A. Introduction

130. The new Labour government pledged to introduce a national minimum wage aimed at ensuring fairness in the labor market. While the decision to set a minimum wage has already been taken, the level has not yet been determined. A Low Pay Commission has been established to sift the evidence and make a recommendation on an appropriate level.⁵²

131. Much of the public discussion of a minimum wage in the United Kingdom has focused on “half average earnings” as a benchmark. This principle, however, implies a wide range of actual levels, depending on whether a median or a mean is used; whether it is based on all workers or only males; whether it is limited to regular earnings or includes average overtime hours; and whether it is based on data from the Family Expenditure Survey (FES) or from the 1996 New Earnings Survey (NES) which undersamples part-time and lower-paid workers (Gosling, 1996). The figure favored by the trade unions is calculated by taking half male median full-time weekly earnings (including overtime) from the NES, and dividing by average hours excluding overtime—yielding a minimum wage of around £4.50. In contrast, a median based on FES data on regular pay for both male and female workers would yield a figure slightly over £3.⁵³

132. The main concern over a minimum wage is its possible effects in aggravating unemployment, particularly among young and unskilled workers whose present earnings are low. For instance, the OECD Jobs Study argued that statutory minimum wages could have serious adverse effects on employment opportunities for some groups; evidence reported in a review process for some countries (France, Canada, and the Netherlands) confirmed that minimum wages constrained employment for the young and less-skilled. Thus, if the minimum were set too high, it could frustrate the efforts to alleviate youth unemployment that are a central feature of the Welfare-to-Work program (see Chapter IV). Employment effects of the minimum wage have been debated, with some recent empirical studies supporting the view that these effects are small or even positive.

133. A second key issue is whether a minimum wage would achieve its intended effect on income distribution—that is, increase the incomes of the poor. Many low-earning individuals are not from low-income households, many of which have no employed members. Moreover, raising the minimum wage would reduce the means-tested benefits to which a worker is entitled, attenuating the net effect on his or her income.⁵⁴

134. A third issue is the possibility of differentiation of the minimum wages: in particular, a lower wage for youth. The government has not ruled out such differentiation, and its likely effects would need to be explored.

135. This chapter discusses each of these issues in turn. It concludes that a minimum wage is most likely to have the desired effect on the incomes of the poor if it is set at a low level—perhaps around £3, at the lower end of the range under discussion. The overall effects on income distribution include a combination of direct wage effects and employment effects, and the latter become increasingly difficult to predict, the higher the level at which the minimum wage is set. Moreover, a lower minimum wage has a relatively greater impact on the incomes of the working poor, while higher levels of the minimum tilts these effects more toward not-so-poor households with multiple wage earners.

136. Moreover, even if a higher minimum wage results in a net increase in the incomes of poor families, it has some important drawbacks compared with other means of achieving this objective—such as direct income transfers or refundable tax credits. Its adverse employment effects and the lack of targeting of income effects to poor families implies that its effects are not evenly distributed. It also understates the true cost of government intervention in the economy by allocating resources for public purposes through regulation on the private sector rather than more transparently through the budget.

B. Employment Effects

137. Any assessment of both the distributional and the efficiency effects of a minimum wage depends on its effects on employment. The levels being contemplated for the United Kingdom, based on alternative definitions of half median earnings, are in a similar range to those in other European countries (Table 1): a level of £4 per hour (half male median hourly wage) would be in a similar range to that in France and well above that in the United States. It may also be noteworthy that about 12 percent of workers now earn less than £3.50 per hour (Table 2)—similar to the percentage in France earning at or near the minimum wage; a minimum wage of £4 would affect one fifth of the workforce. These comparisons prove nothing in themselves, but do suggest that the employment effects of the minimum wage are potentially of concern.

^{Table 1.}

Statutory Minimum Wages: Selected OECD Countries, 1995

(In percent of wages per working hour, manufacturing)

Source: OECD.

^{Table 1.}

Statutory Minimum Wages: Selected OECD Countries, 1995

(In percent of wages per working hour, manufacturing)

Belgium	65
Canada	38
France	55
Netherlands	58
New Zealand	43
United States	34

Source: OECD.

View Table

^{Table 1.}

Statutory Minimum Wages: Selected OECD Countries, 1995

(In percent of wages per working hour, manufacturing)

Belgium	65
Canada	38
France	55
Netherlands	58
New Zealand	43
United States	34

Source: OECD.

^{Table 2.}

Workers Aged 19-59 (22-59) Affected by Given Levels of Minimum Wage

(In percent)

Source: IFS, Family Expenditure Survey for calendar year 1995 (in December 1996 prices).

