Unemployment Benefits. What are we talking about? Unemployment benefits offer replacement income to...
-
Upload
elinor-reynolds -
Category
Documents
-
view
216 -
download
0
Transcript of Unemployment Benefits. What are we talking about? Unemployment benefits offer replacement income to...
Unemployment Benefits
What are we talking about?
• Unemployment benefits offer replacement income to workers experiencing unemployment spells. In principle should protect jobseekers rather than jobholders (as EPL, Chapter 10).
• The first UB system was introduced in the UK in 1911. Beneficiaries considered “on the dole”.
• Complex design to discourage opportunistic behavior
Outline
• Measures and cross-country comparisons• Theory
– A Competitive Labor Market– An Imperfect Labor Market
• Empirical evidence– effects of UBs on reservation wages– effects of UBs on unemployment duration – effects of unemployment on UBs
Outline (cont.)
• Policy issues– Why and when should UBs be publicly
provided?– Optimal structure of UBs
• Why do Unemployment Benefits exist?
Multidimensional institution
• Different features characterise a UB system: – level of the income transfer wrt to the previous
(future) wage– maximum duration for which they can be
offered– eligibility conditions (conditions for access)– entitlement (rules for duration including
sanctions after assessment of search intensity)
Measures and cross country comparisons
Measures of the generosity of UBs
• Replacement rates (rr): subsidies as a fraction of the previous (backward looking) or potential (forward looking) earnings
• Replacement rate can be computed net or gross of taxes
• At different unemployment durations• For different household characteristics
Measures and cross country comparisons
Many numbers, one single indicator?
“Summary measure of benefit generosity” (OECD, Jobs Study): average of replacement rates in the first two years of unemployment for Average Production Worker (APW) with seniority sufficiently long to yield maximum duration of UBs
Measures and cross country comparisons
Shortcomings of replacement rate measures
• Neglect the coverage of the subsidies (fraction of unemployed receiving the benefit)
• However coverage is partly endogenous (% of youngsters, without work experience)
• Do not consider the entitlement conditions (categorical vs. means-tested)
Measures and cross country comparisons
Summary generosity at 5 yearsN
orw
ayB
elgi
umA
ustri
aD
enm
ark
Irela
ndPo
rtuga
lG
erm
any
Fran
ceFi
nlan
dA
ustra
liaSp
ain
New
Zea
land
Swed
enIc
elan
d
UK
Net
herla
nds
Med
ian
Switz
erla
ndLu
xem
bour
gC
anad
aH
unga
ryPo
land
Cze
chJa
pan
Turk
eySl
ovak
Gre
ece
Italy
Kor
eaU
SA
0
10
20
30
40
50
60
70
80
Unemployment BenefitFive-year average
Unemployment Insurance and Unemployment Assistance
Unemployment Insurance (UI) principle/component:
– Benefit depends on payments during past work experience
– Offers provisions proportional to past earnings– The length of the entitlement period is dependent on the
length of the contribution period.– Some “experience-rating” (e.g., in the US) with
employers paying more if they use it (like a fringe)
Measures and cross country comparisons
Unemployment Assistance component of UB (UA)
–Accessible indipendently of (if any) payments during the past working experience
–Flat subsidy: provisions indipendent of past earnings
–Entitlement not conditional on the length of the contribution period.
Measures and cross country comparisons
Economically relevant distinction
• Each UB system includes both components, unemployment insurance and unemployment assistance
• In economics, the most important distinction is between the cathegorical and the means-tested components
• the former can be analyzed mainly in terms of individual incentives and disincentives
• the latter necessarily requires an analysis of labor supply at the household level (Chapter 7)
Measures and cross country comparisons
Net Replacement Rates for four family types at two earnings levels
After tax and including family and housing benefits for long-term benefit recipients (1999-2000)
Note:
1. NNRs are based on SA except in France, Germany, Greece, where NRRs are based on unemployment assistance.
2. Housing benefits are not included due to very small number of recipients. 3. Social assistance (Reddito minimo di inserimento) is not included in Net Replacement Rates due to its
experimental character (on trial in 39 municipalities). NRR are based on family benefits. 4. People in work are not entitled to social assistance.
