ISSN 1183-1057
Department of Economics
Discussion Papers
02-5
Severance Payments &
Unemployment Insurance: A
Commitment Issue
R. Martin
S. Mongrain
S. Parkinson
February, 2002
NO S
US ET
S O M M E S PR
SIMON FRASER UNIVERSITY
Severance Payments and Unemployment Insurance:
A Commitment Issue
by
Richard Martin
Simon Fraser University
Steeve Mongrain
Simon Fraser University and RIIM
and
Sean Parkinson
Queen’s University
February 2002
Abstract
In the event of a job termination, many workers receive severance payments from
their employer, in addition to publicly provided unemployment insurance(UI).
In the absence of a third party enforcer, contracts featuring severance payments
must be supported by an implicit self-enforcing contract. Workers believe em-
ployers will make severance payments only if it is in their best interestex post.
Firms that compete for workers face the incentive to reduce the severance pay-
ment they offer, in order to relax their incentive constraint. Workers are forced
to bear risk, and too many workers are laid-off. We show that publicly provided
UI can correct these distortion.
Key Words: Unemployment Insurance, Severance Payment, Layoffs
JEL: J65 J41
Corresponding author: Steeve Mongrain, Department of Economics, Simon Fraser
University, 8888 University Drive, Burnaby, BC, Canada, V5A 1S6, e-mail: mon-
grain@sfu.ca, Phone: (604) 291-3547, Fax: (604) 291-5944. We would like to
thank Robin Boadway, Patrick Francois, Nicolas Marceau, Joanne Roberts and
for useful comments. Steeve Mongrain would also like to thank Simon Fraser Uni-
versity, RIIM and SSHRC for financial support. All errors are the responsibility
of the authors.
1. Introduction
Being fired is a very costly incident for many workers. Accordingly, unem-
ployment insurance(UI) can improve welfare by partially smoothing the income
streams of risk adverse workers. In the event of a job termination, most workers
are eligible for publicly provided unemployment insurance. However, a sizable
proportion of workers also receive insurance from their employer.1 When work-
ers and firms sign contracts that include a severance payment, they have a
means to further smooth workers’ income between periods of employment and
unemployment. Unfortunately, these contracts are typically difficult to enforce.
McLeod and Malcomson (1989) and Carmichael (1983) have illustrated that,
because it is difficult for a third party to observe who initiated a job separation,
severance payments must be supported by a self-enforcing contract. Consider
a firm that wishes to reduce its work force. If the firm was to honor its con-
tracts, it would have to make severance payments to the terminated workers.
A cheaper alternative for the firm would be to put pressure on the employees
to quit, in which case no severance payments must be made. This implies that
workers will only accept a contract that is self-enforcing for the firm. Firms will
find it in their best interest to make severance payments if they extract sizable
rents from their workers. If this is the case, then not honoring their contracts
jeopardizes this stream of future profits.
In the absence of publicly provided UI, we demonstrate that the equilibrium level
of severance payment is inefficient. To see why, consider firms that competes
to attract worker. Firms face the incentive to reduce the size of the severance
payments they promise to pay. By doing so, firms can relax their incentive
constraint, increasing their ability to compete for workers. Lower severance
payments reduce the profits necessary for the contracts to be self-enforcing,
enabling the firm to offer a larger share of the revenue to their workers. However,
1
In the US, 36% of all full time employer of the large and medium private estab-
lishments receive severance payments, while 15% of all full time employer of the
small private establishments receive severance payments. The same numbers for
professional technical related employees are 48% and 23% respectively.
1
if the equilibrium severance payment provides less than full insurance, then firms
will lay off too many workers. Low severance payments imply that the firms
only bears part of the cost of firing a worker. Furthermore, lower severance
payments makes the income stream of fired employees more volatile. Thus, in
the absence of publicly provided UI, severance payments will be inefficiently low.
The analysis of unemployment insurance program will consequently be done in
a second best world, which leads to some interesting results.
It is well recognized in the literature that publicly provided UI programs can
have a negative impact on employment. In this paper, we show that a well
designed public provision of UI can improve the efficiency of the economy. The
existence of publicly provided UI reduces the need for firms to provide insurance
to their workers. As the size of the severance payment shrinks, the firm’s com-
mitment problem associated with making the payment diminishes. Thus, public
provision of UI can reduce both layoffs and the volatility of the workers’ income.