^{Table 2.}

Workers Aged 19-59 (22-59) Affected by Given Levels of Minimum Wage

(In percent)

	£3.00	£3.50	£4.00	£4.50
Men	2.6 (1.9)	5.3 (4.0)	9.5 (7.3)	14.1 (11.5)
Women	8.2 (7.6)	18.6(17.1)	29.0 (26.9)	37.6 (35.3)
All	5.4 (4.8)	12.1 (10.7)	19.4 (17.3)	26.1(23.6)

Source: IFS, Family Expenditure Survey for calendar year 1995 (in December 1996 prices).

View Table

^{Table 2.}

Workers Aged 19-59 (22-59) Affected by Given Levels of Minimum Wage

(In percent)

	£3.00	£3.50	£4.00	£4.50
Men	2.6 (1.9)	5.3 (4.0)	9.5 (7.3)	14.1 (11.5)
Women	8.2 (7.6)	18.6(17.1)	29.0 (26.9)	37.6 (35.3)
All	5.4 (4.8)	12.1 (10.7)	19.4 (17.3)	26.1(23.6)

Source: IFS, Family Expenditure Survey for calendar year 1995 (in December 1996 prices).

138. If a minimum wage does have an adverse effect on employment (or on the hours of part-time workers), this may outweigh the direct effect of higher wages on poverty. Unemployment and inactivity are already the leading reasons for poverty: in 1991, roughly 30 percent of the bottom decile of households had no employed member, and fewer than 5 percent had at least one member employed full-time (Goodman and Webb, 1994). Thus the scope for raising the incomes of the poorest through higher minimum wages may be limited, while the poorest may bear the brunt of any associated reduction in employment.

139. The predicted effect of minimum wage laws is based on the standard neoclassical analysis of labor markets. The minimum wage fixes the price of labor (or of a category of labor) above its market-clearing level. The resulting employment effect then depends on the elasticities of labor supply and demand. This may only, in part, be reflected in a rise in unemployment; it may also result in a decline in participation rates (as displaced workers become discouraged).⁵⁵

140. However, the predicted effect of minimum wages on employment does not show up very strongly in the empirical literature (see e.g., Brown, Gilroy, and Kohen, 1982). Studies of the effects on teenage unemployment have typically uncovered elasticities of the order of 0.1 to 0.3—implying that a 10 percent increase in the minimum wage would lead to a 1-3 percent decrease in employment. Lower figures have been estimated for young adults and for low-wage industries and regions. Few significant estimates have been obtained for aggregate employment or for adults in general. A study of the United Kingdom’s system of Wage Boards and Councils—a set of industry-specific minimum wage regulations that in 1990 covered 7 percent of employees, and were finally abolished in August 1993—found that eliminating this system lowered the wages of those affected but had little effect on employment (Machin and Manning, 1994).⁵⁶

141. Some recent studies in the United States have gone further: in particular, controversial studies by Card and Krueger (1994, 1995) have found positive effects on employment. One of their studies, in particular, found that after New Jersey’s minimum wage was raised in April 1992, fast-food chains in the state raised their employment in the state relative to Pennsylvania (where there was no change in minimum wage). The methodology followed by Card and Krueger has, however, been extensively criticized (e.g., Neumark and Wascher, 1995).

142. Empirical results suggesting zero or positive employment effects have led to a search for an explanation. If employers have market power in the labor market, however, a minimum wage may actually increase employment and economic efficiency. In the case of monopsony, employers face an upward-sloping supply of labor, and may artificially restrict employment (below the competitive market-clearing level) in order to reduce wages. In this case, an appropriately set minimum wage can create the right incentives for employers to increase employment toward the level that would emerge in a competitive market. Economists have typically viewed monopsony as a curiosum, whose relevance (if any) is limited to special cases such as “company towns.” Recently, however, some prominent labor economists have argued that it should be taken more seriously (Card and Krueger, 1995; Machin and Manning, 1996), particularly in a dynamic context where it is costly for workers to change jobs. In this setting, an employer could have room to lower a worker’s wage somewhat below the prevailing level without inducing him/her to quit.⁵⁷ It seems implausible a priori that such costs would be large enough in the case of low-wage (that is, unskilled and predominantly part-time) workers to account for large discrepancies in wages, but this is an empirical question.

143. Empirical studies examining monopsony have typically found effects that are quite small (Boal and Ransom, 1997). Elasticity of labor supply facing the individual firm has been estimated at (at most) about 0.03 to 0.04, implying effects on wages of a similar magnitude (i.e., some 3-4 percent). In classic cases such as mining company towns, more substantial effects are found (up to about 15 percent), but even these are still not large in relation to the range of minimum wages now being contemplated in the United Kingdom.

