3 Cf. later, p. 162. 1 The ratio of employed persons of all kinds (workers, salaried employees, public officials, domestic employees) to the total occupied population may be considered as expressing the degree to which t h a t population has been proletarianised. The ratio is about 20 per cent, in purely agricultural countries (such as Bulgaria, Greece and India). I t is between 30 and 45 per cent, in agricultural countries t h a t are beginning to be industrialised (Poland, 27; Spain, 34; Yugoslavia, 38; Rumania, 43; Italy, 43 per cent.). In countries where small peasant holdings are the rule, the figure is usually about 50 per cent., even when the countries have a highly developed capitalist system (France, 49 ; Norway and Sweden, 51 per cent.). In industrial countries the ratio is between 60 and 80 per cent. (Czechoslovakia and Switzerland, 59; Germany, 66; Denmark, 67; Netherlands, 68; Belgium, 70; United States, 73; Great Britain, 79 per cent.); all these figures refer to the post-war period. (Cf. Wl. WOYTINSKY : Die Welt in Zahlen, Vol. I I , pp. 1 et seq. Berlin, 1926). — 8 — ments, while the other concerns the temporary fluctuations in its social structure. The simplest solution for the first of these problems is that used in the official German statistics, which are based on the assumption that the number of persons engaged in independent economic activity in the country does not change from year to year (since 1925) or fluctuates only to a negligible extent. Consequently, the full effect of any increase or decrease in the occupied population as a result of the natural movement of the whole population would be felt on the labour market. This is, of course, only a working hypothesis, but it seems on the whole to be a happy one 1. The statistics compiled by the National Statistical Office on this basis have often been used in German statistical literature and have proved consistently useful and illuminating. DIAGRAM I. INDEX NUMBERS OF POPULATION AND OF EMPLOYED PERSONS IN GERMANY 2 (1927 = 100) (From official estimates) ioa * g>- 106 y ^ r w IUQ 106 " • ^ 104 * A 102 too 102 J%9) w 96 1925 26 98 o¿. 11 26 Î9 30 SI 32 54 34 35 36 37 38 39 40 •———-— Population. • Number of employed persons. 1 The writer's only quarrel with the calculations of the German National Statistical Office is that public officials are not included in the number of employed persons. 2 The index number for the probable movement of population (after 1932) is reckoned on the basis of the official estimate, which assumes t h a t the birth and mortality rates will remain unchanged during the period under con- — 9 — Among the most interesting results of t h e calculations of t h e National Statistical Office is the conclusion t h a t the variations in t h e occupied population (or in t h e number of employed persons) m a y follow quite a different r h y t h m from the movement of t h e total population. According to the advance estimates of the National Statistical Office for the period u p t o 1940, there was reason to expect t h e population t o increase a t a more or less steady rate (falling off shghtly), whereas the number of employed persons was expected to fluctuate considerably, reflecting the high birth rate of t h e immediate pre-war years, t h e drop in births during the war and t h e higher birth rate immediately after. I n America also it has sometimes been assumed t h a t the influx of dependent workers on t h e labour market is t h e same as t h e increase in the occupied population. I t is a hypothesis t h a t is perfectly acceptable for calculations covering short periods (say, u p to 10 years), b u t it would lead t o errors if used for longer periods. W i t h regard to the temporary fluctuations in the social structure of the population, it must not be forgotten t h a t in a capitalist society there is no hard and fast boundary between independent and dependent work, and t h a t m a n y a man does not know himself whether he is—statistically speaking—independent or not. I t is b y no means surprising, when a change takes place in t h e economic situation, t h a t many a craftsman should become a wage earner, while his neighbour, formerly a factory worker, endeavours to earn a livelihood as a hawker—i.e., as an independent worker. B u t t h e difficult question is how to show such cases in statistics of t h e labour market. I n the writer's opinion, the problem is insoluble, a t least as statistics are a t present; we have absolutely no idea whether the depression accelerates the flow of new workers to t h e labour market, or whether, on the contrary, it causes t h e surplus of unemployed wage earners to change over to " independent " work. I t is probable t h a t the former reaction is more marked t h a n the latter, b u t it is impossible to prove it. sideration (Statistik des Deutschen Reiches, Vol. 401, II; Cf. Friedrich BUBG: Vorausberechnungen über die deutsche Bevölkerungsentwicklung bis zum Ende des 20. Jahrhunderts " in the reports to the World Population Congress, Rome, 1931). The index number for the number of employed persons is based on the official estimate, the results of which are given in Statistisches Jahrbuch für das Deutsche Reich, 1932, p. 17. DÖBFER — 1 0 In any case, these are only temporary encroachments of one social group upon another, and as they are temporary they can be ignored. That being so, the increase in the number of wage earners in a country may be taken to be approximately the same as the increase in the gainfully employed population. For the whole of a country, as well as for branches of its economic system, such as industry, mining, commerce, etc., the hypothesis AS = APa holds good, where AS represents the increase in the number of wage earners (salariés) and APa the increase in the occupied population (population active) 1 . Over a comparatively short period, in a country that has not an advanced proletarian character, it may be assumed also that the increase in the occupied population is distributed evenly over the various social grades, the formula then being AS : S = APa : Pa. In the following pages each of these formulae will be used, according to circumstances. If the figure for the occupied population is taken from the census, the result of migration movements will be included in the term APa. If on the other hand the figure is an estimate based on the data of the last census and the mortality tables, allowance must still be made for the balance of migration movements. It will thus be seen that variations in the labour supply can easily be traced back to demographic causes and that they can be assessed statistically with reasonable accuracy 2 . The various conditions determining the demand of society for labour, more especially for wage-paid labour, are not so easily analysed. For Malthus, the deciding factor was the rigid limit of the means of subsistence : in his view, the available quantities of wheat, meat and potatoes determined the amount of labour that could be fed and usefully employed by society. That was the law of population in the eighteenth century. Nowadays it is not the lack of foodstuffs and raw materials that limits production (and therefore also employment); it is the absence of sufficient markets. Technical progress has made such strides that not 1 As was pointed out above this method is used in Germany and the United States. 2 The effects of legislative measures (prohibiting the employment of young persons, raising the school-leaving age, fixing an upper age limit for employment, etc.) can also be easily expressed in statistical form. A reduction in the statutory working hours, for instance, can be expressed as a decline in the supply of hours of labour. — 11 — only industrial production but also the provision of raw materials and foodstuffs can be enormously increased. Somewhere there is a limit of production that cannot, for material and technical reasons, be exceeded, but it is still far off and is hidden from sight by other and much closer obstacles. It is possible for industry to go on for some little time producing for stock; it is even possible for a part of the output to be destroyed on occasion so as to prevent a glut. But in the long run the volume of production is determined by market possibilities—in the widest sense of the term, including the current needs of society, investments of all kinds, exports, etc. As market openings develop, the labour requirements of undertakings will, ceteris paribus, increase; there will be room for more guests at " Nature's feast ". When, on the other hand, markets shrink, the number of guests must fall. But the labour requirements of a country are determined not only by the extent of production; they depend also on the number of workers necessary to carry out a given programme of production—i.e., on the individual output per worker. This may sound like tautology, for the average output per head is, of course, reckoned by dividing the volume of production by the number of persons employed. In reality, however, the productivity of labour is one of the most important of the factors that determine the demand for labour by any undertaking or branch of production. Frank recognition of this fact does not imply that the spectre of technological unemployment is to be held responsible for all economic ills. It must be noted that the volume of production and the output per worker are not unconnected variables, for the market for any commodity depends to no slight extent on the cost of production, which in turn depends in part on the productivity of labour. It is always so in the realm of economic and social phenomena, which form a network of reciprocal actions and reactions. Notwithstanding their reciprocal relationship, therefore, marketing possibilities and labour productivity must be considered as two distinct factors acting from different sides and in opposite directions on society's labour requirements. If the volume of production be denoted by V and the average output per individual worker by T, then the number of workers that can find employment will be expressed by the ratio V : T. Everything that leads to an expansion of markets at home and — 12 — abroad, and thus to an expansion of production, tends to increase the demand for labour. Everything that increases the productivity of labour without at the same time opening up new market possibilities reduces the demand for labour. There are therefore three groups of factors, the interplay of which determines the trend of employment possibilities or of unemployment : (a) change of population ; (b) economic conditions; (c) technical progress. In calling these the three sources of unemployment the author is fully aware of the looseness of this expression. As was pointed out above, unemployment really springs from one single source— the absence of balance between certain economic and social phenomena. But every scientific plan of investigation requires the breaking up of the phenomena under observation into their component parts. The threefold division here adopted has, it is thought, the advantage of doing the least violence to the actual facts, while at the same time it can be adapted to the existing statistics and will be found to facilitate their co-ordination and use. But there is no need to argue the merits of the method; the reader must judge it for himself in the light of the results it gives. Before applying this threefold formula to the investigation of unemployment in various countries at different periods, it will be necessary to explain it in somewhat greater detail. And as the statistical data will have to be dealt with by mathematical processes, formula of investigation will first of all be converted into mathematical terms. CHAPTER II THE MATHEMATICAL FORMULA OF UNEMPLOYMENT Let S E represent the number of wage earners (salariés), the number of wage earners -in employment (salariés employés), Ch the number of unemployed (chômeurs). Then S = E + Ch. Further, let V be the volume of production, T the output per head of the wage earners in employment (the technical factor). Then E = |- The number of unemployed can therefore be expressed by the formula : (i) Ch = S-^- That is the basic formula for unemployment. It is of course impossible, by means of this formula, to determine to what extent the unemployment in any country at any given time is due to an excess of wage earners (S) or to an insufficient volume of production (V) or to excessive technical development (T). But the formula can be used to follow the course of unemployment, starting from a date at which unemployment was comparatively slight. The symbol A (difference) will be taken as marking the extent to which, at any later date, each of the terms of equation (i) has changed from its initial value (A may be positive, negative or nil). J — 14 — As equation (i) remains true for all values of its components, the following equation can be formed : V + AV (ii) ch + ACh = S + A S - i r ^ F This second equation subtracted from equation (i) gives : 4-AV V\ T + AT Ty V AT AS + T T + AT AV.T—AT.V (T + AT).T AV T 4- AT ACh = AS-f ' + 1—?=) = *S- or (iii) ACh = AS + E AT T + AT AV T + AT Each of the three terms of this last equation has a definite economic meaning : AS represents the increase in the number of wage earners ; AT represents the number of wage earners who, T 4- AT assuming that the volume of production had remained constant, would have had to be dismissed in consequence of the increased output per head, which rose from T to T + AT; AV represents the number of wage earners who, at T -f- AT the new level of technical efficiency, were able to find employment as a result of the increase in production AV. In this way the increase in unemployment is resolved into its component parts, corresponding to the three sources of unemployment. The relative importance of each of these components can at once be determined. Two examples will serve to show the practical application of equation (iii). Example A. Let it be assumed that, initially : S = E = Ch = 1,000,000 990,000 10,000 — 15 — Let it be assumed further t h a t t h e output per head of t h e wage earners in employment T = 1, and t h a t consequently V = E T = 990,000. After a certain time (say, 10 years) it is found t h a t S = 1,200,000, so t h a t AS = 200,000; T has increased b y 20 per cent., so t h a t AT = 0.2; V has increased by 30 per cent., so t h a t AV = 297,000. Then formula (iii) becomes : ACh = 200,000 -f 990,000 0.2 1.2 1 297,000 • — 1.2 = 200,000 + 165,000 — 247,500 = 117,500. Example B. Assume t h a t the initial values are the same b u t t h a t they develop differently, so t h a t : AS = 100,000 AT = 0.1 AV = 0.08 V = 79,200 I n t h a t case formula (iii) gives : ACh = 100,000 4- 990,000 0.1 — 79,200 1 1.1 1.1 = 100,000 + 90,000 — 72,000 = 118,000. The increase in unemployment is practically the same in t h e two cases. I n t h e first example it is due to a marked influx of new wage earners combined with very considerable technical progress, which is not entirely balanced b y the 30 per cent. increase in production. I n the second case, on the other hand, where there is a comparatively small influx of workers, the increased unemployment is due to t h e fact t h a t the rise in production has not kept pace with technical advances, slight as t h e latter have been. Two cases are shown schematically in diagram I I , in which the E . AT value — — and the increase in the number of wage earners & T + AT AV AS have been added and the value — — deducted from the T + AT t o t a l thus obtained. — 16 — DIAGRAM II. THE INFLUENCE ON THE NUMBER OF UNEMPLOYED ( C h ) OF AN INCREASE IN THE NUMBER OF WAGE EARNERS ( S ) , IN TECHNICAL PROGRESS ( T ) I N THE VOLUME OF PRODUCTION Number of wage earners or ol unemployed (thousands) Hypothesis A AND (V) Hypothesis B Number of wage earners or of unemployed (thousands) 400 -¿.CO 500 300 200 100 The same method can be used to follow the development of unemployment from one year to another; if necessary, it is quite easy to represent diagrammaticaUy a negative value of AS (due to a fall in natality at an earlier date) or of AV (decline in production caused by market slump). It will simplify the calculations if the initial values of the various factors be taken as : Ch = 0, S = 100 and V = 100 ; in that case E = 100, and the formula becomes : (iv) ACh = AS + 100 AT T + AT AV T + AT where ACh, AS and AV are expressed as percentages of S or of V. On the basis of equations (iii) and (iv), the conditions for the maintenance or loss of economic equilibrium can be expressed as follows : Unemployment remains unchanged in volume when AV = AS + E T + AT AT T + AT — 17 — or (ignoring the difference between E and S at the initial point of the observation period) when AV T + AT AS + 100 AT T + AT Unemployment increases when the left-hand term of these last two equations is smaller than the right-hand term; it diminishes when the left-hand term is the larger. The boundaries of the economic system to which these formulas are applied may be as narrow or as wide as is desired. If good, comprehensive statistics were available they could be applied to the whole economic system of a country. But that would necessitate production figures that included every type of economic activity in a single index. It is true that attempts have been made to compile such indices, but as a rule there are grave doubts as to their reliability. Moreover, it is extremely difficult to establish a true relationship between agricultural production (within the general index figure) and the number of wage earners, for in agriculture independent work (including the assistance of members of the family) usually plays a more important part than wage-paid work 1 . In practice, therefore, the application of the method described above is more or less limited to industrial production. But the concept " industrial production " should be made as wide as possible, including manufactures, handicrafts, light, water and mining. With regard to the building trade, which sometimes has its own special rhythm of development, independent of the rhythm of industry in general, it is best to follow the statistical practice of the country in question : if building is not included 1 In most capitalist countries the " degree of prolétarisation " of agriculture shows no tendency to increase with the advance of economic progress ; it remains far below the proportion of wage earners in industry or in commerce and transport. Before the depression the " degree of prolétarisation " (as a percentage) was : Country United States I n industry I n commerce and transport 85 60 87 76 93 69 47 55 71 75 I n agriculture 28 33 53 24 39 — 18 — in the index of industrial production, then building workers will not be reckoned among industrial wage earners. In interpreting the concept " industrial occupations " one is also bound by the available statistics. When using census results, the volume of production will have to be correlated sometimes with the number of persons occupied (engaged in gainful activity) in industrial production, and at other times with the number of wage earners. In every case, the difficulty is to determine what exactly is V the E that is to be related to V in the equation T = — • E It is obvious that, for any given value of V, the wider the interpretation given to E (the number of persons in employment) the smaller will become the value of T. But in so far as index figures are employed and E (or S) and V are taken as being 100 at the initial point, the question is of no importance. The important point to bear in mind is that the value AT is influenced by any extension of the concept " persons in employment " (E). Technical progress leads to a relatively more rapid increase in the number of salaried employees in industry, than in the number of wage earners engaged in production processes. But this latter figure in turn increases more rapidly than the total number of persons gainfully engaged in industrial production. AccorV dingly, the value of — will rise most rapidly if the denominator E includes all persons engaged in industrial occupations. It will rise less rapidly if E denotes wage earners only, and more slowly still if E includes both workers and salaried employees in industrial occupations. Before this formula can be applied, the statistics on which it is to be employed must first be carefully examined. The results obtained will be comparable only in so far as the statistics used in the calculations were compiled by the same method. CHAPTER III UNEMPLOYMENT BEFORE THE WAR General Survey Before the war, unemployment was already known in every capitalist country. But the number of unemployed persons was quite small, judged by present-day standards : unemployment was a chronic disease in industrial States, but it never reached calamitous proportions. The cyclical fluctuations in the demand for labour at that time were not very strongly marked. British statistics provide a picture of the course of unemployment, from year to year, over a century. TABLE I . UNEMPLOYMENT IN GREAT BRITAIN FROM 1 8 3 1 TO 1 9 3 0 Percentage unemployed among trade union members.1 Year Unemployment Year 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 5.2 7.1 8.1 6.2 5.4 5.0 12.4 10.5 11.1 14.8 18.5 11.0 7.4 5.1 3.9 19.3 15.7 33.4 22.3 13.2 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 Unemployment 3.9 6.0 1.7 2.9 5.4 4.7 6.0 11.9 3.4 1.9 5.2 8.4 6.0 2.7 2.1 3.3 7.4 7.9 6.7 3.9 Year Unemployment Year 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1.6 1.0 1.1 1.6 2.2 3.4 4.4 6.2 10.7 5.2 3.5 2.4 2.6 7.1 8.5 9.5 7.1 4.1 2.0 2.1 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 Unemployment 3.4 6.2 7.7 7.2 6.0 3.3 3.4 2.9 2.0 2.4 3.3 4.0 4.7 6.0 5.0 3.6 3.7 7.8 7.7 4.7 Year Unemployment 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 3.0 3.2 2.1 3.3 1.1 0.4 1.4 1.2 2.4 2.4 15.3 15.4 11.5 8.2 10.5 12.5 9.7 10.8 10.4 16.1 i For 1831-1850: unemployed among the members of the Union of Ironlounders of England, Ireland and Wales. For 1851-1870 : information concerning all unions (Statistical Tables and Reports on Trade Unions, Fourth Report, London, 1891, pp. 523-524). For 18711890 : averages for metal working industries and other branches (Fifteenth Abstract of Labour Statistics of the United Kingdom, London, 1912, p. 2). For 1901-1925 : average figures for all unions. From 1926 onwards : percentage of insured workpeople unemployed (including temporary stoppage) (The Ministri) of Labour Gazette). — 20 — DIAGRAM III. UNEMPLOYMENT IN GREAT BRITAIN Percentage unemployed among trade union members.1 I ÒSO 1340 ISSO I860 1670 I860 1890 1900 1910 1920 1950 i Cf. table I, p. 19. The disadvantage of this table is that the further back the statistics go the narrower becomes the circle of persons covered. The data for the years 1831-1850, which refer to a single occupation that is particularly subject to cyclical influences, show excessive fluctuations ; these would certainly be considerably less marked if statistics for other occupations were available for the same period. On the other hand, the unemployment insurance figures (1926 onwards) are lower and more balanced than those of the trade unions for the preceding period. Apart from these defects, diagram III would seem to give a faithful picture of the movement of unemployment in Great Britain from 1831 to 1933. In the forties of last century there was a slump in employment that is almost comparable with that of recent years—with the difference that the disastrous phenomenon was restricted to a limited section of the national economic system and was of only short duration. Nevertheless it left in British history the memory of that period of sensation and unrest that marked the growth of Chartism. After that time unemployment only twice reached 10 per cent.—in 1858 and 1879. — 21 — The limit of 8.3 per cent, (an average of one month's unemployment per worker in the year) was passed on only five occasions between 1851 and 1920. Since 1920, on the other hand, the figure has only once fallen as low as 8.2 per cent. The difference between the period 1851-1920 on the one hand and the forties of last century and the last decade on the other is very striking : over a whole century the balance of the labour market has only twice been disturbed—at the beginning and the end. Between lies a period of 70 years during which the economic and social system remained in a state of equilibrium broken only by slight temporary fluctuations. The percentage of trade union members who were unemployed throughout this period was as follows : 1851 1861 1871 1881 1891 1901 1911 3.9 5.2 1.6 3.5 3.4 3.3 3.0 There were, it is true, lean years, in which the demand for labour was slight, but they were invariably succeeded by fat years, when employment was plentiful. The same holds good for other capitalist countries for which more or less reliable unemployment statistics are available. In every case, large-scale unemployment was merely a passing phenomenon occurring approximately every 10 years, in periods of depression. It declined again at the first signs of business recovery, reaching the usual minimum figure during the boom period. How this balance was achieved will now be examined, using as examples Great Britain, the United States and Germany. Great Britain It is unfortunately impossible to go back beyond 1861 in a detailed examination of the evolution of the British labour market; there are too many deficiencies in the earlier statistics. The observation period will therefore be the 50 years from 1861 to 1911. During that period there was a tremendous increase in the population of Great Britain, more especially in the occupied population and most of all in its army of wage earners. The population of the United Kingdom (in millions) grew as follows : 1861 1871 1881 1891 1901 1911 28.9 31.5 34.9 37.7 41.5 45.2 — 22 — The size of the occupied population is shown in the statistics from 1881 onwards only. The ratio of the occupied to the total population in each decade rose as follows : 1881 1891 1901 1911 36.5% 38.5% 39.3% 40.5% By extrapolation, the ratio of the occupied to the total population may be estimated at 35.5 per cent, in 1871 and 34.5 per cent. in 1861. The growth of the occupied population in Great Britain (in millions) was therefore the following (the figures in brackets being estimates) : 1861 1871 1881 1891 1901 1911 (10.0) (11.2) 12.7 14.5 16.3 18.3 British statistics do not show the social distribution of the population and the number of wage earners for the period under consideration. But the distribution of the occupied population over the main branches of the economic system is known for England and Wales : Occupied population (millions). Number of above in industrial production (millions) 1881 1891 1901 1911 11.2 12.8 14.3 16.3 6.4 7.4 8.5 9.6 The percentage of the occupied population claimed by industry was thus : 1881 1891 1901 1911 57.8 58.3 59.0 59.5 The regularity with which the figures in this series rose is so striking that it is quite possible, by extrapolation, to obtain figures for the two preceding decades. The ratio of the industrial to the occupied population may therefore be put at 57.3 per cent. for 1871 and 56.8 per cent, for 1861. If these percentages for England and Wales be taken as holding good for the United Kingdom as a whole, the total occupied population and the total number of persons engaged in industry may be estimated as follows : Year 1861 1871 1881 1891 1901 1911 Increase 1861-1911 Occupied population (millions) 10.0 11.2 12.7 14.5 16.3 18.3 8.3 Persons engaged in i n d u s t r y (per cent.) (millions) 56.8 57.3 57.8 58.3 59.0 59.5 — 5.68 6.42 7.34 8.45 9.62 10.89 5.2 — 23 — It would thus appear that of the increase of 8.3 millions in the occupied population (including independent and wage-earning workers indiscriminately), 5.2 millions, or about 63 per cent, were absorbed by industry. That does not mean that they all found employment in industrial occupations. On the contrary, there was always a fraction of the occupied population—wage earners and independent workers alike—that was unemployed. If this fraction be assumed to be roughly the same as the percentage of unemployed workers among trade union members, and if 3.5 per cent, be added to cover cases of temporary unemployment on account of sickness, etc., the number of members of the occupied population actually in employment in British industry in the census years may be assessed at the following figures (in millions) : 1861 1871 1881 1891 1901 1911 5.19 6.10 6.83 7.88 8.99 10.28 Taking 1861 = 100, it will be found that the number of persons engaged in industrial occupations (without deducting the unemployed, the sick, etc.) must have developed thus : 1861 1871 1881 1891 1901 1911 100 113.0 129.2 148.8 169.4 191.7 The same index numbers may be presumed to be correct for 1860, 1870, 1880, etc., in which case the number of those in employment and those out of employment would be as follows 1 : Unemployed(Ch) . . Wage earners in employment^) I860 1870 1880 1890 1900 1910 1.9 4.4 6.7 3.1 4.1 9.0 98.1 108.6 122.5 145.7 165.3 182.7 Over against this last series of figures must be placed those showing the development of industrial production in Great Britain. According to calculations made by the Berlin Institute for Market Research 2, the growth of industrial production in Britain was as follows : 1913 = 100 1860 = 100 (V) I860 1870 1880 1890 1900 1910 34 100 44 129.4 53 155.9 62 182.4 79 232.4 85 250 This set of figures will be used because no better ones are available, but it must be clearly understood that they are exact 1 2 Cf. table I, p . 19. Die Industriewirtschaft, Special No. 31 of Vierteljahrshefte zur turforschung, Berlin, 1933, p. 69. Konjunk- — 24 — only in so far as they reflect the general r h y t h m of development; one or other of t h e terms m a y easily be too high or too low. W i t h this reservation, t h e above figures will now be used to establish index numbers of industrial production per head of those actually in employment (independent workers and wage earners) : I860 T = 1.019 1870 1880 1890 1900 1910 1.192 1.273 1.252 1.406 1.368 I t is now possible to show in diagrammatic form t h e interplay of the three factors t h a t determine the course of development on the labour market : t h e influx of new members t o t h e occupied population; technical progress, which renders a fraction of the hitherto employed workers superfluous; t h e increase in t h e volume of production 1 . DIAGRAM IV.