The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
S T R U C T U R A L U N E M P L O Y M E N T AND P U B L IC P O L IC Y IN IN T E R W A R B R IT A IN : A R E V IE W ESSA Y Prakash Loungani W orking Paper Series M acro Econom ic Issues Research D epartm ent Federal Reserve B ank o f Chicago February, 1991 (W P-91-1) Structural U nem ploym ent and Public Policy in Interw ar Britain: A Review Essay1 by Prakash Loungani D ecem ber 1990 1. Introduction L ong accustom ed to m ysteries in which a single individual em erges as the perpetrator o f the crim e, m ost readers o f A gatha C hristie’s 1930 novel "M urder on the O rient Express" w ere caught unaw ares by the em ergence o f as m any as tw elve perpetrators o f a single crime. A lm ost as m any perpetrators are needed to explain the m ystery o f British interw ar unem ploym ent. As shown in Figure 1, the aggregate unem ploym ent rate hovered around 8% for m ost o f the 1920’s, jum ped to nearly 16% between 1929 and 1932, and then fell back gradually to 8% betw een 1932 and 1937. It is com m on to group the factors responsible for the persistently high unem ploym ent o f this period into three categories: (1) structural factors, stem ming from the decline o f several staple industries; (2) cyclical factors, stem m ing from the stringent m onetary policy follow ed by the Treasury in order to return to the gold standard at the pre-w ar parity; (3) "voluntary" unem ploym ent, attributable to the generosity o f unem ploym ent benefits relative to wages. These factors interacted with one another in ways that w orsened the problem . The tight m oney policy was a further blow to industries experiencing structural decline. The generosity o f the dole m ay w ell have inhibited labor m obility from declining sectors to expanding ones. In this essay I describe the response o f B ritish policymakers to the em ergence and persistence o f mass unem ploym ent. M y description is based largely on W.R. G arside’s m asterly analysis o f 1 I thank Charles Calomiris and M ark Rush for very useful com ments. The evidence presented in Section 4 o f this paper is based on "Sectoral Shifts in Interw ar Britain" by Loungani and Rush (Federal Reserve Bank o f Chicago W orking Paper No. W P-90-7, July 1990). 2 the subject in his recent book British Unemployment, 1919-1939: A Study in Public Policy (Cam bridge U niversity Press, 1990). I also discuss recent research on the sources o f British interw ar unem ploym ent and provide new evidence on the im portance o f structural factors. 2. The Emergence of Mass Unemployment In 1860, Britain accounted for nearly a fifth o f the w orld’s industrial output. Exports accounted for a large fraction o f B ritain’s output and they were concentrated in a few sectors o f the econom y. B ritain’s com parative advantage was "rooted in coal and unskilled labour. Textiles and iron were m ajor users o f steam power, and cotton textiles in particular relied on low -grade labour" [Crafts (1985, p. 146)]. B ritain’s dom inance slipped in the last quarter o f the nineteenth century. A set o f innovations, sometimes referred to as the second industrial revolution, raised the return to investm ent in hum an capital. The innovations altered techniques o f production in existing industries and led to the em ergence o f several new industries that were ’hum an-capital intensive.’ Britain appears to have under-invested in technical and scientific education, relative to the U.S. and Germany. B ritain’s share o f industrial output was surpassed by the U .S. around 1895 and by Germ any around 1910. Foreign com petition intensified during W orld W ar I, while the British econom y was geared towards the production o f w ar materials: "Com petition was particularly felt in the staple export industries. Shortages o f B ritish coal exports during the w ar had encouraged the opening or expansion o f m ines in G erm any, Poland, the Netherlands, Spain and the Far East. M ajor textile industries grew during the w ar in Japan and India, im portant British markets before 1914. R ival shipyards had opened in the U nited States, Japan, H olland and Scandinavia. W orld iron and steel m aking capacity also expanded during the war, especially in the U nited States and challenged B ritish com panies later" [Constantine (1980, p. 11)]. A fter the war, large scale unem ploym ent em erged in these industries w hich now faced a 3 perm anent decline in the dem and for their products. Several factors im peded the reallocation o f labor from the staple industries to the new industries. G ood jobs in the expanding