144. While the Card and Krueger study suggested zero or positive employment effects, other studies have indicated a much larger negative impact. A recent study (Abowd et al., 1997) uses longitudinal individual wage and employment data to model flows into and out of employment for the United States and France. This study found significant and large employment effects: in France, a 1 percent increase in the minimum wage resulted in 214 percent decline in the probability of employment of a young man currently employed at the minimum wage; in the United States, it had a 2.2 percent effect on the probability that an employed person came from nonemployment.

145. The main conclusion of the empirical literature appears to be that the employment effects of the minimum wage are difficult to identify empirically and may be fairly small. However, as regards adult minimum wages, the range of historical variation may be too small to permit its effects to be distinguished from those of other determinants of employment (Dolado et al, 1996). Uncertainty about the employment effects would not be a reason for ignoring such effects, but rather militates in favor of caution in setting its level, at least initially.

C. Income Distribution

146. Since the aim of setting a minimum wage is usually to raise the incomes of the poor, it is important to consider which workers and households would likely benefit directly from a minimum wage. There are several dimensions to this issue.

147. One important conclusion is that most of the benefits of a national minimum wage would not accrue to the poorest households due to the prevalence of unemployment and non-participation. As already mentioned, most families in the bottom decile have no full-time wage earners, and only a third have any wage earners at all. For this reason, a minimum wage of £4.50, for instance (which would affect 30 percent of the workforce as a whole) would raise fewer than ½ percent of households (benefit units) out of poverty (defined as less than one third of mean income). Only about 3 percent of households in the bottom decile would gain anything at all from a minimum wage of £3, while about 8 percent would gain from a minimum wage of £4.50. A larger proportion of households in any of the richest five deciles would receive higher incomes as a result of a minimum wage (Gosling, 1996).⁵⁸ Moreover, the incidence on family incomes would depend on the level of the minimum wage: a £3 minimum wage would have its greatest effect on the lowest paid, while a wage of £4 or £4.50 would have greater benefit for higher deciles—since families in the upper deciles are more likely to include multiple wage earners, some of whom would benefit from a higher minimum wage.

148. Many of the workers earning a minimum wage would be part-timers, and fewer than one quarter of low-wage earners are heads of household. For a wage set at £4, some 45 percent of beneficiaries would be part-time women and 8 percent part-time men (Freeman, 1996); many of these part-timers are in households with multiple earners, most of which are not poor. A large percentage of low-wage earners are women, however, so that a minimum wage would militate in the direction of gender pay equality (Machin and Manning, 1996).

149. The transient nature of much low-paying work spreads the impact of the minimum wage more widely, but also more thinly. A recent study by Sloane and Theodossiou (1996) found that much of low-paid employment is transitory, and analyzed the determinant of transition. Of the households in the British Household Panel Survey, only 55 percent of those in low-paying jobs (third decile or lower) in 1991 were still in low-paying jobs in 1993, with 15 percent moving into higher-paying jobs and the remaining 30 percent moving into other categories. High mobility implies that relatively more households would benefit somewhat from a minimum wage over a longer period (Gosling, 1996). It would also, however, further reduce the extent to which a minimum wage would raise the incomes of the poor.

150. In-work benefits further attenuate the effect of a minimum wage on poverty. Many of the low-paid are receiving in-work benefits (notably family credit), which is withdrawn as family income increases. Such benefits result in an effective marginal tax rate of over 70 percent for about 650,000 people, so that much of the effect of a higher minimum wage would be absorbed in a withdrawal of benefits rather than in an increase of the incomes of the working poor. This would essentially transfer some of the responsibility for alleviating poverty from the government to private employers—a less transparent form of government involvement—with limited effects on poverty. This suggests that if the minimum wage is to have its intended effect on poverty, it must be coordinated with reform of the tax and in-work benefits system—which the new government is planning to undertake.

151. More generally, the minimum wage should be compared with other means of achieving distributional objectives, such as direct income transfers or refundable tax credits. In principle, the latter can be better targeted and have fewer distortionary effects than the minimum wage. Moreover, they are more transparent: they place the costs of income redistribution directly on the budget, rather than hiding it in the form of a regulatory burden on private employers.

D. Differentiation of Minimum Wages

152. A national minimum wage may exacerbate regional unemployment problems, as it limits scope for wage differentials to reflect productivity differences. In principle, this would be a case for allowing regional differentiation in the minimum wage. However, in the United Kingdom the only sizeable regional wage variations are for London and the South-East (on the high side) and for Northern Ireland (on the low side); regional differentiation on such a basis is not on the agenda. However, in the absence of regional differentiation, it will be important to set the national minimum at a level that does not impose undue strain on disadvantaged regions.

