1

Designing Optimal Designing Optimal Unemployment Unemployment

InsuranceInsurance

Bertil HolmlundBertil Holmlund

Department of Economics Department of Economics

Uppsala UniversityUppsala University

2

OutlineOutline

1.1. Unemployment benefits around the world Unemployment benefits around the world 2.2. Why public UI?Why public UI?3.3. Theoretical backgroundTheoretical background4.4. Optimal UI:Optimal UI:

Time limits of benefit receiptTime limits of benefit receiptMonitoring and sanctionsMonitoring and sanctionsWorkfare Workfare

5.5. The financing of UIThe financing of UI6.6. UI savings accounts UI savings accounts 7.7. ConclusionsConclusions

3

Unemployment benefits around Unemployment benefits around the worldthe world

Unemployment insurance (UI)Unemployment insurance (UI) Mainly in developed countriesMainly in developed countries Typically mandatoryTypically mandatory Percent of past wage (up to a ceiling)Percent of past wage (up to a ceiling) Time limitedTime limited

Severance pay Severance pay One-time payment One-time payment Collective agreements or mandated by Collective agreements or mandated by

gov’tsgov’ts Depend on years of serviceDepend on years of service

4

Unemployment benefits around the Unemployment benefits around the world, cont.world, cont.

Unemployment assistance (UA)Unemployment assistance (UA) Minimum income guaranteeMinimum income guarantee Often without time limitOften without time limit Sometimes means testedSometimes means tested

Social assistanceSocial assistance Means tested, directed at poor in generalMeans tested, directed at poor in general

Unemployment insurance savings accounts Unemployment insurance savings accounts (UISA)(UISA) Forced savings in individual savings accountsForced savings in individual savings accounts Surplus at retirement age converted into Surplus at retirement age converted into

retirement incomeretirement income

5

WhyWhy public UI? public UI?

Informational asymmetriesInformational asymmetries Moral hazard (hidden action)Moral hazard (hidden action) Adverse selection (hidden Adverse selection (hidden

characteristics)characteristics)

Macro shocks – private insurance Macro shocks – private insurance cannot deal with economic slumpscannot deal with economic slumps Public UI an instrument for Public UI an instrument for stabilizationstabilization

policypolicy

6

Theoretical background Theoretical background Effects of time limits in UIEffects of time limits in UI

Seminal paper: Mortensen (1977)Seminal paper: Mortensen (1977)

Sequential searchSequential search Maximization of lifetime utilityMaximization of lifetime utility Fixed duration of benefit paymentsFixed duration of benefit payments Stochastic duration of employment spellsStochastic duration of employment spells Eligibility condition: Eligibility condition:

work must precede benefit receiptwork must precede benefit receipt

7

Effects of time limits in UIEffects of time limits in UI

The reservation wage The reservation wage declinesdeclines as as the insured worker gets closer to the insured worker gets closer to the date at which benefits expirethe date at which benefits expire

The exit rate The exit rate increasesincreases with with elapsed durationelapsed duration

8

Job finding over the spell of Job finding over the spell of unemploymentunemploymentExit rate

Benefits expire

Benefits exhausted

Benefit period

Duration

9

The theory implies that a rise in the The theory implies that a rise in the benefit levelbenefit level

increases exit rates to employment among increases exit rates to employment among workers not eligible for UI (workers not eligible for UI (entitlement effectentitlement effect))

causes a newly unemployed worker to increase causes a newly unemployed worker to increase the reservation wagethe reservation wage

but induces an insured worker close to benefit but induces an insured worker close to benefit expiration to expiration to reducereduce the reservation wage the reservation wage

10

Effects of a rise in the benefit levelEffects of a rise in the benefit level

Exit rate

Duration

Benefits exhausted

Higher benefits

Entitlement effect

11

Strong empirical evidence on the Strong empirical evidence on the impact of the potential duration of impact of the potential duration of benefits:benefits: Job finding increases as the worker gets Job finding increases as the worker gets

closer to benefit expirationcloser to benefit expiration

12

Time Time limitedlimited UI and UI and jobjob findingfinding in in SwedenSweden

0

0.04

0.08

0.12

0.16

0.2

0.24

1 3 5 7 9 11 13 15 17

Period (4 weeks, exhaustion in period 15)