—DEVELOPMENT OF UNEMPLOYMENT IN GREAT BRITAIN (S and V in 1860 = 100) Manufactures and Mines -too •loo -, AS *• T.AT 80 60 aV 60 40 20 I860 1670 1880 1890 I900 I9IO The period from 1860 to 1910—or a t least up t o 1900—was one of continuous industrial expansion in Great Britain, while technical progress advanced b u t slowly. I n the space of these 50 years production rose by 150 per cent, (an annual increase of 1.85 per cent.), whereas t h e output per head of those in employment (independent and wage-earning workers) during the same period increased b y only 34 per cent., representing 0.59 per cent. annually. This figure would be slightly lower still if the output per head were reckoned per head of the wage earners actually in 1 Cf. Appendix, p. 165. — 25 — employment, for t h e proportion of wage earners among those engaged in industrial occupations would appear t o have increased during t h e period in question. There are several reasons for t h e comparatively slow rate of technical progress in British industry. In addition to the conservative tendency t h a t characterises economic organisation in t h a t country, there is t h e fact t h a t labour productivity declined in coal-mining, which is one of Britain's key industries. The following survey, although it covers only the period from 1881 to 1911, will show the marked difference between t h e trend of development in the British mining industry and t h a t in other branches of production, such as t h e textile and metal industries. Number of persons in the occupation 1 : 1881 Coal and shale mining Metals, machines, etc Textiles Mining Metals Textiles Production 2 1891 1901 (thousands) 437 599 752 927 1,095 1,447 1,191 1,253 1,169 Index numbers (1881 = 100 137 172 100 118 156 100 106 97 1911 1,021 1,765 1,294 100) 234 190 109 : Index numbers (1880 = 100) 100 123 148 173 100 156 202 244 100 111 125 139 Mining Metals Textiles i COJIMITTEK ON INDUSTRY ASD TRADE : Survey ol Industrial Relations, 1926, p. 416. The figures are based on the census results and inelude unemployed persons belonging to the2 occupations in question. Die Industrietoirtfchalt, p. 69. If it be assumed t h a t t h e apply t o approximately t h e given in the census figures for following index numbers of employed in the branches calculated : last index numbers of production same occupational groups as are 1881, 1891, 1901 and 1911, then t h e output per head of the persons of industry concerned can b e Output per head (1880-1881 = 100) : Mining Metals Textiles 1880-1881 1890-1891 1900-1901 1910-1911 100 90 86 74 100 132 130 128 100 105 129 128 The index of individual output for manufacturing industries alone (excluding mining) would show greater progress t h a n t h e combined index for manufactures and mining together. In any case, t h e contrast between t h e slow increase of the output per 3 — 26 — head and the much more rapid growth of the volume of production is characteristic of British economic development as a whole. Thanks to this relationship between t h e relative speed of t h e two processes, unemployment in British industry remained insignificant, although t h e number of persons in industrial occupations was nearly doubled in 50 years, as is brought o u t in diagram IV. No further proof is required of the divergent trends of industrial expansion and technical progress in Great Britain. From 1881 onwards the increase in the number of persons employed in industry (in England and Wales) can be proved from the census returns. Whether it was or was not permissible to assume t h e same tendency for the period 1860-1880 is immaterial. There can be no doubt t h a t before t h e war the population of Britain was becoming more and more industrialised, i.e., t h a t the number of persons employed in industry increased from decade to decade, not only as an absolute figure b u t also relatively to the t o t a l population. I t is obvious t h a t this must have led t o an increase in the numbers employed in other occupations (commerce, transport, etc.), for the distribution and transport of the constantly increasing volume of commodities naturally called for the services of a growing number of workers and employees. Notwithstanding these facts, Great Britain would have suffered from chronic unemployment for several decades before the war if emigration had not provided a safety valve. Table I I shows t h e importance of emigration in the development of population in t h e United Kingdom. TABLE II.—MOVEMENT OF POPULATION IN THE UNITED KINGDOM,. 1871-1911 1 (in millions) Period 1871-1881 1881-1891 1891-1901 1901-1911 1871-1911 Population at beginning of period 31.5 34.9 37.7 41.5 — Excess of births over deaths + + + + Net loss by migration Actual increase 4.3 4.4 4.3 4.8 — 0.9 — 1.6 — 0.6 — 1.1 + + + + + 17.8 — 4.2 + 13.7 3.4 2.8 3.7 3.7 i " Statistical Abstract for the United Kingdom for each of the fifteen years 1913 and 1919 to 1932 ", London, 1934, p. 8. — 27 — But for emigration, Britain would have had to feed, in 1911, not merely 45.2 -f 4.2 = 49.4 millions, but a much higher figure, for there would have been the natural increase in the fraction of the population that emigrated; the excess of births over deaths for this group would presumably have been relatively higher than for the population remaining in Britain, which would comprise a larger proportion of children and of aged and infirm persons. As no comprehensive statistics are available concerning the age of the emigrants, the probable increase in the population of the United Kingdom, had the safety valve of emigration not been operative, can be only roughly estimated : Year 1861 1871 1881 1891 1901 1911 Actual population Probable population (excluding emigrants and (had emigrants remained their descendants) at home) Millions Index numbers Millions Index numbers 28.9 31.5 34.9 37.7 41.5 45.2 100 108 121 130 144 156 28.9 32.5 37.1 42.1 47.4 53.5 100 113 129 145 164 186 During this period of fifty years (1861-1911) the population of Great Britain would probably have increased by some 24.6 millions, but 8.3 millions of those (approximately 35 per cent.) were transferred abroad by emigration. The economic system of the country had therefore to absorb only the remaining 16.3 millions, of whom 8.3 belonged to the occupied population. How this was done has been seen above. Whether the problem could have been solved without the help of emigration is questionable. As the majority of those who emigrated were persons of working age, a stoppage of emigration would naturally have raised the ratio of the middle-age groups to the total population. In 1911 the occupied population would have been about 22 or 23 millions instead of the actual figure of 18.3 millions. What would have happened to the surplus 4 or 5 million workers if they had not found work and a livelihood in America and the British colonies? The inquirer can only conjecture, but there seems to be little doubt that the combined action of the various factors analysed above would have given entirely different results : instead of a constant return to the position of equilibrium the country would certainly have had permanent unemployment. It is equally certain, too, that the interplay of these factors — 28 — would still have had the same result if the excess of births over deaths in t h e United Kingdom had only been, say, 12 millions instead of 17.9 millions from 1871 to 1911 and if the whole of this increased population had remained in t h e country. The one factor of decisive importance is t h e actual net increase in t h e population, a n d more especially in the occupied population, the annual growth of which is indicated by the following figures for the period under consideration : Per cent. 1861-1871 1871-1881 1881-1891 1891-1901 1901-1911 1.14 1.26 1.33 1.17 1.16 1861-1911 1.21 Germany The development of the German labour market before the war can be followed with t h e aid of the occupational censuses of 1882, 1895 and 1907. This period of 25 years was one of rapid expansion for Germany. The years t h a t mark the limits of the period, 1882 and 1907, were both subject to considerable fluctuations in market conditions, coming, as they did, at the turning point from a boom period to a slump. I n 1895, on the other hand, the economic situation was satisfactory, although no better t h a n the average for the preceding or succeeding years. I t will be remembered t h a t about 1890 was t h e lowest point in the long-period curve of economic development, which began to drop in t h e seventies and rose again steadily until the outbreak of war. Only isolated statistics of unemployment in Germany during this period are available. The most useful are the results of two special enquiries carried out on 14 J u n e and 2 December 1895. The number of unemployed persons (including those who were absent from work a t the moment on account of temporary incapacity) was found to be extraordinarily low 1 : on 14 J u n e , 299,000; on 2 December, 771,000. I n so far as these statistics are reliable, then, the number of unemployed persons during the summer of 1895 would appear t o have been little more t h a n the usual minimum number of people temporarily out of employment through illness or a change of 1 Vierteljahrshefte zur Statistik to No. 4, pp. 11* and 12*. des Deutschen Reichs, 1896, supplement — 29 — job. During the winter the number rose by about 480,000, representing the extent of seasonal unemployment. In 1882 and 1907 there was little sign of large-scale unemployment either 1. It may therefore be concluded that during the period under consideration unemployment—apart from temporary fluctuations of which there are no exact records—did not increase, so that ACh = 0 for this 25-year period. We must now try to discover how the balance of the economic system was maintained in this instance. Table III shows the movement of population in Germany during the period in question. TABLE III. MOVEMENT OF POPULATION IN GERMANY FROM 1882 TO 1907 1882 1895 1907 39.8 100 45.9 115.3 55 0 138.2 42.4 16.9 100 43.0 19.8 117.2 45.7 25 2 149.1 33.8 5.7 100 37.7 7.5 131.6 39.1 9 8 171.9 Total population : Index (1882 - 100) Occupied population : As percentage of total popuIn millions Index (1882 - 100) Population occupied in industry : As percentage of occupied popuIn millions Index (1882 — 100) As in Britain, the increase in the occupied population in Germany was more rapid than the growth of the total population, and the ratio of the occupied population in industrial production to the whole occupied population grew with each census. In the space of 25 years the total population increased by 38 per cent., the occupied population by 49 per cent., and the number of persons engaged in industry by 72 per cent. The annual increase was therefore as follows : For the total population For the occupied population For those occupied in industrial production 1.3 per cent. 1.6 „ „ 2.2 „ „ 1 The number of unemployed persons among the members of trade unions t h a t compiled unemployment statistics was 1.6 per cent, on the average during 1907. — 30 — Of the 8.3 millions that represent the influx of new workers to the German economic system, 4.1 millions were absorbed by industry. As practically all the available workers in Germany were in employment during the observation period, it may be concluded that at the three census dates the volume of industrial production (V) was equal to the number of available workers (Pa) multiplied by the average productivity of labour (T). The development of industrial production in Germany is shown in the following index figures : 1913 = 100 1882 = 100 28.4 47.6 82.9 100 167.6 291.9 1882 1895 1907 In 25 years, then, the increase in industrial production in Germany was relatively greater than it had been in Great Britain over a fifty-year period (1860-1910). Nor is this surprising, for at the beginning of the observation period Germany was a young capitalist country; in other words, it was at the stage where economic expansion proceeds most rapidly—both relatively and absolutely. A comparison of the index of industrial production with the index for the number of persons engaged in industrial occupations provides the following index figure for individual output in German industry : 1882 1895 100 127.5 1907 169.8 The average annual increase over this period would therefore be: For industrial production For individual output 4 . 3 per cent. 2 . 1 ,, ,, 1 The growth of industrial production in Germany is illustrated by diagram V, which follows the usual formula, but is simplified by the fact that both Ch at the initial date and ACh are considered as being nil. 1 These figures were obtained by the same methods as were used above when dealing with the economic development of Britain. The two sets of results are therefore comparable. — 31 — DIAGRAM V. THE G E R M A N LABOUR M A R K E T (S and V in 1882 = 100) Manufactures and Mines 120 IOO IÔÔ2 1695 1907 There remains the question of migration. The wave of emigration from Germany had reached its highest point before the period to which the above figures refer. From 1881 to 1885 some 857,000 Germans left the country to seek new homes across the sea. In the succeeding years, which, as was mentioned, were years of industrial expansion in Germany, the tide of emigration ebbed, although the numbers were still quite considerable : 1886-1890 1891-1895 1896-1900 1901-1905 1906-1910 485,000 402,000 127,000 146,000 134,000 Total... 1,294,000 The number of members of the occupied population who left Germany between 1880 and 1910 cannot have fallen far short of 1.5 millions. For Germany, therefore, as for the United Kingdom, the same question arises : but for emigration, how would the country have been able to maintain its economic system in stable equilibrium? Not only would Germany have had to import more of the food and raw materials it lacked; it would also—and this would have been the more serious problem—have had to find fresh markets for its industrial products. Since it is extremely unlikely that the increased pressure of the surplus population would have — 32 — acted as a brake on technical progress, it may be concluded that, but for emigration, Germany would long before the war have been face to face with the problem of chronic large-scale unemployment. United States The population of the United States rose as follows during the second half of the nineteenth century (in millions) : 1850 1859 1869 1879 1889 1899 23.3 30.7 37.9 49.1 61.8 74.8 A very important factor in this demographic revolution was the tremendous immigration from Europe; the United States received approximately 17 million immigrants during those 50 years. The occupied population increased certainly more rapidly than the total population. If it be assumed that the ratio of the occupied to the total population developed more or less as it did in most European countries, for instance in Great Britain, then the occupied population in the United States must have risen from about 8 millions in 1850 to 38 millions in 1910. At the outset the number of persons in industrial occupations represented only a small proportion of the occupied population, but the ratio rose remarkably quick. The American censuses give the number of workers employed in manufactures (excluding mining and construction, but including handicrafts in so far as wage-paid workers are employed) as follows x : In thousands Index 1859 1869 1879 1889 1899 1,311 100 2,054 157 2,733 208 4,252 324 5,306 405 It is probable that the rise in the total number of persons occupied in manufacturing industries was less spectacularly rapid. But it is certain that the industrial population of the country increased more rapidly during this period than did the occupied population as a whole; in other words, the movement of population was characterised by progressive industrialisation. The volume of industrial production can only be very roughly assessed on the basis of the value of the goods produced, using the 1 These figures are not comparable with the data given above for Great Britain. They do not include independent workers or members of the family assisting the head of the household (handicrafts and home work); they refer moreover only to manufacturing industries. — 33 — wholesale price index as a corrective to allow for changing prices. The method loses still more of its possible claim to accuracy through the fact that the index of wholesale prices, which has to be used, refers to raw materials rather than to manufactured articles. There are other sources of error which need not be mentioned here. Subject to that reservation, the development of the volume of industrial production in the United States may be estimated as having been (1859 = 100) : Value of production... Wholesale prices Volume of production . 1859 1869 1879 1889 1899 100 100 100 180 113 159 285 97 294 497 94 529 689 84 820 The output per head of the workers in industrial undertakings would thus appear to have doubled during the period under consideration. If 1859 = 100, the output can be expressed thus : 1859 1809 1879 1889 1899 100 102 141 163 202 The position in manufacturing industries in the United States accordingly developed as follows : Workers employed (E) Volume of production (V) Individual output ( T = g") 1859 1869 1879 1889 1899 100 100 1.00 157 159 1.02 208 294 1.41 324 529 1.63 405 820 2.02 The average annual increase over this period of 40 years would therefore be : Per cent. For production For the number of workers For individual output 5.40 3.56 1.77 These results would seem to be influenced by the fact that the year 1859 was taken as the starting point for the calculations. The beginning of the observation period thus coincides with the Civil War, which, it is well known, interrupted the economic progress of the United States and even fundamentally altered the economic system of the country. For the period 1869-1899, the average annual increase would be : For production For the number of workers For individual output Per cent. 5.63 3.21 2.34 — 34 — For the following period, from 1899 onwards, a basis is provided by the investigations of the group of American economists that were endeavouring to analyse the recent economic changes in their country in 1929, just before the depression began. The following table showing the development of manufacturing industries in the United States before the war is taken from their report. 1 TABLE IV. DEVELOPMENT OF MANUFACTURING I N THE UNITED STATES, INDUSTRY 1899-1913 (1899 = 100) Persons employed Year Volume of production Output per head 100 101 112 122 124 122 144 154 153 129 159 162 155 179 185 100 96.2 101.8 103.4 100.7 104.0 113.3 114.0 108.5 101.5 109.6 108.7 103.4 114.6 116.3 100 105 110 118 123 117.5 127 135 141 127 145 149 150 156 159 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 The figures in table IV cannot be compared with those for the earlier period, because they refer to other categories of undertakings and of persons. They do not, for instance, include hand and similar industries and establishments with products valued at less than 500 dollars; on the other hand, salaried employees and workers in the undertakings covered are both included in the figures. But even if it is not comparable with the results for earlier years, the table is none the less instructive in many other respects. For one thing, it throws light on the mechanism of cyclical fluctuations in production. 1 COMMITTEE ON RECENT ECONOMIC CHANGES C O N F E R E N C E O N U N E M P L O Y M E N T : Recent Economic States, 1929, Vol. I I , p . 454. OF THE PRESIDENT'S Changes in the United — 35 — DIAGRAM VI. INDUSTRIAL PRODUCTION IN THE UNITED STATES (1899 = 100) Manufacturing 1899 1900 01 02 03 04 03 industry 06 07 06 09 10 II 12 15 ———^-^ Volume of production. Number of wage earners. *m^m*^mmmm Individual output. The observation period begins in 1899, when the economic situation was favourable. The boom continued until 1903; in 1904 there was a downward tendency, followed by a further rise in the next year. The year 1908 brought a very severe, but short, slump, and in 1911 there was another drop in the curve, but only for a short period. All these changes are very clearly brought out in diagram VI by the curve of production (V). They are reflected less clearly and on a reduced scale in the curve of persons in employment (E), but the curve of individual output (T) shows an unexpected sensitiveness to cyclical fluctuations. It is true that some of the minor variations in this last curve may be attributed, especially during the earlier years, to inaccuracies in the statistical methods employed. But these possible sources of error cannot explain away the general rhythm of the curve, which repeats with amazing regularity the movement of the business cycle. In periods of depression, the output per worker falls very considerably, while in periods of prosperity it rises again. This is due mainly to changes in hours of work —overtime during boom periods; short time during bad trade. — 36 — The total increase in individual output in t h e manufacturing industries of t h e United States from 1899 to 1913 was about 16.3 per cent., being an average of 1.1 per cent, annually over these 14 years. F r o m this one can calculate how violent would have been the fluctuations in employment in industry if the shock had not been absorbed to some extent by the elasticity of working hours. Diagram V I I shows how employment would have varied in t h a t case, assuming the volume of production to have been the same. DIAGRAM VII. NUMBER OF WAGE EARNERS IN MANUFACTURING ESTABLISHMENTS IN THE UNITED STATES (1899 = 100) 60 .. 1A0 - Í 5r- 120 i . 1vV / * B 02 •'••.. .140 •\ p •• 100 _Ä J 1699 1900 01 -^ / ; g «• * 07 03 09 • *P • • 10 II I5.0 ' • 05 04 05 06 12 loo 13 = = = = = ..._ Quantity of labour employed. Quantity of labour that would have been employed had individual output developed at an even rate. "• ™ ™ ™ Development oí individual output at an even rate. With regard t o t h e trend of the series of figures given in table IV, the most noteworthy feature is the rapidity with which production grew before the war and the increase in the quantity of labour employed in industrial establishments. I n the 14 years, production rose b y about 85 per cent., representing an annual average increase of 4.5 per cent., while the number of those employed rose by 59 per cent., or 3.4 per cent, as an annual average. The output per head, on the other hand, increased b u t slightly—only 16.3 per cent, in the 14 years, as was seen above. I t should also be noted t h a t from 1899 to 1913 the influx of immigrants to the United States was greater t h a n a t any earlier date. I n the 14 years (from the beginning of 1900 to the end of 1913) the country absorbed more t h a n 12 million immigrants without showing any appreciable increase in unemployment. — 37 — This great power of absorption of the American labour market is in keeping with its rate of industrial development. In the 13 t o 15 years preceding the war the United States doubled not only t h e production of their manufactures but also their output of coal and pig iron ; in t h e same period they trebled their output of steel a n d increased their extraction of petroleum fourfold. I t would have been quite impossible to achieve economic expansion on such a scale with the labour available within the country. The liberal immigration policy of the United States was a necessity forced upon the country by the relative rapidity of its demographic, economic and technical progress, the mutual relations of which were described above. Summary In the three cases studied—old England, Germany in its period of youthful economic development and the United States in their prodigiously rapid expansion—the economic system was successfully maintained ,in equilibrium before the war. The additions to t h e occupied population and those who became available on t h e labour market as a result of technical progress in these three States were absorbed into the production process, so t h a t unemployment on a large scale had no chance to develop. The reserve of labour t h a t gathered during periods of depression returned to active economic service with the next boom. But each of the three countries solved the problem of economic balance in its own way. The profound difference in their development is clearly brought out b y the following table : Average annual increase Great Britain 1861-1911 °/o Total population +0.9 Occupied population +1.2 Occupied in industry +1.3 Industrial production . . . . +1.85 Individual output +0.59 1 Workers in manufacturing industries only. 2 Workers and salaried employees. Germany 1882-1907 °/o +1. +1. +2. +4. +2.1 United States 1869-1899 1899-1913 °/o 3 6 2 3 +2.3 — + 3.21 +5.6 +2.3 °/o +3.2 — + 3.42 +4.5 +1.1 Although the rates of increase of the population, of industrial expansion and of technical progress were quite different in the three cases, large-scale unemployment could be avoided in all three because the expansion of production not only kept pace with, but actually outstripped technical progress and therefore provided employment for an ever larger section of the population. — 3S — Industrial occupations not only absorbed the natural increase in the families of industrial workers b u t also provided openings for the sons and daughters of peasants who left the country for the towns. I n America, in addition, employment could be offered to millions of immigrants, and this enabled the European countries to dispose of their surplus population. Taking the world economic system as a whole, one may attribute the equilibrium of the pre-war labour market to two facts : (a) I n industry, the position was always : A V > E - A T . AV (b) The difference AT —E • » which represents t h e power of industry to absorb fresh elements in the occupied population, was always so large t h a t t h e relative increase in t h e AE number of wage earners employed in industrial undertakings —E AS or the number of wage earners in industrial occupations —remained constantly higher t h a n the relative increase in the total occupied population I n other words, the ratio between the speed of industrial expansion and the rate of technical progress was such as to render possible the progressive industrialisation of the rapidly increasing population. That is the secret of the balance t h a t was maintained on t h e labour market of t h e world before t h e war. I t is true t h a t neither Germany nor Great Britain could have escaped mass unemployment if they had not been able to send a steady stream of emigrants to America. And this migration was possible just because a t t h a t time the population of t h e United States was becoming industrialised. The importance of emigration for t h e labour markets of the European countries before t h e war is sufficiently well known. The purpose of the above remarks was merely to modify in one respect the current notion of the p a r t played by emigration. I t is incorrect to consider t h a t the freedom of movement of labour before t h e war was the governor in the mechanism of the labour market and t h a t its failure is t h e cause of the present catastrophic unemployment. International migration could continue to play this part only so long as there were countries in the world whose — 39 — native working population was not large enough for their economic development, and these countries were therefore obliged to call in the help of foreign workers. But before the war even the emigration countries were still in a position to absorb in their own economic systems quite a considerable increase in population. The reasons why Europe, America and Asia are now all unable to absorb in the production process the increase in their population are not to be sought in the stoppage of migration ; they lie in deeper disturbances affecting the interplay of economic and demographic factors. These disturbances must be studied separately in each case. CHAPTER IV U N E M P L O Y M E N T A F T E R T H E WAR United States Profound changes took place in the economic system of the United States during the war years. F r o m the demographic point of view, the first effect of the war on the labour market was a decline in t h e supply of labour. The influx of immigrants stopped immediately, the net immigration being as follows (in thousands) : 1913 1914 1915 1916 1917 1918 1919 890 915 123 169 229 16 18 I n five years t h e aggregate decrease in immigration as compared with the pre-war figure was 3.5 millions, and this meant a considerable relative decline in the supply of labour on the market. F r o m 1917 onwards, this shortage was accentuated by the departure of millions of the best elements in the occupied population on military service. Thus the United States labour market was depleted just a t a time when there was a rapid increase in the demand for labour. The United States suddenly found themselves undisputed masters of foreign markets for which they had long been contending. They supplied the European belligerents with all sorts of raw materials and manufactured articles, a n d this necessitated the hurried creation of their own war industry, the building of gigantic plant and the constant supply of fleets of vessels to all parts of the world to replace those t h a t were sunk. The potentialities of development of the American economic system and its adaptability were p u t to a severe test, from which they emerged triumphant. B u t a t the outset this expansion of industrial production was not accompanied either b y technical progress or b y t h e usual selection of t h e new labour t h a t was engaged. At a time when there was a market for any kind of commodity, any kind of worker was good enough. Hence the characteristic feature in the growth of industry in the United States during the war period is — 41 — that the number of persons employed in industry increased more rapidly than did the volume of production, while the individual output fell as if the country had been passing through a depression instead of a boom period. These circumstances helped to pave the way for the sudden, miraculous advance in the technique of production that was to astonish the world a few years later. The situation continued unchanged through 1919 into 1920. Europe was suffering from a shortage of commodities that her industries could not make good. The United States dominated the world's markets, and their industrial undertakings were packed with workers whose average output per head was far below the 1913 level, being nearer to the state of affairs at the beginning of the century. In 1921 came the depression : production fell, the surplus labour was paid off, but the output per head—in contrast to what had occurred in earlier depressions 1—remained about the same. In the United States the 1921 depression was essentially one of rationalisation, and its effects made themselves felt as soon as the economic situation revived, in 1922 and 1923. Individual output shot up rapidly, production expanded enormously, but the number of persons employed rose only to a limited extent. TABLE V. DEVELOPMENT OF MANUFACTURING INDUSTRY IN THE UNITED STATES, 1 9 1 3 - 1 9 2 7 * (1899 = 100) Annual average 1913 1914 1915 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 Persona employed 159.0 156.2 160.0 187.0 204.0 210.0 204.4 205.0 158.2 172.9 196.7 184.0 188.9 191.1 186.4 i Recent Economic Changes, Vol. II, p. 454. 