industries required a set o f skills different from those possessed by the unemployed. H ence, m any o f them were faced with the prospect o f m oving from their (old) skilled job to unskilled work. The geographic concentration o f declining industries m eant that the search for new jobs involved inter-regional migration. The norm al social and psychological costs associated with migration were exacerbated by housing shortages in the expanding regions. M obility may also have been reduced because "the unem ployed w ere offered the dole which prevented them from being faced with a forced choice betw een m igration and starvation. A rm ed with this guarantee o f u subsistence income, the unem ployed could hope for an im provem ent in the prospects o f the basic industries and fall back on their closely-knit w orking-class com m unities and family connections" [Booth and Glynn (1975, p. 623)]. The tight m oney policy followed by the Treasury, discussed in greater detail later in this essay, led to unem ploym ent even in expanding regions and industries. This further discouraged m igration out o f the staple industries because workers suspected that they w ould sim ply be relegated to the end o f the unem ploym ent queue at the new industries. Calomiris (1990, p.10) m akes a sim ilar point in his discussion o f the behavior o f relief workers in the U.S during the D epression.2 2 "... these (relief) w orkers were in line behind a buffer o f non-relief unem ployed w ith superior opportunities...The high unem ploym ent o f non-relief workers m ay have been a sufficient disincentive for relief w orkers’ search, given differences in the two types o f workers. A ccording to this interpretation, only very large im provements in m arket opportunities w ould have rem oved the buffer o f non-relief unem ployed and encouraged search by relief workers." 4 3. The Policy Response A. 1919 to 1925 In the period im m ediately following the war, the restoration o f the gold standard w as the prim ary objective o f British public policy. It was widely believed that the dom inance o f British industry in the pre-w ar period and the pre-em inence o f London as an international financial center were attributable to price stability engendered by the gold standard regim e. The C unliffe Com m ittee recom m ended balanced governm ent budgets and increases in the B ank Rate "to check a foreign drain o f gold in order to create the conditions necessary to the m aintenance o f an effective gold standard. The Cunliffe Com m ittee was not unaw are o f the fact that high interest rates, apart from attracting foreign capital to help stabilize the reserve position, could also depress output and employment. But it rem ained convinced that the trade-off between internal and external balance in the short run was the essential price to be paid for stability and prosperity in the long term " [Garside, p. 116]. The B ank Rate was raised from 5 % in N ovem ber 1919 to 7% in A pril 1920 and was held at that level for a year. W holesale prices tumbled, but unem ploym ent rose from 2% in 1920 to 11.3% in 1921. The em ergence o f unem ploym ent on this scale led to some soul-searching am ong policym akers, but by and large they succum bed to orthodox econom ic and financial opinion and a tight m oney policy was pursued again after July 1923. Britain returned to the gold standard in A pril 1925 at the pre-w ar parity.3 The successive governm ents over this period were inclined to largely ignore the structural aspects o f B ritish unem ploym ent over this period. Garside (p. 204) states that "governments looked to sound currency and to the revival o f trade to foster econom ic recovery and refused to be drawn into any direct interventionist policy, least o f all on 3 "G.D.H. Cole likened the em phasis on m onetary policy to a G reat G od nam ed P ar w ho is w orshipped daily at the Treasury. P ar likes unemployment; it is his form o f hum an sacrifice" [Garside, p. 122]. 