UI

No UI

13

Strong empirical evidence on the Strong empirical evidence on the impact of the potential duration of impact of the potential duration of benefits:benefits: Job finding increases as the worker gets Job finding increases as the worker gets

closer to benefit expirationcloser to benefit expiration

Not much evidence on entitlement Not much evidence on entitlement effects effects

14

Optimal UI: Time limitsOptimal UI: Time limits

Seminal paper: Shavell and Weiss Seminal paper: Shavell and Weiss (1979)(1979)

a case for declining time profilea case for declining time profile

A wage tax as complement: Hopenhayn A wage tax as complement: Hopenhayn and Nicolini (1997):and Nicolini (1997):

Benefits should decrease over the elapsed Benefits should decrease over the elapsed duration of unemploymentduration of unemployment

The wage tax should increase with the length The wage tax should increase with the length of the previous unemployment spellof the previous unemployment spell

Numerical examples: large welfare gains, Numerical examples: large welfare gains, high replacement rateshigh replacement rates

15

Endogenous work effort: Wang and Endogenous work effort: Wang and WilliamsonWilliamson (1996) (1996)

Endogenous Endogenous inflowinflow into unemployment: the into unemployment: the probability of remaining employed is increasing in probability of remaining employed is increasing in work effort.work effort.

Endogenous Endogenous outflowoutflow from unemployment depends from unemployment depends on search efforton search effort

Implications for optimal UI: Implications for optimal UI: A large drop in consumption in the first period of A large drop in consumption in the first period of

unemployment (discourages shirking)unemployment (discourages shirking) A large reemployment bonus (encourages search)A large reemployment bonus (encourages search) Optimal UI: compensation increases initially and then falls Optimal UI: compensation increases initially and then falls

throughout the spellthroughout the spell

16

Collective bargaining in general Collective bargaining in general equilibrium:equilibrium:

Cahuc and LehmannCahuc and Lehmann (1997, 2000)(1997, 2000)

The fall-back position of the union is The fall-back position of the union is the welfare of the short-term the welfare of the short-term unemployedunemployed

A declining time profile increases the A declining time profile increases the welfare of the short-term unemployed welfare of the short-term unemployed relative to the long-term unemployedrelative to the long-term unemployed and may therefore increase wage and may therefore increase wage

pressure and unemploymentpressure and unemployment

17

Search equilibriumSearch equilibrium::Fredriksson and Holmlund (2001)Fredriksson and Holmlund (2001)

Endogenous search effort, wage Endogenous search effort, wage bargainingbargaining

Two states of unemployment:Two states of unemployment: Insured (I) and Non-insured (N)Insured (I) and Non-insured (N)

Stochastic benefit durationStochastic benefit duration Benefit entitlement through employmentBenefit entitlement through employment

18

Labor market flowsLabor market flows

19

Value functions for Value functions for IInsured, nsured, NNon-insured and on-insured and EEmployed workermployed worker

( , ) ( ) ( )I I I I N IrU B s s E w U U U

( , ) ( )N N N NrU b s s E w U

( ) ( ) ( )IrE w w U E w

, , ;j js j I N search effort s transition rate

20

Welfare objective: expected utility of Welfare objective: expected utility of the workerthe worker

Is a two-tiered benefit system better Is a two-tiered benefit system better than a uniform one?than a uniform one? Yes, in generalYes, in general Entitlement effect: benefit differentiation Entitlement effect: benefit differentiation

encourages search among those who encourages search among those who have run out of benefitshave run out of benefits

The welfare improvement relative to a The welfare improvement relative to a uniform system is non-trivialuniform system is non-trivial

21

Summary: time profileSummary: time profile

A reasonably strong case for a A reasonably strong case for a declining time profiledeclining time profile

A case for a waiting period?A case for a waiting period? Discourage shirkingDiscourage shirking Discourage temporary layoffsDiscourage temporary layoffs Private savings as a substitutePrivate savings as a substitute

22

Monitoring and sanctionsMonitoring and sanctions

In practice: UI systems condition In practice: UI systems condition benefits on performance criteria:benefits on performance criteria: availability for workavailability for work actively searching for workactively searching for work

Monitoring through benefit Monitoring through benefit administration (public employment administration (public employment service).service).

