United Kingdom: Recent Economic Developments
Author: International Monetary Fund
Publication Date:
16 Dec 1996
eISBN:
9781451814064
Language:
English
Keywords:
ISCR; CR; investment; income; PFI project; distribution data; household income; income distribution; income inequality; investment income; Income distribution; Wages; Labor markets; Europe
Download PDF (2.5 MB)

This paper reviews economic developments in the United Kingdom during 1991–96. The paper examines two sets of possible reasons why the United Kingdom's savings and investment rates are lower than elsewhere. The first consists of factors leading to lower optimal rates of savings and investment: these include demographics, economic structure and technology, a liberalized financial environment, and so on. The second consists of distortions leading to lower-than-optimal savings and investment rates. The paper also presents new estimates for the potential growth rate, the output gap, and the natural rate of unemployment.

Abstract

This paper reviews economic developments in the United Kingdom during 1991–96. The paper examines two sets of possible reasons why the United Kingdom's savings and investment rates are lower than elsewhere. The first consists of factors leading to lower optimal rates of savings and investment: these include demographics, economic structure and technology, a liberalized financial environment, and so on. The second consists of distortions leading to lower-than-optimal savings and investment rates. The paper also presents new estimates for the potential growth rate, the output gap, and the natural rate of unemployment.

VI. INCOME DISTRIBUTION SINCE THE LABOR MARKET REFORMS OF THE 1980s81

A. Introduction

192. Since the early 1980s, the United Kingdom’s labor market has been transformed. Changes in labor laws and regulations, and other aspects of the framework in which the markets operate (Annex 1), brought a striking increase in the market’s flexibility, with favorable consequences for productivity growth and the natural rate of unemployment (Chapter II). At the same time, this period has witnessed a significant increase in income inequality, albeit accompanied by rising real incomes for most of the population. This chapter will describe how measures of inequality of household incomes have evolved in the United Kingdom since the 1980s, compare this to other industrialized countries, and discuss the reasons.

B. Some Facts82

193. This section reviews what has happened to the United Kingdom’s income distribution since the early 1980s. While the main focus is on inequality of household incomes, other elements—including inequality of consumption expenditures, (financial) wealth, and earnings also come into play.

194. Household incomes in the United Kingdom have become increasingly unequal since the early 1980s. After being more or less stable for 20 years, the Gini coefficient (see Box) rose from 24.5 at the end of the 1970s to almost 32.5 in 1991 (Chart 1). Three sub-periods can be distinguished: inequality was diminishing during 1970-77, then started rising during 1978-84, with a sharper increase after 1984. The United Kingdom is not alone in having experienced a rise in income inequality during the 1980s: increases are common to many OECD countries. Moreover, the United Kingdom started from a position of greater equality of household income than in other OECD countries at the end of the 1970s; its Gini coefficient thus only converged to levels elsewhere in the EU, and remained well below that for the United States. However, in the United Kingdom, unlike in most other OECD countries, income inequality continued to increase in the early 1990s. Moreover the ratio of the income of the richest 20 percent of the population to that of the poorest 20 percent is now greater in the United Kingdom than in any other big Western industrialized nation, including the U.S. (see Chart 2).

CHART 1
CHART 1

SELECTED COUNTRIES: Gini Coefficients for Household Income

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Deininger and Squire (1996) and staff estimates
CHART 2
CHART 2

SELECTED COUNTRIES: Ratio of Highest to Lowest Quintile Incomes, Average 1980-92

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: UNDP Human Development Report (1995).

195. This picture can be rounded out by examining the evolution of the income shares of different groups. Chart 3 shows that in the United Kingdom real incomes of the richest 10 percent increased substantially (by about 30 percent) from 1979 to 1991, while incomes of those of lower down the distribution increased by much less or, for the poorest 10 percent, declined.

CHART 3
CHART 3

UNITED KINGDOM: Income Shares by Decile, 1979 and 1990/91

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Jenkins (1996)

196. Shifts in income distribution net of housing cost are even more striking (Chart 4): on this basis, incomes of the poorest 5 percent of the population declined by 10 percent from 1981 to 1991, after holding their own or rising relative to the rest of the population over the preceding two decades—and in stark contrast to the richest 5 percent, whose incomes grew about 60 percent during the 1981-91 period. The bottom 10 percent also experienced declining incomes over the 1979-91 period, with a 14 percent decline net of housing cost. For the second decile real income rose, but net of housing cost it did not change. For the bottom fifth of the population as a whole, real incomes thus rose over this period, but this rise in incomes was more than fully absorbed by increased housing costs.83

CHART 4
CHART 4

UNITED KINGDOM: Changes in Income by Group

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Johnson (1996) and HBAI (1993)