1 Cf. above, p. 34 and diagram VI. Volume of production Output per bead 185.0 169.4 188.0 223.0 224.0 220.0 213.7 221.4 169.7 222.2 260.7 244.7 274.6 284.2 278.7 116.3 108.5 117.4 119.2 109.8 104.7 104.5 107.9 107.3 128.5 132.5 133.0 145.4 148.7 149.5 — 42 — The sudden rise in the average output per head from 107.3 in 1921 to 132.5 in 1923—an increase of 25 per cent, in two years— impressed public opinion as being a technical revolution. Since then it has become the habit, in dealing with the latest economic developments in America, to take the year 1920 or 1921 as a basis for comparison. It would be just as correct to take 1917 or 1918 as a standard for comparing the birthrates. The low individual output in the industries of the United States from 1918 to 1921 did not reflect either the technical equipment of the country or the real quality of its workers; it was simply artificially lowered by the peculiar economic conditions arising out of the war. As soon as conditions returned to normal, the index of individual output returned to a level corresponding more or less closely to the general trend of its development. DIAGRAM VIII.—INDUSTRIAL PRODUCTION IN THE UNITED STATES (1899 = 100) 280 -**- 260 240 220 200 180 160 140 120 \oa 100 1915 14 16 16 17 16 19 20 21 «•••••anaa» Volume of production. = = = = = Number of wage earners. • .......~.~ Individual output. 22 26 24 2» 26 27 Output per person in industry in 1927 was 28.5 per cent. higher than in 1913, the annual average increase being about 1.8 per cent. This rate of progress is more rapid than that for — 43 — the period 1899-1913, but it does not reach the rate of that period of feverish expansion, 1869-1899. Since the end of the war, however, an important new tendency has manifested itself in the economic and demographic evolution of the United States : industry, notwithstanding its steady expansion, is not absorbing any new workers. This phenomenon, which is not confined to the United States, must now be studied in some detail. According to Recent Economic Changes, the occupied population of the United States was distributed as follows at different dates : TABLE VI. GAINFULLY OCCUPIED PERSONS IN THE U N I T E D STATES 1 (In thousands) 1920 Total population. 106,422 Total gainfully 40,008 Total non-agricultural gainfully employed 31,137 Total employees attached to nonagricultural pur27,558 Including : Mines, quarries, Manufacturing . Construction . . Transportation. Mercantile emPublic " em- Miscellaneous groups 2 . . . . 1921 1922 1923 192a 1925 1926 1927 108,370 109,742 111,478 113,466 115,004 116,442 117,980 40,429 40,701 41,313 42,095 42,659 43,218 43,943 31,681 32,382 33,024 33,909 34,621 36,491 36,372 27,989 28,505 29,293 30,234 30,941 31,808 32,695 1,217 1,234 1,250 1,254 1,196 1,182 1,278 1,285 11,183 10,754 10,737 10,713 10,487 10,488 10,677 10,598 932 932 1,199 1,277 1,352 1,613 1,594 1,563 4,235 4,151 4,431 4,691 4,658 4,582 4,744 5,204 3,215 " " 3,298 3,694 4,237 2,719 2,689 2,618 2,633 4,057 4,931 4,576 4,488 4,015 4,297 4,412 4,623 2,674 2,736 2,785 2,819 5,852 6,043 6,318 6,603 -- - i Recent Economic Changes, Vol. II, p. 474. The figures are based on King's estimates and are reproduced here with some minor alterations. * Including banking, the professions, etc. According to this table, the occupied population of the United States increased by about 4 millions in seven years. At the same time, there was a large influx of workers from agricultural to urban occupations, so that the latter had to absorb in 1927 some — 44 — 5.2 million more gainfully occupied persons than in 1920. And that number was made up entirely of wage earners. Yet the number of wage earners in mines and manufacturing industry fell quite appreciably during that period, having been 12,400,000 in 1920 and 11,883,000 in 1927—a decline of 517,000. Consequently the rest of the non-agricultural occupations received, during the period in question, an influx of 5.7 million new workers, distributed as follows over the main groups : Construction Transportation Mercantile employees Public employees Miscellaneous" 630,000 970,000 1,410,000 100,000 2,550,000 Total... 5,660,000 During this period, then, the great influx of new workers to the towns in search of employment was accompanied by a sort of " de -industrialisation " of the population 1 . Employment in industry had reached saturation point, as had long been the case in agriculture, and industrial occupations, with the exception of the building industry, were closed to the new members of the occupied population. It must be admitted that the results of the 1930 census do not concord entirely with the conclusions of the authors of Recent Economic Changes. For the years 1920-1930 the census shows no absolute decrease in the number of persons occupied in mines and manufactures. But it does provide striking confirmation of the relative decline in this part of the population. TABLE VII. CLASSIFICATION OF THE OCCUPIED POPULATION OF THE UNITED STATES BY ECONOMIC BRANCHES 1 Percentage distribution Actual n u m b e r s (thousands) Economic branches 1910 1920 1930 1910 1920 12,630 10,936 10,722 33.1 26.3 Mines and manufactures... 11,622 13,922 15,095 30.4 33.4 Commerce and transport .. 6,363 7,410 9,981 16.7 17.8 7,552 Total... 9,346 13,032 19.8 22.5 38,167 41,614 48,830 100 ' Statistical Year-Book of the League of Nations, 1933-1934, p. 39 1 See note 2, p. 32. 100 1930 22.0 30.9 20.5 26.6 100 — 45 — The divergence between the estimates made in Recent Economic Changes (table VI) and the census results (table VII) can be explained in part by the fact that the latter include building along with industry. Of the 1,173,000 gainfully occupied persons who, according to the census, migrated to industrial occupations, the majority selected some occupation in the building industry. The absolute increase in the number of persons in industrial occupations in the narrower sense was insignificant. It may possibly have occurred during the period 1927-1929, when the revival of industry awakened among many of the unemployed the hope of finding employment in some manufacturing industry. The disparity between the two sources is therefore less than it appeared at first sight. The important point is that of the 7.2 million new workers added to the occupied population of the United States between 1920 and 1930, only about 5 or 6 per cent. were absorbed by industrial production in the strict sense, which amounts, for all practical purposes, to saying that these occupations were closed. The estimates given in Recent Economic Changes for the years 1920-1927 show that there were considerable fluctuations in employment in various branches of industry and occupational groups, more especially in mining and manufactures. TABLE VIII. CLASSIFICATION OF WAGE EARNERS IN EMPLOYMENT IN THE UNITED STATES BY ECONOMIC BRANCHES 1 (In thousands) Economic branches Mining Manufacturing Construction Transportation and communication.... Public service, mercantile employees, etc Total... 1920 1921 1922 1923 1924 1925 1926 1927 874 955 925 905 943 764 730 870 10,696 8,200 8,976 10,281 9,563 9,910 10,125 9,871 702 684 969 1,057 1,002 1,268 1,314 1,141 4,065 3,553 3,851 4,440 4,318 4,398 4,600 5,052 11,058 12,166 12,716 13,145 13,671 9,751 10,518 HO,538 26,157 23,719 25,064 27,761 27,919 29,166 30,139 30,640 i Recent Economic Changes, Vol. I I , p p . 475, 477, 478. After the 1921 depression, which led to a sudden drop in employment in manufacturing industries (23 per cent.), mining (20 per cent.), construction (3 per cent.) and commerce (12 per — 46 — cent.), employment improved from year to year, so that in 1927 about 4.5 million more workers were in employment than seven years earlier. But this was achieved only at the cost of overcrowding in non-industrial occupations, for the number of those employed in manufacturing industries and in mines had declined by approximately 800,000. This redistribution of the wageearning population was due in part to new economic trends in the United States. Progress in motor transport had created new occupations and stimulated building activity; the growing interest taken in sport by wide circles of the population after the war accentuated the same tendencies ; at the same time, modern commercial methods demanded the employment of ever increasing staffs. But it is clear that the closing of industrial production to all new workers was bound to create a dangerous situation on the United States labour market : the country thereby entered the phase of increasing chronic unemployment. American economists are not agreed as to the extent of unemployment during the years under consideration. The estimate of the authors of Recent Economic Changes are reproduced in table IX. TABLE I X . ESTIMATED AVERAGE MINIMUM VOLUME OF UNEMPLOYMENT IN THE UNITED STATES x (In thousands) Economic branches Total Including : A Total... 1920 1921 1922 1923 1924 1923J 1926 1,401 4,270 3,441 1,532 2,315 1,775 1,669 2,055 520 274 470 487 2 , 5 5 4 1,761 329 432 326 924 308 578 323 552 380 727 761 3 , 0 2 4 2,281 761 1,250 886 875 1,107 230 248 230 220 350 345 280 422 170 598 580 251 340 184 144 152 240 . 400 350 300 375 360 370 374 771 1,065 889 794 948 1927' B Transportation and communications . . P u b l i c service, m e r cantile, miscellaTotal... 640 1,246 1,160 i Recent Economic Changes, Vol. II, p. 478. — 47 — No claim to strict accuracy can be made for the figures in the above table; they are merely estimated minimum figures. But even as such they would seem to be misleading in one respect : they give the impression that unemployment developed more or less evenly in every section of the economic system. The figures given can be shown as follows : Mining and manufacturing . Other occupational groups . 1920 192T (iu thousands) Increase (per cent.) 761 640 1,107 948 -f- 45 -j- 48 But it is a well-known fact that the various sections of the labour market do not constitute watertight compartments, and in the United States labour is in any case particularly mobile. It is therefore possible that unemployment was not recorded in the occupations in which it actually occurred, but was shown as being in quite different branches, to which the workers from distressed occupations went in search of employment. Take, for example, the building industry, in so far as tables VI, VIII and IX throw light on the situation. The number of unemployed, in thousands, is shown as : Workers in the occupation Including : In employment Unemployed 1920 1927 932 1,563 702 230 1,141 422 In seven years, 631,000 workers entered this occupational group. Of those, 439,000 found employment, while 192,000 went to swell the army of unemployed. Consequently, the increase in unemployment among building workers is not due to the economic development of this particular industry, but must be explained by the situation in other branches of the economic system which were unable to absorb any new workers during the observation period, reduced the number of workers they employed from year to year and therefore forced those who became unemployed to seek employment in other occupations. In other words, the one source of unemployment in the United States during this period was manufacturing and mining. The development of American industry from 1920 to 1927 can be expressed in terms of the formulae used above. In 1920, the occupied population of the United States num- — 48 — bered 41.6 millions, of whom 11.2 millions were wage earners in manufacturing industries 1. In 1930, the occupied population had risen to 48.8 millions. If this increase was evenly distributed over every branch of the economic system, the number of wage earners in manufacturing occupations would have risen by about 1.7 millions in 10 years, or 170,000 annually. If there was to be no unemployment in this branch, the number of persons actually in employment in manufactures would have to evolve more or less as follows : 1920 = bands')" Index = U ' 1921 2 0 0 1922 1923 1924 1925 1926 1927 370 H) 11,540 11,710 11,880 12,050 12,220 12,390 100 101.46 102.92 104.38 105.84 107.30 108.76 110.22 But in reality the number of wage earners in employment in manufacturing industry in the United States 2 was : 1920 = 10 700 1921 8 1922 200 8 980 1923 280 1924 9 bands'?" > > ' 10, ,560 Index = 95.5 73.2 80.2 91.8 85.4 Basis : Number of wage earners in 1920 (11,200) = 100. 1925 1926 1927 9,910 10,120 88.5 90.4 9,870 88.1 The volume of industrial production (V) and the ratio of this volume to the number of persons in employment ( T = — 1 fluctuated as follows during the same period : 1920 V = 100 V T = - = 1,047 1921 1922 1923 1924 1925 76 100 117 110 124 1926 128 1927 126 1,038 1,247 1,274 1,288 1,401 1,416 1.420 Diagram IX shows the development of unemployment in the manufactures of the United States during this period, which was considered by public opinion as being one of rapidly increasing and untroubled prosperity, whereas in reality profound disturbances were spreading under the surface of the social and the economic system. 1 For the occupied population in 1920 and 1930, cf. table VII. The number of wage earners in industrial occupations (in the narrow sense) is taken from table VI. » Cf. table VIII, p. 45. — 49 — DIAGRAM IX. DEVELOPMENT OF UNEMPLOYMENT IN THE MANUFACTURES OF THE UNITED STATES (S a n d V for 1920 = 100) 1920 1921 1922 1929 1924 192« 192Ó 1927 In the years covered by this diagram, the heads of the arrows representing the (theoretical) volume of unemployment in industrial occupations as a percentage of the number of persons in these occupations rise much higher than the gradually ascending curve AS. The ordinates of this curve represent the sum of two quantities : (a) the number of unemployed at the beginning of the observation period (4.5 per cent, of the number of persons in industrial occupations in 1920) and (b) the probable increase in the number of persons in these occupations (approximately 170,000 or 1.46 per cent, annually). In none of these years was industry able to absorb even a fraction of this mass of unemployed persons. On the contrary, industrial undertaking had to dismiss a fraction of their staffs—in 1921 because of the falling off in production, in 1922 because individual output had increased so quickly while production remained stagnant at the 1920 level. It is true that the volume of production rose in the following years, but its increase as compared with 1920 could not keep pace with the march of technical progress. In other words, during the years 1920 to 1927, human labour was crowded out of industrial production in the United States by machinery. But this statement calls for some further explanation. — 50 — The elimination—or the release—of human labour as a result of technical progress is by no means peculiar to very recent times. It is, indeed, one of the essential processes in the course of economic and social development. But for the spread of machinery it would scarcely have been possible to raise the standard of living of the world. The nineteenth century was truly an age of miracles in the realm of technical progress. But so long as the liberation of workers from certain branches of production is more than balanced by the creation of new branches and the expansion of production as a whole, there is no such thing as a problem of technological unemployment. Even if the balance is temporarily disturbed, the results may be considered as part of the price nations have to pay if they wish to progress. The novel feature in the industrial development of the United States after the war was that the counterbalancing mechanism broke down right in the middle of a boom period, and the expansion of production created no extra demand for labour. The biennial census of production in the United States gives only a faint reflection of this new trend. TABLE X. INDUSTRIAL Establishments PRODUCTION IN with an annual production Number of workers Year THE UNITED STATES x of not less than §5,000 Number of salaried employees H.P. Production Absolute figures (thousands) 1923 1925 1927 1929 - 8,778 8,384 8,350 8,839 1,269 1,256 1,301 1,359 33,094 35,773 38,826 42,931 — Index numbers (1923 = 100) 1923 1925 1927 1929 100 95.5 95.1 100.7 100 99.0 102.5 106.4 100 105.0 117.3 129.7 100 104 106 119 i Statistical Abstract of the United States, 1932, p. 730. According to these figures, industrial undertakings in the United States were employing rather more workers in 1929 than they had six years earlier, the whole of the improvement having taken place in the last two years. — 51 - In that case, the development of industry would have been quite favourable. There would, it is true, have been no appreciable increase in the number employed, but at least there would have been no dismissals. The following coefficients would therefore represent the progress of American industry : Volume of production Output per person N u m b e r of workers employed . . Percentage increase over 6 years annually +19.0 + 2.9 + 17.6 + 2.7 + 1.5 +0.2 It should be noted, however, that the figure given by the census for the number of those in employment in 1929 was affected by the fact that the peak of the business cycle was reached that summer. Quite a different picture is presented by the current employment statistics of the United States. The figures from that source for the years up to 1927 were given in table V. For the subsequent period the Bureau of Labor Statistics has compiled new series going back as far as 1923. TABLE X I . DEVELOPMENT OF MANUFACTURING INDUSTRY IN THE UNITED STATES, 1 9 2 3 - 1 9 3 3 x (1923 = 100) Average tor year 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 Persons employed 100 90.3 91.4 92.0 88.7 86.4 89.8 78.0 66.4 55.4 59.5 Volume of production 100 94 105 108 106 112 119 95 80 63 76 Output per head 100 104 115 117 120 129 133 122 120 113 128 i Monthly Labor Review, February 1934, p. 381. Ct. diagram X. A comparison of diagram X, in which these figures are represented graphically, with the corresponding graphs for the pre-war period (diagram VI) is sufficient to show the profound difference between the development of United States industry during these two periods. — 52 — DIAGRAM X . — I N D U S T R I A L PRODUCTION I N THE UNITED STATES (1923 = 100) Manufacturing industry 192* 24 íb 26 27 78 29 — — • — — - Volume oí production. - Number of wage earners. Individual output. SO 31 32 3i For the pre-war period, the broad curve marking individual output lay well below the other two curves ; in the later period it dominates the upper p a r t of the graph. Leaving out of account for the moment the developments after 1929, it will be well t o analyse first the growth t h a t took place from 1923 to 1929. I n t h a t period the volume of production increased on t h e average by 2.9 per cent, annually. Contrary to what is generally held, this must be considered as very slight progress when compared with t h a t of preceding years. This will appear at once from a comparison with the average rate at which industrial production increased from 1869 to 1889 and from 1899 to 1913 1 : 1869-1899 1899-1913 5.6 per cent, annually, 4.5 „ „ The annual increase in the output per earners in the manufacturing establishments was : 1869-1899 1899-1913 1923-1929 1 Cf. above, p. 37. head of the wage on the other hand 2.3 per cent. 1.1 „ „ 4.8 „ „ — 53 — It must be remembered that the first two figures refer to long periods covering several business cycles, whereas the period 19231929 covers only the boom section of a cycle. But notwithstanding this reservation an increase of 4.8 per cent, must be considered very high. Yet even this rate of technical progress would not have destroyed the balance of the social system in the United States if the volume of production had continued to rise, as formerly, by 4.5 or 5.6 per cent, annually. The whole danger lay in the ratio between these two rates of development—i.e., in the fact that E •AT>A V and that the increase in production was not sufficient to absorb the workers released by technical progress. DIAGRAM X I . PRODUCTION AND JL THE DEPRESSION IN THE UNITED STATES : EMPLOYMENT IN MANUFACTURING too w 90 (*-!) 100 — — 113.5 109.3 114.4 114.7 112.8 114.1 113.9 118.4 The evolution of the British labour market can now be summed up by means of the usual formula. It is assumed at the outset that in view of the annual increase of 1.4 per cent, in the wage-earning population, industry should have increased to some extent the number of persons it employed. In so far as this increase did not take place, it follows that unemployment developed in that branch of the economic system, although that unemployment may have been absorbed, in whole or in part, by increased employment in other branches. The year 1924 will be taken as a starting point, and the number of insured workers in industrial occupations in July 1924 (S = 7,156,000) will be considered as = 100. As there were at that date 738,000 unemployed persons in industry, the index number of wage earners in employment (E) must be put at 89.8 (there is no need, in this case, to allow for absence through sickness, etc.). The increase that should have taken place in the number of those employed in industrial occupations will be put at the minimum rate of 0.5 per cent, of the 1924 figure annually. 1 I t is of importance, however, for the year 1926 (coal strike), but t h a t year has been omitted from the following table. — 76 — The following diagram will then represent the development of unemployment in British industry from 1924 to 1934. DIAGRAM XVI. DEVELOPMENT IN BRITISH OF UNEMPLOYMENT INDUSTRY (S and V for 1924 = 100) This diagram brings out the fact that the volume of production rose more slowly than the level of technical progress from 1924 to 1929. In so far as technical progress is represented by the individual output of those in employment, it was much more rapid than before the war. The increase of 14.4 per cent, in 5 years means an annual rise of 2.7 per cent., as against 0.52 per cent, for the period 1861-1911. The increase in the volume of production, on the other hand, was not much more rapid than before the war : 2.2 per cent, as against 1.85 per cent. There is perhaps a close connection between these two trends. It is possible that it was the difficulties of industrial expansion that forced British industry to improve its plant and its technical processes and bring them up to date. This created a situation in which, notwithstanding the steady improvement in business, unemployment continued to spread in industrial occupations, although its growth was, for a time, obscured by the displacement of labour towards other branches of the economic system. The situation became still worse during the depression. Capital investment came practically to a standstill, and as the undertakings had no longer any reason to replace their workers by machines, individual output has not risen at all since the — 77 — depression began. It has even fallen slightly—perhaps as a result of shorter working hours. However that may be, the decline in production was alone responsible for the decrease in employment in industry from 1930 to 1932, and the two decreases followed almost parallel courses. During this period labour continued to flow from industry to other branches of the economic system, and this influx soon made itself felt. The effects of this displacement and the distribution of unemployment over the two main branches of the labour market are shown in diagram XVII. DIAGRAM XVII.—DISTRIBUTION OF UNEMPLOYMENT IN GREAT BRITAIN Millions Millions 5 ' = 1924 2 5 26 27 20 29 BO 51 32 —5 53 Unemployment arising in non-industrial occupations. äjfi Unemployment passed on from industry to other occupations. • Unemployment registered in industrial occupations. When industry revived in 1933 the situation gradually improved, but in July 1934 the number of unemployed persons in Great Britain and Northern Ireland was still 2.1 millions, of whom 1.2 millions were in industrial occupations. If unemployment was to be brought back to what was considered the "normal" pre-war level, the number of persons employed in industry would have to be raised by about 17 per cent, on the average above the 1934 summer figure. But that would require — 78 — some considerable time, and in the meanwhile neither the increase of the occupied population nor the march of technical progress could be held up. Consequently, if Great Britain wishes to restore its economic equilibrium it will have to increase production by between 20 and 25 per cent., or else reduce hours of work in the same proportion, or combine these two measures. Theoretically, each of the three solutions is possible, but in practice the only one of real importance is the last one, which can, of course, be varied in countless ways. There is little hope of success for any attempt to restore the balance of the labour market by transferring the "surplus" workers from industry to non-industrial occupations. Commerce and transport may not so far be overcrowded occupations in Britain, but the time will inevitably come when the mechanism of distribution will have to be rationalised. And then hundreds of thousands or even millions of commercial employees and workers will be thrown out of employment. Germany The first post-war years in Germany were marked by great economic activity. The demand for commodities was so strong that the industries for consumers' goods had no difficulty in marketing their products, while the industries for producers' goods worked feverishly to provide new industrial plant. Inflation stimulated capital investment, which rapidly degenerated into unbridled company promotion. Consequently, and more especially in view of the low level of wages, German employers could afford to employ an ample quantity of labour. The demobilisation of the army and the reintegration of ex-service men into economic life passed off without much friction. In 1919, 1920 and 1921, unemployment among trade union members fluctuated round about an annual average of 4 per cent. In 1922, when inflation was at its height, the number of unemployed persons actually fell below 1 per cent. There is no means of knowing the exact number of wage earners or even the size of the occupied industrial population in general during these years. But as unemployment was practically non-existent at that period and all the available workers were — 79 — fully employed, the total amount of work performed cannot have fluctuated much or increased greatly from year to year. The growth of industrial production is therefore all the more striking. On the basis of 1913 = 100, the volume of production was as follows : 1919 1920 1921 1922 37 54 65 70 The rise in the level of individual output was more or less parallel. This was no "miracle of technical progress" but simply the process of reconstruction : the economic system had been thrown completely out of gear and was now gradually returning to normal. In 1922 individual output had not yet returned to its pre-war level. The daily output per person was still about 20 or 30 per cent, lower than in 1913. This was partly the result of the reduction of working hours—from 9.5 to 8 hours a day on the average. But even the hourly output in 1922 still fell short of the normal figure. The corollary of this low level of output was an unusually low standard of living. Real wages were brought still lower by the fact that a large fraction of the yield of German industry was devoted to capital expenditure on a large scale; this was the more easily financed because of the continued inflation. The stabilisation of the currency was followed by a brief but violent depression. The unemployment figure soared upwards at the end of 1923, and short time suddenly became extremely widespread. But the spring of 1924 brought recovery with it. From that date the German labour market entered on a new and chequered phase of its history. In view of the marked seasonal fluctuations that are characteristic of employment conditions in Germany, and in view of the curious fluctuations in the market situation from 1924 to 1934, it will be advisable to study the development of unemployment during that period from month to month and not merely from year to year. Table XIX shows the number of unemployed persons registered with the employment exchanges at the end of each month from 1925 to 1934. The question of the reliability and completeness of the employment exchange statistics may be ignored for the moment, while the general trend indicated by these figures is examined. — 80 — TABLE XIX.—UNEMPLOYMENT IN G E R M A N Y 1 (Number of unemployed persons registered) (In thousands) End of month 1925 1926 1927 1928 1929 1930 1931 1932 January . February March . . . April . . . . May . . . . June . . . . July August . . September October . November December 800 731 648 523 431 401 406 459 503 636 997 1,712 2,221 2,269 2,243 2,113 2,090 2,081 2,004 1,911 1,781 1,709 1,786 2,127 2,257 2,167 1,708 1,462 1,213 1,061 927 841 772 787 1,117 1,714 1,791 1,718 1,489 1,234 1,112 1,075 1,028 1,034 1,030 1,164 1,569 2,265 2,850 3,050 2,484 1,712 1,350 1,260 1,252 1,272 1,324 1,557 2,036 2,851 3,218 3,366 3,041 2,787 2,635 2,641 2,765 2,883 3,004 3,252 3,699 4,384 4,887 4,972 4,744 4,358 4,053 3,954 3,990 4,215 4,355 4,623 5,060 5,668 6,042 6,128 6,034 5,739 5,583 5,476 5,392 5,224 5,103 5,109 5,355 5,773 687 2 , 0 2 8 1,336 1,376 1,916 3,140 4,573 Annual average 1933 6,014 6,001 5,599 5,331 5,039 4,857 4,464 4,124 3,849 3,745 3,715 4,059 5,575 4,804 1934 s 3,773 3,373 2,798 2,609 2,529 2,481 2,426 2,398 2,282 2,268 2,353 2,604 2,658 i2 From reports ot the employment exchanges. From July 1933 onwards the figures do not include persons employed In the Labour Service (150,000 in July 1933). DIAGRAM XVIII.