5 behalf o f industry. M inisters believed that downturns in trade such as occurred in 1920/21 would, like pre-w ar depressions, be o f relatively short duration; the m ost appropriate response, therefore, was to offer industry only such assistance as w ould enable it to overcom e its temporary difficulties...For much o f the first post-w ar decade, it was difficult to discern an industrial policy as such." G overnm ents were also largely im m une to pleas o f special assistance for the ’distressed’ regions o f the econom y. Some steps were taken to im prove B ritain’s export perform ance through the provision o f guarantees and credit insurance, but Garside concludes (p. 146) that exporters found such schemes "to be overly cautious and bureaucratic." In sharp contrast to their ’m inim alist’ approach to industrial policy and regional assistance, interw ar governm ents over this period were fairly generous in the provision o f cash benefits to those unem ployed. Provision o f unem ploym ent insurance had been instituted as early as 1911. H ow ever, the provisions extended to only about a quarter of the total male labor force, benefits were fairly low and definitions o f eligibility were stringent. W ith the em ergence o f large scale unem ploym ent after the war, the unem ploym ent insurance scheme was revised. In a series o f steps between 1920 and 1921, coverage was extended to m ost m anual workers, w eekly benefits for males were tripled, benefits were instituted for women, the num ber o f contributions that had to be m ade before claim ing benefits was reduced, and the num ber o f weeks for w hich benefits could be claim ed was increased. In addition, unem ployed workers who could show that they had been ’genuinely seeking w ork’ could draw uncovenanted benefits—benefits paid in advance o f the required num ber o f contributions. B. 1925 to 1931 The years 1924 to 1929 were m arked by a w orld boom in w hich Britain did not share, suggesting to m any observers that structural problem s lay at the root o f B ritain’s continuing 6 unem ploym ent. This led to the increasing popularity o f schemes o f ’rationalization’ o f industries even though there was wide disagreem ent as to w hat rationalization actually involved.4 As in the earlier period, governm ents essentially followed an industrial policy o f enquiry and consultation, preferring that the ailing industries form ulate and finance their own rationalization plans. "Industrialists on the other hand..were often reluctant to bear the cost o f reviving w eaker firms for the benefit o f industry as a whole. The intense individualism w hich characterized producers in coalm ining, shipbuilding, iron and steel and textiles seriously retarded plans for industrial reorganization or the rationalization o f capacity. Conflicts between rival interests w ithin firm s and the diffusion o f decision-m aking powers proved to be a serious hindrance in steel. Elsew here, in coal and cotton textiles for exam ple, there was no effective way o f securing jo in t action by industry as a whole...As a result, traditional methods o f operation w ithin such industries rem ained virtually intact throughout the interw ar period" (p. 236). Some steps were taken to aid distressed areas through the creation o f an Industrial Transference B oard to foster labor mobility from these areas. Again, the steps were fairly cautious and took the form o f coordination o f the activities o f em ploym ent exchanges and training centers and the provision o f m odest transfer grants. A som ew hat om inously-nam ed H ousehold R em oval Scheme was also introduced to transfer households and thus avoid the break up o f families. 4 "..few observers agreed on w hat precisely was m eant by rationalization. To some, the concept involved the application to industry o f a greater degree o f scientific organization and managem ent; to others, it im plied w idespread m erger and am algam ation aim ed at altering the scale and efficiency o f industrial enterprise; to others again, it involved a com m itm ent to technical advance and the scrapping o f obsolete plant" (Garside, p.210). 7 W ith regard to m onetary policy, the persistence o f unem ploym ent did keep the Treasury from increasing the B ank Rate any further. H owever, the B ank Rate that prevailed over this rate was neither high enough "to secure sufficient leverage in the international m oney m arket nor did it fall low enough to afford m aterial help to jobs by stim ulating industrial enterprise" (p. 130). A new U nem ploym ent Insurance A ct in 1927 granted all insured workers w ho had exhausted their standard benefits the right to claim extended benefits for as long as they w ere unem ployed. The only requirem ent was that claimants had to prove that they were genuinely seeking work; but even this requirem ent was abolished by a 1930 Act. This A ct also transferred the financial burden o f providing benefits from the unem ploym ent insurance fund to the Treasury. C. 1931 to 1939 The return to the gold standard had failed to revive British exports and industry. H ence governm ent revenues were low while expenditures were m ounting due to the increased generosity o f unem ploym ent