A. Wage Rigidity and the Exchange Rate Regime

1. An important consideration in the United Kingdom’s participation in EMU focuses on the potential cost of losing the nominal exchange rate as a tool for macroeconomic stabilization. In the face of wage rigidities, a flexible exchange rate can potentially reduce unemployment and output loss from real asymmetric shocks, both by buffering the economy in response to temporary shocks and by speeding adjustment in the case of permanent shocks. One of the five “tests” on EMU entry stated by the Chancellor relates to the U.K. economy having sufficient flexibility to cope with asymmetric shocks among the euro area economies. Flexibility in labor markets is considered essential in this regard because of the implications for employment and output growth.²

2. The flexible or fixed nature of a country’s exchange rate regime can significantly influence the costs of adjustment to a shock in the presence of nominal and real wage rigidities. In a flexible exchange rate regime, if the United Kingdom suffers a permanent real shock and nominal wages are sticky, a depreciation of the exchange rate will reduce real wages through higher inflation and, in principle, mitigate negative effects on unemployment and output.³ Under a fixed exchange rate regime or a single currency, nominal wage inertia will initially lead to unemployment and output loss, and the resulting adjustment costs will be higher the less responsive are real wages to unemployment. The costs of giving up exchange rate flexibility under EMU would thus depend in part on how flexibly both nominal and real wages respond to changes in output and labor demand.

3. Against this background, this paper surveys a broad range of indicators of wage flexibility, both macroeconomic and microeconomic. The paper argues that a number of institutional and structural changes in U.K. labor markets during 1980s may have played a role in increasing both aggregate and relative wage flexibility. In particular, deunionization has led to increasingly decentralized wage bargaining which, together with welfare reform, appears to have intensified competition at the lower end of the wage distribution and may have contributed to the better cyclical adjustment of labor markets in the 1990s. The accompanying rise in wage dispersion and income inequality, while indicative of relative wage flexibility, raises concerns about equity, and has prompted measures to strengthen incentives and social safety nets and raise the skill levels of low-income workers.

B. Labor Market Performance in the 1980s

4. Much of the concern over wage rigidity in the United Kingdom arises from the labor market experiences of the 1980s. Following the recession in the early 1980s, real wages in the United Kingdom rose rapidly despite high and rising unemployment; between 1979 and 1984 real wages rose by about 8 percent, while total employment fell by a similar amount. Unemployment rose from 5 percent to 11 percent. The resilience of real wages, given the deepest recession in the post-war period, was surprising. In their 1985 annual report, the OECD noted:

The perceived rigidity of the labor market in the United Kingdom has been held to be one of the major impediments to the solution of the unemployment problem. In particular, the Government sees the failure of real wages to adjust sufficiently to deteriorating employment conditions as the most important reason why more jobs have not been created during the last five years.

5. Consistent with this, cross-country estimates of the response of aggregate real and nominal wages to unemployment showed the United Kingdom to have a high degree of wage rigidity relative to other OECD countries. In a frequently cited study, Layard, Nickell, and Jackman (1991) estimated structural wage and price equations based on data through the mid-1980s and found that real wage rigidity in the United Kingdom exceeded that in Japan, the United States, and all of the other eleven European countries included in the study. Likewise, Coe (1985) estimated that the responsiveness of nominal wages to unemployment was significantly lower for the United Kingdom than for the United States, Japan, Australia, Canada, and the six other European countries examined.

6. The resilience of wages in the 1980s was attributed in a large part to the industrial relations system and the monopoly power of the trade unions. It was sometimes argued that the United Kingdom’s wage formation process was the least desirable from a macroeconomic perspective; with much of the wage bargaining in the early 1980s taking place at the industry level, the United Kingdom’s labor market lacked both the coordination of highly-centralized wage bargaining systems (as in Austria, Belgium, and some Scandinavian countries) and the market-clearing aspects of a highly-decentralized bargaining system (as in the United States). Weakening the power of unions in the UK was seen as one way of achieving greater wage flexibility. An alternative view held that as long as unions played an important role in wage formation, centrally-coordinated wage bargaining (namely through centralized positions adopted by the unions and employers) could secure a greater degree of aggregate wage moderation.⁴