197. In assessing the importance of these changes in household income distribution, it is useful to examine whether they can be explained by any substantial changes in the size and structure of households. Chart 5 shows the bottom and top deciles classified by family type, and Chart 6 by economic status. Within the bottom decile “couples with children” is by far the biggest group, now making up more than half of the poorest decile, compared with only a third three decades ago. Thus, the reduction in incomes of the poorest does not reflect growing numbers of single people in this group—on the contrary, to an increasing extent the poor consist of families. The salient aspect of the economic status of the lowest decile is the increasing share of this decile who were unemployed—reflecting the rise in unemployment rates from some 4½ percent in 1979 to 8½ percent in 1991. A small and declining share of the lowest decile consisted of pensioners. In contrast, the top decile consisted mainly of couples with or without children, with the majority of these families containing two full-time workers.

CHART 5
CHART 5

UNITED KINGDOM: Income Groups Classified by Family Type

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Goodman sad Webb (1994)
CHART 6
CHART 6

UNITED KINGDOM: Income Groups Classified by Economic Status

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Goodman and Webb (1994)

198. Finally Chart 7 shows that the relative share of labor income in total income has decreased substantially over the last three decades, while those of social security benefits, investment incomes and pensions have increased.

CHART 7
CHART 7

UNITED KINGDOM: Household Income by Source

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Goodman and Webb (1994)

199. The distribution of expenditure has also widened considerably since the mid-1980s but less rapidly than that of income (Chart 8). Whereas for the mean and median changes expenditure were similar to those in income, for the poorest 5 percent, real expenditure increased during 1979-92 despite the reported decrease in income; for the richest 5 percent, expenditure rose less than in proportion to their income. This may be attributable in part to consumption smoothing. However, the low share of pensioners in the bottom decile evidence suggests that such consumption smoothing does not mainly reflect life cycle considerations. Instead, it suggests that many of those at the bottom of the income distribution were there temporarily, due to adverse shocks. This is consistent with the view that the difference between changes in income and expenditure inequality is indicative of increased short-term income risk (Blundell and Preston, 1995)—which in turn may be the counterpart of the greater labor market flexibility of the 1980s and 1990s.

CHART 8
CHART 8

UNITED KINGDOM: Growth in Income and Expenditure, 1979-1992

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: Johnson (1996)

200. The conflicting behavior of income and expenditure of the lowest decile probably partly also reflects the underground economy: income may be underreported for reasons of tax evasion and unemployment benefits fraud, and this underreporting may have become more prevalent with increasing numbers of unemployed and self-employed in the labor force. To the extent that income of the poorest is disproportionately underreported, measured incomes overstate the true extent to which income inequality has risen.

201. Distributional mobility—the rate at which people move up or down the earnings ladder—is another important aspect of the picture: greater mobility across income groups might mitigate social concerns over inequality—as it would confirm that there is equality of opportunity even if not of outcome. Increased mobility would be consistent with evidence of rising expenditures of the lowest income households, as discussed. However, empirical studies do not support the view that the United Kingdom is characterized by greater social mobility than other countries. For instance, the latest OECD Employment Outlook (1996) reports surprisingly uniform distributional mobility across the eight countries examined (United States, United Kingdom, Denmark, Finland, France, Germany, Italy, and Sweden), despite major differences in labor market regulations: around half of those who were still employed in 1991 stayed in the same earnings decile as in 1986, while about a third moved up or down one band, and only the remaining sixth moved two bands or more. This indicates that the more liberalized labor markets of the United Kingdom and the United States are not characterized by greater earnings mobility.

C. Possible Explanations

202. In accounting for a change in income inequality, the “usual suspects” include changes in any of the following:

  • 1. age distribution;

  • 2. household composition;

  • 3. employment structure (e.g. female labor force participation);

  • 4. industrial structure (e.g. decline of traditional industries);

  • 5. unemployment;

  • 6. the business cycle;

  • 7. taxation and social benefits;

  • 8. wage inequality; and

  • 9. the share of non-wage income.

203. A study by Jenkins (1995) evaluates the relative importance of these factors by decomposing changes in inequality indices between 1971 and 1986 in two ways. One approach involves accounting for changes in the mean logarithmic deviation (an index of inequality) in terms of “pure inequality changes”, changes in the number of people in each group, and changes in the relative incomes of these groups. A second approach captures the influences of “income packages” in which the percent change in the inequality index is explained by the percent changes in incomes from different sources. The conclusion drawn was that the aggregate income inequality trends were due to a combination of changes in earnings inequality, employment structure and unemployment, and to a lesser extent, income from capital. On the other hand, this study indicated that secular changes in age distribution, household composition, the distribution of cash social security benefits, income tax payments and employee national insurance contributions, and industrial structure were not to any significant extent responsible for the increase in inequality of family incomes.84

204. Another recent study (Johnson and Webb, 1993) applies a simulation model of the tax and benefit system to quantify how far tax and social security policy had contributed to rising inequality. It also assesses the role of differential rates of earnings growth at different income levels and of changes in the size and composition of the employed and non-employed populations. The study concludes that changes to the personal tax and benefit system, the pretax earning distribution and the pattern of economic activity together explain most of the increase in inequality, with tax and benefit changes having the largest impact.