benefits. This im balance came to a point o f crisis in 1931 and it becam e apparent that some cuts in benefits w ould be needed to balance the b u d g et The Labor governm ent resigned rather than accept the required cuts and over the next decade Britain was governed by a sequence o f National governments under whom som e reform s were at last carried out. One area o f policy that saw significant reform was the provision o f unem ploym ent insurance. A 1934 Act m ade a sharp distinction between the needs o f the short-term unem ployed and those o f the long-term unem ployed. Part I o f the Act restored an actuarially balanced schem e o f contributory insurance for the short-term unemployed. Part II o f the A ct established an U nem ploym ent Assistance B oard to provide m eans-tested benefits to the long-term unem ployed, 8 that is, those who had exhausted their 26-w eek benefits. The B oard drew up proposed scales o f benefit, which were subm itted for approval to Parliam ent, and the paym ents were financed from tax revenues. D ram atic changes were also forthcom ing in the conduct o f m onetary policy, the first step being the abandonm ent o f the gold standard. As a replacem ent, the Treasury and the B ank o f England established an Exchange Equalization Account, which by intervening in the foreign exchange markets was able to maintain a fairly stable sterling exchange rate. L iberated from the need to keep the Bank Rate high enough to attract foreign funds and thus m aintain the parity o f sterling, the N ational G overnm ent kept the Bank Rate an at historically low level o f 2% from June 1932 to the end o f the decade. The unem ploym ent rate did fall, from 15.6% in 1932 to 9.4% in 1936. M any ’radical’ econom ists, such as Keynes and D. H. Robertson, felt that m onetary policy w ould w ork at a very slow pace in alleviating unem ploym ent and advocated an expansionary fiscal policy. However, despite their U -tum in monetary policy, the National governm ents rem ained w edded to fiscal conservatism. The turnaround in fiscal policy cam e about, o f course, as a result o f the increases in defence expenditures after 1937. Through a m ixture o f deliberate policy and serendipity, these increased expenditures finally alleviated structural unem ploym ent.5 5 "At first, defence contracts were allocated to the depressed areas m ore as a w ay o f relieving pressure on capital and labor resources elsewhere in the econom y rather than w ith the intention o f reducing interregional unem ploym ent percentages per se...U nem ploym ent in the coal, iron & steel, engineering and shipbuilding industries, all heavily concentrated in the depressed areas, declined from 27.2, 23.5, 13.6 and 44.4 respectively in 1935 to 1 6 .7 ,1 9 .5 ,7 .0 and 21.4 in 1938, w ith further falls in the follow ing year. Rearm am ent, in other words, provided a stimulus to increased expenditure and em ploym ent in the areas o f chronic unem ploym ent beyond anything that governm ent had achieved or contem plated on their behalf in previous years" (p. 360-362). 9 4. New Evidence on the Sources of Interwar Unemployment (i) The Role o f Structural Factors As N.F.R. Crafts has pointed out, econom ic historians tend to stress the structural aspects o f interw ar unem ploym ent whereas macroeconomists have a tendency to dow nplay the role o f such factors. In this section I present some new time series evid en ce-u sin g regressions o f the sort presented by Benjam in and K ochin (1979)—which highlights the im portance o f structural factors. Based as they are on a small num ber o f annual observations, these regressions are best thought o f as broad characterizations o f the data. The dependent variable in the regressions is the aggregate unem ploym ent rate. The set o f independent variables is picked on the basis o f the discussion in the previous tw o sections o f the essay. First, as discussed in Section 3, monetary policy was fairly tight through m ost o f the 1920’s, w ith a substantial easing after 1932. The Bank Rate and the growth rate o f the m onetary base turn out to perform equally w ell in capturing the im pact o f this policy on unem ploym ent. In this essay I report results using the latter variable, denoted DB. Second, as discussed in Section 2, there were large negative dem and shocks to m any staple industries due to increased foreign com petition. The tight m oney policy and the provision o f the dole ham pered the re-em ploym ent o f labor displaced from the contracting industries and hence these structural shocks led to a very persistent increase in unem