7. In addition to institutional factors, the increase in the share of long-term unemployed was thought to explain partly why wages were unresponsive to high levels of total unemployment. Long-term unemployment, which had been rising for some time, increased sharply during the recession of the early 1980s; the number of men unemployed for more than a year rose from about 50,000 in the mid-60s, to around 250,000 by the late-70s, and exceeded 1 million by 1985 (OECD, 1986). Since the long-term unemployed are generally less active in looking for work, they are thought to exert less downward pressure on wages than short-term unemployed. The absolute number of short-term unemployed was actually falling by the mid-80s, which might partly explain the lack of wage moderation during the recovery despite persistently high total unemployment (Figure 1)

View Full Size

United Kingdom: Unemployment Rates

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: Office for National Statistics1/ Those unemployed for less than 26 weeks.

• Download Figure
• Download figure as PowerPoint slide

C. Structural and Institutional Changes in Labor Markets

8. Since the early 1980s labor market institutions have undergone a number of changes which, along with structural shifts in the United Kingdom economy, may have increased wage flexibility. Two changes in particular—the weakening of trade unions and, to a lesser extent, the weakening of Wages Councils—are thought to have had a significant impact on wage formation.

9. A series of legislative acts during the 1980s and early 1990s limited picketing rights, substantially weakened closed shop rules, imposed greater legal responsibility on unions for their actions, and required pre-strike ballots. With these changes, employers could choose not to recognize unions and bargain directly with individuals. The resulting shift in bargaining power substantially weakened the monopoly power of trade unions (Machin and Stewart, 1996), and led to a sharp decline in union density during that period (Table 1).⁵

^{Table 1.}

Union Density by Sector in Britain ^1/

(in percent)

Source: Nickell (1996).

^1/Defined as union membership (excluding retired and unemployed members) divided by employment.

^{Table 1.}

Union Density by Sector in Britain ^1/

(in percent)

		1968	1979	1987
Total		43.9	55.8	46.3
Public Sector		68.0	83.7	81.7
	Utilities	73.8	94.9	96.4
	Public service	59.7	79.6	78.9
Manufacturing		50.7	72.9	60.1
	Manual	62.8	83.4	65.5
	Non-manual	15.9	46.7	45.3
Construction		32.0	41.4	29.8
Agriculture, forestry, fishing		25.9	24.1	13.9
Private service		12.8	18.1	14.4

Source: Nickell (1996).

^1/Defined as union membership (excluding retired and unemployed members) divided by employment.

View Table

^{Table 1.}

Union Density by Sector in Britain ^1/

(in percent)

		1968	1979	1987
Total		43.9	55.8	46.3
Public Sector		68.0	83.7	81.7
	Utilities	73.8	94.9	96.4
	Public service	59.7	79.6	78.9
Manufacturing		50.7	72.9	60.1
	Manual	62.8	83.4	65.5
	Non-manual	15.9	46.7	45.3
Construction		32.0	41.4	29.8
Agriculture, forestry, fishing		25.9	24.1	13.9
Private service		12.8	18.1	14.4

Source: Nickell (1996).

^1/Defined as union membership (excluding retired and unemployed members) divided by employment.

10. The role of unions in wage formation was further reduced in the 1980s by the significant shift in the structure of employment away from manufacturing towards the service sector. In some sectors, the relocation has been quite large; between 1975 and 1995, the engineering sector experienced a decline of nearly 1.6 million workers, or nearly 50 percent, while the banking sector nearly doubled, expanding by 1.7 million workers (Greenaway et al, 1999). Reduced union powers and the shift away from manufacturing contributed to increasingly decentralized pay determination in the United Kingdom, reflected in fewer contracts covered by multi-employer arrangements and a parallel fall in collective bargaining. (OECD, 1996).