205. Yet another recent paper (Jenkins, 1996) shows, using a non-parametric technique, that the changes in income distribution resulted from a shift away from employment towards both unemployment and new forms of work (such as self-employment), and a shift in income distribution within each economic status sub-group (e.g. increased inequality growth among working and among non-working families).

206. Although their precise diagnoses differ, these studies agree that increased earnings inequality and rising unemployment are key factors explaining inequality of household incomes. Widening pay inequality in the United Kingdom since the early 1980s can partly be explained in terms of shifts of supply and demand for different types of labor: in the 1980s, the supply of qualified manpower did not rise quickly enough to match the increased demand for educated labor and this excess demand caused the rise in earnings for the high decile groups in the earnings distribution (see Chart 9); at the same time, the supply of less educated labor did not decline fast enough to match declining demand for unskilled labor, explaining the fall in earnings for the lower deciles.

CHART 9
CHART 9

UNITED KINGDOM: Earnings of Top and Bottom Deciles

(Ratio to Median)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: OECD (Employment Outlook) (1993) and Robinson (1994)

207. Most authors believe that structural factors, such as the decline in unionization, must also be brought into account in explaining the widening earnings inequality. Robins (1994) argues that the United Kingdom has instead witnessed a unique set of institutional changes which, in combination with the functioning of labor markets, may have generated greater pay inequality since the late 1970s. The narrowing of the earnings dispersion in the United Kingdom in the 1970s was in part attributable to wage compression associated with incomes policies. Moreover, unions tended to promote equalization of pay among members, and the density of unionization rose through the 1970s. This changed in the 1980s, as the United Kingdom witnessed simultaneously a sharp decline in union influence due to restriction by successive acts of legislation; a reduction in the scope and bite of minimum wage protection (the Wages Councils whose coverage was eroded throughout the 1980s, with minimum pay rates lagging behind the growth in average earnings); a decentralization of pay determination away from the industry level towards the firm; and the abandonment of any attempt by the state to control pay trends in the private sector while maintaining more or less tight controls over public sector pay. Robinson (1994) notes that these changes restored the United Kingdom’s labor market institutions to their pre-1906 state. Moreover, these institutional changes were accompanied by rising aggregate unemployment—albeit a trend that was shared by many other industrialized countries.

208. However, while earnings inequality more or less explain the trend in overall income inequality for much of the 1960s and 1970s, other factors, notably the distribution of investment income, income from self-employment and pensions became increasingly important beginning in the late 1970s (Goodman and Webb, 1994). In the second half of the 1980s in particular, inequality of self-employment income increased greatly and contributed significantly to the rise in overall inequality. Self-employment income makes up a disproportionate share of total income among both the richest and poorest households. Furthermore, the growth of income from dividends and property, especially amongst the richest households, contributed markedly to growing inequality.

209. The United Kingdom’s recent labor market policies has been narrowly targeted to address long-term unemployment (Annex 1), eschewing the broad-based wage subsidization and job creation initiatives introduced elsewhere. According to the OECD (1996), this approach is cost-effective, with a comparatively low share of GDP spent on labor market programs yielding comparatively favorable results: the United Kingdom’s unemployment rate is now more than 2 percentage points below the EU average; and long-term unemployment, a particular target of the programs, declined from a peak of 50 percent of total unemployment in 1985 to 45 percent in 1994 (although still comparing unfavorably with rates below 20 percent in the early 1970s). As the OECD report also notes, however, greater labor market efficiency has been accompanied by a widening of income inequality.

210. The link between labor market programs and inequality is likely to be quite complex, with effects on patterns of employment and activity, earnings distribution, and the incidence of the resulting taxation. However, there is some empirical support for the simplistic proposition that merely spending more on labor market programs can bring about a more equal income distribution. Chart 10 shows amounts spent on labor market programs and the Gini coefficient: the top panel, showing 1985-91 averages for OECD countries, suggests a negative relationship; the bottom panel shows a corresponding time series for the United Kingdom highlighting the fact that, since the mid-1980s, declining spending on active labor market policies has gone hand in hand with rising inequality. Annex II presents some formal empirical tests for a cross-section of countries; the results support a negative relationship between a country’s income inequality, as measured by a Gini coefficient, and the amount spent on active labor market policies.