Source: OECD tax-benefit models
APW - level 66.7% of APW – level
Single Married couple
Couple
2 children Lone parent 2 children
Single Married couple
Couple
2 children
Lone parent
2 children
Canada 24 41 62 60 35 57 81 80 France1 30 28 42 43 43 41 59 60 Germany1 54 52 65 63 63 61 71 71 Greece1 8 8 10 11 8 8 11 12 Ireland2 31 43 56 56 41 59 66 64 Italy3 0 4 18 14 0 5 21 17 Luxembourg 50 67 75 59 70 92 93 82 Norway 66 67 74 83 65 67 82 90 Sweden4 54 71 85 59 79 102 110 70 United Kingdom 46 57 80 71 66 80 88 81 United States 7 12 46 38 10 17 59 48
UBs often operate in connection with..
Non-employment benefits (other income transfers to non-employed individuals in working age) such as:
– Social assistance of the last resort (different from unemployment assistance)
– Early retirement (Chapter 6)– Liberal access to invalidity pensions – Sickness benefits.
Summarizing evolution of UBs
• Increasing generosity up to the 1980s, especially in Europe. Levelling off or small decline in the 1990s
• Net rr on average 2/3 higher than gross • Increasing sanctions for refusal of jobs or
ALMP• Relatively low coverage notably in
Southern Europe
Measures and cross country comparisons
Effects on individual labour supply
• Labor/leisure choice affected by non-work income• Budget constraint with spike in correspondence to
0 earnings (under cathegorical conditions)• Substitution effect discourages work• Negative net wage at low hours• Increase in the reservation wage of unemployed
benefit recipients
Theory
Reservation wage without (Left Panel) and with (Right Panel) unemployment benefits
l l
c c
A m+b
B
l0
wr
hA
w m
cA
wr < w
u (c,l)=k2
u (c,l)=k1
l0
w wr
u (c,l)=k3
wl0
m
wr > w
Effects on the reservation wage
• Unemployment benefits increase the reservation wage of individuals receiving UBs (may reduce the one of jobseekers not receiving UBs)
• If minimum guaranteed income UA scheme, then wage rate of zero at the participation margin
Theory
In Imperfect Labor Markets
• 3 effects• Job search effect (on dynamic reservation wage)• Wage effect (on the bargaining outcome and via an
increase of efficiency wages)• Entitlement effect (increase in participation of
those not receiving UBs)
Also tax effect (Chapter 4) related to funding of UBs
Theory
Job search effect
• Jobseekers become more choosy. Longer duration of unemployment among UB recipients.
• They only accept job offers involving a higher wage
• This higher (dynamic) reservation wage discriminates between unemployment and employment (unlike the static reservation wage separating employment and non-employment)
Theory
Why links with SA are so important
Wage effect
• Higher outside option of workers at the bargaining table (pure bargaining effect)
• Higher wage is required to deter shirking (efficiency wage effect). The penalty associated with unemployment (the disciplining device, see box 11.2) is reduced in presence of UBs
Theory
Entitlement effect
• UBs increase the value of employment• More participation in the labor market
(shifts across participation margins)• Lower reservation wage of jobseekers not
receiving UBs. Higher job finding rates of unemployed not eligible to UBs.