11. Another set of reforms aimed at reducing the importance of sectoral minimum wages. Throughout the 1980s the Wages Councils continued to set minimum wages for workers in a number of low-wage industries covering about 2.5 million workers, primarily in clothing manufacture, retail trade, and other services. The importance of the Councils—as measured by the mean ratio of minimum wages to average wages—declined throughout the 1980s from a peak in 1981 (Figure 2); the system was finally abolished in 1993 (Dickens et al, 1993). While the impact of the weakening and elimination of Wage Councils on wage formation is not as important as the decline in trade unions, the effect would have been concentrated at the bottom end of the skill distribution. As these workers are most affected during downturns, eliminating wage floors could enable the less-skilled unemployed to exert greater pressure on wages, possibly increasing aggregate wage flexibility.

View Full Size

United Kingdom: Indicators of Labor Market Institutional Change

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: Nickell (1996); and Dickens et al (1993).

• Download Figure
• Download figure as PowerPoint slide

12. While these changes affect the flexibility of wage formation directly, a number of changes may have also increased wage flexibility by increasing the competitive pressure of unemployment on wages. Perhaps the most significant is the reduction in coverage and generosity of social security benefits. Unemployment benefits in the United Kingdom fell sharply during the 1980s (Figure 2), eligibility requirements were tightened, and the unemployed were put under greater pressure to find work.⁶ There is ample evidence that a generous benefit system leads to higher unemployment (Nickell, 1997), but it can also reduce the responsiveness of wages to a given level of unemployment by reducing the incentive for the unemployed to compete for vacancies. Lowering and tightening benefits could increase wage flexibility by intensifying competition for jobs during cyclical downturns.

D. Labor Market Performance in the 1990s

13. While the direct effect of these and other changes is difficult to gauge, labor market outcomes in the 1990s have generally been consistent with greater wage flexibility. Not only has the cyclical performance of the labor market improved, but there is ample evidence that relative wages have adjusted flexibly to clear labor markets in response to considerable structural changes in the economy. The two developments are related, and can be attributed in part to changes to labor market institutions which took place in the 1980s.

Aggregate labor market performance

14. Greater wage flexibility may be one reason for the improved cyclical performance of the labor market. The recession of the early 1990s was followed by more moderate real wage growth and a more rapid turnaround in unemployment in comparison with the 1980s (Morgan, 1996). As discussed above, in the 1980s both long- and short-term unemployment continued to grow for a surprisingly long period into the recovery, with the unemployment rate peaking at over 11 percent in 1986. In the 1990s, however, unemployment began falling significantly in the second year of the recovery (Figure 3), and now stands at 4 percent.⁷ Nearly ¾ of the improvement in unemployment performance during the initial output recovery came from a faster recovery in employment (Figure 3). In addition to the faster recovery in employment, wage growth has been more moderate in the later recovery. Greater wage moderation is particularly apparent in the manufacturing sector, with wages growing at an annual rate of 2 percent in the first four years of the recovery compared to 5 ½ percent in the earlier recovery (Figure 3).

View Full Size

United Kingdom: Labor Market Cyclical Performance

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: Office for National Statistics; Obstfeld and Peri (1998); and staff calculations.1/ An index of average earnings deflated by the retail price index excluding mortgage costs.

• Download Figure
• Download figure as PowerPoint slide

15. Along with the improved aggregate performance, there has been a marked narrowing of the dispersion of unemployment rates across regions (McCormick, 1997). Most of the narrowing occurred during the first years of the 1990s recovery, when the fall in the unemployment rate tended to be strongest in the regions with above-average unemployment (Figure 3). The fall in the dispersion of unemployment in the latest recovery has been sharper in the United Kingdom than any of the other major European countries, with the result being that the United Kingdom currently ranks among the countries with the lowest regional disparity of unemployment rates (Table 2).

^{Table 2.}

Cross-Country Comparison of Unemployment Disparities by Region

Source: Soltwedel et al (1999).

^1/Regions according to the Eurostat regional classification NUTS level 2.

^2/State level.