CHART 10
CHART 10

SELECTED COUNTRIES: Gini Coefficients and Spending on Labor Market Programs

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: OECD (Employment Outlook) and Deininger and Squire (1996)

D. CONCLUSION

211. This chapter confirms that the United Kingdom’s distribution of household income has become more unequal since the labor market reforms of the early 1980s. The rise in inequality—from a more equal starting point—was sharper than in other OECD countries. However, the resulting degree of inequality is difficult to assess: as measured by the Gini coefficient, the United Kingdom’s income distribution is not very different from that of other OECD countries especially in Europe, but the ratio of the income of the richest 20 percent of the population to that of the poorest 20 percent is now greater in the United Kingdom than in any other large Western industrialized nation, including the United States. This indicates that income distribution among middle-income people is comparatively equal in the United Kingdom, but the very rich are richer and the very poor are poorer than in other countries.

212. Household incomes of the very poorest have failed to keep pace with those of the rest of society. For the poorest tenth of the population, real incomes after housing cost decreased by 14 percent between 1979 and 1991, in stark contrast with the 30 percent increase enjoyed by the richest tenth of the population. However, the poorest groups were somehow able to increase their expenditure despite ostensibly declining incomes; this likely reflects a combination of two things: consumption smoothing in the face of temporary shortfalls in income; and underreporting of income, particularly given the increased share of self-employed and unemployed.

213. The composition of the bottom decile has changed—but not in a way that explains away the increasing inequality. The share of the unemployed has risen—reflecting the more than eightfold increase in unemployment between the early 1960s and mid-1980s, and the relatively parsimonious social benefits; that of retired people has meanwhile decreased. Families with children now make up more than half of the poorest decile, compared with only around a third three decades ago.

214. Mobility among earning groups is no higher in the United Kingdom than in other OECD countries.

215. A shift away from employment towards both unemployment and self-employment, increased earnings inequality, tax and benefit regimes, and the pattern of economic activity explain most of the overall change in income inequality. Demographic issues and changes in the family size and structure, in contrast, played only a minor role. The contribution of labor market policies to changes in inequality is complex: the United Kingdom’s labor market reforms emphasizing the flexibility and efficiency of the labor market may have contributed—via widening of the earnings distribution and tightening of unemployment benefits—to the increased income inequality evident in the data for 1991. The upward trend in unemployment between the late 1970s and early 1990s was also an important factor. More recently, however, the United Kingdom’s labor policies have contributed to reductions in the NATJRU (Chapter II) and in actual unemployment; lower unemployment would in turn, to the extent that it reflects higher employment rather than a decline in labor force participation, be expected to reduce income inequality. More recent income distribution data, as they become available, will permit a fuller assessment of the consequences of labor market policies for income inequality.

The Gini Coefficient

216. The Gini coefficient is the most widely-used summary measure of inequality. It is related to the Lorenz curve, the cumulative distribution of income across individuals. The Gini coefficient measures the area between the Lorenz curve and the line of perfect equality (the surface labeled A in the figure) relative to the surface of the entire lower triangle (labeled B). Thus a higher Gini coefficient denotes greater inequality; perfect equality would yield a Gini coefficient of zero, and perfect inequality (where one individual receives all the income) unity.

217. No single number can capture all relevant aspects of income distribution, and alternative summary measures have been devised. In particular, the Gini coefficient is most sensitive to inequality changes around the mean, another measure, the Atkinson index, provides explicit weights for different parts of the income distributions. The sensitivity of the lower part of the income in this index rises with a weighting coefficient (conventionally used weights are 0.5 or 1.0). This weighting reflects social judgements as well as considerations related to the robustness of the results with regard to errors in data.

ANNEX I: Trends in Labor Market Policy

218. The United Kingdom’s labor market policies underwent a dramatic change in the 1980s which have been followed up in the 1990s. The early steps focused on liberalizing the labor market: easing restrictions on hiring and firing, eliminating minimum wages, and scaling back social benefits. More tellingly, the prerogatives of trade unions were sharply curtailed: the Employment Acts of 1980 and 1982 and the Trade Union Act of 1984 ushered in the decentralization of bargaining over wages and working conditions at the firm level, rather than national or industrial level. The legislation also penalized unions for unofficial strikes, banned “secondary actions” so that workers could not participate in disputes to which they were not party, and mandating compulsory balloting of union members in industrial disputes involving stoppages (Mayhew, 1991; Ramaswamy and Prasad, 1994). As a result, the number of working days lost due to industrial disputes declined dramatically, with marked improvements in the wage determination process and the functioning of labor markets more generally. The major black mark on the record was the share of long-term unemployment, which rose from below 20 percent of total unemployment in the early 1970s to about SO percent in 1985.