ployment. The variable used to capture the intensity o f structural shifts is the cross-section standard deviation o f industry stock returns, denoted S. That is, St = [(X(Rit - Rt)2)/n J 1/2 10 where Rjt is the stock m arket return in industry i at time t, R, is the mean stock m arket return in period t and n, is the num ber o f industries6 included in the sample at tim e t. W hile m otivated by L ilien’s use o f the dispersion in employment growth as a proxy for the intensity o f structural shifts [Lilien (1982)], this variable avoids m any o f the pitfalls associated with L ilien’s m easure. First, stock prices depend in large part on expectations about the future. Inform ation about an industry’s profitability will be rapidly incorporated into stock prices, whereas the response o f industry em ploym ent m ay be staggered over time. Second, m ore resources w ill be transferred between industries, the more persistent is the divergence in the fortunes o f industries. Since stock prices represent the present value o f profits over a long horizon, the im pact o f innovations in industry profits on its stock price will depend on how long the shocks are expected to persist. A dispersion measure constructed from sectoral stock prices therefore assigns greater w eight to shocks that lead to a persistent divergence in industry fortunes. U sing annual and quarterly U .S. data, Loungani, Rush and Tave (1990a, 1990b) present em pirical evidence that strongly supports this view.7 6 The num ber o f industries ranges from 9 to 17 and com prises coal, iron, steel and engineering, shipping, textiles, electric light, telephone and telegraph, brew eries, theatres, hom erails, hotels, motors & cycles, banks, insurance, newspapers, cem ent, groceries & provisions and dry goods & stores. 7 U sing VAR systems w hich include a com prehensive set o f aggregate dem and m easures, Loungani, Rush and Tave (1990b) show that there is significant feedback from aggregate dem and to employment dispersion. Furtherm ore, once the aggregate dem and measures are included, innovations in em ploym ent dispersion account for a very small fraction o f the variance o f unem ploym ent. On the other hand, there is very little feedback from aggregate dem and to stock market dispersion. Even in V AR systems that include monetary base growth, interest rate spreads, the m ean stock m arket return and government spending as indicators o f aggregate dem and, stock m arket dispersion continues to account for a third o f the variance o f unem ploym ent at long horizons. 11 The final explanatory variable is the benefits-to-wages ratio (BW), as reported in Benjam in and Kochin. The estim ated regressions are presented in Table 1. The first regression shows that 70% o f the variance o f unem ploym ent can be explained by the current and tw o lagged values o f the stock m arket dispersion index. M oreover, the dispersion measures account fairly w ell for the persistence o f unem ployment; the D urbin-W atson statistic is 2.3. Colum n (2) reports a regression o f unem ploym ent on m onetary base growth, DB, and a lagged dependent variable. The coefficient estim ate is negative, as expected, but is not significantly different from zero at conventional levels o f significance. Lagged values o f m onetary base growth were not significant. D espite the inclusion o f lagged unem ploym ent, the adjusted-R2 of this equation—0.42—is m uch low er than that o f the first. Column (3) shows that the benefits-to-wages ratio explains 16% o f the variance o f unem ploym ent and that the coefficient estim ate o f BW is significant at a 5% level. H ow ever, this variable does not capture the serial correlation in unem ployment: the D.W . statistic is only 0.9. Finally, colum n (4) reports the results o f an all-inclusive regression. M onetary base growth and the dispersion variables em erge as highly significant,8 while the B W variable has the hypothesized sign but does not attain standard levels o f significance. The adjusted-R2 is 0.80 and the D .W statistic is 2.04. Figure 1 provides a plot o f actual unem ploym ent and predicted unem ploym ent from tw o o f the regressions reported above, those in colum n (2)—the regression based on stock m arket dispersion only—and colum n (4). The figure clearly shows that the current and lagged values o f the stock m arket dispersion index do a fairly good job of tracking unem ploym ent rates over this 8 The significance o f the stock m arket dispersion variables holds over a longer sample period, 1910 to 1938. It is also robust to the inclusion o f other explanatory variables, such as the stock m arket m ean return. 