—UNEMPLOYMENT IN GERMANY World economie depression K. Millions . 6 Millions Depression due to rationalisation < Extreme cold *• 1925 26 27 28 29 30 31 32 35 34 . _ _ . . . . . - . . . . . . Basic trend between the two depressions. 1 From employment exchange reports. — 81 — If seasonal fluctuations (which reach their peak in January or February and their lowest point in summer) are left out of account, it will be seen that the curve of unemployment in Germany for the period 1925-1934 reflects two unequal waves : the first began to rise towards the end of 1925, falling again in the spring of 1927; the second rose from the winter of 1929, reaching its peak in 1932 and falling gradually in 1933 and 1934. Between these two waves the basic trend of the curve is not horizontal, but would seem to betray a tendency to rise. It is not difficult to determine the relative significance of these fluctuations, which are in part superimposed on each other. The peaks in the winter months reflect the usual dismissal of about a million workers (mostly from the building trade). The peak is particularly high in the winter of 1928-1929, which was exceptionally cold in the whole of Central Europe, and especially in Germany. During the depression, on the other hand, winter unemployment was relatively less marked on account of the slackness in building even in summer. The extent of the first wave of unemployment—the so-called " rationalisation unemployment " of 1926—may be estimated at about 1.5 or 1.6 million persons. During the recent depression, the increase in the number of unemployed workers during the summer months exceeded 4 millions. During the recovery period between the two depressions the annual increase in the number of unemployed persons registered with the employment exchanges may be put at 200,000 or 300,000. The same trend of the labour market is shown by the unemployment statistics compiled by the trade unions l . These figures have the advantage of eliminating the sudden winter variations by making a distinction between occupations in the " seasonal " and the " cyclical groups " 2. The dotted curve in diagram X I X shows the same peaks as the curve in diagram XVIII; every winter, between 5 and 8 per cent, of the members of trade unions were thrown out of employment by the stoppage of building activity; this proportion rose in the winter of 1928-1929, but fell during the depression. The 1 2 Cf. Table XX, p. 83. The " seasonal " group comprises building occupations, brickworks and horticulture; the cyclical " group includes all other occupations. Cf. Wl. WOYTINSKY : " Konjunktur und Saison " in Arbeit, 1929, No. 2, Berlin, and Der deutsche Arbeitsmarkt, published by the Allgemeiner Deutscher Gewerkschaftsbund, 1930, Berlin. Cf. diagram X I X , p. 82. — 82 — curve for the seasonal group reveals whence these peaks come, whereas the curve for the other group reflects the cyclical fluctuations over the same period : a brief wave of unemployment resulting from rationalisation, a gradually rising curve for the years 1927-1929 and then a rapid rise from the beginning of the world depression. DIAGEAM XIX.—UNEMPLOYMENT AMONG TRADE UNION MEMBERS IN GERMANY Per cent Per cent 1001 1 «100 K 90 \ , 80 1 6o 50 1v 50 1 \\ t f i 1 VA \ «V % V 1924 2 5 M 1 I 1 V V 26 27 20 1* }'• -\ // ^^ 10 ^ß\ 2d 40 SO J, \J J i 70 50 t ^•~^V V 80 60 V \ 40 IO / \J i 70 20 90 29 SO 32 5» Cyclical group. Seasonal group. All occupations. The development of the labour market during the " rationalisation depression" may be taken as a typical example of successful rationalisation, since a certain quantity of labour was temporarily displaced but was immediately provided with fresh employment as a result of the extension of production K There is one serious source of error in table XXI. The production figures refer to " industrial " production in the widest sense, whereas the employment figures include " all groups " of wage earners indiscriminately—not only those in industry, but also 1 Cf. table X X I , p. 84 and diagram X X , p. 85. — 83 — TABLE XX.—PERCENTAGE UNEMPLOYED AMONG MEMBERS IN G E R M A N Y Year Jan. Feb. Mar. April May June July TRADE UNION 1 Aug. Sept. Oct. Nov. Dec. 4.5 15.6 4.7 6.6 9.6 22.8 35.5 44.1 7.3 5.8 14.5 4.6 7.3 11.0 24.0 37.2 43.4 8.1 10.7 14.5 7.5 9.4 13.8 26.3 39.5 43.8 13.1 19.4 17.2 12.9 16.7 20.3 31.8 42.8 45.6 4.5 15.6 4.8 6.4 8.9 18.8 28.8 37.6 5.7 14.0 4.3 6.5 9.2 19.4 29.7 36.4 8.2 13.1 4.7 7.3 10.2 20.6 31.1 36.4 14.7 13.3 6.2 9.5 12.8 24.3 33.9 36.6 4.7 15.6 3.9 7.4 12.8 39.7 66.4 75.0 6.8 16.8 5.5 10.3 17.7 43.3 71.1 76.8 25.4 21.0 20.6 18.5 28.6 51.1 78.2 79.4 42.8 35.6 44.0 46.7 51.1 64.3 94.2 83.8 All Occupations 1924 1925 1926 1927 1928 1929 1930 1931 1932 26.5 8.1 22.6 16.9 11.4 19.4 22.2 34.5 44.3 25.1 7.3 22.1 15.9 10.5 22.3 23.7 34.8 44.9 16.6 5.8 21.6 11.8 9.3 16.8 21.9 34.0 45.2 10.4 4.3 18.7 9.0 6.9 11.1 20.5 32.3 44.5 8.6 3.6 18.3 7.1 6.3 9.1 19.8 30.4 43.9 12.5 10.5 10. 3. 3.7 4 . 3 17.9 17.0 18. 6.4 5.6 5.1 6.2 6.3 6.5 8.6 8.6 9.0 19.8 20.8 22.0 30.2 31.6 34.1 43.6 44.4 44.5 8.4 Cyclical Group 1925 1926 1927 1928 1929 1930 1931 1932 6.0 18.2 12.4 6.4 10.3 14.2 25.8 35.2 5.4 19.1 11.7 6.2 11.4 15.0 25.8 35.7 4.7 19.7 10.0 5.9 10.6 15.1 25.6 36.4 4.3 17.6 8.5 5.8 9.2 15.2 25.2 36.7 3.7 17.8 7.3 5.9 8.6 15.4 24.7 36.8 3.5 18.0 6.7 6.0 8.4 16.0 24.8 37.0 3.7 17.9 6.0 6.3 8.6 17.0 26.1 38.0 4.2 17.1 5.4 6.5 8.6 18.2 27.8 38.1 Seasonal Group 1925 1926 1927 1928 1929 1930 1931 1932 21.1 48.4 39.4 34.1 58.4 55.5 72.8 87.0 18.7 38.9 37.3 30.1 68.1 59.5 74.7 88.4 12.0 32.0 20.8 24.6 43.3 50.2 71.4 86.3 5.0 24.9 11.9 12.0 19.2 42.9 63.7 81.4 3.1 21.1 6.4 8.1 11.0 38.1 55.5 77.1 2.8 19.8 4.9 7.0 9.2 36.4 54.1 75.0 3.3 18.2 3.9 6.0 8.9 36.8 55.8 74.5 4.7 16.5 3.7 6.5 10.4 38.1 62.4 74.8 i Jahrbuch des Allgemeinen Deutschen Gewerkschaftsblindes, 1931, Statisticher Anhang, p. 2. those in agriculture, commerce and transport, the liberal professions, domestic service, etc. B u t it is a well-known fact t h a t in 1926 rationalisation meant, in Germany, the technical and administrative reorganisation of industrial undertakings only. Consequently the employment figures in table X X I contain two — 84 — TABLE X X I . — T H E " RATIONALISATION DEPRESSION IN GERMANY 1 (Average for 1925 = 100) Year and month Number of wage earners employed Amount of work performed Industrial production Individual output 1925 January . . February . March April May June July August . . . September October . . November. December 95.7 96.7 98.2 102.0 104.1 104.1 103.0 102.5 103.0 101.5 98.2 91.4 96.3 97.4 98.8 102.8 104.9 104.8 103.6 102.8 102.9 100.4 96.2 88.9 101.2 103.7 104.6 101.1 103.3 100.6 99.4 96.5 100.1 97.5 98.5 92.2 1926 January . . February . March . . . . April May June July August . . . September October . . November. December 88.3 88.3 89.3 92.5 93.4 93.5 94.1 94.6 95.1 95.6 95.6 91.5 84.8 85.0 86.0 89.9 91.2 91.4 92.1 93.0 94.1 95.2 95.6 91.7 88.2 88.4 88.1 87.5 88.9 93.2 92.1 98.0 102.5 106.3 111.1 113.0 104 104 103 97 97 102 100 105 108 111 117 123 1927 January . . February . March . . . . April May June July August . . . September October . . November. December 90.5 91.5 96.7 100.4 103.5 104 105 105 106 106 104.1 98.8 90.9 92.1 97.6 101.5 104.8 105.9 106.5 106.9 107.5 108.1 105.6 100.7 112.4 111.2 115.5 118.8 123.5 120.8 125.4 127.5 130.1 128.8 131.1 124.4 123 121 118 117 119 115 119 120 122 120 125 124 105.1 106.4 105.9 98.3 98.4 96.0 96.0 93.9 97.3 97.1 102.3 103.7 i The index of the number of wage earners in employment is calculated on the basis of the reports of the sickness insurance funds (cf. Wirtschaftszahlen 1925 bis 1931, published by the " Institut für Konjunkturforschung ", Berlin, 1932). The index of the amount of work performed is calculated with the help of the index of wage earners in employment, taking into account the average hours of work. The index of the volume of production is that of the " Institut für Konjunkturforschung ". The index of individual output is obtained by dividing the index of production by the index of the amount of work performed. — 85 — elements, one of which was directly influenced by the depression, while the other was not affected to any appreciable extent. The sickness insurance fund statistics that were used in calculating the index of employment do not permit of a distinction being made between these two elements. If it were possible to work out the index of employment for mines and manufactures only it would certainly show much more marked fluctuations than those in the general employment index given in table XXI. With this one reservation the table may be taken as giving a satisfactory general picture of the economic development of Germany from 1925 to 1927. DIAGRAM XX. THE "RATIONALISATION" DEPRESSION IN GERMANY (Average for 1925 = ' I925 ' .' I I •—I 1 l—i 1 1 •—I—_-J I926 100) _J—I —-I 1 • 1 1 1 1 1 1 J ÖO I927 o-o-O-O-o-o-oo Individual output. • _ _ _ • _ • _ _ _ • Production. _ _ _ _ _ _ _ _ _ _ Amount ol work performed. In the second quarter of 1925 it will be noticed that the volume of industrial production in Germany fell from month to month, although the number of wage earners in employment tended to rise. The output per head of those employed must therefore have been falling steadily. It fell from 106 in February and March 1925 to 94 in August of the same year. Then came the rationalisation of production, with, as its first — 86 — consequence, the better organisation of work in industrial undertakings and the dismissal of superfluous workers. After allowing for winter unemployment, it will be found that the number of wage earners dismissed was about 2.5 millions in round figures, or between 12 and 13 per cent, of the total number employed 1 . At the same time the amount of work performed declined to an even greater extent because of the reduction of hours of work. It is true that these phenomena were accompanied by a decrease in the volume of production, but this decrease fell far short of the fall in the amount of work performed. Between the third quarter of 1925 and the month of April 1926 the index of industrial production fell by about 11 per cent., while the amount of work performed fell by 13 per cent. When it is remembered that some 6 or 7 million wage earners in non-industrial occupations were not affected by the depression, whereas the number of those employed in mining and manufacturing was cut down from 13 to 11 millions, it will be seen that the decrease in the amount of work performed in industry may be assessed at 20 per cent. But the increase in individual output enabled industry to expand its sales and production beyond the previous level. In the summer of 1926 the number of persons in employment began to rise again. A year later their number was half a million higher than it had been before the depression. As a result of the reorganisation of industry, the output per head of those employed had risen by 24 per cent. One special feature in the development of the " rationalisation depression " in German industry is worthy of note. When the depression began, recourse was had to shorter working hours before the dismissal of workers began. It would appear that the industrialists considered the decline in the amount of work as a temporary phenomenon, and therefore decided to distribute the available work over their existing staffs 2. It will be seen in diagram XXI that the curve of short time rose a few months before the curve of complete unemployment in 1925. The two phenomena occurred in the same order within 1 The process of " weeding out " must therefore have been much more complete than might be imagined from the reports of the employment exchanges, which recorded an increase in unemployment of from 1.5 to 1.6 millions. The main reason for the difference between the two sets of figures is the fact t h a t some of those who were temporarily thrown out of employment did not register with the employment exchanges. a Cf. table X X I I and diagram X X I , p. 87. — 87 — the various groups of occupations and branches of the economic system during this depression brought about by rationalisation K TABLE X X I I . — P E R C E N T A G E OF WORKERS ON SHORT TIME AMONG TRADE UNION MEMBERS IN GERMANY Cyclical 1 group At the end of the month of : Year Jan. Feb. Mar. April May June July Auf?. Sept. Oct. Nov. 1925 1926 1927 1928 1929 1930 1931 1932 5.6 25.8 7.5 4.1 9.3 12.6 22.6 26.7 5.4 24.9 6.6 4.2 10.0 15.1 23.1 26.5 5.1 24.9 5.0 4.3 9.1 14.7 22.4 26.6 4.9 21.4 4.3 5.0 8.0 14.3 21.5 25.8 5.0 20.4 3.4 5.8 7.8 14.1 20.9 26.1 5.2 19.2 3.2 6.8 7.6 14.8 20.9 26.1 5.9 18.8 3.1 7.5 8.0 16.3 22.7 26.7 7.1 17.2 3.4 8.1 8.2 17.1 25.1 26.9 8.9 14.5 2.9 7.7 8.0 17.5 26.0 26.3 12.9 11.6 2.4 7.7 8.1 17.8 25.7 26.1 17.8 9.5 2.5 8.2 8.6 18.5 25.5 25.6 Dec. 21.5 8.3 3.5 8.1 9.4 19.5 26.3 26.4 i Jahrbuch des Allgemeinen Deutschen Qewerkschaftsbundes, 1931, Statistischer Anhang, p. 3. Only the figures for the " cyclical " group are given. In the " seasonal " group the amount of short time is negligible. DIAGRAM XXI.—COMPLETE UNEMPLOYMENT AMONG TRADE UNION MEMBERS IN Cyclical % ft. AND SHORT TIME GERMANY Group % ¿5 20 IS IO b 192» 1926 1927 1926 1929 " ^ • • " ^ Complete unemployment. - — — = = Short time. 1 für 1950 Cf. W l . WoYTiNSKY : " Arbeitslosigkeit u n d K u r z a r b e i t " i n Jahrbücher Nationalökonomie und Statistik, Series I I I , vol 79 (1931), p p . 13 et seq. — 88 — In 1929, on the other hand, the reduction in hours of work tended to come after the reduction in staff; the first reaction was to cut down staff, and it was only later that the idea of distributing the work over those who were still kept on was put into practice. The reason for this would seem to be that the heads of undertakings took quite a different view of the situation in 1929 from that taken in 1925. It has since been asserted in many quarters that the rationalisation movement in German industry was of doubtful economic value; it has been described as " irrational rationalisation ". It is true that the reorganisation and modernisation of production was the latest fashion in 1925-1926, and there is no doubt that some individual industrialists and some whole branches of production went too far in their efforts to be up-to-date. On the whole, however, the rationalisation movement achieved its aim, and by 1927 the German economic system had regained its equilibrium1. This conclusion is not invalidated by the fact that unemployment increased in Germany, as was mentioned above, from 1927 to 1929. This tendency existed before rationalisation brought on the depression. The average number of unemployed persons registered with the employment exchanges during the third quarter of the year (when seasonal unemployment is at its lowest) was as follows (in thousands) : 192S 1926 1927 1928 1929 increase 1925-1929 458 (depression) 847 1,031 1,283 825 But during these four years there had been an influx of 1,433,000 new workers on the labour market, representing an increase of 350,000, or 1.8 per cent., annually. This rate was, it is true, more or less the same as before the war 2 . But conditions had changed, and the capacity of the German economic system to expand could not be the same as it had been at the beginning of the century, for example. It was able to absorb just over 600,000 of the new wage earners representing the younger generation; the remaining 800,000 went to swell the reserve army of labour. 1 There is no foundation for the view that Germany was able to overcome the depression in 1927 only by a lucky accident—the British coal strike. The field of industrial expansion after rationalisation was in production for the home market far more than in production for export. 3 Cf. above, p . 29. — 89 — It is interesting to consider whither the further development of economic life in Germany might have led if its course had not been interrupted by the depression in the autumn of 1929. If it is assumed that nothing happened to check its normal course, it must follow that the demand for labour would have increased so as to provide employment for about 150,000 persons annually. For a country like Germany, that is a low figure, but it would, in a few years, have proved sufficient to relieve considerably the overcrowding on the labour market. It must be remembered that the small contingents of young persons born during the war years were just then reaching working age 1 . The German National Statistical Office had calculated the probable number of workers for the years 1929 to 1933 as follows (in thousands at the beginning of each year) : 1929 1930 1931 1932 1933 20,993 21,127 21,094 20,963 20,832 The labour force of the country was therefore expected to remain stationary, after having increased rapidly during the preceding period. As the supply of labour was therefore constant and the demand for labour was increasing, unemployment should have dwindled from year to year. On this assumption it may be calculated that the number of unemployed persons registered with the employment exchanges would have moved as follows (in thousands for the third quarter of each year) : 1929 1930 1931 1932 1933 1,283 1,270 1,100 830 550 But the depression upset all these calculations. A study of employment in large and medium-sized industrial undertakings confirms the impression that the structural development of the German economic system before the depression had on the whole been healthy. The number of workers and salaried employees in employment in undertakings with five or more workpeople fluctuated as follows from 1926 to 1930 (in thousands): Industry * 1926 1927 1928 1929 1930 7,583 8,892 9,099 8,877 7,523 Commerce and transport 1,134 1,330 1,425 1,515 1,470 Theatre, education, public health Total 166 169 185 205 213 8,883 10,391 10,709 10,597 9,206 i Including mining, manufactures, building a n d water, gas a n d electricity services. 1 Cf. diagram I, p . 8. 7 — 90 — Of the increase of 1,500,000 persons employed in large and medium-sized industrial undertakings from 1926 t o 1927, industry in the wide sense absorbed 1,300,000, including 230,000 in the building trade. The following year the number of those in employment rose again by about 320,000. Then came a decline in 1929, marking the beginning of the depression. I t is of interest to consider the ratio of the volume of industrial production to the number of persons employed in these large and medium industrial undertakings. If 1928 = 100, the index numbers for the annual averages will be as follows : 1926 Persons employed in large and medium industrial undertakings (E) Volume of industrial production (V) Individual output f l O O g ) 1927 1928 1929 83.3 80.8 97.7 101.1 100 100 97.5 100.4 97.0 103.5 100 103.0 The period of intensive rationalisation had thus been- followed by one of less rapid technical progress. If production had continued to expand steadily by 3 or 4 per cent, annually, the undertakings would have had to increase their staffs every year. I t may therefore be concluded t h a t the collapse of the German labour market cannot be attributed t o the use of machinery; i t was due t o a decline in sales and in production. The statistics of the employment exchanges (table X I X , p. 80) give only a faint impression of the extent of this collapse. According to them, the number of unemployed persons during t h e depression was only about 4.2 millions more t h a n in 1928-1929. B u t in reality the situation was more serious, as can easily be seen from the fluctuations, from the beginning of 1929 onwards, in the number of wage earners in employment, as recorded b y the sickness insurance funds. (Cf. table X X I I I , 91.) According t o these statistics, the number of wage earners in employment fell from 18,638,000 in J u n e 1929 to 12,779,000 in J u n e 1932. As the total number of wage earners in Germany remained unchanged during this period, the only possible conclusion is t h a t the number of unemployed persons increased by more t h a n 5.8 millions, and not merely by 4.2 millions. The difference can be explained by the invisible unemployment t h a t developed during the depression. The lists compiled by the employment exchanges contain only the unemployed workers who have reported to t h e exchanges. So long as it pays the unemployed to report, and so long as t h e y — 91 — have some prospect of obtaining work through the exchanges, the lists may be presumed to be more or less complete 1 . But the depression entirely disorganised both the labour market and the statistics of the employment exchanges. Unemployment benefit was reduced; relief was refused altogether to certain groups; the unemployed gradually lost all hope of finding jobs through the exchanges, and they came to feel that they must rely on their own efforts and trust to casual jobs to keep them alive. They therefore deserted the employment exchanges, and the number of unemployed persons who were not registered with the exchanges increased rapidly. TABLE XXIII.—NUMBER OF WAGE EARNERS IN EMPLOYMENT IN GERMANY 1 (In thousands) Month January . . February . March April May June July ...... August . . . September October . . . November . December . 1929 15,849 15,473 16,669 18,061 18,490 18,638 18,539 18,538 18,427 18,232 17,714 16,535 1930 16,159 15,934 16,293 16,794 17,120 17,033 16,843 16,687 16,540 16,230 15,693 14,617 1931 13,970 13,765 14,092 14,813 15,197 15,253 15,020 14,618 14,370 13,978 13,433 12,440 1932 12,085 11,928 11,974 12,535 12,744 12,779 12,756 12,755 12,834 12,915 12,699 11,983 1933 11,487 11,533 12,193 12,698 13,180 13,307 13,436 13,716 13,921 14,062 14,020 13,287 1934 13,518 13,967 14,687 15,322 15,560 15,530 15,533 15,599 15,621 15,636 15,476 14,772 i From the statistics of the sickness insurance funds. The only way to determine the real extent of unemployment in Germany is to apply the British method of calculation to the German statistics of the labour market. In British statistics the number of workers in employment is determined by calculating the difference between the total number of workers and the number of those who are unemployed or incapacitated by sickness. For Germany, it is possible to determine the number of unem1 Complete agreement between the statistics of unemployment and those of workers in employment is possible only when, as in Great Britain, the two sets are compiled by the same organisations, which can thus from the outset adapt the various figures so as to ensure correspondence. Failing this, divergences are bound to arise, if only through differences in the interpretation of such terms as " wage earners ", " wage earners in employment ", " unemployed persons ", etc. — 92 — ployed persons by deducting from the total number of workers the number of those who were in employment or absent through sickness x. Table XXIV gives the results of this calculation for each quarter of the years under consideration. TABLE XXIV.—VISIBLE AND INVISIBLE UNEMPLOYMENT IN GERMANY (In thousands) Distributed as Hollows Dates Number of wage earners Absent through sickness or child birth Unemployed In employment Not Registered registered with with exchanges exchanges 1929 January April . . July . . . October 20,993 21,026 21,060 21,093 ,197 778 779 802 15,849 18,061 18,539 18,232 2,933 1,712 1,251 1,557 ,014 475 491 502 1930 January April . . July . . . October 21,127 21,119 21,111 21,103 887 718 718 654 16,159 16,794 16,843 16,230 3,218 2,787 2,765 3,252 863 820 785 967 1931 January April . . July . . . October 21,094 21,061 21,028 20,995 ,007 674 631 609 13,970 14,813 15,020 13,978 4,887 4,358 3,990 4,623 1,230 1,216 1,387 1,785 1932 January April . . July . . . October 20,963 20,930 20,897 20,864 712 586 522 540 12,085 12,535 12,756 12,915 6,042 5,739 5,392 5,109 2,124 2,070 2,227 2,298 But there is another form of invisible unemployment—short time. The German official statistics of employment and unemployment give no indications on this point, but they are supplemented in this respect by the trade union statistics, which show every month the percentage of trade union members in employment who are working short time. By applying this same percentage to the total number of wage earners in employment, 1 Cf. Wl. WOYTINSKY : " Der deutsche Arbeitsmarkt in der Krise " in Schmollers Jahrbuch, 1933, pp. 415 et seq. — 93 — one can divide the latter into "full-time employed" and "workers on short time". In order to determine the total decrease in employment during the depression, all that remains to be done is to convert the short time figures into the corresponding number of full-time employed and unemployed wage-earners1, and to place the number of completely unemployed persons thus obtained over against the total number of wage earners. The results of these calculations are given in table XXV. (Cf. diagram XXII, p. 94.) TABLE XXV. THE DECREASE IN EMPLOYMENT IN GERMANY DURING THE DEPRESSION Dates Unemployed Conversion of The workers in to complete employment included short-time unemployment (in thousands) gives (in thousands) Full-time Short-time Full-time workers workers workers Completely unemployed 1929 January April . . July . . . October 3,947 2,187 1,742 2,059 14,232 16,724 17,223 16,865 1,617 1,337 1,316 1,367 15,445 17,713 18,171 17,863 4,351 2,535 2,110 2,428 1930 January April . . July . . . October 4,081 3,607 3,550 4,219 13,961 14,308 13,980 13,065 2,198 2,486 2,863 3,165 15,588 16,098 16,303 15,387 4,652 4,303 4,090 5,064 1931 January April . . July . . . October 6,117 5,574 5,377 6,408 9,961 10,873 10,905 9,198 4,009 3,940 4,115 4,780 12,767 13,749 13,868 12,640 7,320 6,638 6,529 7,746 1932 January April . . July . . . October 8,166 7,809 7,619 7,407 7,263 7,684 7,641 7,871 4,822 4,851 5,115 5,055 10,632 11,080 11,252 11,579 9,619 9,264 9,123 8.754 1 The method employed in this operation is the following : the trade union statistics of unemployment classify short-time workers according to their hours of work : less than 24 hours a week ; from 24 to 32 hours ; from 32 to 40 hours; from 40 to 48 hours a week. From this it is possible to calculate the average decrease in working hours per short-time worker (as a percentage of the normal 48-hour week: 6 hours = 12.5; 9.6 hours = 20, etc.). If the average decrease in working hours per short-time worker is 20 per cent., it may be concluded t h a t 100 short-time workers = 80 full-time workers + 20 completely unemployed persons. — 94 — DIAGRAM -THE XXII. THE Millions 20 ia GERMAN u 16; 14 Tfmrt MARKET DURING DEPRESSION Millions ftfíjffjfi^fo;;£ IÍiffö^«^^ 'iï&iwy~;ï& — r á LABOUR / v illlllllllMllllllllií illiu 1 v i J xlUP \ ^s Hut 1 1929 10 8 6 ^ I. II. III. IV. V. 16 14 "" s. 10 Id pplll 12 J ® 12 20 4 I93O Fall-time workers. Workers on short time. Unemployed registered with exchanges. Invisible unemployment. Temporarily incapacitated. I n this way, the total visible and invisible unemployment in the summer of 1932 is found to have been about 9 million completely unemployed persons (as against 2.1 millions in J u l y 1929). If the average for the year 1929 be taken as a basis (17,300,000 = 100), the index of employment during the depression will be found to have fluctuated as follows : Jan. Feb. Mar. April May June 1929. 1930. 1931. 1932. 90 91 74 62 87 89 73 61 94 91 75 61 103 93 80 64 105 95 82 65 106 94 82 66 July Aug. Sept. Oct. Nov. Dec.^ v e r *?^ 105 94 80 65 105 92 77 65 103 89 73 67 105 91 75 66 100 79 64 62 93 79 64 62 for y6i*r 100 90 76 64 As unemployment spread, the scope of the benefit paid by the State became, as was mentioned, more and more restricted. I n August 1930, 40.2 per cent, of the unemployed were in receipt of benefit at the full statutory rate ("Alu"), 11.8 per cent, were in receipt of reduced emergency relief ("Kru"), and 14.4 per cent. had to depend on welfare offices ("We"). I n August 1932, the proportion in receipt of ("Alu") benefits was only 9.2 per cent.; — 95 — ("Kru") relief, 17.1 per cent.; assistance from welfare offices, 26.7 per cent. There was 47 per cent, of the unemployed who received no official relief of any kind. DIAGEAM X X I I I . CLASSIFICATION OF UNEMPLOYED IN PERSONS GERMANY A : Ahi. B : Kru. C : We. D : Unemployed persona registered with exchanges but not in receipt ol relief. E : Unemployed persons not registered with exchanges. The index numbers of employment given above include not only wage earners and employees in industrial undertakings, but also those in commerce and transport, domestic service and the liberal professions. But in Germany as in other countries the real seat of the depression was in mines and manufactures, whereas in other branches the decline in employment had been but slight. It will therefore be clear that the index of employment in industry would fall even lower than that for the economic system as a whole. Therein lies the explanation of the divergence between the index of employment and the index of production. During the depression the decline in German industrial production was more marked than the fall in the index number of employment as calculated above. But the divergence between the two is not really so very great; if the average for 1929 = 100, the index numbers will be as follows : Employment (E) Industrial production (V) 1929 100 100 1930 90.3 89.7 1931 75.8 73.3 1932 64.3 61.0 Individual output f 100 ^) 100 99 97 95 / — 96 — DIAGRAM XXIV.