Theory
Empirical evidence
• Receipt of benefits increase reported reservation wages
• Longer duration of benefits correlated with longer duration of unemployment
• Unemployment outflows increase in proximity of the maximum duration of benefits
• Presence of spillovers between recipients and non-recipients of UB: also labor supply enhancing effects (as predicted by entitlement effect)
Empirical evidence
Reservation Wages of the Unemployed; 1995(Estimates of Mincer-type reservation wage equations)
coeff sig st. error coeff sig st. error coeff sig st. error gender (M=1) 0.0593 ** 0.0127 0.0585 ** 0.0234 age 0.0200 *** 0.0030 0.0196 *** 0.0030 0.0274 *** 0.0056 age^2 -0.0003 *** 0.0000 -0.0002 *** 0.0000 -0.0003 *** 0.0001level of education: primary or lower -0.1370 ** 0.0272 -0.1385 *** 0.0271 -0.1265 *** 0.0445 tertiary -0.2023 ** 0.0274 -0.2053 *** 0.0274 -0.2108 *** 0.0446type of job being seeked in the private sector -0.0432 ** 0.0183 -0.0435 ** 0.0182 -0.0269 0.0308 part-time -0.2300 ** 0.0153 -0.2298 *** 0.0153 -0.2237 *** 0.0247 within commuting distance -0.0123 ** 0.0106 -0.0135 0.0106 -0.0134 0.0108labour market status first-time jobseeker 0.0039 ** 0.0127 -0.0010 0.0127 0.0005 0.0129 unemployment benefit recipient 0.0489 ** 0.0223 0.0517 ** 0.0222 0.0483 ** 0.0216relation with head of household husband/wife -0.0894 ** 0.0196 -0.0887 *** 0.0195 -0.0589 * 0.0335 son/daughter -0.0694 ** 0.0195 -0.0691 *** 0.0195 -0.0574 * 0.0318 relative -0.0854 ** 0.0295 -0.0836 *** 0.0295 -0.1093 ** 0.0495 nr of family members 0.0021 *** 0.0044 0.0013 0.0044 0.0060 0.0074local conditions unemployment rate (district-level) 0.4793 *** 0.1075 dummy North-West 0.0358 *** 0.0199 0.0319 0.0199 0.1136 *** 0.0416 dummy Center 0.0435 ** 0.0199 0.0269 0.0202 0.0340 0.0325 dummy South 0.1115 ** 0.0182 0.0460 ** 0.0234 0.0744 ** 0.0321 constant 13.9764 * 0.0731 13.9493 *** 0.0732 13.8686 *** 0.1247
Mills lambda -1.7710 *** 0.6008
R2 0.1242 0.1276n 5112 5112 5112
UB and unemployment duration
• Level of benefits – elasticity w.r.t. duration– Layard et al. (1991) 0.2-0.9– Carling et al. (2001) Sweden: 1.7– Roed and Zhang (2003) Norway: 0.35-0.95– Van Ours and Vodopivec (2004): 1.4
• Potential benefit duration 1 week longer– Katz and Meyer (1990) US: 0.16-0.20 weeks more
unemployment– Ham et al. (1998) Czech-Slovak Republics: 0.30-0.93
weeks more unemployment– Van Ours and Vodopivec (2004): 0.86 weeks more
Empirical evidence
Duration analysis
• Retrospective data, matched records across LFS or administrative (social security) records
• Problems with survey data: recall bias, lenght-biased sampling, right-censoring
• Problems with administrative records: recording affected by regulations (e.g., coverage)
Empirical evidence
Unemployment hazard rates
• The hazard rate, λ , is the conditional probability of leaving unemployment (e.g., probability that an individual leaves U in the 10th week given that she has been U for 9 weeks) after a certain period of time
• If constant, then the (conditional) (survival) probability of leaving U at the 10th day isf(10)=λ(1-λ)9 where λ is the hazard
• More generally , λ(i) is the hazard function
Empirical evidence
Natural Experiments
• (vanOurs-Vodopivec) Reform in Slovenia reducing potential benefit duration
• Maximum benefit duration dependent on previous work experience (months):– 3 to 3, 6 to 3, 9 to 6, 12 to 6, 18 to 9, 24 to 9, 24 to 12,
24 to 18, 24 to 24. • October 1987• Examples 12 to 6 – both outflow to job and to
other destinations increases
Empirical evidence
Eligibility 12 months before - 6 months after
0
0,05
0,1
0,15
0,2
0,25
0,3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Months of unemployment
Mon
thly
exi
t ra
te
Before After
Entitlement effects
• Incentives to accept also risky jobs (precarious or with temporary spells) for the outsiders
• May improve mobility in economies experiencing structural change if in the declining sector there is wage compression
• May also decrease the reservation wage and reduce unemployment
Empirical evidence
Policy endogeneity
• Extended duration of unemployment. Benefits often granted as policy response to crises
• Regionally adjusted UBs in the US (Card and Levine, 2000)
• Austrian Regional Extended Benefits Program (Lalive-Zweimueller, 2002): benefits extended from 30 up to 209 weeks
Empirical evidence
Empirical findings
• Policy endogeneity is significant• Estimates of the effects of UB duration on LTU
likely to be biased upwards• Yet it is still there: in Austria increase in benefit
duration from 30 to 209 weeks reduces the transition to jobs by 17% (40% without correcting for endogeneity), increasing expected unemployment duration by 9 weeks
Empirical evidence
Trade-offs in the provision of UB
• Reduced incentives to work• Fiscal costs• Better risk sharing (with risk-averse workers) Increase in
welfare• Spillovers: workers encouraged to take risky, high-
productivity, jobs• Subsidy to job search, matching efficiency. Acemoglu-
Shimer: there can be productivity gains by raising UB in the US to European levels
Policy issues
Possible private provision of unemployment insurance?