Of course, UI programs can introduce distortions of their own. Unemployment
benefits subsidize layoffs, making layoffs more attractive to the firm. A conven-
tional way to correct this problem is to introduce an experience rating in the
2
way the system is financed. Experience rating forces firms to internalize part
of the subsidy provided by unemployment benefits, and consequently reduces
the distortion in the layoffs decisions. We demonstrate that publicly provided
UI, featuring a sufficiently high experience rating can improve the efficiency of
an economy. The remainder of the paper is organized as follows. In next section,
we present an overview of the model. In Section 3, we analyze the equilibrium,
both when contracts are enforceable, and when they are not. Publicly provided
UI is discussed in Section 4. We conclude in Section 5. All proofs are in the
appendix.
2. The model
The economy is composed of an infinite number of subsequent generations. Each
2
Topel and Welch (1980) has a good survey on the topic.
2
generation lives for two periods. M workers are born at the beginning of each
generation and live for two periods. We call the workers young in their first
period of life, and old in their second period. Workers are risk averse, do not
have access to financial market, and do not discount the future. The per period,
twice continuously differentiable utility function is given by U = U (W + R)
where W is the outside income and R is home production. This home production
takes a value of zero if the worker is working, and a value of r if the worker is
unemployed. It is also assumed that U ≥ 0 and U ≤ 0.
The first sector of the two-sector economy which is denoted sector 1 is composed
of a large number of infinitely lived identical firms, which are not exposed to
productivity shocks. It is assumed that firms in this sector have constant returns
to scale and act competitively. Worker’s productivity per period is given by x
where x > r. It is assumed that workers who accept jobs in the other sector
cannot move back to sector 1 if they are laid off.3
The other sector denoted sector 2 is composed of K identical firms that are
infinitely lived. Firms discount each generation at a rate of δ, but do not discount
within a generation. Firms take prices as given and are risk neutral. The output
price is normalized to one. In the first period of each generation, production
function is F [N ] where N is the number of workers working for the firm. It is
assumed that F > 0 and F −1 for
every value of s between zero and full insurance.
Assumption I says that the elasticity of the number of layoffs with respect with
changes in s has to be less that one in absolute value. If it was not the case, the
firm could lower the total cost of the severance payment by increasing the size
of the severance payments.
We now solve for the contract {wy , wo , s} that will be offered by firms to workers.
Firms will choose a contract that will maximize their profits subject to the
participation constraint of the workers. Every worker can choose to work in
sector 1 for a wage of x − τ for two periods, which will give them a lifetime
utility of 2U (x − τ ). On the other hand, if a worker accepts a long-term contract
in sector 2, the worker will receive wy for the first period and wo for the second
period, if the worker is not laid off. However, because of the productivity shock
in the second period, firms will lay-off some workers. Because all workers have
7
the same productivity, the probability that a worker been laid-off is simply
N(wy ,s)−n(wo ,s)
N(wy ,s) . The participation constraint of a worker is given by:
n(wo , s) N (wy , s) − n(wo , s)
U (wy ) + U (wo ) + U (r + s + b) ≥ 2U (x − τ ).
N (wy , s) N (wy , s)
Firms offer workers a contract that maximizes the firm’s expected profit, subject
to the worker’s participation constraint.13 For future reference we will call this
contract type I.
Lemma 1: When severance payments are enforceable, firm fully insure workers
by offering a long term contract where wy = wo = x − τ and s = x − τ − r − b.
The intuition behind lemma 1 is straightforward. Because workers are risk
averse, it is optimal for firms in sector 2 to fully insure workers. Given this
contract, the number of workers retained n is given by pF [n ] = r + b − τ −
eb. An increase in b reduces severance payments in a proportion one to one,
since workers are fully insured. However, because of this reduction in severance
payments, firms will retain fewer workers. On the other hand, an increase in e
increases the number of workers who are kept by increasing the cost of layoffs.
The number of workers hired N is given by F [N ] = 2x − τ − r − b + eb. The
number of workers hired is increasing in b, but is decreasing with e. The per
periods profits under this contract are:
π = F [N ] − xN + pF [n ] − xn − (x − τ − r − b + eb)[N − n ].
Under this contract, if the unemployment insurance system features a full expe-
rience rating (e = 1 and τ = 0) then hiring and layoff decision are made in a way
13
By doing so we implicitly assume that there is a very large number of workers
and a relatively smaller number of job in sector 2, so the bargaining power is
all in the hand of the firm in sector 2. Obviously this is not the only possible
labor market structure, in particular we could introduce bargaining over wages
with different bargaining power. In such environment participation constraint will
be more difficult to satisfy, and workers would achieve a higher level of utility.
