The Minimum Wage by Alan Manning Minimum Wages

The Minimum Wage (by Alan Manning)

Minimum Wages • Systems of minimum wages vary across countries – 2 most common systems are: – statutory minimum wage (set by govt or in national collective bargaining) – set by sectoral collective bargains with extension to non-signatory employers • Some countries have a single minimum (e. g. US), others have variation by age, region, industry, occupation • Measure of bite of minimum wage: – Kaitz index = minimum/median – Spike – percentage of workers at minimum

Kaitz Index – selected countries

Kaitz Index – Selected Countries

Kaitz Index (net, at mean) – OECD

Kaitz Index in Labor Costs – OECD

Summary • Minimum wage lowest in US, highest in France – but age variation in FR, not US • For teenagers US Kaitz 80 -90% • Most countries have Kaitz index of 40 -50% • No big increases in recent years – most countries have falls • CZ in 2007: 8000 / 19300 = 42% • CZ in 2003: 6200 / 15000 = 41%

Introduction • Have introduced in context of institutions that might affect wage inequality • But will also have discussion about impact on employment as this is often regarded as most interesting question • Effect on labor supply (and wage gains of low-skilled) often ignored.

Minimum Wage and Employment • Competitive model has a very clear prediction • Minimum wage above market-clearing wage will cause job losses (unemployment) • Follows from the fact that factor demand curves slope downwards • As w=MRPL any increase in wage makes marginal worker unprofitable

A Picture supply wage Minimum wage MRPL employment

Any models with a different prediction? • Monopsony can give a different prediction: Starting from wage chosen by monopsonist, an increase wage will raise employment. Intuition: – MRPL=MCL>w so marginal worker still profitable after rise in wage and more workers want to work Y’(N)=MRPL=MCL=w(N)+w’(N)N – Employment is supply-determined and increased wage increases labour supply

A Picture MCL supply wage Minimum wage MRPL employment

Can one raise the minimum wage and employment without limit? • Does not sound very plausible • Will not be possible – there comes a point where employment demand determined – can think of N=min(Ns(w), Nd(w)) • Employment will be maximized at wage where Ns(w)=Nd(w) i. e. market-clearing wage (intersection of supply and MRPL) • This is efficient minimum wage to set (with no involuntary unemployment)

How useful is this in practice? • Market-clearing wage different in different labour markets – by age, education, region • Typically minimum wage does not have much variation – too high in some markets, too low in others. • It is a blunt policy instrument • Also have only considered single employer – interactions are likely to be important

Models of Oligopsony • May have very different prediction about employment effect of minimum wage • E. g. suppose labour supply curve is: Ni=Bi(Wi/W)ε • Where W is average wage • Then each employer has some monopsony power but raising minimum wage does not raise employment

Conclusion on Theory • • Competitive model has clear prediction Monopsony prediction ambiguous => should look at evidence with open mind Lee & Saez (2008) if society cares about equity (wages of low-skilled) MW welfare improving despite disemployment effects. • Until Card-Krueger ‘Myth and Measurement’ consensus in US was small negative employment effect especially for teenagers

Card-Krueger Myth and Measurement • Re-examined all evidence for negative employment effects of minimum wage • Look at variety of natural experiments • Concluded no evidence for view that minimum wage causes job loss • Will focus on NJ/PA study as that is most famous – also Card-Krueger, AER 94 – Neumark-Wascher, + Card-Krueger, AER 00

The NJ/PA Study • US system of minimum wages is a federal minimum with individual states choosing higher minimum if they want • in 1992 NJ raised its minimum wage to $5. 05 above the federal minimum of $4. 25 • NJ fast food restaurants the treatment group, restaurants in eastern PA the control group • Data collected by phone interview before and after rise in NJ minimum wage

A Map

Effect on Wages

Basic Results – Difference in Difference Estimator

Neumark-Wascher Criticism • They argued data was of very poor quality, especially on dependent variable – does this matter? • Got hold of payroll data and claimed to find evidence of negative employment effects • Unfortunately some of this data was supplied by noted opponent of minimum wage so perhaps not random sample • Results strongest in this sub-sample • Perhaps some evidence of reduction in hours per worker • See AER 2000 for exchange and make your own mind up

Longer Time Series Using Administrative Data

Evidence on Employment Effects for other Countries • The UK: – Studies of introduction of NMW in 1999 – Aggregate studies failed to find any impact – Machin, Manning, Rahman did find small negative effect among care workers where 30% affected • Problem for many other countries is lack of big change to be basis of natural experiment • E. g. France – SMIC seems very high but lack of much variation in recent years means that hard to evaluate

Machin, Manning, Rahman JEEA, 2003 – Research Design • Sample of care workers in retirement homes for elderly – very low paid job • Surveyed both before and after introduction of NMW • Some homes unaffected as initially paid above NMW – these are effectively the control group • Look at change in hours and employment

Machin, Manning, Rahman JEEA, 2003 - Results

The Minimum Wage and Wage Inequality • Yet again, most research for US • Consensus was that minimum wage unimportant for wage inequality as <5% of workers paid the minimum wage • This was challenged by: – Dinardo, Fortin, Lemiuex, Ecta, 1996 – Lee, QJE 1999

di. Nardo, Fortin, Lemiuex • Pointed out that minimum wage had a very obvious effect on wage distribution in 1979 • Because it did not change in nominal terms in period until 1990, declined in real terms so seemed unimportant by the end • But can help to explain rise in lower-end wage inequality • Especially true for women

A Picture to give flavour of results

Lee, QJE 1999 Basic idea • Federal minimum wage does not vary across states but average level of wages does so minimum wage more important in AK than in NY • If minimum wage important for wage inequality should see bigger rise in wage inequality in low-wage states • This is what he finds

A Picture to Summarize Results

Interpretation • Low-end wage inequality initially much smaller in low-wage states in 1979 – consistent with minimum wage being important • Low-end wage inequality then rises much faster in low-wage states • Top-end wage inequality similar in low- and highwage states and shows no trend • Concludes that min wage can explain almost all of rise in low-end wage inequality in 1980 s • Implies substantial spill-overs

Evidence from UK • Initial studies of impact effect of introduction of NMW suggested modest effect because only 5% directly affected and there seemed no spilll-overs e. g. Dickens-Manning, EJ 2004 • But perhaps some indication that more powerful in longer-run • Perhaps can explain most or all or reduction in low-end wage inequality in UK – but can’t explain the top • CZ: little data available on low-wage sectors