• No because moral hazard and adverse selection. Asymmetric information.
• Workers can alter the probability of losing a job• Private insurance would ask for premia selecting
only workers with higher than average risk• Risk pooling problem: risks are correlated (e.g.,
during recession)
Policy issues
Optimal design of UBs dealing with the agency problem
• Public provider faces the same moral-hazard problems (as compulsory contributions, less adverse selection), related to the non-verifiability of search effort.
• Ways to reduce disincentives to seek jobs.• Low replacement rates, declining with
unemployment duration. Administrative pressure on recipients (“help and hassle”). Offer of slots in ALMPs as a way to elicit effort
• Financial incentives to the take-up of jobs: premia in terms of residual benefit claims and ECI
Policy issues
Why do UBs exist?
• Properly designed UBs improve the allocation of human capital and thus, foster economic growth.
• However, UBs should not be too generous in order not to discourage job search altogether and generate stagnant unemployment pools.
• The most relevant issues do not concern whether or not a country should have a UB system, but how the system should be designed along its several dimensions. Difficult to reform once in place.
Policy issues
Politically feasible reforms
• Exploiting the UB/EPL tradeoff, e.g., increasing the degree of experience-rating
• Grandfather existing entitlements• Change enforcement more than rules• Combine benefit cuts with employment
conditional benefits or wage subsidies; way to win support by employees and reduce the opposition of the unemployed
Policy issues
Date of First Introduction of UBs
1905-1944 1945-1979 1980-1990 1991-1992 1993-2002
Australia Austria Brazil Argentina AlbaniaBelgium Bangladesh Bulgaria Armenia AlgeriaCanada Croatia China Azerbaijan South KoreaChile Ecuador Colombia Belarus Taiwan
Denmark EgyptCzech Republic
Estonia TunisiaFinland Greece Hungary
GeorgiaTurkey
France Hong Kong Iran KazakhstanGermany Iraq Poland KyrgyzstanIreland Israel
Romania
LatviaItaly Japan
Slovak Republic
LithuaniaNew Zealand Netherlands
Sri Lanka
MoldovaNorway Nigeria
Uruguay
RussiaSpain Portugal Turkmenistan
Sweden Serbia UkraineUnited Kingdom Slovenia Uzbekistan
USA South Africa
Source: Social Security Programs Throughout the World and our own research based on administrative sources.
Dynamic Effect of UBs on Job Destruction -- Reallocation
Dynamic effect of UBs on Job Destruction (JD) (coefficients and confidence intervals)
Review questions
• Why do replacement rates offer an incomplete measure of the generosity of unemployment benefits?
• Do unemployment benefits redistribute in favor of low-skilled workers?
• What effects should we expect UBs to have on first-time jobseekers when benefits are conditional on previous work experience?
• How does the introduction of a UB system affect labor force participation?
Review questions (cont.)
• What is the hazard rate?• What type of relation do we expect to
observe between generosity of UBs and structural change?
• Why is unemployment insurance not provided by private insurance companies?
• Is there a socially optimal replacement rate?• Should UBs be experience rated?
Practicing with real data
• Box 11.3:How shortering the potential duration of
unemployment benefits affects the duration of
unemployment: evidence from a natural experiment
(pag. 241-242).• A Stata data file with the Van Ours and Vodopivec
(2006) dataset, a do file and a log file are available at the website:
http://www.frdb.org/images/customer/vanours.zip
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
Practicing with real data
• Box 11.4: The effect of benefit sanctions on the
duration of unemployment (pag. 242-243).• A Stata data file with the Lalive, van Ours, and
Zweimuller (2006) dataset, a do file and a log file are available at the website:
http://www.frdb.org/images/customer/lalive.zip
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.