12 period. H owever, the addition o f the other explanatory variables im proves the fit considerably in some years. (ii) The Role o f the G old Standard M uch o f the recent work on the interw ar period suggests that due to technical problem s associated with the operation o f the gold standard, m onetary contractions in the U.S. and France were transm itted to the other countries on the gold standard, leading to w orldw ide deflation.9 B em anke and James (1990) argue that this deflation then led to a num ber o f banking panics—or to a substantial w eakening o f the financial position o f banks—w hich adversely affected the real econom y by interfering with the flows o f credit to industry. Since Britain w ent o ff the gold standard in 1931, it avoided the big price declines that occurred in the gold-standard countries. N evertheless, to test the potential im portance o f the ’deflation’ view, I included the grow th rate o f the m oney m ultiplier between the monetary base and M 3—a proxy for credit creation—as an additional explanatory variable in the regressions reported above. H owever, the coefficients o f the m ultiplier variables were never close to significance, while the other coefficient estim ates were unaffected. Similarly, replacing m onetary base growth by the growth rate o f M3 had little im pact on the results reported above. Hence, the im pact o f the deflation is probably m ore im portant in explaining cross-country differentials—as in the B em anke and Jam es paper—rather than the tim e series variation in U.K. unem ploym ent rates.10 9 See Eichengreen and Sachs (1985), H am ilton (1988) and Tem in (1989) for a fuller discussion. The ’technical problem s’ alluded to include the asym m etry between surplus and deficit countries in the required monetary response to gold flows. 10 A review essay by D eLong (1990) in an earlier issue o f this journal discusses the cross country evidence on unem ploym ent rates in greater detail. 13 References Benjam in, D aniel and Levis Kochin. Searching for an Explanation o f U nem ploym ent in Interw ar Britain. Journal o f Political Econom y, 87, 1979, 441-78. B em anke, Ben and H arold James. The Gold Standard, Deflation, and Financial Crisis in the G reat D epression: An International Comparison. N BER W orking Paper No. 3488, O ctober 1990. Booth, Alan and Sean Glynn. U nem ploym ent in the Interw ar Period: A M ultiple Problem . Journal o f Contem porary H istory, 10, 1975, 611-36. Calom iris, Charles. Comm ents on G reat D epression Papers—1990 EH A meetings. Constantine, Stephen. U nem ploym ent in Britain between the W ars. Longm an Group Lim ited, London, 1980. Eichengreen, Barry and Jeffrey Sachs. Exchange Rates and Econom ic Recovery in the 1930’s. Journal o f Econom ic H istory, 45, 1985, 925-46. Crafts, N.F.R. British Econom ic Growth during the Industrial Revolution. C larendon Press, O xford, 1985. D eLong, Bradford. Facets o f Interw ar Unemployment: A Review Essay. Journal o f M onetary Econom ics, 25, 1990, 305-311. G arside, W .R. British U nem ploym ent, 1919-1939: A Study in Public Policy. C am bridge U niversity Press, Camridge, 1990. H am ilton, James. The R ole o f the International G old Standard in Propagating the G reat D epression. Contem porary Policy Issues, 6, 1988, 67-89. Lilien, David. Sectoral Shifts and Cyclical Unemployment. Journal o f Political Econom y, 90, 1982, 777-793. Loungani, Prakash, M ark Rush and W illiam Tave. Stock M arket D ispersion and U nem ploym ent. Journal o f M onetary Econom ics, 25, 1990, 367-388. Loungani, Prakash, M ark Rush and W illiam Tave. Stock M arket D ispersion and R eal Econom ic Activity: Evidence from Quarterly Data. Federal Reserve B ank o f Chicago W orking P aper No. W P-90-15, Septem ber 1990. Tem in, Peter. Lessons from the G reat Depression. M IT Press, Cam bridge, 1989. T a b le 1 Unemployment Equations, 1920 to 1938 (1) (2) (3) (4) S 27.16 (5.76) [.0003] • • 25.10 (4.78) [.0002] SI 11.98 (5.94) [.0619] • • 11.33 (4.79) [.0344] S2 10.73 (5.68) [.0784] • • 18.01 (5.56) [.0064] DB - -20.94 (15.56) [.1971] • -33.24 (10.59) [.0078] BW • • 13.30 (6.36) [.0517] 4.45 (3.58) [.2358] U1 • 0.67 (0.17) [.0013] • Intercept 0.17 (1.54) [.9113] 3.52 (1.66) [.0503] 3.37 (3.04) [.2832] -2.53 (1.70) [.1600] 2.28 1.81 0.89 2.04 0.70 0.42 0.16 0.80 Independent Variables D.W. Adj. R2 • Notes: The numbers in parentheses (...) are standard errors. The numbers in brackets [...] are p-values. Unemployment Rate, 1920 to 1938 Actual and Predicted Actual Pred.-Structural Pred.-AII Factors