153. Another option, which the government has not ruled out, is to set a lower minimum wage for youth. This would take account of the fact that young workers may be less productive while they are acquiring job skills. Setting a uniform national minimum wage could contribute to youth unemployment, as corroborated by empirical evidence indicating that the adverse impact of the minimum wage on employment is much stronger and well-identified for teenagers and young adults than for adults (Brown, Gilroy, and Kohen, 1982). Moreover, since youths in many cases continue to benefit from their parental family units, the distributional implications of a minimum wage may be less important.

154. In order for the youth minimum to be effective in alleviating the employment impact on young inexperienced workers, it must be far enough below the adult minimum to compensate for productivity differences associated with inexperience. At the same time, if the gap between youth and adult minimum wages is too large, this creates an incentive for misallocation of labor. For instance, employers may try to lay off workers as they pass the age threshold, replacing them with less expensive younger workers. More generally, to the extent that the minimum wage differential between youths and adults does not correspond to productivity differentials, this differential may displace older workers who are better suited to particular jobs. Thus, the existence of a lower minimum for youths, while it may attenuate the distortionary features of a minimum wage, does not give free rein for setting the adult minimum.

155. Several countries have set lower youth minimum wages (Table 3). In several cases, results have been favorable. In the Netherlands, for instance, minimum wages for youth were lowest in the early 1980; the minimum for 16-year-olds is set at about ⅓ of the adult minimum, rising to 84 percent for 22-year-olds. This appears to be one of the features that accounts for the favorable labor market performance of the Netherlands—in particular, low youth unemployment—in relation to other EU countries.⁵⁹

^{Table 3.}

Youth Minimum Wages

(As percent of adult minimum)

Source: Dolado et al, 1996.

^{Table 3.}

Youth Minimum Wages

(As percent of adult minimum)

Belgium (1992)	small reduction for age under 23
Denmark (1994)	40 percent (for under 18)
France (1993)	80 percent (age 16); 90 percent (age 17); 30-75 for trainees
Germany (1991)	embodied in industry agreements
Ireland (1993)	63 percent (under 18); 81 percent (under 21 in hotels)
Italy (1991)	embodied in industry agreements
Netherlands (1993)	34.5 percent (age 16) rising to 84 percent (age 22)
Portugal (1993)	75 percent (under 18)
Sweden (1992)	85 percent (under 24)
United States (1993)	No reduction

Source: Dolado et al, 1996.

View Table

^{Table 3.}

Youth Minimum Wages

(As percent of adult minimum)

Belgium (1992)	small reduction for age under 23
Denmark (1994)	40 percent (for under 18)
France (1993)	80 percent (age 16); 90 percent (age 17); 30-75 for trainees
Germany (1991)	embodied in industry agreements
Ireland (1993)	63 percent (under 18); 81 percent (under 21 in hotels)
Italy (1991)	embodied in industry agreements
Netherlands (1993)	34.5 percent (age 16) rising to 84 percent (age 22)
Portugal (1993)	75 percent (under 18)
Sweden (1992)	85 percent (under 24)
United States (1993)	No reduction

Source: Dolado et al, 1996.

E. Conclusions

156. The issues discussed in this chapter indicate a number of reasons for uncertainty about the impact of a minimum wage. Most empirical estimates suggest that minimum wages may have a fairly small effect on employment but the preponderance of empirical evidence suggests it is negative. Moreover, most of the empirical studies of the employment effects pertain to levels of minimum wage in the United States, which are well below those being contemplated in the United Kingdom; the effects are likely to be stronger for higher levels of the minimum wage now being considered as more workers would be affected. In particular, at levels of the minimum wage as high, as £4 per hour, the employment effects are quite uncertain but are likely to be negative.

157. A minimum wage would have little effect on poverty, or on the incomes of the poorest, since most of the poorest are not employed, and most of the lower-paid are part-time workers in multiple-earner households who are not poor. The effect of a minimum wage on poverty would be further attenuated by the existing benefits system, which implies a marginal tax rate of over 70 percent on many of the working poor families. Thus, a higher minimum wage would result in a loss of fiscal transparency: it would shift the cost of achieving social objectives from the budget to private employers.

158. A lower youth minimum wage could help attenuate the employment effects of the minimum wage on youth unemployment. International experience indicates some success with lower minimum wages for youth. However, a large differential between youth and adult minima may be distortionary, leading to substitution of younger for older workers. This suggests that a lower youth minimum, while useful, does not permit the adult minimum to be set with impunity. These considerations suggest that it would be safest, at least initially, to set the United Kingdom’s new national minimum wage at a cautious level—perhaps between £3 and £3.50.