However, the incentive side of the problem, and the need to provide insurance
will remain present, generating similar results.
8
which is consistent with what a surplus maximizing planer will choose. Wages
will be equal to x and severance payment will be equal to x − r − b. However if
e r + sII + b).
If we compare a self-enforcing contract of type II with one of type I, we observe
the following differences. First, as mentioned in Lemma 2, type II contracts do
not fully insure workers.15 Under both contracts the lifetime expected utility of
a worker in sector 2 is the same. In order to relax their commitment constraint,
14
We are looking for a steady state level of profits which is incentive compatible.
Obviously since the game is infinitely repeated, variable profits which are differ-
ent from the steady state but yield the same present value could be incentive
compatible. However to keep the analysis simple we abstract form that issue.
15
One peculiarity of this self-enforcing contracts is that the wage contracts will
10
firms reduce the severance payments. Since firms in sector 2 do not fully insure
workers, the equilibrium contract implies that firms will retain fewer workers
than in the full commitment equilibrium.
Consider the ”laissez-faire” economy where b = 0 and τ = 0. If the contract
fully insured workers, then wo = s + r. In this case, the number of worker
retained by the firm is given by pF (n) = wo − s = r. Under the type II
II
contract wo > sII +r, and the number of workers retained nII is determined by
II
pF (nII ) = wo −sII > r. Therefore, by concavity of the production function we
know that nII x and sII N . Similarly, if wy + sII > 2x − r,
then N II 0. Because nII = −nIIo , and because of assumption I,
s w
we know that the above inequality will be satisfied. If the right hand side of
II
(B6) is positive, by concavity we know that wo > r + sII + b.
Proof of Theorem 1:
Under full experience rating τ = 0 and e = 1. If b = x − r we know that for
¯
any given δ, the contract is of type I since δ = 0. Consequently, substituting
the values of b, e and τ , we get that s = 0 and wo = wy = x. Under these
conditions n is given by pF (n ) = x − b = r, which is efficient. Similarly, N
is given by F (N ) = x + b = 2x − r, which is also efficient. The contract is first
best.
16
Proof of Corollary 1:
The proof goes in the same line that Theorem 1, under full experience rating
τ = 0 and e = 1. For a given δ, the contract is of type I is self-enforcing if:
¯ [N (bf ) − n (bf )](x − r − bf )
δ=δ= .
[N − n ](x − r − bf ) + π (bf )
Consequently, substituting the values of b, e and τ , we get that s = x−r−bf
and wo = wy = x. Under these conditions n is given by pF (n ) = r, which is
efficient. Similarly, N is given by F (N ) = 2x − r, which is also efficient. The
contract is first best.
17
6. References.
Boadway R. and A.J. Oswald (1983), “Unemployment Insurance and Redis-
tributive Taxation”, Journal of Economics 20, 193–210.
Boadway R. and N. Marceau (1994), “Time Inconsistency as a Rationale for
Public Unemployment Insurance”, International Tax and Public Finance
1, 107–126.
Burdett, K. and R. Wright (1989), “Optimal Firm Size, Taxes, and Unemploy-
ment”, Journal of Public Economics 39, 275–287.
Carmichael L. (1983), “Firm Specific Human Capital and Promotion Ladders”,
Bell Journal of Economics 14, 251–358.
Greif A., P. Milgrom and B.R. Weingast (1994), “Coordination, Commitment,
and Enforcement: The Case of the Merchant Guild”, Journal of Political
Economy 102, 745–776.
Macloed, W.B. and J. Malcomson (1989), “Implicit Contracts, Incentive Com-
patibility, and Involuntary Unemployment”, Econometrica 57, 447–480.
Marceau N. (1993), “Unemployment Insurance and Market Structure”, Journal
of Public Economics 52, 237–249.
Oswald A.J. (1986), “Unemployment Insurance and Labor Contracts under
Asymmetric Information: Theory and Facts”, The American Economic Re-
view 76, 365–377.
Shapiro C. and J.E. Stiglitz (1984), “Equilibrium Unemployment as a Worker
Discipline Device”, The American Economic Review 74, 433–444.
Thomas J. and T. Worrall (1988), “Self-Enforcing Wage Contracts”, Review of
Economics Studies 55, 341–353.
Topel R. and F. Welch (1980), “Unemployment Insurance: Survey and Exten-
sion”, Economica 47, 351–379.
18