219. Another important policy goal was to reduce the tax and benefits wedge between employment and unemployment, by reducing national insurance charges for the low-paid and expanding in-work benefits. More recently, employers’ national insurance contributions (NICs) were further altered to favor employment of the part-time, the low paid and the long-term unemployed: as from April 1996, employers are exempt from NICs for up to a year, when hiring someone unemployed for more than two years. Of the in-work benefits, Family Credit (FC) is the most important, adding an average of some £50 per week to family earnings for families on low wages with children. FC is currently received by almost 600,000 families representing a theoretical take-up rate of 81 percent. Some two-thirds of families move off FC because they take on full-time or multiple part time jobs, with an average duration on FC of six months. However, to prevent the loss of FC from becoming a disincentive to increase hours of work, a FC bonus of £10 per week was recently introduced for those working 30 hours a week or more. An extension of in-work benefits to include lower-paid workers without dependants is being piloted in eight areas of the United Kingdom in a scheme called “Earnings Top-Up”. Finally, the financial incentive to work has been strengthened by lowering replacement rates for unemployment benefits, which are set at a flat rate and indexed to retail prices rather than average earnings. Receipt of unemployment benefit has also been made more strictly conditional upon claimants’ actively seeking work.

220. Active labor market policies in the United Kingdom were first introduced in the late 1970s to address high youth unemployment. The focus of policy has been sharpened since the mid-1980s, with a selection from the menu of programs adopted in other countries (Table 1). Spending has been lower than in many other EU countries (Chart 11). The long-term unemployed have been a particular target: programs have sought to reintegrate them in the labor market by helping them find jobs and develop work skills on the job. Long-term unemployment accounting for 40 percent of the total (Chart 12), is concentrated among older males with few labor-market skills and low educational attainment, lone parents, the handicapped and other disadvantaged groups.

Table 1:

Active vs Passive Labor Market Policies.

article image
Source: OECD Employment Outlook.
CHART 11
CHART 11

SELECTED COUNTRIES: Spending on Labor Market Programs, Average 1985-94

(In Percent of GDP)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: OECD
CHART 12
CHART 12

SELECTED COUNTRIES: Standardized Unemployment Rates, Average 1985-94

(In Percent)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: OECD

221. Starting in 1986 with the introduction of Restart, active labor market policies increasingly focused on search and matching. Benefit payments and job search activity were re-integrated and the Employment Service (ES) became the gateway to a menu of programs aimed primarily at helping long-term unemployed by dealing with their individual circumstances.85 These centers devote most of their resources to placing and advisory efforts (e.g. placing claimants with employers and advising claimants on job search), matching job applicants with existing job openings. The United Kingdom has been piloting a wide range of active labor market policies. Some of these, employment subsidies (Workstart), trial employment period with employers (Work Trials) and job introduction methods (Job Introduction Guarantee) are primarily aimed at overcoming employer prejudice towards the long-term unemployed.

222. Most of the money spent on employment programs goes on 10 national schemes (Table 2 and Chart 13). The places on the schemes are largely in the gift of the ES rather than being demand led and the programs are often difficult to divide from the daily placing and advisory efforts which make the programs more efficient. Programs become available at different durations of unemployment, with the intensity of the program (as measured by the commitment of the claimant’s time) increasing with the duration of the employment.

CHART 13
CHART 13

UNITED KINGDOM: Expenditures on Active and Passive Labor Market Policies, 1985-94

(In Percent of GDP)

Citation: IMF Staff Country Reports 1996, 130; 10.5089/9781451814064.002.A006

Sources: OECD
Table II.

Employment Service (ES) Programs and Pilot Programs

article image
article image

223. For young people, the primary focus has been on education and training, with income support for youth being made conditional on participation in training. From 1976 to 1992, the proportion of 16-18 year olds in training schemes rose from zero to 11 percent, while enrollments in full-time education rose from 27 to 47 percent. The main measures in place to help train young people are Youth Training, Youth Credits, and Modern Apprenticeships: Youth Training applies to all 16- and 17-year-olds not in full-time education, training or a job, and guarantees an offer of suitable training for work to at least National Vocational Qualification (NVQ) level 2; Youth Credits are training “vouchers” aimed at 16-17 year olds who have left full-time education; and Modern Apprenticeships are a major reform of the apprenticeship system. The scope of these youth measures is limited by a lack of training places; there are also lingering doubts—confirmed by international studies (OECD Jobs Study, 1996)—about the effectiveness of government-sponsored training programs in improving long-term employment prospects.