^{Table 2.}

Cross-Country Comparison of Unemployment Disparities by Region

	Standard deviation			Coefficient of variation
	1985	1990	1995	1985	1990	1995
Belgium	2.3	2.7	3.2	0.19	0.36	0.34
Greece	2.2	2.2	2.5	0.37	0.37	0.33
France	1.8	1.7	2.0	0.17	0.19	0.18
West Germany	2.3	1.9	1.8	0.30	0.36	0.26
Whole Germany	…	…	3.5	…	…	0.42
Italy	3.5	6.5	6.8	0.38	0.64	0.55
Spain	4.8	6.1	5.7	0.23	0.38	0.26
Netherlands	1.7	1.8	1.0	0,16	0.23	0.13
United Kingdom	2.9	3.5	1.6	0.23	0.43	0.18
Europe 1/	5.0	5.1	5.9	0.47	0.59	0.55
United States 2/	1.9	1.1	1.3	0.27	0.21	0.24

Source: Soltwedel et al (1999).

^1/Regions according to the Eurostat regional classification NUTS level 2.

^2/State level.

View Table

^{Table 2.}

Cross-Country Comparison of Unemployment Disparities by Region

	Standard deviation			Coefficient of variation
	1985	1990	1995	1985	1990	1995
Belgium	2.3	2.7	3.2	0.19	0.36	0.34
Greece	2.2	2.2	2.5	0.37	0.37	0.33
France	1.8	1.7	2.0	0.17	0.19	0.18
West Germany	2.3	1.9	1.8	0.30	0.36	0.26
Whole Germany	…	…	3.5	…	…	0.42
Italy	3.5	6.5	6.8	0.38	0.64	0.55
Spain	4.8	6.1	5.7	0.23	0.38	0.26
Netherlands	1.7	1.8	1.0	0,16	0.23	0.13
United Kingdom	2.9	3.5	1.6	0.23	0.43	0.18
Europe 1/	5.0	5.1	5.9	0.47	0.59	0.55
United States 2/	1.9	1.1	1.3	0.27	0.21	0.24

Source: Soltwedel et al (1999).

^1/Regions according to the Eurostat regional classification NUTS level 2.

^2/State level.

16. Consistent with the better cyclical performance of the labor market in the 1990s, recent estimates of the responsiveness of aggregate real wages to unemployment show a greater degree of real wage flexibility for the United Kingdom than previous estimates. Using data through the mid-90s and employing structural vector autoregression techniques which attempt to address some of the econometric problems of earlier approaches, more recent studies have found the United Kingdom to rank in the top half of European Union countries in terms of real wage flexibility. (Vinals and Jimeno, 1996, Berthold et al, 1999). Although highly sensitive to the modeling assumptions, these more recent estimates contrast the relative inflexibility found in the Layard et al (1991) study referred to earlier, and are consistent with some increase in wage responsiveness to unemployment. Another indicator of the United Kingdom’s wage flexibility in comparison to other countries is the relatively rapid speed of adjustment of real wages. The OECD (1999) estimates that in response to a demand shock, half the labor market adjustment towards long-run equilibrium is completed in 1¾ years in the United Kingdom, compared to 1¼ years for the United States and 4 ½ years for the euro-area; the finding of relatively rapid adjustment in the United Kingdom, according to the OECD, is robust across different estimates of aggregate wage equations.

Widening earnings dispersion and relative wage flexibility

Widening earnings dispersion

17. In addition to better macroeconomic performance, the dispersion of wages has increased sharply since the early 1980s across skill levels, occupations, sectors, and regions. This has corresponded with greater diversity of work patterns, particularly a wider dispersion of working hours and a greater proportion of short-term jobs (OECD, 1996). These changes are one sign that labor markets are able to respond flexibly to changes in relative demand for different types of workers.

18. One of the consequences of greater wage dispersion has been an increase in income inequality. It should be noted that in principle, greater income inequality by itself is not a sign of greater wage flexibility, nor does an increase in flexibility necessarily require a rise in inequality. However, the sorts of changes in relative labor demand which have raised income inequality in the United Kingdom have also occurred in many other industrial countries. As in Germany’s case which will be discussed below, countries that have limited the rise in inequality by preventing relative wages from adjusting have often experienced greater unemployment, particularly for low-skilled workers. In light of this, the rise in inequality, while indicative of relative wage flexibility, underscores the need for strengthening incentives and social safety nets and raising the skill levels of low-income workers.