—THE DEPRESSION IN GERMANY : EMPLOYMENT AND INDUSTRIAL PRODUCTION (1929 = 100) 100 V 90 80 90 > ^ 60 ^ 70 70 ^ 60 ^ 6o fio 1929 1950 50 1952 19&I • • • • " » • • » Industrial production. • Employment. • • « • « * • • • • • Individual output. German statistics offer yet another means of determining the decrease in employment during the depression : the figures of the Reports on Industry (Industrieberichterstattung), which are published monthly and indicate, for a series of typical undertakings and branches of industry, the number of workers actually employed as a percentage of the employment capacity of the undertaking or branch, and the number of hours actually worked as a percentage of the possible hours 1 . This method shows an even greater decline in employment than was estimated above. The hours actually worked, as a percentage of the possible hours, are g i v e n as being : Percentage of employment Index number (1929 = 100) 1929 1930 Annual average 1931 1932 67.4 100 56.2 83.4 44.5 68.1 35.7 53.0 It should be noted that these statistics of the utilisation of plant are no substitute for direct statistics of the employment of labour. But they would appear to be really instructive for the period under consideration here, for one is entitled to assume 1 " Employment capacity " is used here in a technical and not in an economic sense; it means the normal number of posts in the works. The number of " possible hours " is obtained by multiplying the normal number of posts by 48 (the normal working week). — 97 — that the capacity of the undertakings remained unchanged over this period. But one reservation must be made : these figures are based on selected specimen undertakings, and such a method is very unreliable in a period of flux. For the purpose of the present study, therefore, the index of employment calculated above will be preferred. The apparent decline in individual output (from 100 to 95) is due in this case not so much to the disorganisation of production as to the fact that in a period of depression commercial and transport undertakings generally do not reduce their staffs to the same extent as industrial concerns. The conclusion to be drawn from diagram XXIV is that the decrease in employment in Germany was due to the decline in industrial production. Obviously this conclusion does not explain the phenomenon, for it merely substitutes one unknown (the cause of the economic depression) for another (the cause of the widespread unemployment). But it seems nevertheless to be an interesting conclusion. The collapse of the German labour market occurred in quite a different way from the corresponding phenomenon in Great Britain and the United States. The way was not paved for it by a steady increase in structural or technological unemployment. Its causes were of a purely economic character 1. 1 The author has shown elsewhere (Die 40-Stunden Woche, Berlin, 1931, p. 218) t h a t the army of unemployed persons in Germany in the spring of 1931 could be classified as follows : Wage earners in search of work, registered with the employment exchanges Unemployment concealed by short time 5,000,000 1,000,000 Total... 6,000,000 This total could be subdivided : Normal unemployment, about Seasonal „ Structural „ Cyclical „ 400,000 800,000 800,000 4,000,000 No account has been taken here of invisible unemployment. calculation for the summer of 1932 would give these figures : The same Unemployed persons registered with exchanges... 5,700,000 „ not registered with exchanges. 2,100,000 Persons working short time (converted to complete unemployment) 1,500,000 Total... 9,300,000 This total could be subdivided : Normal unemployment, about Structural „ Cyclical „ 400,000 800,000 8,100,000 In both cases the estimate for structural unemployment is a maximum; no account has been taken of the absorption of this unemployment as a result of the decrease of the population. — 98 — About the middle of 1932 there was a revival of activity, and the number of unemployed persons began to fall steadily, while the number of wage earners in employment increased1. According to the official statistics, the situation developed as follows (in thousands) : June 1932 „ 1933 „ 1934 Increase or decrease from June 1932 to June 1934 Wage earners In employment Wage earners registered with employment exchanges 12,779 13,307 15,530 5,476 4,857 2,481 +2,751 —2,995 In two years, the unemployment figure had fallen by practically 3 millions, but the number of wage earners in employment had risen by only 2 3/4 millions. The difference between these two figures proves either that invisible unemployment had increased by a quarter of a million during this period or that the total number of workers had decreased. It is true that the increase in the number of wage earners in employment is due in part to a change in the method of compilation. Since Germany began to feel the depression, the normal opportunities of employment in public or private undertakings have been supplemented by various forms of subsidiary employment offered to unemployed persons by the State or the local authorities in lieu of unemployment benefits. During 1929 and 1930, this subsidiary employment was not of very great importance : a few tens of thousands of unemployed persons were helped in this way during the summer months. As the work in question was in the building trade and did not differ from other work in that industry except that it was subsidised by the State, the workers employed on it were reckoned as wage earners in employment. In 1932, the voluntary labour service for young persons was introduced, but those who enrolled for such service were not counted as wage earners in employment. The total number of unemployed persons engaged in subsidiary employment in autumn 1932 was 440,000, of whom only 70,000 were shown in the monthly statistics of the insurance funds as being employed wage earners. In 1933 the labour service was greatly extended and was supplemented by a new form of subsidiary employment, hundreds of thousands 1 Cf. tables X I X , p. 80, and X X I I I , p . 91. — 99 — of unemployed persons being sent to the villages to help in agricultural work. According t o the rule adhered to earlier, all these people should have been considered as unemployed. B u t this practice was abandoned in J u l y 1933. The following figures (in thousands) show how the consequent error can be eliminated 1 : Wage earners In regular employment June 1929. 1930. 1931. 1932. 1933. 1934. 18,810 17,100 15,180 12,730 13,100 15,010 Officially Wage earners counted aa in subsidiary wage earners employment in employment 100 18,910 40 17,140 15,250 130 180 12,780 13,310 530 15,530 800 The number of wage earners in regular normal employment increased, it will be seen, by 2,280,000 (or 17 per cent.) and not by 2,751,000 during the last two years o f t h a t period. The proportion in which staffs were increased varied greatly from one branch of the economic system to another, being highest in industry. According to the Industrieberichterstattung, the number of workers in employment, expressed as a percentage of the employment capacity of the undertakings, developed as follows : June 1932 „ 1933 „ 1934 Workers Salaried employees 41.8 46.5 59.6 61.2 60.3 68.6 These figures indicate an increase of 40 per cent, in the number of workers in employment, and 12 per cent, in the number of salaried employees. The average hours of work also rose : June 1932 „ 1933 » 1034 6.94 hours per day 7.26 „ „ „ 7.46 „ „ „ The total amount of work performed must therefore have risen by more t h a n 40 per cent. The degree of activity in industrial undertakings is shown to have improved t o a still greater 1 Wochenbericht des Instituts für Konjunkturforschung, 1934, No. 36, p. 166. Berlin. — 100 — extent by the official index number of production, which rose as follows (average for 1928 = 100) : August 1932 (lowest point) June-July 1933 June-July 1934 58.5 70.0 89.1 In other words, production would appear to have increased by 52 per cent. The discrepancy between these two percentages can hardly be attributed to an increase in the rate of individual output. It is due rather to certain inaccuracies in the figures used. The German economic system is undergoing a transformation, and the various branches are developing very unevenly during the transition period. Under such conditions, great caution must be exercised in using index numbers that are based more or less on estimates or on a certain number of random samples. France France is one of the few countries in which the population has scarcely increased at all for several decades. The number of inhabitants was as follows : 1880 1890 1900 1910 37.7 millions 38.3 39.0 39.6 Increase from 1880 to 1910 . . . + 1.9 millions This insignificant increase of 0.17 per cent, annually on the average was due in part to immigration from abroad. After the war the working population in France declined (in thousands) 1 : 1921 1926 Men Women Total 13,114.5 13,556.3 8,606.1 7,837.8 21,720.6 21,394.1 1 The decline in the number of women in the occupied population is a reaction after the war years. Many women who had not formerly held jobs were drawn into occupational activity during the war. After the war, some of these women returned to domestic tasks. — 101 — The number of foreigners who immigrated into France or who left the country again during this period was as follows (in thousands) : Foreign immigrants Foreign emigrants 62.5 80.1 50.3 193.1 60.0 273.5 265.4 47.8 176.3 54.4 162.9 48.7 1921 1922 1923 1924 .1925 1926 1921-1926 1,151.3 323.7 For five years, from the middle of 1921 to the middle of 1926, the excess of immigrants over emigrants was more than 700,000. Most of these went to swell the occupied population, being mainly workers in mines or in the building trades and especially agricultural workers from Italy and Poland. But for this influx of foreign workers, the occupied population of France would have fallen by about two million during these five years. The number of persons employed in agriculture fell during the same period, in actual figures and relatively, whereas the number in industrial occupations (including mines) rose rapidly. On the eve of the depression, the population of France differed from those of the United States and Great Britain in that it was in course of rapid industrialisation. TABLE XXVI. CLASSIFICATION OF THE OCCUPIED IN FRANCE BY ECONOMIC BRANCHES Year Total Agrioccupied population culture Mines Industry POPULATION Commerce Other and occutransport pations In thousands 1921 1926 21,721 21,394 9,024 8,199 318 434 6,181 6,681 3,592 3,641 2,606 2,439 16.6 17.1 12.0 11.4 In percentages 1921 1926 100.0 100.0 41.5 38.3 1.5 2.0 28.4 31.2 — 102 — This was the effect of the rapid expansion of French industry. The official index number of production in manufactures and mines moved as follows (1921 = 100) : Mines Manufacturing industries 1921 1922 1923 1924 1925 1926 100 114 136 160 178 194 100 143 160 198 195 228 Individual output made great strides during the same period. On the basis of 1921 = 100, the index of output per head of occupied workers in 1926 was : In mines In manufacturing industries 145 211 I t is true t h a t this increase is due in the main to the reconstruction of the economic system, which had been disorganised by the war. B u t the return t o normal conditions was accompanied by the rationalisation and bringing up t o date of industrial undertakings. This is clearly brought out by the fact t h a t the total volume of industrial production, including mining, in France in 1921 (within the present frontiers) was 45 per cent, below the 1913 level, whereas in 1926 it was 26 per cent, above t h a t level. But it must not be forgotten t h a t in 1926 France was passing through a feverish period of inflation. I n the following year there was a reaction (depression accompanying stabilisation), changing to a fresh upward trend in 1928. I n 1929 and into 1930, the country experienced great prosperity. I n the second half of 1930 there was a slackening of activity in some branches of industry, but on the whole the economic situation was still very satisfactory, so t h a t France was an oasis of prosperity in the desert of world depression. I t was not until 1931 t h a t it began to lose this favoured position. On the basis of 1913 = 100, the volume of industrial production (including mines) developed as follows from 1924 onwards : 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 109 108 126 110 127 140 140 124 96 107 99 I n its years of prosperity, the French economic system had, as was mentioned above, to obtain the assistance of foreign labour. When the situation became worse, some of these foreign workers were sent out of the country—a method of protecting the national labour market t h a t had proved very effective in dealing with the — 103 — depression incidental to stabilisation in 1927. The number of foreigners arriving in France in search of work since t h a t time or leaving the country has varied as follows (thousands) : 1927 1928 1929 1930 1931 1932 1933 Immigrants Emigrants 64.3 97.7 179.3 221.6 102.3 69.4 74.6 90.0 53.8 38.9 43.8 92.9 108.5 49.0 The fact of using foreign labour during years of prosperity gives France considerable advantages in times of depression : as its reserve army of workers is kept outside t h e country, it can draw on it t o the extent of its requirements at a n y given moment, without being responsible for the maintenance of the reserves it does not require—an appreciable relief in periods of depression. This accounts for the relatively small extent of unemploymentin France. French unemployment statistics show : (a) the number of unemployed persons in receipt of relief; (b) the number of applicants for employment for whom t h e communal employment exchanges could find no job. B u t as France has no unemployment insurance and does not make it compulsory to register with the exchanges, both sets of figures may contain serious lacunae. The relief provided by t h e communes and departments is governed by the Act of 19 April 1918 and its several amendments. The system is highly decentralised : the communes may, out of their own resources, supplement t h e prescribed rates of relief. They follow t h e principles laid down by the central authorities as regards the categories of persons entitled to relief, b u t they do not always interpret these principles uniformly. As unemployed workers must register with the employment exchanges in order t o qualify for relief, the number of unplaced applicants for employment is always higher t h a n the number of those in receipt of relief. But the relatively small difference between the two figures proves t h a t unemployed persons who are not entitled t o relief rarely register with the exchanges in the hope of finding a job. I t also happens quite frequently t h a t unemployed persons who could claim relief under the Act do not apply to the exchanges because they feel t h a t a certain social humiliation attaches to t h e acceptance of relief. — 104 — For all these reasons, the official statistics can indicate the general trend of the development of unemployment in France, but not its exact extent. TABLE XXVII. Year Jan. Feb. Mar. April UNEMPLOYMENT I N FRANCE May June July A. Unplaced applicants (in thousands) 1926 1927 1928 1929 1930 1931 1932 1933 1934 . . . 13 .. . 36 . . . 32 .. . 12 .. . 13 .. . 37 .. . 255 . . . 353 . . . 370 12 93 30 13 14 56 321 370 383 11 89 25 11 13 68 352 350 379 10 73 20 10 12 70 346 345 369 9 56 15 9 12 61 328 308 352 9 37 10 9 10 54 305 282 345 Aug. Sept. Oct. Nov. Dec. 9 29 9 8 11 53 301 264 358 12 26 11 11 15 68 290 261 382 15 27 12 12 18 107 292 287 417 21 28 10 11 21 165 303 345 455 1 8 30 8 9 10 50 295 270 350 B. Unemployed in receipt of relief2 (in thousands) 10 28 10 9 12 57 298 252 357 • 1926 . . . 0.5 0.8 0.6 0.4 0.4 0.4 0.4 0.4 0.3 0.4 0.5 13 1927 .. . 56 81 75 58 40 24 17 15 13 8.6 10 13 1928 . . .. 18 15 10 7.2 3.7 1.7 1.1 1.0 0.6 0.5 0.5 0.9 1929 .. . 1.7 3.5 1.1 0.7 0.6 0.4 0.4 0.4 0.4 0.4 0.6 0.8 1930 .. . 1.5 1.7 1.6 1.2 0.9 1.0 0.9 1.0 1.0 1.7 4.9 12 1931 . . .. 29 41 51 50 41 36 36 38 39 56 92 147 1932 .. .. 241 293 303 282 262 232 262 263 260 248 255 277 1933 . ... 316 332 314 310 277 252 240 234 227 233 258 313 1934 .... 336 349 346 335 318 311 320 326 323 348 375 419 i Applicants for whom no employment was found within a week. figures include unemployed persons in receipt of relief. 2 Excluding those receiving relief from welfare offices. From February 1927 onwards, these The curve of unemployment in France x shows two unequal waves in the period 1926-1934 : the first, and shorter, one during the stabilisation depression of 1927, and the second, the longer one, beginning in 1931, rising rapidly in 1932, falling in the second half of 1933 and rising once more in 1934. In comparison with the United States, Great Britain and Germany, France would appear to have suffered only moderately 1 Cf. diagram XXV, p. 105. - - 105 DIAGRAM XXV.—UNEMPLOYMENT IN FBANCE 29 30 SI 52 45 54 Unplaced applicants for employment. Unemployed In receipt of relief. from unemployment. During the depression caused by stabilisation there were never more than 100,000 unplaced applicants for employment, and during recent years the figure only once (in winter 1934-1935) exceeded 500,000. When one thinks of the fluctuations in industrial production during the same period, these figures seem extraordinarily low. It was found above that the index of production for 1927 (average over the year) was 13 per cent, lower than the index for the preceding year. This fall in production should have meant a decline of from 500,000 to 600,000 in the number of wage earners employed in industrial undertakings, and the expulsion of some tens of thousands of foreign workers from France could scarcely make good this decrease. Notwithstanding this situation, the average number of registered applications for employment over the year was barely 35,000. The figures for the last depression are still more surprising. The fall in the index of production from 140 on the average for the years 1929-1930 to 96 on the average for 1932 (1913 = 100) — 106 — should have been accompanied by a reduction of about 1.7 millions in the number of wage earners in employment. The decline in employment in non-industrial occupations should have further increased the number of those in search of work. It is of course true that the growth of unemployment was restricted by two facts : (1) the decline in immigration from 1931 onwards and the expulsion of foreign workers ; (2) the arrival at working age of those born during the war years, when the birth rate was low. These two factors possibly reduced the number of applicants for employment by some 400,000 or 500,000. Even allowing for that correction, however, the divergence between the official unemployment statistics and the increase one would normally expect is too great. An explanation of this discrepancy may be sought for in other statistics, more especially the statistics of employment compiled by the factory inspection service since 1931. They cover all establishments employing 100 persons or over in mining, manufacturing industry, commerce and transport. At the beginning of the depression these establishments employed about 2.5 million wage earners, of whom almost 90 per cent, were in minesand manufacturing industries. The statistics show, monthly : (1) the ratio of the number of wage earners in the various occupational groups to the corresponding number for the same month a year previously; (2) the percentage distribution of those in employment according to their hours of work. The disadvantage of the employment figures of the factory inspection service is that the basis of comparison changes from month to month. For example, the number of wage earners employed in June 1932 was 86.6 per cent, of the number in employment in June 1931, and the number in September 1932 was 87.7 per cent, of the figure for September 1931. But that does not show whether employment increased or decreased from June to September 1932. To know that, one has to refer to the figures for the previous year. In this particular case, the figure for June 1931 was 93.5, and that for September 91.1. But 86.6 per cent, of 93.5 = 81.0, and 87.7 per cent, of 91.1 = 79.9. Thus employment fell from June to September 1932 as compared with 1930, but it is still impossible to state whether the employment figure in September 1930 was higher or lower than in June 1930. It is thus rather difficult to make use of these figures, but they can be simplified by reducing them all to a single basis of — 107 — comparison1. The level of employment for the corresponding month in 1980 has been selected as a base. TABLE XXVIII.—INDEX NUMBER OF WAGE EARNERS EMPLOYED IN LARGE ESTABLISHMENTS IN FRANCE (Number of wage earners for the corresponding month in 1930 = 100) Beginning of month of : January . February , March April May June July August . . , September October . . , November . December , 1931 1932 1933 1934 96.3 95.3 94.6 94.1 94.0 93.5 92.8 92.2 91.1 90.5 88.9 87.4 83.0 81.3 80.9 80.1 80.8 81.0 82.0 80.5 79.9 78.6 80.3 80.5 78.3 77.9 78.3 78.7 78.9 79.1 80.9 80.1 79.5 78.5 79.9 79.5 77.7 77.3 77.0 76.7 76.8 77.1 78.4 77.0 76.2 75.1 76.1 75.0 The bases of comparison for the figures in table XXVIII are the various months of 1930. But unfortunately this is not a sufficiently sound basis, for there was a slight depression in French industry in the second half of that year, and industrial activity declined during that period as compared with the first half of the year. If the average for 1930 = 100, the index numbers of production throughout the year will be : January February March April 102.6 102.6 102.6 102.6 May June July August 102.6 101.8 100.4 99.0 September .. November... December . . . 97.6 96.9 96.2 95.5 If it be assumed that employment in large establishments varied in more or less the same proportions as the volume of . industrial production and that it was therefore about 7 per cent. lower at the end of the year than in spring, the index number of employment can be corrected as follows. 1 All that has to be done is to multiply by each other the original figures for each month in successive years. — 108 — TABLE XXIX.—CORRECTED INDEX NUMBER OF WAGE EARNERS EMPLOYED IN LARGE ESTABLISHMENTS IN FRANCE (Average for 1030 = 100) Beginning of month oí : January . . February . March April May June July August . . . September October . . . November . December . 1930 1 1931 1932 1933 1934 102.6 102.6 102.6 102.6 102.6 101.8 100.4 99.0 97.6 96.9 96.2 95.5 98.9 97.8 97.2 96.6 96.5 95.3 93.2 91.3 89.0 88.3 85.6 83.5 85 83 83 82 82 82 82 79 78.0 75.8 77.3 77.9 80.3 79.9 80.3 80.7 80.9 80.6 81.3 79.3 77.6 76.1 76.9 76.0 79.8 79.3 79.0 78.7 78.8 78.6 78.8 76.2 74.3 72.7 73.1 71.5 i Movement oí industrial production. Diagram XXVI shows the fluctuations in the two index numbers of employment calculated above (tables XXVIII and XXIX) together with the variations in industrial production over the same period. DIAGRAM X X V I . — I N D E X NUMBERS OF WAGE EARNERS EMPLOYED IN LARGE ESTABLISHMENTS AND OF THE VOLUME OF INDUSTRIAL PRODUCTION IN FRANCE (Average for 1930 = 100) 100 ^ v 90 \ 90 ""••^W t>'0) 80 - \ ao y ^ ^*C¿ V 70 in 70 i i i i i i i i i i i i i i i i M i l l i i i i i i t i i i t 1 1 1 1 bo Original index number ) of wage earners employed Corrected index number j i n l a rge establishment Industrial production. — 109 — The discrepancy between the corrected and the original index numbers is very slight, but the former does give a clearer idea of the change that took place in the second half of 1932. The difference between the two employment curves on the one hand and the production curve on the other is very significant. During 1931 and the first half of 1932 production dropped to a far greater extent, proportionally, than the number of wage earners employed; the increase in production in the second half of 1932 and the first half of 1933, on the other hand, did not bring with it any appreciable improvement in employment. This divergence cannot be explained away by a possible error arising out of the difference in the methods of calculation employed or the fact that the field of observation was not exactly the same in the two cases. This is another example of the phenomenon already noted in the case of Germany and the United States : the spread-over of employment. DIAGRAM XXVII.—CLASSIFICATION OF PERSONS EMPLOYED IN LARGE ESTABLISHMENTS IN FRANCE BY HOURS OF WORK % IOO -\oo ¡ffl"~ BO Jfl III I I l I I i l 60 - ÔO 60 60 AO 40 20 20 I U 111 IV I95O I I [ 48 hours or over. II III IV I I95I IH II III I V I I952 \ 40 to 48 hours. • II III IV I 1953 34 Less than 40 hours. It is interesting, in this connection, to note the information collected by the factory inspection service concerning the extent of short time (Cf. table XXX, p. 110, and diagram XXVII above). — 110 — TABLE X X X . — S H O R T TIME IN LARGE ESTABLISHMENTS IN Percentage of wage earners working Index number of average hours of work i Quarter FRANCE Index ot a m o u n t of work 1 performed m 48 h. = 100 Average for 1930 = 100 Average for 1930 = 100 0.2 0.2 0.5 0.5 99.6 99.7 99.6 99.4 100.0 100.1 100.0 99.8 102.6 102.4 99.0 96.0 4.3 6.1 7.3 11.7 96.7 95.7 95.2 93.5 97.1 96.1 95.6 93.9 95.0 92.3 87.2 80.6 28.9 26.9 27.7 27.3 25.8 25.3 21.8 15.4 89.4 89.8 90.7 92.6 89.8 90.2 91.1 93.0 75.4 . 74.4 72.9 71.8 59.3 61.0 63.6 63.6 28.1 26.0 23.6 24.5 12.6 13.0 12.8 11.9 93.4 93.5 93.8 94.0 93.8 93.9 94.2 94.4 75.2 75.8 74.8 72.2 60.1 58.3 55.1 53.5 31.2 28.2 29.4 30.6 8.7 13.5 15.5 15.9 93.8 92.5 91.7 91.5 94.2 93.2 92.1 91.9 74.8 73.3 70.4 66.5 48 h. or over 40 to 48 h. I II Ill IV 96.5 97.0 96.6 94.5 3.3 2.8 2.9 5.0 1931 I II Ill IV 75.7 68.8 66.7 58.8 20.0 25.1 26.0 29.5 1932 I II Ill IV 45.3 47.8 50.5 57.3 1933 I II Ill IV 1934 I II Ill IV 1930 i Calculated by the author. As a result of the extension of short time, the average hours of work in the first quarter of 1932 was 10 per cent, below the average for 1930. As the number of wage earners in employment at that time had fallen by about 16 per cent., the total decrease in the amount of work performed may be estimated at 25 per cent. But in the second half of 1932, with the first signs of a revival in economic activity, hours of work became longer again, so that in the spring of 1933 the undertakings were able to increase —Ill — production without adding to their reduced staffs. The curves in diagram XXVIII show the simultaneous fluctuations in the numbers of persons employed and in hours of work. DIAGRAM XXVIII.—EMPLOYMENT IN LARGE ESTABLISHMENTS IN FRANCE : INDEX NUMBERS OF PERSONS EMPLOYED, HOURS OF WORK AND AMOUNT OF WORK PERFORMED (Average for 1930 = 100) JOO 100 9° -. 60 00 70 II III 1950 . IV I 0 III IV 1951 i II III 1932 IV 11 m 1955 1» 1 70 54- Persons employed. Hours of work. Amount ol work performed. The index number of the amount of work performed agrees on the whole with the index of production. The only notable divergence is in 1932, when the decrease in the amount of work as compared with 1930 was 27 per cent, (average for the year), whereas industrial production had fallen by 31 per cent. This difference is due in part to the fact that the figures compiled by the factory inspectorate include commerce and transport, in which market fluctuations are less acutely felt than in industry. The average fall in the amount of work performed over the year 1932 was only about 10 per cent, in commerce and transport as against 29 per cent, in industry. In the later phase of the depression, industry employed more workers, while commercial and transport undertakings continued to reduce staff. The gap between production and employment figures therefore narrowed, so that the curves of the two index numbers meet in the first — 112 ^quarter of 1934 at a point 25 per cent, below the average level for 1930. It may be concluded that the hourly output per person did not vary during the period under consideration. DIAGRAM XXIX.—INDUSTRIAL PRODUCTION IN FRANCE : INDEX NUMBERS OF PRODUCTION AND OF THE AMOUNT OF WORK PERFORMED IN LARGE ESTABLISHMENTS (Average for 1930 = 100) 100 •100 90 90 ÔO ÔO *-**. 70 60 II III I95O IV I II III 1951 IV I II III IV 1952 70 I II III I95Ô IV 6O 1 SA . . . . . . . . . . . . . Amount of work performed. •———«—• Industrial production. From what has been said above, two conclusions can be drawn concerning the origin of unemployment in France : (1) Neither during recent years nor earlier was there any real overcrowding on the French labour market as a result of the influx of new members to the occupied population. (2) Machinery did not supplant human labour. During the most prosperous periods, when the output per head was rising extremely rapidly, the demand for labour exceeded the supply; industrial expansion forced the country to depend to an increasing extent on foreign labour. Unemployment on a large scale developed only at a time when technical progress was at a standstill. — 113 — Unemployment in France is therefore due to one single cause : the slackening of production as a result of the decline in sales. One point remains to be considered : the absolute extent of unemployment in France. The results of the 1926 census may be taken as a basis for estimating this. That census showed that there were in France 12 1 /, million workers and salaried employees in all, distributed over various groups as follows 1 (in thousands) : Men 1,712 3,916 1,483 167 128 949 Women 669 1,392 514 194 667 216 Total 2,381 5,308 1,997 361 795 1,165 Total . . . 8,355 169 3,652 74 12,007 243 Grand t o t a l . . . 8,524 3,726 12,250 Agriculture, fishing, forestry Mining and manufacturing industry. Commerce and transport Liberal professions Personal and domestic service Public services Unemployed In 1926, when the official statistics showed about 10,000 unplaced applicants for employment, there were 240,000 unemployed wage earners in France. This is probably the normal level of " invisible " unemployment in that country. On the eve of the depression the number of wage earners was practically the same as in 1926 : 5.3 to 5.4 millions in industry and about 2 millions in commerce and transport. These were the two groups of wage earners that suffered most severely from the decline in employment. The fall of over 31 per cent, in production should, as has just been seen, have led to the dismissal of 1.7 million workers, but the extension of short time proved a substitute for some of this reduction of staff 2. The reduction in the average hours of work, saved some 500,000 workers from dismissal. In commerce and transport, on the other hand, there was scarcely any recourse to short time, but the decline in employment was less marked than in industrial occupations : staffs were cut down by about 10 per cent, only, representing the dismissal 1 Annuaire statistique 1933, Paris, 1934, p. 138. * There were a t t h a t time in France between 3.5 and 4 million workers on short time, as against about 3 million full-time wage earners. — 114 — of some 200,000 wage earners. To this must be added the dismissals from other branches of the economic system, but they were not very numerous. But for the various factors that restricted its extent, unemployment would probably have reached the figure of 2.2 millions in 1932, distributed as follows : Normal unemployment before the depression Decrease in employment in industry ,, „ ,, in commerce and transport ,, „ ,, in other occupations 240,000 1,700,000 200,000 60,000 Total... 