224. Currently, there are also pilot schemes to subsidize or create jobs, mostly inspired by the “workfare” ideas. Such programs (e.g. the British Community Program) are designed to strengthen incentives for other employment and develop job skills. Additional benefits would accrue to the extent that the employment provided is productive, although studies suggest that in many such programs the value added is negligible (Normington et al, 1986) and the costs of supervision must also be borne in mind.

225. Chart 9 shows some unemployment measures together with the related labor market policies. Unemployment among the 15-24 age group increased to 15.5 percent in 1995 from 10 percent in 1990,86 while long-term unemployment rose from 2.4 percent of the labor force in 1990 to 4.5 percent in 1993, then declining to 4.0 percent in 1994. This record is not stellar, but nonetheless compares favorably with that of other EU countries that spend much more on labor market programs; it avoids the substantial deadweight costs associated with large-scale government training and employment programs and wage subsidization, which studies also suggest are limited in their effectiveness (OECD Jobs Study, 1996).

ANNEX II: Labor Market Policies and Inequality: Is There a Connection?

226. This Annex reports some empirical findings on the relationship between labor market policies and income inequality. The equation estimated is a reduced form derived from a model in which inequality depends on unemployment and per capita income.87 The latter, in turn, is assumed to evolve over time according to a Neoclassical growth model—incorporating the (endogenous) accumulation of physical capital and technological know-how—and employment is determined by Okun’s law. In this setting, the government can influence unemployment through the share of GDP spent on labor market programs and changes in taxation—specifically, the tax wedge between the cost and the remuneration of labor.

The resulting reduced-form equation for the degree of income inequality is:

log[inequality]=constant+α1log[amount spent on labor market programsGDP]+α2log[investment in physical capitalGDP]+α3log[investment in R&DGDP]+α4log[pop.growth+growth rate of‘technology’+depreciationrate]

227. In estimating this equation, several alternative measures of inequality were examined: the Gini coefficient, the income shares of the top and bottom quintiles, and the ratio of incomes of top to bottom quintile. The amount spent on labor market programs is specified as either mean total amount spent on active programs, or amount spent on passive policies.

228. The equation was estimated for a cross section of OECD countries; the dependent variable is for the most recent observation (1991 in most cases) while the explanatory variables are for averages of the preceding period (generally 1985–91).88 Data on inequality were taken from the databank assembled by Deininger and Squire (1996). Investment shares in physical capital came from the Penn World Tables. For R&D investment shares the OECD Science and Technology Indicators were used. Figures on the tax wedge and labor force growth came from the OECD Jobs Study, and percentages of GDP spent on labor market policies were drawn from the OECD Employment Outlook.

229. Regression results are presented in Table 2. First, spending on labor market policies does not have a significant effect on the Gini coefficient, but they do affect other measures of inequality: spending on active labor market policies significantly improves the income share at the bottom (partly) at the expense of the top. The effect of passive labor market policies, however, is not statistically significant in any of the specifications.

230. The tax wedge has a significant effect on inequality in almost all of the estimated regression equations. A decrease in the tax wedge has the predicted positive effect on the income share of the bottom part of the distribution; this supports the hypothesis that reducing tax distortions has a favorable effect on employment, particularly for lower-paid workers. The tax wedge also has a positive effect on the share of the top quintile; this may either be due to a adding-up effect (a larger share for the bottom quintile necessarily means a lower one for someone else) or possibly that reducing the tax wedge on labor induces substitution away from physical and human capital with which the top income group is more heavily endowed.

231. The results also show that a large part of the difference in inequality are due to broader economic forces, including the accumulation of physical and human capital, the growth rate of the labor force, and technical change; indeed, these variables have a much larger role than labor market policies and taxation in accounting for differences in income inequality across the countries in the sample. First, accumulating physical capital significantly increases inequality, as reflected in a high Gini coefficient, a low income share of the bottom quintile, a high income share of the richest quintile and a high ratio of richest to poorest quintile of the income distribution.89

232. The accumulation of technological know-how—a key factor in recent theories of economic growth—is also found significantly to affect inequality in all equations.90 This result is consistent with the notion of “biased technological change”: new technology makes the unskilled relatively more unskilled, reducing their employment opportunities and their relative wage; while the rewards are to a greater extent reaped by those in the upper part of the income distribution, who hold a disproportionate share of the physical and human capital.

233. The effect of variation in the growth rates of the labor force could not be estimated precisely, and was nowhere significant at the 5 percent level.

234. The simple model was able to explain up to about 90 percent of the variation in the log of the top-to-bottom income ratio. On average the regressions explained about three quarters of the variation in measures for inequality in OECD countries. Caution is needed in interpreting these results, given the small sample size (a cross section of 13 or at most 15 observations), as well as compromises dictated by the limitations of the data set. Taking these caveats into account, the estimates provide an initial assessment of the determinants of inequality in a broader macroeconomic context. The results suggest that labor market policies and taxation are important in accounting for income inequality, but that the underlying determinants of economic growth—accumulation of capital and know-how—are more important in explaining cross-country differences in income distribution.