19. The widening of the income distribution in the United Kingdom since the early 1980s has been particularly striking (Figure 4).⁸ Despite rising real earnings, inequality as measured by the dispersion between the lower and upper deciles of the earnings distribution grew sharply, with the most rapid growth occurring between 1983 and 1986. Although a similar increase in inequality has occurred in number of industrial countries in the 1980s (Figure 4), the increase has been significantly greater in the United Kingdom despite the fact that, unlike the United States, real earnings in the United Kingdom increased in absolute terms in the bottom deciles during that period.

View Full Size

Growth in Earnings Inequality

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and Morgan (1996).1/ Five-year change in ratio of the upper limit of the ninth and fifth deciles.2/ Five-year change in the ratio of the upper limit of the first and fifth deciles.

• Download Figure
• Download figure as PowerPoint slide

20. The growth in earnings inequality has been accompanied by a significant increase in the dispersion of wages across occupational categories, sectors, and regions. There has been a clear decline in relative wages for the lowest paid occupations between the early 1980s and the mid-1990s (right panel, Figure 5), despite real wages for most these groups which have risen.⁹ Moreover, the distribution has become more skewed as relative wages in the upper tail of the distribution have become less compressed. The widening wage dispersion appears to be strongest for manual workers, where the decline in relative wages of the lowest-paid occupations has been particularly pronounced (Figure 6). For non-manual workers, the growth in dispersion has been less pronounced, although the high end of the distribution has grown fairly rapidly (Figure 7); earnings for specialized managers, such as treasurers, marketing and sales managers, computer systems and data processing managers, and investment analysts, have experienced particularly rapid growth in the 1990s.

View Full Size

United Kingdom: Earnings Distribution Across Occupations ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all workers.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

United Kingdom: Earnings Distribution Across Occupations ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all manual workers.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

United Kingdom: Earnings Distribution Across Occupations ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all non-manual workers.

• Download Figure
• Download figure as PowerPoint slide

21. A similar pattern can be seen for wages across sectors. There has been a clear widening of the earnings distribution for manual workers, with relative wages falling for low-wage sectors and rising for high-wage sectors. This widening has occurred despite the fact that the absolute level of real wages for manual workers has risen in the lowest-paying sectors (Figure 8). Within manufacturing, the increase in absolute real wages for manual workers is particularly apparent, although there has been less of an increase in earnings dispersion across manufacturing sectors (Figure 9). For manual workers in nonmanufacturing sectors, the increase in dispersion is mainly accounted for by rising relative wages in the high-wage sectors such as transport and communications (Figure 10). The absolute wage increase for non-manual workers across sectors is also quite striking, even though the dispersion has risen only modestly (Figure 11). Rising wages in the non-manufacturing sectors appears to account for a large part of the wage growth for non-manual workers (Figure 12).

View Full Size

United Kingdom: Earnings Distribution Across Sectors ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all manual workers.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

United Kingdom: Earnings Distribution Across Sectors ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all manual workers in manufacturing.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

United Kingdom: Earnings Distribution Across Sectors ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey, and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all manual workers in non-manufacturing sectors.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

United Kingdom: Earnings Distribution Across Sectors ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all non-manual workers in all sectors.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

United Kingdom: Earnings Distribution Across Sectors ^1/

Citation: IMF Staff Country Reports 2000, 106; 10.5089/9781451814149.002.A004

Source: New Earnings Survey; and staff calculations.^1/ Earnings distribution for full-time adult males. Wages for the 1980’s are the average for the period 1981-1985, and for the 1990’s, the average for the period of 1995-1998. Category definitions change between periods, so the figures are not strictly comparable.2/ Weekly earnings including overtime in 1990 pounds deflated by the RPIX.3/ Absolute wage relative to the wage for all non-manual workers in non-manufacturing sectors.