2,200,000 The reduction in hours of work enabled 500,000 of these workers to be kept on; the total number of wage earners fell at the same time by about 400,000 or 500,000. There would therefore be 1.3 million wage earners actually unemployed. In 1932, then, " visible " unemployment represented only a fraction of the whole drop in employment. During the following phase of the depression the situation changed somewhat, and the comparative number of officially registered unemployed persons increased. Italy Before the war Italy was in the forefront of the countries that suffered from a surplus population and were obliged to send their sons abroad. The number of emigrants (in thousands) was : 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 252 447 512 416 239 399 403 263 405 560 1904 , .^13 . 3,896 These figures, it must be remembered, include the seasonal workers who went every summer to work abroad, returning home for the winter. But there were also large numbers of emigrants who left Italy for good, thus relieving the labour market of the country to a very appreciable extent. During the war emigration ceased, but this did not cause an excessive supply of labour, since more than 5 million persons had been mobilised. Subsequently, Italy's war losses—500,000 — 115 — killed and a million wounded—counterbalanced the stoppage of emigration, so that when peace came there was no overcrowding on the labour markets. Demobilisation in Italy was accomplished as easily as in the other belligerent countries, but the economic system of the country required some little time to adapt itself to the new postwar conditions. The severe slump of 1921-1922 was followed by a boom period that continued until 1926. Progress slackened off before the approach of the depression, which, however, was not felt in Italy until some little time later than in Germany and the United States. The growth of unemployment has been a feature of Italian economic development since the war. TABLE XXXI.—UNEMPLOYMENT IN ITALY : UNPLACED APPLICANTS FOR EMPLOYMENT 1 (In thousands) Year I II III IV V 270 262 235 202 115 1920 .... — — — 250 — 1021 607 576 499 432 410 1922 392 328 281 270 244 1923 281 259 219 177 156 1924 156 157 143 127 101 1925 156 126 109 98 98 1926 225 259 228 215 216 1927 439 413 412 357 307 1928 462 489 293 258 228 1929 466 457 385 372 367 1930 1931 723 765 707 670 635 1,051 1,148 1,053 1,000 968 1932 1,225 1,229 1,082 1,026 1,000 1933 1934 1,158 1,104 1,057 996 941 VI VII VIII IX 106 389 372 216 131 86 83 215 247 193 322 574 905 884 831 88 435 304 183 118 80 80 263 234 202 342 638 931 824 887 93 471 318 179 119 72 83 292 248 217 376 693 946 889 867 116 473 313 181 116 83 89 306 269 229 395 748 949 907 887 X XI XII 101 107 102 492 512 542 354 382 321 225 259 200 117 136 150 112 122 86 149 181 113 332 376 414 282 321 364 297 333 409 534 642 446 800 878 982 956 1,039 1,130 963 1,066 1,132 905 970 962 1 The figures in table XXXI are taken from the official estimates, compiled from various sources. The methods of calculation have varied many times during the period under consideration, and the figures for different years are therefore not on a uniform basis. When the powers of the communal employment exchanges were limited in 1922 and 1923, the figure compiled by them fell, and the number of unemployed persons for the years 1924-1926 is therefore underestimated as compared with the preceding years. Again, the figures for recent years do not include all the unemployed persons in the country. Indeed it may be admitted that, for a variety of reasons, a certain number of permanently vmemployed persons have not registered with the employment exchanges during this period. Subject to these reservations, there can be no doubt that the Italian statistics reflect faithfully the general course of employment in the country. In the curve in diagram XXX—based, for the sake of simplicity, on the quarterly and not on the monthly figures in — 116 — 1 table XXXI —it is possible to distinguish three successive phases in the evolution of the Italian labour market since 1921. DIAGRAM XXX.—UNEMPLOYMENT IN ITALY : UNPLACED APPLICANTS FOB EMPLOYMENT In thousands 1200 MOO In thousands 1000 900 n. I pno/a 800 OOO 1 m. aoo PHAIC moa t I 700 700 600 600 l„ SOO soo 400 4O0 500 200 100 0 I v soo è 200 100 w l >-!• I920 21 _i_u_ J-UL • > 1 111 -111 23 2» 24 25 26 t i i 27 .l_U- 1 1 1 1 1 . 1 1 it» 2« 2? SO 91 1 1 1 II 1 mi 52 34 3» o 1st phase (1922-1926) : Economic recovery; unemployment falls. 2nd phase (1926-1930) : Expansion stops; unemployment increases slowly. 3rd phase (1930-1934) : Depression; widespread unemployment. Before going on to study these three phases more closely, it should be noted that, quite apart from the inexactitudes resulting from changes in methods of calculation or compilation, the Italian statistics do not cover all the unemployed workers. The data concerning unemployment in agriculture are particularly defective. It is generally difficult to register such unemployment accurately in statistics, and the task is made still more difficult in Italy because there are hundreds of thousands of small peasant proprietors who also work as wage earners on the estates of large landowners. The sources of error in the official unemployment statistics may be reduced by excluding agriculture and separating the unemployed in industry from those in other occupations. 1 Taking the first month of each quarter. — 117 — TABLE XXXII.—UNEMPLOYED IN ITALY BY OCCUPATIONAL GROUPS l (In thousands) Total Year unemployed persons 2 Agriculture Industry Other occupations^ 92.8 121.6 62.3 50.1 20.6 24.4 75.6 80.3 89.5 103.5 165.8 201.9 213.4 187.0 241.3 279.4 175.6 110.1 62.1 64.9 172.3 209.2 177.4 267.3 479.2 679.4 702.6 669.5 41.9 54.3 47.4 38.0 27.6 24.5 30.6 34.9 33.9 54.6 89.5 125.1 103.0 107.2 376.0 455.3 285.3 198.2 110.3 113.9 278.5 324.4 300.8 425.4 734.5 1,006.4 1,019.0 963.7 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 * Annuario Statistico Italiano, 1927-1933. Sindacato e Corporazione, Vols. LVIII and LXIII. 2 For the years 1921, 1922, 1923 and 192a, the figures represent the average of the two months In which unemployment was at its highest and its lowest in each of these years. These figures are generally slightly higher than the average for the years in question. a Public services, commercial establishments and private transport undertakings. The development of unemployment must now be compared with the growth of the population during the same period. The population of Italy increased from 38.7 millions in 1921 to 41.2 millions in 1931—an increase of 6.5 per cent. The- rate of growth is about the same as before the war. The population of working age increased more rapidly than the totaj population. The number of persons between the ages of 15 and 65 years in Italy rose as follows (in thousands) : Men Women Total... 1921 10,904 11,559 1931 11,772 12,708 Increase +8.0% + 9.9% 22,463 24,480 +9.0% One might expect the occupied population to increase more rapidly than the total population but less rapidly than the middle age groups 1. On this point the results of the Italian 1 I t was to be expected t h a t the number of women in employment would rise more slowly or even t h a t the number would fall, since the proportion of women in employment in 1921 must be considered unduly high (a relic of the war years). — 118 — census of 1931 are surprising. The number of persons found to be engaged in occupational activity was (in thousands) : Men Women Total. 1921. 13,154 5,276 13,369 3,903 Increase or decrease + 1.8% — 26 % 18,431 17,272 - 1931 6.3% Ci. table XXXIII. The results of the two censuses as regards agriculture are not comparable, as they obviously rest on different interpretations of the term " occupation ". But the development of the nonagricultural population of Italy during this period is sufficiently clearly brought out by the census figures. The number of persons occupied in industry (including mines) rose by 14.6 per cent., and the number of those in commerce and transport by 16.9 per cent. TABLE XXXIII.—DISTRIBUTION OF THE OCCUPIED POPULATION IN ITALY * (In thousands) Year Total Mines Agriculture, and hunting, manufacfishing tures Public services, Transport, liberal commerce, professions, finance etc. 1921 Total... 13,154 5,276 7,147 3,117 3,309 1,250 1,647 250 1,051 659 18,431 10,264 4,560 1,897 1,710 13,369 3,903 6,630 1,539 3,972 1,252 1,897 320 870 792 17,272 8,169 5,224 2,217 1,662 1931 Total... 1 Annuario Statistico Italiano, 1927, p. 27, and 1934, p. 13. The natural increase in the occupied population exceeded the power of absorption of the urban labour markets, and a considerable fraction of the new members of the occupied population had to seek work abroad. — 119 — TABLE X X X I V . — I T A L I A N EMIGRATION AND IMMIGRATION (In thousands) Emigration Year 1921.. 1922.. 1923.. 1924.. 1925.. 1926.. 1927.. 19281. 1929.. 1930.. 1931.. 1932.. 1933.. 1934.. Intercon- Continentinental tal migration migration 117 126 185 125 102 122 136 71 62 59 41 25 22 26 84 156 205 239 178 141 92 80 88 221 125 59 61 42 Immigration Total 201 282 390 364 280 263 228 151 150 280 166 83 83 68 Intercon- Continentinental tal migration migration 94 55 40 65 67 71 73 50 44 47 43 34 26 21 30 56 79 107 122 106 67 49 65 82 64 39 40 29 Total Excess of emigration 124 111 119 172 189 J 77 140 99 109 129 107 73 66 50 77 171 271 192 91 86 88 52 41 151 59 10 17 18 i From 1928 onwards the figures include only persons emigrating in search of employment abroad. During the ten years from 1921 to 1930 the excess of emigration from Italy was 1,220,000 persons—a figure that is higher than the whole increase, from one census to another, in the number of persons engaged in urban occupations. It is true that up to 1927 the emigration figures include persons not engaged in any occupational activity, but even when allowance is made for that fact the excess of emigration over immigration may be estimated at about a million of the occupied population. But for the safety valve of emigration and for the crisis, the increase in the number of persons trying to earn a living in urban occupations would probably have been not l 1 ^ but 21/i millions (30 per cent.). The development of the occupied population in Italy (apart from agriculture) may therefore be considered as the resultant of two forces acting in opposite directions : (1) The influx of new members to the occupied population as a result of the natural increase in population, together with the drift of workers from the country to the towns in search of work : altogether 2 */« million persons in ten years; — 120 — (2) Emigration, which removed a million persons during the same period. The first of these factors was more or less steady ( + 230,000 persons annually); the second fluctuated from year to year. The proportion of the occupied population (excluding agriculture) belonging to industrial occupations was practically the same at the two census dates : 56 per cent, in 1921 and 57 per cent, in 1931 ; it may therefore be presumed that it did not vary much in the interval. The development of the occupied industrial population of Italy may therefore be estimated as follows : TABLE XXXV.—ESTIMATED OCCUPIED POPULATION OF ITALY (In thousands) Year 1921................::.::. 1923 1925 1926 1927 1928 1929 . : 1930 1931 ¿ Total occupied population, excluding agriculture Occupied population in mines and manufactures 8,020 8,190 8,270 8,250 8,310 8,470 8,630 8,790 8,990 9,200 9,300 4,490 4,590 4,640 4,650 4,700 4,800 4,900 5,000 5,120 5,240 5,310 iCf. diagram XXXI, p. 121. But allowance must be made for the fact that these figures include, in addition to wage earners, employers and independent craftsmen, who are of no account from the point of view of the labour market. It would therefore be necessary to know the proportion of wage earners in the total occupied industrial population of Italy in 1921 and succeeding years. The 1931 census puts the proportion of workers (including apprentices) in the occupied industrial population at 75.6 per cent. This ratio seems higher than would have been expected from the statistics of other countries. It may possibly be a consequence of the depression. In any case, the figure must be accepted. 121 — DIAGRAM XXXI.—GROWTH OF THE OCCUPIED POPULATION IN ITALY (EXCLUDING AGRICULTURE) Increase as compared with 1921 1921 22 25 T Emigration. i 24 ^_ • Hi o. a. Natural increase. 25 26 27 Increase in mines and manufactures. 26 _-. I I—I ¿9 SO 31 Increase in non-industrial occupations. &. 5. Remained in the country. If that ratio is applied to the figures in table XXXV it becomes possible to assess the number of dependent persons available for employment in Italian industry each year. By deducting from these figures the number of unemployed and temporarily incapacitated persons 1 in industrial occupations, one is left with the number of those actually employed in industry 2. It is obvious that employment figures calculated in this way cannot claim to be completely accurate. They are probably rather high. More exact figures could be obtained by making allowance for the inevitable gaps in the unemployment estimates and increasing the official statistics in question by a certain 1 In accordance w i t h t h e British method, t h e number of persons temporarily incapacitated b y sickness is taken as being 4 per cent, of t h e number of wage earners w h o are not recorded as being unemployed. 2 Cf. table X X X V I , p . 122. 9 — 122 — percentage. In table XXXVI it is assumed that the official statistics register about 80 per cent, of the unemployed persons in industrial occupations. TABLE XXXVI.—ESTIMATED NUMBEB OF WAGE EARNERS IN EMPLOYMENT IN ITALIAN INDUSTRY (In thousands) Year Total n u m b e r of wage earners (S) 3,394 3,470 3,508 3,515 3,553 3,629 3,704 3,780 3,871 3,961 4,014 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 Unemployed persons (Ch) 304 352 222 139 78 82 217 255 223 336 607 Temporarily incapacitated b y sickness 108 109 115 118 122 124 122 123 128 127 119 Wage earners in employment (E) 2,982 3,009 3,171 3,258 3,353 3,423 3,365 3,402 3,520 3,498 3,288 The margin of error in these calculations can be considerably reduced by using index numbers in place of the absolute figures. These index numbers can then be compared with the index of production so as to determine, by the usual formula, the development of individual output. y tear 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 Number of wage earners in industry i (S) 100 102.3 103.3 103.4 104.4 106.7 108.9 111.3 114.0 116.7 118.5 Number of Volume wage earners of in employment2 production a (E) (V) 90.8 91.5 96.5 99.2 102.0 104.5 102.4 103.5 107.1 106.4 100.0 100 111 118 134 150 167 156 166 180 167 141 Individual output „„„_ , „ A V 100 T = 1 0 0 - 110.1 121.3 122.3 135.1 147.1 159.8 152.3 160.4 168.1 157.0 141.0 i2 Not including those absent through sickness. S = 100 for 1921. a Index number of the " Institut für Konjunkturforschung ", Berlin. It will thus be seen that the number of wage earners in Italian industry and their output per head both rose rapidly and steadily. — 123 — But the economic expansion of Italy during these years must not be considered merely as post-war reconstruction. Italy had not suffered very severely economically during the war, and its situation at the beginning of the observation period was comparatively favourable 1 . As early as 1922, industrial production in Italy was above the 1913 figure, and in 1923 the output per head of the population was also higher than before the war. But this period of prosperity was of short duration. In 1927, as was mentioned above, the increase in unemployment foreshadowed the coming of the depression2. There was a slight improvement in 1929, but in the following year unemployment again spread rapidly. DIAGRAM XXXII.—DEVELOPMENT OF UNEMPLOYMENT IN ITALIAN INDUSTRY (S and V for 1920 = 100) 50 4.0 50 20 10 _ AS • 1921 1922 1923 1924 I925 1926 I927 1920 <929 199O ICSI One of the outstanding traits of Italian economic development after 1926 was the considerable fluctuations in the volume of industrial production, accompanied by parallel fluctuations in individual output (spreadover of employment, overtime to meet rushes of orders, short time in slack periods). During this time 1 In 1921, industrial production was 33 per cent, below the 1913 level in Germany, 32 per cent, in Great Britain and 46 per cent, in France, whereas in Italy it had fallen by only 3 per cent. » Cf. table X X X I , p. 115 and diagram X X X , p. 116. — 124 — dismissals of workmen on account of machinery were not parti cularly frequent, but industry, which had lost its power of expansion, was no longer able to absorb the additions to the occupied population. With regard to the evolution of the Italian labour market in more recent years, official statistics show that unemployment reached its peak in the spring of 1933, when about 1.2 million workers were unemployed. The average for 1932 was slightly lower—about 1 million as against 300,000 in 1929. The number of unemployed persons had increased by 500,000 in industrial occupations, by 110,000 in agriculture and by 90,000 in urban non industrial-occupations *. At first sight these figures appear very low, and it will be well to check them by reference to other statistical sources of information. Emigration from Italy practically ceased after 1930, and the influx of new persons in search of employment from that time onwards was determined solely by the natural growth of the population. During the depression the influence of the war years on population began to be felt : the decline of some 400,000 in the annual number of births during the war meant a decrease of almost 200,000 in the number of persons entering occupational life every year from 1931 to 1934. The stream of workers from the country to the towns also dried up during the same period. Perhaps there was even—partly under the pressure of administrative measures—a return movement of unemployed persons towards the villages. In any case, the number of wage earners in industrial occupations did not increase during the depression, but would seem rather to have declined. During this period some of those who were previously in employment must have been dismissed in consequence of the fall in production. Their number may be determined with the aid of the index numbers of employment, of which two exist in Italy : the first, compiled by the Ministry of Corporations, covers some 6,500 large establishments that employed about a million wage earners in 1926; the second, compiled by the General Fascist Confederation of Industry, has the advantage of including all industrial undertakings. 1 Cf. table X X X I I , p. 117. — 125 — On the basis of 1929 = 100, the number of wage earners in employment in Italian industry varied as follows 1 : Average lor year Index oí Ministry of Corporations Index of Confederation of Industry 100 93.0 81.2 70.5 71.4 100 97.3 88.6 78.5 79.4 1929 1930 1931 1932 1933 The difference between the two series proves that large-scale industry felt the influence of market fluctuations more keenly than did industry as a whole. In trying to estimate the actual number of unemployed persons, therefore, it is preferable to use as a basis the index of the Confederation of Industry, The data published by the Confederation of Industry also show to what extent the total amount of work performed decreased as a result of reductions of staff, the abolition of overtime and the extension of short time. The number of hours worked in industrial undertakings is shown by the following index : 1929 1930 1931 1932 1933 100 94.2 83.8 72.4 73 .9 This corresponds quite closely to the fall in the volume of production. Two commonly used index numbers of industrial production in Italy are given below : 1929 1930 1931 1932 1933 Ministry of Corporations i Institut für Konjunkturforschung, Berlin 8 100 92 78 67 74 100 92 83 74 78 1 Thirty-two series of figures, including 9 for the textile industry, 11 for the heavy metal industry, 6 for mechanical engineering, etc. s Forty-six series of figures, including 15 for mines, 6 for food and drink trades, 8 for the heavy metal industry, 5 for the textile industry, 14 for the chemical industry, etc. (Cf. Vierteljahrshefte zur KoniunMurforschung, special number 31.) The figures of the Confederation of Industry for the hours actually worked fall, as will be noticed, between the two index numbers of production; this confirms their accuracy. Moreover, the fact that the number of those employed in industrial under1 Bollettino Mensile di Statistica dell' Istituto Centrale di Statistica del Regno d'Italia. — 126 — takings fell by 21.6 per cent, from 1929 to 1932 means that nearly 800,000 wage earners must have been dismissed. This figure is only 300,000 higher than the increase in the number of unemployed persons in industrial occupations as given by the official unemployment statistics \ The divergence can be explained in part by the fact that a certain number of unemployed persons have left industrial occupations altogether. Therefore, while there is no doubt that there is " invisible " unemployment in Italy, not registered by the official statistics, its comparative extent would not appear to have increased appreciably during the depression. In view of what has been said, there is no need to dwell at length on the causes of Italian unemployment in recent years. Widespread unemployment developed in spite of the almost complete cessation of technical progress as reflected in the output per head of those employed in factories, and in spite of the relatively small afflux of workers on the market as a result of the natural increase in population. The source of the evil is to be found in the slower rate of industrial expansion first of all (19261929) and subsequently in the decline in production (1929-1933). Gzechoslova kia The spread of unemployment in Czechoslovakia since the war is shown by the curve in diagram XXXIII, which covers the period 1921-1934 and is based on the number of unplaced applicants for employment at the end of each quarter 2. The increase in the number of unemployed persons in the second half of 1922 was due to the decline in economic activity consequent upon the stabilisation of the currency. Although Czechoslovakia did not officially stabilise its currency until October 1929, it put a stop to inflation as early as the autumn of 1922 and considerably raised the value of the crown at that time 3. 1 Cf. table X X X I I , p. 117. This is the only series of figures covering the whole period in question and giving a complete general picture (Cf. Statistische Übersicht der Tschechoslowakischen Republik, 1930, pp. 221-222, Prague). 3 The dollar rate in Prague was (parity = 100) : 2 Mar. 1922 1,171 June 1922 1,053 Sept. 1922 627 Dec. 1922 653 Mar. 1923 682 The new parity was fixed at 684. June 1923 677 Sept. 1923 677 Dec. 1923 692 — 127 — DIAGRAM X X X I I I . — U N E M P L O Y M E N T UNPLACED APPLICANTS THE E N D I N CZECHOSLOVAKIA : FOE EMPLOYMENT OF EACH AT QUARTER In thousands OOO I n thousands \i * 600 700 é>0O soo SOO 400 500 iOO 200 200 I0O too 0 19a 1 2 2 2 5 2 4 2 5 2 6 27 28 2? »O 31 52 5 5 »A- 0 The depression in Czechoslovakia in 1922-1923 was therefore one of stabilisation, s u c h a s inevitably came to every country t h a t practised inflation for a more or less lengthy period. The very low unemployment figure from summer 1924 to summer 1930 shows t h a t the country had been able t o restore its economic equilibrium. Notwithstanding the exceptionally difficult conditions with which the country had t o contend in its economic reconstruction, it did not suffer from chronic unemployment on a large scale until the beginning of the world economic depression. Table X X X V I I shows the development of the occupied population in Czechoslovakia during the period under consideration. I n just under ten years (February 1921 to December 1930) the occupied population increased by more t h a n 10 per cent., b u t the increase was limited to urban occupations : the number of persons in those occupations rose from 3,554,000 to 4,249,000, or b y about 20 per cent., whereas the number of persons occupied in agriculture fell by 7l/t per cent. — 128 — TABLE XXXVII. DISTRIBUTION OF THE OCCUPIED POPULATION OF CZECHOSLOVAKIA BY OCCUPATIONAL GROUPS 1 (In thousands) Occupational Groups 1921 1930 Increase ( + ) or decrease (—) Occupied population Total... 1,809 2,169 585 800 1,673 2,502 805 942 — + + + 136 333 220 142 5,363 5,922 + 559 Wage, earners 2 Total... 974 1,866 416 396 786 2,189 580 691 — + + + 188 323 164 295 3,652 4,246 + 594 i Manuel statistique de la République tchécoslovaque, IV. Prague, 1932, p. 16; Annuaire statistique de la République tchécoslovaque. Prague, 1934, p. 15. 3 Workers, salaried employees, apprentices and domestic servants. During this period the development of the country was marked by increasing industrialisation and " prolétarisation ". The number of wage earners in the towns rose in ten years from 2,678,000 to 3,460,000—an increase of 29 per cent. Emigration from Czechoslovakia has been slight since the war. The excess of emigrants over immigrants for the decade 19211930 was 200,000 at most, and this number was more or less evenly distributed over the whole period. In estimating the afflux of new members to the occupied population, therefore, it will be assumed that there was no appreciable variation from year to year. This hypothesis is admissible both for all wage earners outside agriculture and for workers and salaried employees in industry, — 129 — taken alone, but it cannot be assumed that the afflux was the same after 1930. The year 1931 saw the first of the generations from the war years entering the labour market, and the decline in the birth rate from 1915 to 1919 substantially relieved the labour market from 1931 to 1935. But it is very improbable that the number of wage earners in Czechoslovakia fell to any great extent during these years. The distribution of the population by age groups in the 1930 census suggests rather that the number of wage earners did not change appreciably one way or the other. In the light of these considerations, the number of occupied persons in dependent positions (wage earners in the wide sense) in Czechoslovakia from 1921 to 1934 may be estimated as follows (in thousands) : Average for 1921 . 1922 . 1923 . 1924 1925 1926 1927 1928 1929 . 1930 . 1931, unchanged until 1934 AH occupations except agriculture 2,710 2,790 2,870 2,950 3,030 3,110 3,190 3,270 3,350 3,420 3,450 Industry only 1,880 1,915 1,950 1,985 2,020 2,055 2,090 2,125 2,160 2,180 2,190 No reasonably complete statistics of employment were compiled in Czechoslovakia until 1928. The figures given here for earlier years are based in part on estimates. A comparison of these data with those for unplaced applicants for employment gives the following results (in thousands) : Average for : 1926 1927 1928 1929 1930 1931 1932 1933 1934 ... ... Wage earners in employment (statistics of sickness funds) Unemployed (unplaced applicants for employment) Total (2,612) (2,619) 2,488 2,506 2,446 2,313 2,069 1,889 1,878 68 53 39 42 105 291 554 738 677 (2,680) (2,672) 2,527 2,548 2,551 2,604 2,623 2,625 2,555 — 130 — DIAGRAM X X X I V . — N U M B E R OF WAGE EARNERS AND EMPLOYMENT IN CZECHOSLOVAKIA In millions. In millions. The discrepancy between the sum of the number of wage earners in employment and those unemployed and the figures given above for the number of dependent workers may be attributed to a variety of causes : (a) I n the census, certain persons are counted as wage earners who do not fall within this category for social insurance purposes. (b) Some groups of wage earners (higher employees, migrant workers, etc.) are not counted in the statistics of employment. (c) Wage earners who are sick or momentarily out of employment while changing jobs are not counted either as being in employment or as being unemployed. (d) A certain number of agricultural workers have been included among wage earners in employment or among the unemployed. An obstacle to the exact study of developments on the Czechoslovak labour market is the fact t h a t its statistics of employment — 131 — and unemployment make no distinction between industrial and other occupations. According to the writer's calculations (p. 129), the index number of wage earners in industrial occupations was as follows : S = 1921 1922 1923 1924 1925 1926 1927 1928 1929 100 101.9 103.7 105.6 107.4 109.3 111.2 113.0 114.9 In order to determine the proportion of unemployed persons, recourse must be had to the official unemployment statistics, although they cover every economic activity and the figures must therefore be considered rather too low for industry taken alone : 1921 Ch = 1922 2.1% 3.9% 1923 1924 1925 1926 1927 1928 1929 6.2% 2.9% 1.5% 2.1% 1.6% 1.1% 1.2% By combining these two series one obtains the index of the number of wage earners in employment, taking the number in 1921 as a base (1921 = 100). 1921 E = 97.9 1924 1925 1926 1927 1928 1929 102.6 105.8 107.0 109.4 111.8 113.5 The combination of this index with the index of industrial production shows the development of individual output 1 : 100 T V 100 V E 1921 1924 1925 1926 192? 100 107 102.1 104.3 1928 1929 113 112 106.8 104.7 130 145 149 118.8 129.7 131.3 According to these calculations, the number of wage earners in industrial occupations rose by 15 per cent, in eight years. During the same period there was an improvement in individual output of 29 per cent. But the influx of new workers and the march of technical progress were counterbalanced by the expansion of industry, so that when the depression caused by stabilisation had been overcome there were only slight fluctuations in unemployment and no marked tendency for it to increase. 