Equilibrium unemployment and employment in Sweden A review

Equilibrium unemployment and employment in Sweden – A review Anders Forslund 2008 -01 -29 1

Outline • • Estimates of equilibrium unemployment in Sweden Government reforms and equilibrium unemployment/employment 2

Estimates of equilibrium unemployment • Six studies (that I am aware of) estimate equilibrium unemployment (EU); only two of them for more recent years • All studies indicate that EU has gone up since the early 1980 s, most of them indicate that the rise took place in the early 1990 s (one exception) • The most recent estimates (NIER) suggest that EU has gone up by slightly more than 2 percentage points and approx. equals 6% (ILO definition) 3

Government reforms and EU Two main areas where reforms may be expected to have affected EU: 1. Unemployment insurance 2. Income taxes 4

Unemployment insurance and EU Reforms: • Replacement rate lowered from day 200 (from 80% to 70%) • Daily allowance cap lowered the first 100 days (by approx. 7%) • The expected number of ”insured” days shorter due to coordination with active labour market programmes • Not completely clear how a comprehensive measure should be constructed and what it would show (different changes for different groups; different importance for different margins) 5

Expected effects at given wages • Carling et al (2001) estimated effects on hazards from unemployment to employment of lower replace rate; elasticity = 1. 6 • If hazard from work to unemployment is unaffected by UI, the elasticity of unemployment w. r. t. UI benefits is proportional to the elasticity of the hazard from unemployment w. r. t. UI benefits; proportionality factor equals employment rate • Hence, elasticity of unemployment w. r. t. UI approx. 1. 5 • This is a lower bound, because wage-setting effects, if anything, will amplify effects at given wages 6

Expected effects at given wages: an example Given • Elasticity=1. 5 • EU=6% • Reduction of generosity in UI=10% Then EU down by 0. 8 percentage points 7

Effects at given wages: Timing • Effects at given wages arise through changes in job search behaviour and may appear rapidly; results in Carling et al (2001) suggest that effects may arise already as a reform is expected 8

Effects of UI using estimated wage-setting model • Forslund et al (2008) estimated wage- and price-setting schedules. Take account of general equilibrium effects • The estimated model can be solved for an elasticity of unemployment w. r. t. UI benefits; this elasticity equals 3 • Hence, if UI generosity is decreased by 10% and EU equals 6%, EU is reduced by 1. 8 percentage points • In contrast to the adjustment at given wages, this adjustment takes considerable time 9

Income tax reforms Reforms: • Lower marginal and average (labour) income tax rates at low income tax brackets • Small changes in marginal tax rates at higher income tax brackets, lower average income tax rates at higher brackets – Income taxes reduced by 6. 5% at annual labour income SEK 100 000, 5. 1% at SEK 150 000 10

Expected effects on (un)employment I disregard effects on hours worked, although they may be important for welfare assessments No direct estimates of effects of taxes on labour force participation/unemployment available for Sweden to my knowledge. What to do? • Use estimated effects of changes in UI – the same margin affected (relative gain from working), or • Use estimates of participation elasticity from other countries 11

Expected effects on (un)employment (cont. ) Effects for unemployed persons: • The tax cut equals approx. 5% • Hence, given the same line of reasoning as for the cuts in UI benefits, EU should be reduced by around. 4 percentage points at given wage rates 12

Expected effects on (un)employment (cont. 2) Effects on labour force non-participants (using estimated effects of UI changes): • Assume that new entrants will be unemployed to the same extent as persons already in the labour force • Then labour supply would increase by approx. 8% and 5% of these would become unemployed; the rest employed 13

Expected effects on (un)employment (cont. 3) Effects on labour force non-participants (using estimated effects from other countries): • In a survey, Krueger & Meyer (2002) claim that the best guess about the participation elasticity is that it equals one • If this ”guesstimate” is correct, labour supply would instead increase by approx. 5% 14

Expected effects on (un)employment (cont. 4) • However, income tax effect will be moderated if wages increase • Previous studies of swedish wage setting suggest that if taxes are lowered by x%, wages will increase by almost x%, leaving wage costs almost unchanged. If so, equilibrium will involve a larger fraction of unemployed than the supply response would suggest • However, unclear how much bearing aggregate studies have on the effects of selective tax cuts • For example, new entrants are ”outsiders” and it is not clear how tax cuts targeted at such groups affect wage setting 15

Concluding remarks • Estimates suggest that EU may be around 6%. However, large uncertainty: – One reliable estimate would suggest a lower number; some of the most recent estimates suggest higher numbers • The effects of recent reforms in UI and income taxes may have non-trivial effects on EU and employment rates. Adding effects under most ”optimistic” assumptions gives a ”guesstimate” of the effect on EU just above two percentage points; employment as a share of working age population would increase by more 16

Concluding remarks (cont. ) • However, once again, large uncertainty: – Small number of studies – Not clear exactly how large the changes have been – Effects of taxes on participation largely unknown territory in terms of empirical studies 17