The Minimum Wage Revisited
The role and impact of the minimum wage have once again become a critical subject of policy debate in the United Kingdom, particularly amidst concerns over persistently high youth unemployment, significant frictions within the labour market, and the broader challenges posed by general economic stagnation. Moreover, sectors with a high concentration of low-wage labour, such as the hospitality industry, face acute pressures in adjusting to recent, substantial increases. This briefing paper revisits the extensive body of economic theory and empirical evidence on the minimum wage, concluding that, despite theoretical ambiguities in imperfect markets, the overall effect is likely to be harmful to employment, prices, and non-monetary worker benefits.
In a competitive labour market, the minimum wage is always a bad idea …
In a competitive labour market, employees are paid their marginal product of labour. Any less and another employer will out-bid their present employer; any more and the employer would rather reduce their hours or lay them off. In such a market, any effective minimum wage will necessarily reduce work. Naturally, the least productive workers – especially the junior, unskilled, or risky to employ, such as ex-convicts – will be hardest hit. It is their marginal product that is going to be below the minimum wage.
… but in a more realistic model, the effects are ambiguous but unlikely to be very positive.
The labour market is not perfectly competitive, of course. Search frictions and employment protection laws both gum up the system, which may give either employers or employees scope to set wages below/above the marginal product. Low-wage labour markets could be ‘monopsonistic’ (i.e., featuring market power on the demand side; cf. monopolistic on the supply side). This means firms have the ability to drive workers’ wages below their marginal product, which depresses both firm output and workers’ labour supply. In this case, a minimum wage could increase wages and hours. This would be unambiguously good for employees, so long as the minimum wage stays at or below the competitive level.
Equally, a minimum wage could be entirely absorbed by reducing the ‘fringe benefits’ that employees receive, without reducing hours (see Clemens 2021). They could, for instance, receive a higher hourly wage but also less civil hours and more unsafe conditions. This would be bad for employees: They could already choose a high-wage, low-benefit package, but they opt for a lower-wage, higher-benefit package, demonstrating their preference for the latter.
Finally, minimum wages can be accommodated through higher prices. In fact, even advocates for the policy find that this is the primary margin of adjustment, rather than firm profits (see Dube and Linder 2024, p. 103). Much of the burden of these higher prices falls on the poor people whom the policy is intended to help (MaCurdy 2015), even as many of the benefits of the high minimum wages accrue to secondary earners in already-wealthy households.
In our reading, severe monopsonistic conditions are unlikely. Only in serfdom can employers really hold total compensation (including benefits) substantially below marginal product in the long-term. Workers can simply move to a better job elsewhere. Nonetheless, small frictions, such as the search costs of finding a new job, may mean workers are paid somewhat below their marginal product and make a moderate minimum wage tolerable.
Empirically, the evidence suggests that minimum wages are moderately harmful to employment …
In the face of such theoretical ambiguity about the employment effects, we must look at the data. The famous quasi-experimental work of David Card and Alan Kreuger (1994) suggested that a moderately higher minimum wage had no effect on employment, when comparing fast-food restaurants in New Jersey to eastern Pennsylvania. Naturally, by picking a sector not subject to trade competition, this paper was pre-disposed to find small effects. Still, a large literature has since grown up around the question. Rival literature reviews reach different conclusions, but a very good review by David Neumark and Peter Shirley (2021) summarises the findings thus:
“(i) there is a clear preponderance of negative estimates in the literature;
(ii) this evidence is stronger for teens and young adults as well as the less-educated;
(iii) the evidence from studies of directly-affected workers points even more strongly to negative employment effects; and
(iv) the evidence from studies of low-wage industries is less one-sided.”
This evidence suggests that the US labour market (at least) is closer to the competitive model than the monopsonistic model. The evidence from the UK is much more difficult to come by, as we lack a natural control group. Still, a review of the evidence by Mike Brewer, Thomas Crossley, and Federico Zilio (2019) published by the IFS reached the following conclusion:
“The data are consistent with both large negative and small positive impacts of the UK National Minimum Wage on employment … [but] … the existing data, combined with difference-in-difference designs, in fact offered very little guidance to policy makers”
These findings suggest that, though the literature is far from unanimous, minimum wages probably do increase unemployment, as common sense and theory would predict. Moreover, they certainly increase prices and likely reduce other amenities that workers care about in their jobs.
… but very high minimum wages – as the UK now has – are untested and likely to be harmful.
Outside of Latin America, only France and New Zealand have higher minimum wages than the UK in the OECD, as a share of the median wage. Moreover, the Latin American minimum wages are less binding, as the informal economy is much larger. This suggests that we should act very cautiously in increasing our minimum wage further, as there is very little evidence of how the market will react. Even Arindrajit Dube, a noted enthusiast for the minimum wage who was appointed by the last Conservative government to review it, suggested some caution once it reached two-thirds of the median.
As an increasing share of the workforce’s productivity falls at – or below – the minimum wage, the employment effects – as well as the price and fringe benefit effects – are likely to grow quite quickly. In light of the unprecedented level and risk, the increase in the minimum wage must be immediately halted or significantly slowed until productivity growth can catch up.