• Download Figure
• Download figure as PowerPoint slide

22. As with occupations and sectors, earnings dispersion across regions has increased, with the ratio of the standard deviation to mean earnings for males rising from 5 percent in the 1980s to 9 percent in the 1990s (Table 3). This appears to be largely driven by rising dispersion for non-manual workers, with little sign of rising dispersion across regions for manual workers. However, rising regional dispersion is mainly accounted for by a growing earnings differential between the South East and the rest of Great Britain; growth in earnings dispersion excluding the South East has been significantly less than overall growth.

^{Table 3.}

United Kingdom: Wage Dispersion Across Regions ^1/

Coefficient of variation

(in percent)

Source: New Earnings Survey; and staff estimates.

^1/Based on Average Weekly Earnings including overtime for full-time adult males.

^{Table 3.}

United Kingdom: Wage Dispersion Across Regions ^1/

Coefficient of variation

(in percent)

		1975-85	1993-96
All regions
	All workers	5.2	8.9
	Manual workers	3.7	3.4
	Non-manual workers	5.2	8.3
Excluding South East
	All workers	1.9	2.2
	Manual workers	3.3	1.5
	Non-manual workers	1.6	2.4

Source: New Earnings Survey; and staff estimates.

^1/Based on Average Weekly Earnings including overtime for full-time adult males.

View Table

^{Table 3.}

United Kingdom: Wage Dispersion Across Regions ^1/

Coefficient of variation

(in percent)

		1975-85	1993-96
All regions
	All workers	5.2	8.9
	Manual workers	3.7	3.4
	Non-manual workers	5.2	8.3
Excluding South East
	All workers	1.9	2.2
	Manual workers	3.3	1.5
	Non-manual workers	1.6	2.4

Source: New Earnings Survey; and staff estimates.

^1/Based on Average Weekly Earnings including overtime for full-time adult males.

Implications for relative real wage flexibility

23. What do these developments say about wage flexibility? First, they show that U.K. labor markets can flexibly respond to changes in relative demand for different types of workers. Second, changes in labor market institutions which contributed to greater wage dispersion may also account for part of the improved macroeconomic performance in the 1990s.

24. It is generally accepted that the main contributor to the increase in earnings inequality in the United States and other industrial countries is an increase in the relative demand for skilled workers, mainly through skill-biased technical change. Berman et al (1998) argue that the shift in demand for skilled workers is pervasive throughout the developed world as similar sectors across countries engage in skill-upgrading related to the use of microprocessors. Evidence on the change in the structure of the U.K. labor market generally supports a shift in demand for skilled workers as an engine for increasing earnings dispersion. Nickell (1996) finds that differences in real wage growth across occupational categories are much larger than across industries, with the real wages of non-manual workers (who are generally more skilled) growing much more rapidly in the 1980s than those of manual workers. Wage growth across occupations is also strongly positively correlated with growth in employment. The widening of wage differentials across skill categories then is one sign that wages have been flexible in response to large changes in relative demand.¹⁰

25. Changes in labor market institutions in the United Kingdom appear to have also increased relative wage flexibility at the lower end of the distribution, and may partly explain why the increase in income inequality has been larger in the United Kingdom than other industrial countries. Both trade unions and Wage Councils can have a compressing effect on the wage distribution; Wage Councils through sector minimum wages, and unions by organizing lower-paid workers and attempting to standardize pay for given job definitions.¹¹ Indeed, Gosling and Machin (1994) find earnings dispersion in the United Kingom is lower within the union sector than the non-union sector, and that dispersion is lower within unionized plants than non-unionized plants. That being the case, a decline in unionization should lead to a widening of the earnings distribution; Gosling and Machin estimate that for semi-skilled workers, the fall in unionization accounts for about ¼ of the increase in earnings dispersion between 1984 and 1990. Most of this increase reflects a decline in unionization rather than rising dispersion within unionized sectors; Machin (1997) finds that dispersion grew much faster in non-unionized plants, and where unions still exist they have much the same effect on earnings distribution as before. Similarly, Machin finds that dispersion rose by less in sectors covered by Wage Council minima than other sectors.