Table 1.

Description of the Data

article image
Table 2.

Regression Results

article image
Standard errors in parenthesis.

significant at the 10% level

significant at the 5% level or better

All regressions have heteroskedasticity consistent covariance matrices

Table 2.

Regression Results (concluded)

article image
Standard errors in parenthesis.

significant at the 10% level

significant at the 5% level or better

All regressions have heteroskedasticity consistent covariance matrices

REFERENCES

  • Atkinson, A., 1995, “Income Distribution in OECD Countries: Evidence from the Luxembourg Income Study”, in OECD Social Policy Studies #18.

    • Search Google Scholar
    • Export Citation

  • Blundell, R. and Preston, I., 1996, “Income, Expenditures and the Living Standards of UK Households”, Fiscal Studies, 16, 40–54.

  • Cole, Jeffrey, and Christopher Towe, 1996, “Income Distribution and Macroeconomic Performance in the United States,” IMF Working Paper WP/96/97, [August].

    • Search Google Scholar
    • Export Citation

  • Department of Social Security, 1993, “Households Below Average Income: A Statistical Analysis 1979-1990/91”, London, HMSO.

  • Deininger, K. and Squire, L., 1996, “Measuring Income Inequality: A new database”, The World Bank, Washington.

  • Goodman, A. and Webb, S., 1992, “The Distribution of the UK Household Expenditure 1979-92”, Institute for Fiscal Studies, London.

  • Goodman, A. and Webb, S., 1994, “For Richer, For Poorer: The Changing Income Distribution in the United Kingdom 1961-91”, Institute for Fiscal Studies, London.

    • Search Google Scholar
    • Export Citation

  • Jenkins S., 1995, “Accounting for Inequality Trends: Decomposition Analyses for the UK: 1971-86” Economica, 62, 29–63.

  • Jenkins, S., 1996, “Recent Trends in the UK Income Distribution: What Happened and Why?”, Oxford Review of Economic Policy, 12, 29–46.

    • Search Google Scholar
    • Export Citation

  • Johnson, P. and Webb, S., 1993, “Explaining the Growth in UK Income Inequality: 1979-1988”, The Economic Journal, 103, 429–435.

  • OECD, Jobs Study, various issues.

  • OECD, “Unemployment Outlook”, various issues.

  • OECD, 1996, United Kingdom 1995 1995 OECD Economic Surveys (Paris: OECD).

  • Robinson, P., 1994, “Is there an Explanation for Rising Pay Inequality in the UK?”, Center for Economic Performance, Discussion Paper #206.

    • Search Google Scholar
    • Export Citation

REFERENCES

  • Mayhew, K., 1991, “Assessment: the UK labour market in the 1980s”, Oxford Review of Economic Policy, 7, 1–17.

  • OECD, “Jobs Study”, various issues.

  • OECD, “Unemployment Outlook”, various issues.

  • OECD, “Economic Surveys: United Kingdom”, various issues.

  • Ramaswamy R. and Prasad E., 1994, “Shocks and structural breaks: labor market reforms in the United Kingdom”, IMF Working Paper WP/94/152, [November].

    • Search Google Scholar
    • Export Citation

REFERENCES

  • Cook, C., 1995, “Savings Rates and Income Distribution: Further Evidence from LDC’s”, Applied Economics, XXVH, 71–82

  • Deininger, K. and Squire, L., 1996, Measuring Income Inequality: A New Database, The World Bank, Washington DC.

  • Delia Valle, P. and Oguchi, N., 1976, “Distribution, the Aggregate Consumption Function, and the Level of Economic Development: Some Cross-Country Results”, Journal of Political Economy, 84, 1325–34.

    • Search Google Scholar
    • Export Citation

  • Gerzovitz, M., 1988, “Savings and Development” in H. Chenery and T. Srinivasan (eds): Handbook of Development Economics, Amsterdam, North-Holland, 381–424.

    • Search Google Scholar
    • Export Citation

  • Kuznets, S., 1966, “Economic Growth and Income Inequality”, American Economic Review, VL, 1–28.

  • Musgrove, P., 1980, “Income Distribution and the Aggregate Consumption Function”, Journal of Political Economy, LXXXYTII, 504–25.

  • OECD, Jobs Study, various issues.

  • OECD, Unemployment Outlook, various issues.

  • OECD, Science and Technology Indicators (report) III (1989), technical annex, 95-132.

  • Vanhoudt, P., 1996, “Inequality Labor Market Policies, and Economic Growth: Is There a Connection? Empirical Evidence for the OECD Countries,” draft, IMF [September].