Benefit sanctions if search criteria are Benefit sanctions if search criteria are not metnot met

23

Monitoring and sanctions in UI: Monitoring and sanctions in UI: theorytheory

Boone -van Ours (2006); Boone, Fredriksson, Holmlund , van Ours Boone -van Ours (2006); Boone, Fredriksson, Holmlund , van Ours (2007)(2007)

Search and matching modelSearch and matching model

Search effort affects Search effort affects

the job finding rate the job finding rate

the risk of a benefit sanctionthe risk of a benefit sanction

24

Value function for Value function for IInsured worker,nsured worker,monitoring and sanctionsmonitoring and sanctions

FOC for optimal search, insured worker:FOC for optimal search, insured worker:

( , ) ( ) ( )( )I I I I I N IrU B s s E w U s U U

( ) ( )( ) 0I I I I N Is s E w U s U U

;I Is search effort s transition rate

25

Monitoring and sanctions in Monitoring and sanctions in UI: theoryUI: theory

Monitoring and sanctions represent a Monitoring and sanctions represent a welfare improvement for reasonable welfare improvement for reasonable values of monitoring costsvalues of monitoring costs

Boone, Fredriksson, Holmlund, van Ours Boone, Fredriksson, Holmlund, van Ours (2007)(2007)

26

Evidence on monitoring and Evidence on monitoring and sanctionssanctions

Social experiments in the United StatesSocial experiments in the United States

Random assignments of unemployed Random assignments of unemployed benefit claimants into groups exposed to benefit claimants into groups exposed to different search requirementsdifferent search requirements

Washington state, 1986-87Washington state, 1986-87 Maryland, 1994Maryland, 1994

27

The Washington studyThe Washington study TreatmentsTreatments: :

Varying degrees of work-search Varying degrees of work-search requirementsrequirements

Results:Results: Workers without search Workers without search

requirements had 3 weeks longer requirements had 3 weeks longer duration of benefit receipt than duration of benefit receipt than those with standard requirementsthose with standard requirements

28

The Maryland studyThe Maryland study Treatments: increased search Treatments: increased search

requirementsrequirements

An increase in the number of An increase in the number of required employer contacts from 2 to required employer contacts from 2 to 4 reduced the duration of benefit 4 reduced the duration of benefit receipt by 6 %receipt by 6 %

Informing claimants that their Informing claimants that their employer contacts would be employer contacts would be monitored reduced the duration of monitored reduced the duration of benefit receipt by 7.5 %benefit receipt by 7.5 %

29

Other experimental evidenceOther experimental evidence

Dolton, O’Neill (1996), UKDolton, O’Neill (1996), UK Risk of losing benefit if not showing up at interviewRisk of losing benefit if not showing up at interview Fairly strong increase in the exit rate (30 %)Fairly strong increase in the exit rate (30 %)

Ashenfelter, Ashmore, Deschenes (1999), USAAshenfelter, Ashmore, Deschenes (1999), USA Stricter job search requirementsStricter job search requirements At most very small effectAt most very small effect

Van den Berg, van der Klaauw (2001), NetherlandsVan den Berg, van der Klaauw (2001), Netherlands CCounseling and monitoringounseling and monitoring No effect on the transition rate from unemployment to No effect on the transition rate from unemployment to

employmentemployment

30

Non-experimental evidence on Non-experimental evidence on sanctions:sanctions:

How does a sanction affect job finding?How does a sanction affect job finding?

Abbring, van den Berg, van Ours (2005),Abbring, van den Berg, van Ours (2005), Van den Berg, van der Klaauw, van Ours (2004)Van den Berg, van der Klaauw, van Ours (2004)

Increase in the transition rate to Increase in the transition rate to employment by 80-100 %employment by 80-100 %

Lalive, van Ours, Zweimuller (2005)Lalive, van Ours, Zweimuller (2005) Swiss evidence on the impact of warnings and Swiss evidence on the impact of warnings and

actual benefit sanctionsactual benefit sanctions Increase in the outflow from unemployment Increase in the outflow from unemployment

31

Workfare: Workfare: work required in exchange for work required in exchange for

benefitsbenefitsThree arguments:Three arguments:

Benefits for the unemployed more Benefits for the unemployed more politically acceptablepolitically acceptable

Workfare as a tax on leisureWorkfare as a tax on leisure

Workfare as a screening deviceWorkfare as a screening device

32

Workfare and UIWorkfare and UI

Informal argument: active labor market Informal argument: active labor market policy as workfarepolicy as workfare Workfare puts a price on workers’ timeWorkfare puts a price on workers’ time Workers with a high value of leisure self-select Workers with a high value of leisure self-select

out of the benefit systemout of the benefit system Formal modeling, example:Formal modeling, example:

Two types of individuals, different preferences Two types of individuals, different preferences for leisurefor leisure

The government doesn’t know individual The government doesn’t know individual preferences, only the distribution preferences, only the distribution

Kreiner and Tranaes (2005), Fredriksson and Kreiner and Tranaes (2005), Fredriksson and Holmlund (2006)Holmlund (2006)

33

Can workfare be Pareto improving?Can workfare be Pareto improving?