1 Cf. appendix, p. 165. — 132 — DIAGRAM X X X V . — D E V E L O P M E N T OF UNEMPLOYMENT IN CZECHOSLOVAK INDUSTRY (S and V in 1921 = 100) 1921 1922 192a l?24 l?25 1926 1927 l?26 1929 After the world depression began (it was only in 1930 that the situation became serious in Czechoslovakia) the unemployment curve rose rapidly, as can be seen from table XXXVIII K TABLE X X X V I I I . — T H E DEPRESSION IN CZECHOSLOVAKIA Average for year 1929 1930 1931 1933 1934 Index of industrial production (1929=100) 100 89.2 80.6 63.5 60.0 66.5 Wage earners in employment Actual Index number number (thousands) (1929 = 100) 2,506 2,446 2,313 2,069 1,887 1,878 100 97.6 92.3 82.6 75.3 74.9 1 Unemployed unplaced (thousands) 42 105 291 554 738 677 1 From the Monthly Bulletin of Statistics of the League of Nations. The course of the production index would indicate that the depression was at its worst in Czechoslovakia in the spring of 1933, when the index had fallen to 58 (average for 1929 = 100). 1 Cf. diagram X X X I I I , p . 127. — 133 — Such a sharp fall was bound to involve the dismissal of staff, but the number of wage earners employed on the average during 1933 was not 40 per cent, but only 25 per cent, below the 1929 level. This comparatively mild reaction to the fall in production was due to two factors : (1) the spread-over of employment; (2) the less marked influence of cyclical fluctuations on nonindustrial occupations. In 1933, when industrial production had fallen to 60 per cent., industrial undertakings could not utilise much more than 60 per cent, of the amount of labour they required in 1929. But by spreading employment over a greater number of workers (and partly as a result of a slackening in technical progress) they were able to keep on a larger proportion of their former staffs : perhaps 66 per cent, or even more. Employment in industrial undertakings thus declined by a little over 30 per cent. But as the decrease in the amount of work performed in commerce and transport was much less marked, the average decline in employment shown in the official statistics seems to be quite credible. This does not mean that there was no invisible unemployment in Czechoslovakia; all that is claimed is that its volume was not very great. With regard to the origin of the widespread unemployment in Czechoslovakia, no trace has been found of overcrowding on the labour market as a result of an excessive influx of new workers or of too rapid technical progress. The disorganisation of the labour market in this case is of purely economic origin, being due to the decrease in industrial production. Belgium Before the war Belgium was already one of the most highly industrialised countries. According to the 1910 census, 46.1 per cent, of the occupied population were engaged in manufactures and mining, as compared with 32.2 per cent, in the United States, 41.3 per cent, in Germany, 45.6 per cent, in Great Britain, 30.1 per cent, in France 1 , 24.6 per cent, in Italy, 37.8 per cent, in Holland, 25.7 per cent, in Sweden, etc. 1 In 1906. — 134 — The census of 1920 showed a decline in the occupied population as compared with 1910. The figures were (in thousands) : End of 1910 End of 1920 x £*£&£>_) Number of inhabitants .... Occupied population : Agriculture Mines and manufactures. Other occupations 7,424 7,406 — 18 783.4 1,609.9 1,098.5 613.6 1,491.3 1,100.3 —169.8 — 118.6 + 1.8 Total occupied population 3,491.8 3,205.2 — 286.6 i Without the districts of Eupen and Malmedy. These figures are all the more surprising when it is remembered t h a t the population of working age increased appreciably during the same period. The number of persons between the ages of 15 and 60 was (in thousands) : Men Women Total... 1910 2,221 2,236 1920 2,370 2,427 4,457 4,797 • Increase +149 +191 + 340 This fact suggests t h a t the 1920 census figures are incomplete, as is quite probably the case in view of the disorganisation of economic life just after the war. F r o m 1920 t o 1930 the increase in the number of persons between the ages of 15 and 60 years has been estimated at 5 per cent. 1 , and t o this must be added the excess of immigration over emigration. From 1921 to 1930 there were 303,000 emigrants, as against 425,000 immigrants or returning migrants. The excess of immigration was therefore 122,000. On the whole, then, the middle age groups increased from 1920 to 1930 in roughly the same proportion as in the preceding decade. B u t the distribution of the increase over the various branches of the economic system is not known, as the results of the 1930 census have not yet been published. A t t h e end of 1930, however, a census of industry and commerce was made in Belgium, and the results are in p a r t compa- 1 Cf. A. L. BOWLEY : Estimates of the Working Population, etc. Geneva, 1926, p. 12. — 135 — rabie with those of the similar census in 1910 1. The occupied population (in thousands) was : 1910 Women Men Industry Commerce 1,319.4 237.5 1930 390.8 285.2 Men Women 1,592.4 321.0 345.8 243.6 According to these statistics, the number of persons in industrial occupations rose by 13 per cent, in 20 years, and the number of those engaged in commerce by 8 per cent. The number of wage earners among those engaged in industrial or commercial occupations in 1930 (in thousands) was : Industry Commerce Total1... Workers Salaried employees Total 1,480.8 63.8 184.9 119.0 1,665.7 182.8 1,545.4 304.9 1,850.3 i Including wage earners who could not be classified under industry or commerce. Of these persons the following numbers were partially or wholly unemployed at that date (in thousands) : Workers Salaried employees Completely unemployed . Onshorttime 147.5 228.8 7.1 3.5 Total 154.6 232.3 It is assumed in the following pages that the number of wage earners in industry increased uniformly from 1910 to 1930, at the rate of about 0.6 per cent, annually. With regard to the degree of employment for the available labour supply, continuous information is available in the Belgian statistics of unemployment and short time among workers insured against unemployment. But this insurance was introduced by stages after the war 2. It is therefore preferable to take, in place of the actual figures given in these statistics, the percentage of unemployment among insured persons. Up to 1923 no distinction was made between complete unemployment and short time; since 1924 the two groups have been kept separate. 1 2 Revue du Travail, Brussels, June 1934, pp. 719-771. In the years 1926-1928 the number of insured persons fluctuated round 600,000; from 1929 to 1930 it was about 650,000, in 1931, 726,000, in 1932, 850,000, in 1933, 980,000; in 1934 it approached the million. — 136 — The insured population comprised : Full-time workers Average for : 1921 1922 1923 78.4 93.5 97.3 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 96.7 94.4 95.8 94.3 95.6 95.7 88.5 72.2 60.3 65.8 63.8 DIAGRAM XXXVI.—UNEMPLOYMENT Completely unemployed On short time 21 6 6 5 2 7 1.0 1.5 1.5 1.8 0.9 1.3 3.6 10.9 19.0 17.0 19.0 IN BELGIUM : 2.3 4.1 2.7 3.9 3.5 3.0 7.9 16.9 20.7 17.2 17.2 PERCENTAGE OF INSURED WAGE EARNERS UNEMPLOYED (Quarterly figures) I921 22 2% 24 25 26 27 28 29 50 31 32 3S 34 (5) Completely unemployed. On short time. In 1921 Belgium passed through a severe depression. The number of unemployed persons, including those on short time, exceeded 30 per cent, of the total number of insured persons between March and May of that year. But in the second half of the year an improvement began, unemployment decreased rapidly and by 1922 the economic machine was again working to capacity. It was not until 1930 that another depression — 137 — visited the country. It led at first to an extension of short time, and towards the end of the year dismissals of staff became frequent. The relatively favourable situation on the Belgian labour market between those two depressions may be attributed to the increase in the volume of industrial production, the index number of which moved as follows (1913 = 100) 1 : Average for : 1921 1922 1923 1924 1925 1926 1927 64 84 96 105 103 118 129 Average for : 1928 1929 1930 1931 1932 1933 1934 137 138 123 114 95 98 93 The degree of disorganisation in Belgian industry in 1921 is revealed by the fact that the level of production was extremely low as compared with the number of persons employed : individual output had fallen to a point 15 or 20 per cent, below the pre-war level. In 1924 the output per head of those employed had returned to the 1913 figure. As hours of work had been reduced in the interval, this means that the output per head was 10 or 15 per cent, higher than in 1913. Since then the volume of production has risen from year to year. In 1929, Belgian industry was employing little more labour than in 1923, although the volume of production had increased by about 42 per cent. The average annual increase in individual output must therefore have been 6 per cent 2 . This rapid technical progress would have caused technological unemployment but for the simultaneous feverish expansion of production. During the depression, production fell by more than 30 per cent. But commerce, transport and the other occupational groups covered by unemployment insurance are less sensitive to cyclical fluctuations than is industry, and it is therefore normal for the average decline in employment, as shown by the unemployment insurance statistics, to be less marked than the fall in production. 1 8 Institut des sciences économiques, Louvain. A more careful calculation, taking into account the changes in hours of work, would give a slightly lower figure for the average annual rate of technical progress, but this would in no wise affect the explanations t h a t follow. 10 — 138 — TABLE XXXIX. Quarters THE DEPRESSION IN BELGIUM Volume of industrial production (1928 = 100) Percentage of insured persons Completely unemployed On short time 1930 I II Ill IV 97 89 80 81 2.8 2.0 3.0 6.6 4.5 6.1 8.5 12.5 1931 I II Ill IV 81 78 78 75 11.4 9.0 9.7 13.6 17.1 14.6 16.6 18.9 69 65 69 20.1 18.4 19.1 17.9 23.6 22.2 19.7 17.0 1933 I II Ill IV 69 70 66 69 21.1 16.3 13.7 18.7 19.8 17.3 16.2 15.3 1934 I II Ill IV 692 672 66 69 20.2 18.0 17.5 20.1 18.1 17.0 17.3 16.4 1932 I II Ill IV 1 2 i Trade dispute (coal mines). Trade dispute (wool). As was mentioned above, the proportion of insured persons who were completely unemployed in 1932 was 19 per cent., and the proportion on short time 20.7 per cent, (as against 1.3 and 3 per cent, in 1929). If allowance is made for a regular increase of 0.6 per cent, annually in the number of wage earners, it will be seen that there was a decrease of about 20 per cent, in the total amount of work performed. The only cause of the unemployment in 1932 was therefore the decline in production. About the middle of 1932 a slight increase in productionoccurred, and unemployment fell to a corresponding extent. But the improvement of the labour market was as brief as the revival of industrial activity; by the second half of 1933 both had slumped again. — 139 — Sweden, Norway and Denmark During the war there was great industrial activity in the Scandinavian countries. In order to cope with the orders from the belligerent countries, their industries (especially in Norway) had to be brought up to date and their plant extended. The improvement in the technique of production did not, however, prevent the supply of work from being ample for all. Even although emigration was at a standstill, the level of unemployment remained exceptionally low. TABLE XL.—INDEX NUMBERS OF INDUSTRIAL PRODUCTION IN THE SCANDINAVIAN COUNTRIES (v) 1 (1913 = 100) Year 1920 1921 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 Sweden Norway 83 80 93 100 109 112 115 119 129 137 131 115 108 112 137 102 72 90 100 108 117 104 106 118 131 132 103 121 124 131 Denmark 128 107 113 125 136 126 126 128 139 149 163 149 136 157 167 i For Sweden : until 1926, index of the " Institut für Konjunkturforschung, " Berlin (Die Industrieurtrtschaft, p. 65); after 1926, official index of production. For Norway : official index of production (Staiistisk Aarbok for Norge, 1934, p. 66). For Denmark : until 1927, index of the " Institut für Konjunkturforschung, " Berlin (Die Industrieunrtschaft, p. 64); after 1927, official index of production. In 1921—even in 1920 in Sweden—the depression came. Judging by the decrease in production and in the amount of work done, it would seem that this post-war slump in the Scandinavian countries was more serious than were the depressions caused by demobilisation and stabilisation in most of the ex-belligerent countries. It was not until 1923 that the depression began to disappear. By that time the process of reconstruction and adaptation seemed to be complete. But in 1926 there was a — 140 — fresh fall in production in Norway and Denmark; Sweden alone was destined t o enjoy more-or-less satisfactory economic conditions for a few years longer. The world depression reached the Scandinavian countries comparatively late : about the middle of 1930 in Sweden and Norway, and towards the end of t h a t year in Denmark. Table X L shows t h a t even during the most recent years of the depression industrial production in t h e Scandinavian countries remained above the 1913 level; they differ in this respect from Germany, Great Britain, the United States, Belgium, Czechoslovakia and most other industrial countries. I t must not be forgotten, however, t h a t while the machinery of production in the neutral countries escaped the destructive effects of the war the same is t r u e of their population, so t h a t the number of persons to be supported by the economic system of the country had considerably increased. I n Sweden and Norway the natural increase in population was all the more marked on account of the decline in emigration 1. I n the case of Denmark, the extension of the territory of the country must also be borne in mind 2. There was at the same time a change in the economic distribution of the population in the Scandinavian countries. Statistics on this point for the period 1920-1930 are available only for Norway. B u t for t h e period 1910-1920 the census results of the three countries are available, and the data they supply are comparable from country t o c o u n t r y 3 . I n the decade between the two censuses the occupied population increased by the following amounts : Sweden 401,000 = 17% Norway 125,000 = 13% Denmark 131,000 = 11% This increase was confined almost entirely to industry, commerce and transport; the number of persons in these occupational groups rose as follows : Commerce Industry and Total transport Sweden 244,000 = 43% 139,000 = 60% 383,000 = 48% Norway 64,000 = 26% 51,000 = 32% 115,000 = 28% Denmark . . . 71,000 = 24% 47,000 = 26% 118,000 = 25% 1 This explains, inter alia, the decrease in the surplus of women in the middle age groups. 2 Cf. table XLI, p. 141, note 2. 3 Cf. table XLII, p. 141. — 141 — TABLE XLI.—POPULATION OF WORKING AGE ( 1 5 - 6 0 IN THE SCANDINAVIAN COUNTRIES Sweden Year Men Denmark Norway Women Men YEARS) 1 Women Men 2 Women Actual numbers (in thousands) 1910/1911 1920/1921 1930/1931 1,598 1,741 1,940 1,513 1,693 1,891 606 726 817 Index numbers 1910/1911 1920/1921 1930/1931 100 109 120 100 112 125 686 781 866 744 922 1,063 809 986 1,124 (1910-1911 = 100) 100 120 135 100 114 12& 100 124 143 100 122 139 i2 Statistical Yearbook of the League of Nations, 1933-1934, pp. 30-34. The change in the territorial extent of the State after the war increased the population by about 13 per cent. TABLE IN XLII.—DISTRIBUTION THE OF THE OCCUPIED POPULATION SCANDINAVIAN COUNTRIES B Y OCCUPATIONAL GROUPS 1 (In thousands) Sweden Denmark Norway Occupational Groups Women Men 758 480 258 85 321 184 188 44 162 Men Women Men Women 53 62 403 231 110 66 120 40 139 41 225 35 130 72 168 1,588 612 660 285 845 386 808 656 251 153 336 58 62 405 302 69 66 261 110 61 174 53 103 261 118 77 216 1,827 774 298 958 404 1910/1911 Commerce and transport Other occupational Total... 1920/1921 Agriculture Commerce and transport Other occupational Total... 248 150 38 772 i Statistical Yearbook of the League of Nations, 1931-1932, p. 40; 1933-1934, pp. 41-43. — 142 — The number of persons engaged in agriculture and other occupations remained practically unchanged. The provisional results of the last Norwegian census are available for t h e decade 1920-1930. They show t h a t the occupied population of the country has increased from 1,070,000 to 1,167,000—i.e., b y 97,000 or 9 per cent. B u t there has been no change in t h e number of persons in industrial occupations; the increase has been in agriculture ( + 16,000), commerce and transport ( + 43,000) and other occupational groups ( + 38,000). The contrast with the previous decade is striking. Whether the phenomenon is purely local, affecting Norway only, or whether the same trend exists in Sweden and Denmark cannot be decided until the results of the latest census in these countries are available. B u t it would appear t h a t in Sweden and Norway the occupied population has not increased so rapidly since the war as it did from 1910 t o 1920, and t h a t the widespread unemployment t h a t exists has held up to some extent the stream of persons entering industrial occupations in search of employment. As the percentages given above for the increase in the number of persons in industrial occupations (Sweden : + 43 per cent. ; Norway : + 26 per cent. ; Denmark : + 24 per cent. ) are not the result of any cyclical change but rather the reflection of a struct u r a l transformation in the population, it may be concluded t h a t they hold good for the decade 1913-1923 as well as for the period 1911-1921. Since then, as has been seen, the industrial population of Norway has not changed. I n Sweden and Denmark, on the other hand, there would seem t o have been a slow b u t steady increase in t h e number of persons in industrial occupations. The index numbers of the development of the industrial population of the three countries m a y therefore be given as follows : 1913 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 Sweden Norway Denmark 100 143 144 145 146 147 148 149 150 151 152 153 100 126 126 126 126 126 126 126 126 126 126 126 100 124 125 126 127 128 129 130 131 132 133 134 — 143 — For 1923, the figures for all three countries are based on the census results 1 . The Norwegian figure for 1933 is also based on the census, while those for Sweden and Denmark are very approximate estimates (an increase of 7 or 8 per cent.). For the sake of simplicity, it has been assumed that the development between those dates was uniform. DIAGRAM X X X V I I . — I N D E X NUMBERS OF INDUSTRIAL PRODUCTION AND OF THE POPULATION I N INDUSTRIAL OCCUPATIONS I N SCANDINAVIAN (1913 = THE COUNTRIES 100) 1913 IÇ25 25 27 29 51 54 ^_^__^_^ vi ^ ^ ^ ^ ^ ^ ^ ^ ™ Industrial production. = = = = = = Number of persons occupied in industry. I. Sweden. II. Norway. III. Denmark. With the reservation that these figures are an indication of trends rather than accurate statistics, the index numbers given above may now be compared with the index numbers of production in the three countries. The differences in the development of the economic situation and in the population of the countries will then become quite apparent. The common feature for all three is the marked increase in the number of persons in industrial occupations. 1 Cf. table XLII, p. 141. — 144 — The estimated index numbers of the occupied population in industry given above may also serve as index numbers of the number of wage earners in industry. But there is always a certain divergence between these figures and the number of wage earners in employment. The number of those in employment has, in recent years, always remained below the number available. Even during the prosperous years 1928 and 1929 there was unemployment on a large scale in Norway and Sweden, and the situation of Sweden during the period under consideration here can be considered satisfactory only in so far as it was better than that of her neighbours. An index number of employment in industry can be calculated from the index number of wage earners in industry and the percentage of unemployment among trade union members 1 . By dividing the index of production by the index of employment in industry, one obtains the index number of individual output 2 . TABLE XLIII.—UNEMPLOYMENT COUNTRIES : PERCENTAGE IN THE SCANDINAVIAN OF TRADE UNION MEMBERS UNEMPLOYED * Average for year 1913 1920 ' 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 Sweden Norway Denmark 4.4 1.5 2.3 17.7 17.1 10.3 8.5 13.2 24.3 25.4 19.1 15.4 16.6 22.3 30.8 33.4 30.7 7.1 5.8 19.7 19.2 12.6 10.8 14.7 20.7 22.5 18.5 15.5 13.7 17.9 31.7 28.8 22.2 — 26.2 21.9 12.5 10.1 11.0 12.2 12.0 10.6 10.7 12.2 17.2 22.8 23.7 18.9 i For Sweden : Statistisk Arsbok for Sverige, 1934, p. 375.—For Norway : Statistisk Aarbork for Norge, 1934, p. 138.—For Denmark : Statistisk Aarbog, various years. 1 There is in Sweden an official index of employment in large and mediumsized industrial undertakings. Later on, this will be compared with the result of the author's calculations, as a check on the latter. Cf. table XLIV, p. 145. 2 Cf. table XLV, p . 146. — 145 — I t is true t h a t the trade union unemployment statistics do n o t embrace all industrial occupations equally, and the aggregate figures include a certain number of unemployed persons belonging to non-industrial occupations. B u t the inclusion of a few thousands of unemployed workers from commerce, transport or public services is not a sufficiently great error t o have any real effect on the general results of these statistics. They may therefore be used t o determine the index of employment in industry. But it would be a mistake to base any conclusions concerning employment in all urban occupations or among the wage earners of the country as a whole on the trade union unemployment statistics. TABLE X L I V . — I N D E X NUMBERS OF WAGE EARNERS EMPLOYMENT ( E ) IN INDUSTRY IN THE COUNTRIES IN SCANDINAVIAN 1 ( N u m b e r of w a g e e a r n e r s in 1923 = 100) Year 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 Sweden Norway Denmark 87.5 90.3 90.2 89.8 90.3 92.4 93.4 93.1 87.5 82.1 81.8 89.7 91.5 86.8 75.7 74.6 80.9 84.6 83.4 77.7 69.2 66.6 87.4 89.9 86.9 81.3 80.1 84.9 88.8 91.4 87.7 73.5 77.3 i The table is drawn up on the basis of the usual formula. The number of wage earners (S) In 1923 being taken as = 100, the number of wage earners In employment (E) for the year in question was obtained by subtracting Ch from S. The value of Ch as a percentage is given in table XLIII, p. 144. Before going further, it will be well to compare the index number of employment calculated in this way for Sweden with t h e official index of employment in large and medium-sized industrial undertakings in t h a t country. The index numbers are (1929 = 100) : I n d e x calculated b y author Official i n d e x 1926 1927 1928 1929 1930 1931 1932 96 93 97 94 99 98 100 100 100 100 94 91 88 86 1933 88 85 — 146 — It will be seen that the two sets of figures follow an identical course, but the official index appears to be more sensitive to market fluctuations both on the up grade and on the down grade. This is exactly what might be expected, since it covers only large and medium-sized undertakings, whereas the author's index covers the whole of industry. This comparison of the two series of figures confirms the accuracy of the index numbers given in table XLIV. The employment figures in the first line of table XLIV (for 1923)—87.5, 89.7 and 87.4—may seem rather high as compared with those for 1932 or 1933. But when they are compared with the pre-war period or with the situation in other countries before the depression began, they reveal a profound disturbance in the economic balance of the Scandinavian countries. In the " fat " years that followed, Sweden was the only one that showed any appreciable increase in the number of persons employed in industry; in Norway, and to some extent also in Denmark, the number of wage earners in employment continued to fall in spite of the improved economic situation. Table XLV shows the development of individual output (per head of those employed in industry) in the three countries. A specially noteworthy feature is the rapidity of technical progress in Norway, which naturally had an influence on the power of absorption of the labour market. TABLE XLV. INDEX NUMBERS OF INDIVIDUAL OUTPUT IN INDUSTRY IN THE SCANDINAVIAN COUNTRIES 1 Sweden Norway Denmark Year 100 T 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 114.3 120.7 124.1 128.1 131.8 139.6 146.7 140.7 131.4 131.5 136.9 1923 = 100 100 105.6 108.6 112.1 115.3 122.1 128.3 123.1 115.0 115.0 119.8 100 T 111.5 118.0 134.8 137.4 142.1 145.9 154.8 158.3 132.6 174.9 186.2 1923 = 100 100 105.8 120.9 123.2 127.4 130.8 138.8 142.0 118.9 156.9 167.0 100 T 114.4 121.0 116.0 124.0 127.8 131.0 134.2 142.7 135.9 148.3 162.5 1923 = 100 100 105.8 101.4 108.4 111.7 114.5 117.3 124.7 118.8 129.6 142.0 i T = —. For the values of V and E, cf. tables XL, p. 139 and XLIV, p. 145. — 147 — The combined action of the three factors in unemployment represented in the usual manner in diagram XXXVIII. DIAGRAM X X X V I I I . — D E V E L O P M E N T OF UNEMPLOYMENT I N INDUSTRY I N THE SCANDINAVIAN COUNTRIES (S a n d V for 1923 = 100) Sweden l?2S 24 SU 26 27 28 29 SO 31 S2 S3 th(i9i») 10 t?23 24 2-3 26 27 23 29 *0 31 32 33 — 148 — I n all three countries, the opportunities of employment were inadequate even long before the war. In the case of Sweden the decisive factor was the increase in the number of persons in search of employment from 1910 to 1920 as a result of a permanent structural change 1 . I n Norway it was the excessive increase in individual output t h a t was one of the main causes of unemployment 2 . I n other words, the overcrowding on the Swedish labour market was due t o demographic causes, whereas in Norway unemployment was of a technological character. I n Denmark, on the other hand, there was no excessive influx of new workers to the occupied population, nor was there any exaggerated technical progress. B u t the expansion of industry 3 was not sufficient to provide employment for all the available labour. The resulting permanent unemployment may be described as economic in the narrow sense. P o l a n d and H u n g a r y Poland and Hungary are definitely agricultural countries, in which the great majority of the population works on the land. Industry is not highly developed, and small undertakings predominate. During the years 1920-1921, the population of these countries was distributed as follows : Poland (1921) In thousands Per cent. Hungary (1920) In thousands Per cent. Agriculture Mines and manufactures Other occupations . . 10,269.9 75.9 2,126.7 58.2 1,263.3 1,990.0 9.4 14.7 719.8 807.3 19.7 22.1 Total... 13,523.2 100.0 3,653.8 100.0 I n countries where agriculture is so preponderant, the urban labour market is bound up with the rural one : labour flows from the country to the towns or back to the country according to the economic situation. 1 8 Cf. table X L I I , p. 141 and diagram X X X V I I I , p. 147. AT Cf. on diagram X X X V I I I the columns E — ——> which rise very high above the line Ch (1923). 1 + A1 3 Cf. the black downwards-pointing columns on the diagram. — 149 — The official statistics of unemployment in Hungary and Poland, which show the number of unplaced applicants for employment registered with the employment exchanges, cannot give an accurate idea of the real extent of unemployment! The number of officially registered unemployed workers in these countries is very small. The average for the year was as follows (in t h o u s a n d s ) 1 : Poland Hungary 1926 1927 1928 1929 1930 1931 1932 1933 1934 190 13 164 14 126 15 129 15 227 44 300 52 256 66 250 61 342 52 For the year 1933, therefore, the number of unemployed would appear t o have increased as a result of the depression by only 120,000 in Poland and 46,000 in H u n g a r y as compared with 1929. B u t the situation appears in quite a different light when studied with the aid of the employment statistics. For Poland these statistics cover large and medium-sized industrial establishments and the public services. I n Hungary the employment statistics are based on the reports of the compulsory insurance scheme, which covers more than a million workers. The index numbers of employment for the two countries, reckoned as an average, for each year, varied as follows (1929 = 100) 2 : Poland Hungary 1927 1928 1929 1930 1931 1932 1933 1934 89.5 98.3 99.7 101.3 100.0 100.0 86.8 94.3 73.9 89.2 63.3 82.0 62.9 81.1 68.0 86.9 According t o these figures, the number of wage earners in employment decreased during the depression by about 37 per cent, in Poland and from 18 to 19 per cent, in Hungary. T h a t means t h a t almost a million workers were dismissed in Poland and some 200,000 or 250,000 in Hungary. I t is interesting to compare these employment figures with those of industrial production. If 1929 = 100, the index numbers of production for Poland and Hungary will be as follows (average for the y e a r ) 3 : Poland Hungary 1927 1928 1929 1930 1931 1932 1933 88 98 100 99 100 100 82 93 69 86 54 74 55 81 1934 63 95 1 Statistical Yearbook of the League of Nations, 1933-1934, pp. 53-54, and Monthly Bulletin of Statistics, 1934. 2 LEAGUE OF NATIONS : Monthly Bulletin of Statistics. 3 Cf. diagram X X X I X , p. 150. — 150 — DIAGRAM X X X I X . — I N D E X IN INDUSTRY IN NUMBERS POLAND AND OF EMPLOYMENT HUNGARY Poland Hungary I0O 90 80 70 60 50 1927 28 29 SO SI 52 S3 54 „ 1927 28 2? SO SI 32 55 54 = ^ = = = Employment. ^mmmimm^^^m Industrial production. The agreement between the index numbers of employment and those of production proves that in both these countries the spread of unemployment during the depression kept pace with the gradual decline in production. In consequence of the economic structure of the countries, however, only a very small fraction of this increase in unemployment was registered by the exchanges, and the amount of " invisible " unemployment far exceeds the amount of officially registered unemployment. Japan Japan, as is well known, is one of the young capitalist countries with a rapidly developing industry. But in spite of its achievements in this field, which have begun to cause serious uneasiness to the large, old-established capitalist countries, Japan remains predominantly an agricultural country. It is a significant fact that there has been no industrialisation of the popu- — 151 — lation during recent years. The distribution of the occupied population in 1920 and 1930 was as follows (in thousands)* : ,qon iqqn 19¿ü 1ÍM0 Agriculture Mines Manufacturing industry Commerce Transport Public services and liberal professions Domestic service Other occupations 14,661 424 5,297 3,290 923 14,724 236 5,291 4,463 1,109 1,482 655 528 2,031 806 561 Total occupied population . . . 