26. These institutional changes, together with welfare reform, may also account for part of the improved macroeconomic performance in the 1990s. As already noted, tightening and reducing benefits could intensify competition for jobs during cyclical downturns, and declining unions and the elimination of sector minimum wages would allow wages at the lower end of the earnings distribution to adjust more flexibly. Flexibility for lower-wage workers is particularly important for aggregate labor market performance, as these workers are likely to be disproportionately affected by cyclical downturns.¹²

27. There is some direct evidence that these institutional changes have led to better labor market performance and contributed to aggregate wage moderation. Anderton (1997) finds evidence that for many sectors the long-term unemployed put more downward pressure on wages from 1986 onward, coinciding with active labor market programs such as “Restart” aimed at integrating the long-term unemployed into the labor market. Gregg and Wadsworth (1995) provide results consistent with intensified competition at the lower end of the wage distribution as they find that the wages for “entry positions” have become concentrated at the lower tail of the aggregate distribution. Also consistent with greater competition, Anderton (1997) reports that since 1986 about 80 percent of the employment growth has been concentrated in jobs paying significantly less than the average wage. Supporting the role of deunionization in increasing wage responsiveness, Nickell and Kong (1992) find that unemployment has a greater impact on wages in industries where union power is weak. Likewise, Blanchflower and Oswald (1994) find that in micro-data-based wage curve estimates, the pay of weak bargaining groups such as the relatively unskilled and non-union workers tend to display the greatest responsiveness to unemployment. In summary, there are reasons to expect that the institutional changes which contributed to the rise in wage dispersion also had the effect of increasing aggregate wage flexibility.

28. It is useful to compare the United Kingdom’s experience with wage dispersion and cyclical performance with that of Germany. It is likely there has been the same sort of increase in relative demand for skilled workers in Germany as in the United Kingdom and other industrial countries (Machin and Reenan, 1998). In contrast to the United Kingdom, however, Germany experienced a fairly steep decline in the dispersion of wages in the latter half of the 1980s (Figure 4), and the wage distribution has remained remarkably stable throughout the 1990s. Much of the difference from the U.K. experience appears to be accounted for by differences in wage bargaining structures; wage bargaining in Germany is more coordinated than in the UK at both the employer and the union level, and unions have traditionally set effective wage floors and negotiated uniform relative increases for workers of all skill levels (Nickel, 1997), thereby constraining the flexibility of the wage structure. Indeed, unemployment for unskilled workers has risen sharply in Germany in the 1990s. While some of the rise no doubt reflects the effects of German unification, it is also the case that employment and retention rates for unskilled workers in Germany have continued to fall during the recent recovery, in sharp contrast to the rising employment rate for skilled workers (IMF, 1999b). This is precisely the outcome one would expect if a rigid wage structure prevented labor market adjustment through changes in relative wages.

E. Summary

29. The U.K. labor market has undergone a number of institutional and structural changes since the 1980s, and there are reasons to believe these changes have contributed to increasing both aggregate and relative wage flexibility. Most importantly, wage bargaining has become increasingly decentralized as a result of deunionization which, together with welfare reform, appears to have intensified competition at the lower end of the wage distribution. Consistent with this, labor markets in the 1990s have shown a high degree of flexibility. Relative wages have adjusted flexibly to large changes in demand for different types of workers, and labor markets have shown better cyclical adjustment to the recession in the 1990s in comparison to the 1980s.

30. How much institutional changes in the labor market contributed to this better performance is not clear; much of the narrowing in the regional dispersion of unemployment, for example, may be attributed to a business cycle upturn that was less biased against manufacturing than in the 1980s. Moreover, structural changes in the U.K. economy, such as the shift away from manufacturing, may themselves contribute to better macroeconomic performance to the extent that wages in the service sector are more flexible and the shocks less concentrated. However, both the improved aggregate performance of the 1990s and changes in the structure of relative wages across skills, sectors, occupations and regions are consistent with the view that institutional changes in the 1980s have increased the flexibility of wages in response to changes in demand for output and employment. This suggests that the potential costs of joining EMU stemming from a lack of sufficient wage flexibility in the United Kingdom may have diminished in recent years.