Yes!Yes!

Absent workfare, searching as well as non-Absent workfare, searching as well as non-searching individuals may claim benefitssearching individuals may claim benefits

Workfare induces non-searching workers to Workfare induces non-searching workers to self-select out of UI (strong preference for self-select out of UI (strong preference for leisure)leisure)

Makes it possible to raise benefits without Makes it possible to raise benefits without making it worse for non-searching individualsmaking it worse for non-searching individuals

34

Evidence on workfareEvidence on workfare

Black, Smith, Berger, Noel (2003), USABlack, Smith, Berger, Noel (2003), USA Random assignment into mandatory Random assignment into mandatory

employment and training servicesemployment and training services Participation required in order to receive Participation required in order to receive

benefitsbenefits

ResultsResults A reduction in the mean UI duration by about 2 A reduction in the mean UI duration by about 2

weeksweeks A marked rise in reemployment A marked rise in reemployment beforebefore the the

scheduled program participationscheduled program participation

35

The The financingfinancing of UI of UI

Current systemsCurrent systems TaxesTaxes

(Incomplete) experience rating on firms (Incomplete) experience rating on firms (USA)(USA)

AlternativesAlternatives Taxes on workers that depend on their Taxes on workers that depend on their

unemployment experiences unemployment experiences (cf. unemployment insurance savings accounts)(cf. unemployment insurance savings accounts)

More stringent experienceMore stringent experience rating rating

36

Pros and cons of Pros and cons of experience ratingexperience rating

ER discourages layoffsER discourages layoffs

But firms have not much influence over But firms have not much influence over workers’ search and job findingworkers’ search and job finding

ER is effectively on tax on layoffs and ER is effectively on tax on layoffs and thereby indirectly also a tax on hiringsthereby indirectly also a tax on hirings

Many countries already “tax” layoffs Many countries already “tax” layoffs through employment protection legislation through employment protection legislation (EPL)(EPL) Optimal mix of ER and EPL?Optimal mix of ER and EPL?

37

Unemployment insurance savings Unemployment insurance savings accountsaccounts

Feldstein-Altman (1998), Bovenberg-Sorensen (2004), Stiglitz -Yun Feldstein-Altman (1998), Bovenberg-Sorensen (2004), Stiglitz -Yun (2005)(2005)

Mandatory savings of wage income into an Mandatory savings of wage income into an unemployment insurance account (UISA)unemployment insurance account (UISA)

Job loss: withdraw from the accountJob loss: withdraw from the account If funds are insufficient, borrow from the If funds are insufficient, borrow from the

governmentgovernment

At retirement age: convert UISA into At retirement age: convert UISA into retirement incomeretirement income The government cancels the debt if the The government cancels the debt if the

balance is negativebalance is negative

38

Pros and cons of UISAPros and cons of UISA

UISA reduces moral hazard problemsUISA reduces moral hazard problems The worker finances her own benefitsThe worker finances her own benefits UISA allows borrowing against future UISA allows borrowing against future

incomeincome Some workers will retire with negative Some workers will retire with negative

balancesbalances How common?How common?

Rules for withdrawalRules for withdrawal Worker heterogeneity and Worker heterogeneity and

distributional issuesdistributional issues

39

ConclusionsConclusions The case for penalizing less active search is The case for penalizing less active search is

solidsolid Indirect penalty: a declining time profileIndirect penalty: a declining time profile More direct penalty via monitoring of searchMore direct penalty via monitoring of search Workfare can be a useful screening deviceWorkfare can be a useful screening device

CaveatsCaveats Low benefits during the first week(s) would Low benefits during the first week(s) would

discourage unemployment entry discourage unemployment entry Precautionary savingsPrecautionary savings

UISA may be a useful complement to UIUISA may be a useful complement to UI