Social Security and Old Age Unemployment Risk with Efficient Wage Bargaining

• Published date: 01 July 2019
• DOI: http://doi.org/10.1111/sjoe.12301
• Date: 01 July 2019

Scand. J. of Economics 121(3), 1125–1153, 2019

DOI: 10.1111/sjoe.12301

Social Security and Old Age Unemployment

Risk with Efficient Wage Bargaining*

Pietro Reichlin†

LUISS Guido Carli, 00197 Rome, Italy

preichlin@luiss.it

Abstract

I analyze a life-cycle economy with old age productivity risk where wages, employment, and

severance payments are set through efficient bargaining between risk averse unions and risk

neutral firms. Allocations with limited union membership are second-best inefficient as they

generate too little labor supply in young age, too much consumption before retirement, too

little employment of older workers (early retirement), and too little insurance against old age

unemployment. Providing publictransfers to early retirees (disability benefits or early pensions)

might help to increase the degree of risk sharing at the cost of lower old age employment.

Depending on whether absolute risk aversion is increasing or decreasing in consumption, these

policies might or might not produce efficiency gains at equilibrium.

Keywords: Labor market; retirement benefits; risk sharing

JEL classification:E21; E24; H2; H21; H55

I. Introduction

In this paper, I study the interplay between public social security programs

and firms’ labor shedding policies in a dynamic economy characterized

by the unions’ strong bargaining power, labor protection laws, and social

rights based on employment entitlements, which are largely prevalent in

the European labor market. The main question I raise is that of which

public policies are best suited to eliminate the potential inefficiencies that

can arise in these type of environments. Should the government abstain

from providing any sort of insurance against the risk of an early exit from

the labor market to older workers who are facing individual productivity

shocks? Are layoff taxes a good instrument to limit firms’ excessive recourse

to labor shedding policies? Three main observations are at the heart of these

questions and motivate this contribution. The first is that, in many countries,

older workers are overrepresented in redundancies (Samorodov, 1999), face

higher unemployment risks, and are less likely to be re-employed or re-

trained. The second is that labor force participation rates of older workers

*I thank two anonymous referees for helpful comments. Support of a grant from MURST,PRIN

2010, is gratefully acknowledged.

†Also affiliated with CEPR.

C

The editors of The Scandinavian Journal of Economics 2018.

1126 Social security and old age unemployment

(i.e., aged 55–64) have fallen dramatically over the past 40 years, most

prominently in Continental Europe (see Gruber and Wise, 2004). The third

is that employers might use public insurance programs (unemployment

and disability benefits) as “an easy way of getting rid of aging workers

with higher wage demands” (Schnalzenberger and Winter-Ebmer, 2009). In

addition, when postponing the statutory retirement age, governments might

produce more unemployment and/or increase the costs of unemployment

insurance faced by public agencies or firms.1

To provide some insight on these questions, I consider a

stylized overlapping generations model with three-period-lived generations

embedded in an economy with the following main features. First,

employment, severance payments, and wages are set through bilateral

contractual agreements between unions and firms. Secondly, the labor

market is segmented into insiders and outsiders. The latter are young and

middle age workers moving between jobs, offering their labor services in a

competitive environment. However, insiders are represented by middle age

workers, who are not moving across jobs, and by old workers. The latter are

assumed to be more risky (i.e., subject to random individual productivity

shocks) and unable to be relocated to different jobs due to skill depreciation.

In my model, there is no market for insurance against unemployment risks

and workers are risk averse. When workers become old, their individual

productivity levels are revealed, and the firm decides whether to keep them

or lay them off with a retirement package. It must be stressed that “laid-off”

workers, in my model, are effectively in a state of early retirement, as they

are receiving both a contractual severance payment and a transfer from a

public insurance scheme (disability benefit or early pension). Hence, an old

laid-off worker is assumed to have exited the labor force and not be looking

for a new job. A labor contract is the solution to the unions’ expected utility

maximization subject to non-negative profits and a no-shirking condition.

The latter requires the old workers’ utility contingent on being employed to

be above a reservation level and prevents the optimal labor contract from

offering full insurance against unemployment risk. I follow Blanchard and

Tirole (2008) by assuming that there is no aggregate uncertainty and that

individual productivity is observed by the firm only (neither the worker nor

the union nor a third party). In particular, the firm’s employment decision

follows from setting a threshold such that all workers whose productivity

falls below it are dismissed (and provided with an early retirement package)

and the remaining workers are kept on the payroll, earning a productivity-

1As noted by Ebbinghaus (2001), “as governmentsseek to close down preretirement programmes

or make public benefits less attractive, private occupational welfare efforts may gain in

importance, as they may counteract welfare cuts by filling replacement gaps”.

C

The editors of The Scandinavian Journal of Economics 2018.