    • Search Google Scholar
    • Export Citation
81

Prepared by Patrick Vanhoudt and Timothy Lane.

82

The raw data on inequality in the United Kingdom come from the household income and expenditure surveys by the Central Statistical Office. A widely used source is the department of social security’s “Households Below Average Income” or HBAI database. This database is considered to be of high quality and -despite the HBAI label- covers persons in all income groups. Each year’s data file includes about 7000 households (about 19,000 individuals). The Institute for Fiscal Studies recently extended the series to cover each of the 31 years between 1961 and 1991. Data sources for other OECD countries are e.g. the Luxemburg Income Study files (which are based on the different surveys held in each country) and a recent database on inequality compiled by the World Bank.

83

Implicitly the focus on real income net of housing costs assumes that the CPI does not adequately reflect the share of housing in expenditures of low income families.

84

This contrasts with evidence for other countries, in particular the United States, where changes in demographics and family structure have been found to be important in explaining increasing inequality; see Cole and Towe (1996).

85

Unemployed people are interviewed every six months to monitor—and where necessary to assist in—their search efforts. The interviews also provide the ES with information on the characteristics of the claimants, thereby allowing the ES to monitor fraudulent behavior and impose sanctions including benefit disqualification or to direct claimants to more appropriate benefits in case of health problems.

86

This compares with 5.8 percent in 1990 and 7.4 percent in 1995 for the 25-54 age group.

87

The role of per capita income is motivated by Kuznets’ (1955) conjecture that income inequality increases with per capita income. The model is developed in greater detail in Vanhoudt (1996).

88

For more details on the data, see Table 1.

89

These findings are consistent with earlier studies for the OECD countries. Two earlier studies—Della Valle and Oguchi (1976) and Musgrove (1980)—reported a positive association between investment and inequality for the OECD countries, but not for developing countries—possibly due to poor data quality (Gersovitz, 1988). More recently, Cook (1995) reports a significant positive relationship between inequality and savings for the developing countries after controlling for other differences affecting savings behavior.

90

This relationship has not previously been tested in the literature.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
United Kingdom: Recent Economic Developments
Author: International Monetary Fund
Volume/Issue:
Volume 1996: Issue 130
Publisher:
International Monetary Fund
ISBN:
9781451814064
ISSN:
1934-7685
Pages:
179
DOI:
https://doi.org/10.5089/9781451814064.002
  • View in gallery

    SELECTED COUNTRIES: Gini Coefficients for Household Income

  • View in gallery

    SELECTED COUNTRIES: Ratio of Highest to Lowest Quintile Incomes, Average 1980-92

  • View in gallery

    UNITED KINGDOM: Income Shares by Decile, 1979 and 1990/91

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Changes in Income by Group

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Income Groups Classified by Family Type

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Income Groups Classified by Economic Status

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Household Income by Source

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Growth in Income and Expenditure, 1979-1992

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Earnings of Top and Bottom Deciles

    (Ratio to Median)

  • View in gallery

    SELECTED COUNTRIES: Gini Coefficients and Spending on Labor Market Programs

  • View in gallery
  • View in gallery

    SELECTED COUNTRIES: Spending on Labor Market Programs, Average 1985-94

    (In Percent of GDP)

  • View in gallery

    SELECTED COUNTRIES: Standardized Unemployment Rates, Average 1985-94

    (In Percent)

  • View in gallery

    UNITED KINGDOM: Expenditures on Active and Passive Labor Market Policies, 1985-94

    (In Percent of GDP)

SELECTED COUNTRIES: Gini Coefficients for Household Income

SELECTED COUNTRIES: Ratio of Highest to Lowest Quintile Incomes, Average 1980-92

UNITED KINGDOM: Income Shares by Decile, 1979 and 1990/91

(In Percent)

UNITED KINGDOM: Changes in Income by Group

(In Percent)

UNITED KINGDOM: Income Groups Classified by Family Type

(In Percent)

UNITED KINGDOM: Income Groups Classified by Economic Status

(In Percent)

UNITED KINGDOM: Household Income by Source

(In Percent)

UNITED KINGDOM: Growth in Income and Expenditure, 1979-1992

(In Percent)

UNITED KINGDOM: Earnings of Top and Bottom Deciles

(Ratio to Median)

SELECTED COUNTRIES: Gini Coefficients and Spending on Labor Market Programs

SELECTED COUNTRIES: Spending on Labor Market Programs, Average 1985-94

(In Percent of GDP)

SELECTED COUNTRIES: Standardized Unemployment Rates, Average 1985-94

(In Percent)

UNITED KINGDOM: Expenditures on Active and Passive Labor Market Policies, 1985-94

(In Percent of GDP)