27,261 29,221 Increase ( + ) or decrease (—) + 63 — 188 — 6 +1,173 + 186 + + + 549 151 33 +1,960 The increase in the occupied population (about 7 per cent. in 10 years) was distributed over every branch of the economic system with the exception of manufacturing industry and mines. But the group of the occupied population classified as belonging to industry is too varied for these figures to be used for the purpose of this study : it includes not only those employed in modern capitalist undertakings but also a large number of craftsmen and home workers. It is therefore better to consider industrial establishments in the narrower sense. Table XLVI shows that the modern type of undertaking is gaining ground in Japan. Although the total number of persons in industrial occupations remained unchanged from 1920 to 1930, the number of workers in capitalist undertakings increased by 20 per cent, from 1919 to 1928. But this increase seems comparatively slight when set over against the development of industrial production, for which the index number has been as follows 2 : Min pa iviixiea 1913 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 100 121 108 114 115 122 129 129 135 139 143 146 124 Manufacturing industries 100 162 188 197 206 221 241 271 274 295 324 300 303 All industry 100 157 176 184 192 206 222 249 251 270 297 278 276 1 Statistical Yearbook of the League of Nations, 1933-1934, p. 40. 2 " Institut für Konjunkturforschung ", Berlin. 152 TABLE XLVI.—INDUSTRIAL ESTABLISHMENTS IN JAPAN 1 Undertakings employing Year from 5 to 9 from 10 to 100 100 or more workers workers workers Total Number of establishments 1909 1914 1919 1924 1928 16,802 14,655 20,118 23,415 29,116 14,306 15,698 21,587 22,531 24,012 1,120 1,364 2,243 2,448 2,820 32,228 31,717 43,949 48,394 55,948 Number of workers (in thousands) 108 94 137 154 195 1909 1914 1919 1924 1928 i INTERNATIONAL LABOUR OFFICE and 23. 344 393 578 570 624 348 461 896 1,066 1,118 801 948 1,612 1,790 1,936 : Industrial Labour in Japan, Geneva, 1934, pp. 22 The index number of production in manufacturing industries m a y be compared with the number of workers in industrial establishments employing 10 or more persons, which varied as follows (in thousands) : 1914 1919 1920 1921 1922 , 1923 . 1924 , 854 1,475 1,304 1,473 1,536 1,600 1925 1926 1927 1928 1929 1930 1,651 1,701 1,607 1,649 1,852 1,687 1,636 The divergence between the index number of production for Japanese industry and the index of the number of workers in industrial occupations shows the progress in industrial technique t h a t J a p a n has made in recent years. The curves in diagram X L show that, up to 1924, the increase in production was due mainly to the increase in the numbers of persons employed. The curve of the volume of production 1 Revised number, Cf. table XLVI. — 153 — follows very closely, in the left half of the graph, the curve of the number of workers employed in industry. Consequently the curve of individual output is almost horizontal in this section of the graph, since technical progress was quite insignificant from 1914 to 1924. From 1924 onwards the situation changed. The number of persons employed rose very slightly, but the output per head increased. In this part of the graph, the production curve follows the curve of individual output, while the curve of those employed is almost horizontal. It may therefore be concluded that before the world depression began the tendency to replace human labour by machinery had manifested itself in Japan. DIAGRAM X L . — I N D U S T R I A L 1-1.P24I = PRODUCTION XVV/ IN JAPAN II ft I40 150 no no 100 I0O 90 9o /Y 60 70 60 *// 7o 6o 1 *_Ä SO ~ SO *Af *Af _. 50 * // 40 40 1915 I? 2 0 21 22 25 2 4 2 5 2 6 27 28 29 3 0 Volume of production. — — — — — — Number oí persons employed. ai^MMHBHM Individual output. In mines, this process had been going on for a long time. The output of coal and ore in 1930 was 20 per cent, more than the quantities extracted in 1920, although the number of persons employed had been considerably reduced 1. But the Japanese statistics of the labour market do not give a clear picture of the effects of technical progress on employment. 1 Cf. census figures for 1930, p . 151. 11 — 154 — The census of t h e unemployed, made every five years, is incomplete, and it scarcely seems possible t o make use of the results. The number of unemployed persons registered in industrial and mining centres was 105,612 in 1925 and 155,575 in 1930 \ I t is nevertheless certain : (1) t h a t t h e number of workers in J a p a n is increasing b y several hundreds of thousands annually; (2) t h a t the number of wage earners employed in industrial undertakings with 10 or more workers did not increase by more t h a n 36,000 from 1925 t o 1930; (3) t h a t in 1930 the number of persons employed in industry was 165,000 lower than in the preceding year. When faced with these facts, one is inclined to cast doubts on the accuracy of statistics t h a t show an increase of only 50,000 in t h e unemployment figure. Even supposing t h a t the method used is in itself faultless, the statistics must be based on an unduly narrow interpretation of t h e t e r m " unemployed ", so t h a t only a small fraction of those who have no work are registered as being unemployed. For recent years, Japanese statistics provide an index number of employment based on figures for a certain number of typical undertakings in different branches of production. Table X L V I I reproduces this index, together with the Mitsubishi index of production, which also covers manufacturing industry. The ratio of these two index numbers t o each other gives the index of individual output (labour productivity). TABLE XLVII.—MANUFACTURING INDUSTRY IN JAPAN (1929 = 100) Year 1926 1927 1928 1929 1930 1931 1932 1933 1 Number of wage earners employed (E) 109.8 104.1 99.2 100.0 90.0 81.7 82.0 89.9 100.1 INTERNATIONAL L A B O U R O F F I C E : Industrial Production (V) Individual output 77.9 83.0 89.4 100.0 94.6 97.2 103.1 119.0 129.7 Labour in Japan, 71 80 90 100 105 119 126 132 130 p. 282. — 155 — The rate at which individual output increased was more or less steady throughout the period in question. If 1929 = 100, the increase in the output per head of those in employment was as follows : From 1926 to 1927. 1927 1928 1929 1930 1931 1932 „ „ „ „ „ „ 1928. 1929. 1930. 1931. 1932. 1933. + + + -r + 10 10 5 14 7 6 -f 9 j - Average for the year DIAGRAM XLI.—INDUSTRIAL PRODUCTION IN JAPAN (1929 = 100) ISO 120 I20 / no \ 9° ao 70 \y n - ^ 1926 1927 ^ d ir^ Jx \ IOO 9o %% % f 1920 loo % ôo 1929 <93o 1931 1952 70 19» = = = = = Volume of production. _ _ • _ • • . - . . . Number of wage earners employed. ^ B I ^ B B Individual output. The rate of increase slackened only in 1930, but it was made good the following year. The curve of individual output dominates diagram XLI, crossing it diagonally from the bottom left to the top right corner. The decline in production in 1930-1931 was very slight, and had there been no technical progress a slight reduction in hours of work, such as took place in other countries, would have sufficed to prevent any extensive dismissal of workers. These dismissals were therefore due, not so much to the fall in production, but — 156 — rather to the elimination of human labour by new technical methods. The unemployment that developed in Japan during this period must therefore be classed as technological. With regard to its extent, there are monthly estimates of unemployment, published by the Japanese statistical services from 1930 onwards. They give the following annual averages : Actual figures Percentages 1930 1931 1932 1933 1934 369 5.3 423 6.1 486 6.8 409 5.6 373 5.0 It is difficult to reconcile these figures with other Japanese statistics. From 1926 to 1931 the index of employment (1929 = 100) fell from 109.8 to 81.7 x . The number of persons employed in industrial undertakings would thus appear to have been reduced by 25 per cent., and that in itself must have meant, at a conservative estimate, the dismissal of 400,000 or 500,000 wage earners. But in the meantime the number of persons in search of work had increased by reason of the natural growth of the population and the migration of surplus agricultural labour to the towns. The question therefore arises : what has become of all those extra workers? Did they become craftsmen, or have they realised that there is no place for them at nature's banquet. In that case, to pursue the metaphor of Malthus, they must be waiting at the door of the banqueting hall—uninvited guests. They have not disappeared from this world, but they have found no regular occupation in the town, and they have probably returned to the 1 village or are reduced to earning a livelihood by picking up casual i jobs here and there, for that is the fate to which hundreds of thousands, or even millions of men and women are condemned in an overpopulated country. Whether or not they are unemployed is a question they themselves would probably be unable to answer. The statistics of the labour market, in any case, know nothing of their existence. If they are unemployed, they must be classified under " invisible unemployment ". 1 Cf. table X LVII, p. 154 and diagram XLI, p. 155. — 157 — CONCLUSIONS The attempt must now be made to draw some conclusions from this analysis of the development of the labour market in various countries. These conclusions fall under two heads : (a) economic phenomena and (b) statistical problems arising out of these phenomena. * * * 1. The development of employment possibilities in the various countries since the war has not been at all uniform, notwithstanding the fact that three depressions have shaken the economic life of the world during these sixteen years and that each of the depressions has been international in character and spread to a number of countries. The first of these depressions in point of time (1920-1922) was essentially the consequence of demobilisation. I t began in the United States about the middle of 1920 and spread successively to Great Britain, the Scandinavian countries, Belgium, Italy and a few other countries. France, Germany, the Danubian States and Japan escaped. When this first wave of depression had passed, there was a spell of reasonable economic activity in most countries. But it was not long before a second depression set in (1926-1927) centring this time in Great Britain and Germany. In Britain this fresh depression was attended by serious social conflicts (the extensive coal strike); in Germany it was connected with the rationalisation of undertakings. This depression spread to the Scandinavian countries (more especially to Norway and Denmark) and to some parts of Eastern Europe (Poland). The period that followed was one of economic recovery, but the boom was not sufficiently marked to absorb all the existing unemployment. Towards the end of 1929 came the world depression. From the United States, where it first made itself felt, it spread like wildfire throughout the world. Unemployment grew beyond all measure, — 158 — reaching its peak in the summer of 1932. Since then, employment has improved slightly in most of the industrial countries; there are only a very few countries in which unemployment has continued to spread. 2. The characteristic feature of the period between the end of the war and the beginning of the world depression was the fact that in many countries the development of employment failed to keep pace with the growth of production. This was the case in the United States in the middle of a period of marked prosperity, as also in Great Britain, with its standing army of unemployed workers, in Germany during the years of economic recovery, in Japan, the Scandinavian countries, etc. It was during 1929 that the economic situation of the world was, comparatively speaking, most prosperous. But even at that date there were more workers unemployed or on short time in many countries than there usually were during periods of depression before the war 1 . Even before the world depression, the labour markets of these countries were much overcrowded, their economic systems could not utilise all the available labour, and vmemployment was gradually spreading. During the war and the period immediately following the armistice, there was practically no unemployment in the world; there was enough work for all. But this did not mean that the supply of and the demand for labour were really in stable equilibrium. And such a position of equilibrium had first to be found if the economic and social progress of the world after the war was to be guaranteed. No such position was found, and in the preceding analysis it has been seen how the balance was disturbed in various countries by the joint action of demographic, technical and economic factors. 3. From the demographic point of view, two separate phases can be distinguished in the post-war period : up to 1930, the age 1 The percentage of trade union members unemployed or on short time in June-July 1929 was : Completely On short unemployed time Germany 8.5 6.8 Great Britain (unemployment insurance)... 7.0 2.4 Sweden 7.2 — Norway 11.3 — Denmark 10.0 — Australia 10.0 — New Zealand 9.3 — — 159 — groups entering occupational life were well stocked; after t h a t date came the age groups from the war years, when the birth-rate had been low. During the first phase, up t o 1930, there were three determining factors in the influx of new workers to the occupied population : (a) the natural increase (at a rather high rate) of the population of working age; (b) the removal of certain vestiges of the war years, when large numbers of women had temporarily engaged in occupational activities; (c) the decline of international migration movements as compared with the pre-war period. Under the influence of these three factors, the occupied population in most countries increased from year to year. The rate of increase was quite high, but as a general rule it did not exceed the pre-war rate \ I n the second phase, which covers the years of world-wide depression, t h e pressure on the labour market in Germany, Great Britain, France, Italy and some other countries was appreciably relieved by the decline in the influx of new labour 2. This to some extent mitigated the effects of the decrease in production on employment possibilities. 4. The distribution of the population over the various occupational groups may be considered as a demographic problem in the wide sense. The gradual industrialisation of the population was a characteristic feature of the pre-war period : the proportion of the occupied population in industrial occupations rose, and these occupations were able to absorb the surplus supply of labour from rural areas. After the war, the industrialisation of the population ceased in 1 In the United States, the annual increase in the occupied population was 0.9 per cent, from 1910 to 1920 and 1.6 per cent, from 1920 to 1930 (as against 2.3 per cent, from 1869 to 1899 and 3.2 per cent, from 1899 to 1913). In Great Britain, the occupied population increased by 1.4 per cent, annually on the average from 1923 to 1929, as compared with 1.2 per cent, from 1861 to 1911. In Germany the annual increase was 1.1 per cent. from the middle of 1925 to the beginning of 1930, and 1.6 per cent, from 1882 to 1907. In France and Italy, on the other hand, the occupied population fell in numbers after 1920. 2 In Germany, for example, the number of wage earners varied as follows : 20,293,000 a t the beginning of 1927; 21,127,000 at the beginning of 1930; 20,832,000 a t the beginning of 1933. — 160 — many countries. In the United States, Great Britain, Japan and Norway, industrial occupations are so crowded that they have ceased to attract the new elements in the occupied population 1. 5. Everywhere the power of industry to absorb the additions to the occupied population has waned. In every country that has been affected by the latest depression and that has statistics of its labour market it has been found that the seat of the disease of unemployment lies in mining and manufacturing industry (not including building). But this phenomenon is always obscured to some extent by the fact that wage earners in search of employment are gradually forced into other occupational groups. The absolute or relative decrease in the ability of the industrial occupations to absorb additional labour must be attributed to the changes that occurred : (1) in the rate of increase of the volume of industrial production; (2) in the output per head of industrial workers. In all the cases studied in the preceding pages, it has been seen that before the war the volume of production increased more rapidly than did the individual output of the occupied population in industry (or of the wage earners employed in industrial undertaKingSj. Percentage annual increase in the volume of individual production (AV) output (AT) Great Britain, 1861-1911 United States, 1869-1899 . . . 1899-1913 . . . Germany, 1882-1907 1.85 5.6 4.5 4.3 0.59 2.3 1.1 2.1 No figures are available for recent years that permit of an accurate comparison, for the period that could be considered would be too short and would cover only part of the economic cycle. Nevertheless, the figures for the years immediately preceding the economic depression are extremely interesting : Percentage annual increase in the volume of individual production (AV) output (AT) United States, 1923-1929 Great Britain, 1924-1925 Germany, 1926-1929 Italy, 1923-1929 Czechoslovakia, 1921-1929 . . . Sweden, 1923-1929 Norway, 1923-1929 Denmark, 1923-1929 Japan, 1926-1929 1 2.9 2.2 9.0 7.3 6.3 5.4 4.7 3.0 8.7 4.8 2.7 3.1 5.5 3.2 4.2 5.6 2.7 12.1 The same phenomenon has been noted in Portugal and India. — 161 — Before the depression, then, industrial production in the United States, Great Britain, Japan and Norway was increasing less rapidly than the individual output of the workers employed 1. In these countries, AT > AV, which meant the elimination of a certain amount of labour from the production process. The unemployment that had grown up in industrial occupations in the United States, Japan and Norway before the depression may therefore be considered technological unemployment. In the case of Great Britain the situation is rather more complex : the unduly slow growth of production was really the cause of the disproportion in the development of the various factors. On the other hand, no evidence of technological unemployment can be found in the case of Germany, France, Italy, Czechoslovakia, Sweden or Denmark. It is true that human labour was displaced by machinery in these countries too, but this was counterbalanced by an increase in production that absorbed the labour thus set free. 6. The unemployment that has come into existence since 1929 is due entirely to the decline in industrial production. In every country this unemployment was at first concentrated in the same branch of the economic system—industry. The other groups of occupations were not affected until later, and then much less acutely. The amount of work to be performed during the depression kept pace with the falling rate of industrial production, but the undertakings were able, by spreading employment over a larger number of workers, to retain in their service a fraction of those who would otherwise have been dismissed as superfluous. When estimating the extent of unemployment during the depression, therefore, one must add to the official number of registered unemployed persons both those who are on short time and those " invisible " unemployed persons who did not apply to the exchanges because they did not expect to get any help from them. 1 The various percentages are not comparable with each other, having been obtained by a variety of methods and referring to different groups of undertakings. But the two percentages for each country cover more or less the same groups of undertakings and occupations. The sources of error involved in comparing the two are no greater than the possibilities of error in the original figures from which the percentages were calculated. — 162 — The technological unemployment of 1929 might be thought to be insignificant and harmless when compared with this disastrous unemployment of economic origin. But nothing could be more mistaken than such a conclusion. The recent depression might never have reached such alarming proportions if the economic equilibrium of the world had not first of all been upset by the growth of unemployment right in the middle of a period of economic recovery and prosperity. * * * This brief summary of the conclusions that may be drawn from the study of the labour market in several countries leads to the further question of how the development of employment or unemployment can best be kept under continuous observation. The development of unemployment in each country was explained in these pages by the conjunction of demographic, technical and economic factors, the combined action of which was expressed in this formula : AT ACh=AS + E . f T AV ^ - - T ^ . The various analyses were intended to test the value of this formula. In so far as it has survived the test, the conditions to be satisfied by statistics of the labour market may be summed up in the following points. 1. The point of departure for any statistics of the labour market must be as accurate information as possible concerning the size and composition of the available labour supply (S) of the country. These data should be available in the census. The census must therefore provide detailed information as to the social distribution of the population over the various branches of economic activity, the distribution of the wage earning population by sex and age and the amount of unemployment at the date of the census. 2. On the basis of these figures, the probable development of the population of working age in future years should be calculated — 163 — for each country with the help of mortality tables 1. The series of figures thus obtained should be constantly corrected and kept up to date by reference to immigration, emigration and other statistics. 3. From the estimated development of the population of working age it is possible to estimate the probable increase in the occupied population and, more particularly, in the number of wage earners (¿\S). A comparison of the development of the available labour supply with the development of employment (E) enables the investigator to assess the total amount of unemployment—a figure that may differ appreciably from the figures obtained by direct observation through sickness or unemployment insurance funds or the factory inspectorate. The statistics of employment and unemployment require to be improved. It is specially necessary to calculate a separate index number of employment in industry (possibly including mines, but with a separate index for building). Permanent unemployment has its roots in the industrial occupations, which have ceased to absorb the influx of workers from the country to the towns ; it is the loss of stable equilibrium in these occupations that reacts on the whole economic system and disorganises its other branches. Exact statistics for industrial occupations can therefore provide a key to the development of the economic system as a whole 2 . An index number of employment or unemployment that embraces indiscriminately industrial occupations, building, commerce and transport, the liberal professions and perhaps even agriculture is really of very little use. The best 1 The author has in mind the study made by Mr. A. L. Bowley for the League of Nations in 1926, to which reference has several times been made in these pages. The results of his work are now out of date, and the calculations should be made again on a broader basis, with a view to assessing the probable growth of the middle age groups from year to year and not merely the estimated number a t a given date (e. g. 1940 or 1945). * I t is for the same reasons t h a t building should be kept separate from industry in the narrower sense. The evolution of these two branches follows different, and sometimes contrary, trends. In the United States and in Great Britain the building trade absorbed some of the labour displaced from mining and manufacturing industries, by machinery or the decline in production ; in Germany, on the other hand, the building trade is the cause of seasonal fluctuations in employment; in France, it was the great building activity during the depression that stimulated employment in various other branches of the economic system. — 164 — solution would be to have four distinct index numbers for : (a) manufactures and mining; (b) building; (c) commerce and transport; (d) other occupational groups. 4. It is very important that there should be exact statistics of short time and of the spread-over of employment. Experience shows that these statistics can equally well be linked up with those of unemployment (Germany) or with those of employment (France, United States). In this case also industry should be kept separate from other branches. 5. As was pointed out above, the statistics of production, employment and unemployment should be so co-ordinated that it is possible at any given moment to determine the development of output per head of those employed and per hour worked, not only for industry as a whole but also for the various branches of industry. In these days it is time to have done with economic and social policy that gropingly seeks its way in the darkness. It must rise to the level of technical progress and be able to base its action on a full knowledge of the real facts of the moment and an adequate forecast of future trends. The task of economic and social policy would be considerably lightened by the existence of unemployment and employment statistics satisfying those conditions. Such statistics would be like a powerful searchlight penetrating the darkness of the future ; they would reveal in time the factors that threaten to upset the equilibrium of the economic and social system of a country, so that the danger could be met while there were still healthy sectors of the economic system to serve as bases for the necessary defensive operations. — 165 — APPENDIX APPLICATION OF THE MATHEMATICAL FORMULA FOR UNEMPLOYMENT TO EMPIRICAL SERIES OF FIGURES The application of the formula AT AV A C h = AS + E ' T + AT T + AT will be demonstrated below, taking as an example the development of unemployment in Great Britain from 1860 to 1910. Three series of figures are given : (a) the index number of persons in industrial occupations (S) ; (b) the percentage of persons in industrial occupations who were unemployed (Ch as a percentage of S) ; (c) the volume of production (V). S = Ch (as percentage of S) = V = 1860 1870 100 1.9 100 113.0 3.9 129.4 1880 1890 1900 1910 129.2 5.2 155.9 148.8 2.1 182.4 169.4 2.4 232.4 191.7 4.7 250.0 The value S for each year is multiplied by the percentage of unemployed persons. This gives the index number of unemployment (Ch) in terms of the occupied population at the beginning of the observation period, in 1860. These figures are then deducted from the figures for the occupied population for the corresponding years. The difference indicates the changes in the number of members of the occupied population actually in employment (E). Ch = E = I860 1870 1880 1890 1900 1910 1.9 98.1 4.4 108.6 6.7 122.5 3.1 145.7 4.1 165.3 9.0 182.7 The index number of production (V) is then divided by the index of the number of persons in employment (E) to obtain the development of individual output, the basis of comparison (which in this case is 1 and not 100) being the production that would have been obtained at the beginning of the observation period (1860) per head of the occupied population engaged in industry if all these persons had been in employment at that time (i.e., if E = S = 100). This gives : T= I860 1870 1880 1890 1900 ,1910 1.019 1.192 1.273 1.252 1.406 1.368 — 166 — In the following calculations, T will be used to indicate the figure for 1860 in the above series — i.e., 1,019. AT will be used to express the difference between the succeeding figures in the series and this initial figure. This gives two new sets of figures : I860 1870 1880 1890 1900 1910 0 0.173 0.254 0.233 0.387 0.349 0 0.145 0.199 0.186 0.275 0.255 AT = AT T + AT The value of the three terms of the formula must now be calculated : E. m I860 1870 1880 1890 1900 1910 AS = 0 13.0 29.2 48.8 69.4 91.7 AT .7^ = 0 14.2 19.5 18.2 27.0 25.1 T + AT AV 0 24.7 43.9 65.8 94.2 109.7 T + AT AS represents the influx of new members of the occupied population to the labour market. If it be assumed that the volume of production remained constant (AV = 0) and that the technique of production remained unchanged (AT = 0), the number of unemployed persons (in 1860, Ch = 1.9) would have been 1.9 + AS for each year. In order to allow for technical progress, one must add to this sum the amount of labour that would have been eliminated from the production process as a result of this progress if the volume of production had not increaAT sed. This amount of labour is represented by the value of E . m . Arp' AT This gives 1.9 + AS + E . m , A™' which represents the number of persons that would have been unemployed if—and the hypothesis is quite impossible—the population and technical progress had developed as they actually did, while the volume of production remained AV constant. From this sum must be deducted the value of 7¿r~;—r^' T + AT which represents the absorption of unemployment as a result of increased production. I860 AT S = 1.9 + AS + E . - - ^ 1870 1880 1890 1900 1910 = 1.9 29.1 50.6 68.9 98.3 118.7 AV Ch = S — -———- = T + AT 1.9 4.4 6.7 3.1 4.1 9.0 This is the method used for calculating the quantities represented in diagrams IV, V, IX, XVI, XXXII, XXXV and XXXVIII.