Pension Choices and Job Mobility in the UK.

Preparing to load PDF file. please wait...

0 of 0
100%
Pension Choices and Job Mobility in the UK.

Transcript Of Pension Choices and Job Mobility in the UK.

Pension Choices and Job Mobility in the UK.
Vincenzo Andrietti∗ Universidad Carlos III de Madrid, Departamento de Economia
January 17, 2003
Abstract Using data from the British Household Panel Survey we analyze the impact of second tier pension scheme choices on job mobility within a discrete time hazard rate framework. We find that workers either offered and participating or offered and not participating to an occupational pension plan have significantly lower quit rates. However, once the endogeneity of pension scheme status is accounted for through an instrumental variable procedure, these coefficients are no more significant. Alternatively, the effect of pension portability losses on quits’ hazards is never significant. The additional finding that workers participating to occupational pension plans are significantly less likely to quit for a non pension job can be interpreted as indirect evidence that they are in ”good” jobs. Keywords: Labour mobility, Occupational Pension Plans, Duration Analysis, Instrumental Variables. JEL classification: C41, J31, J32, J41, J63, J68. Wold Count: approx. 8.000.
∗Correspondence address: Universidad Carlos III de Madrid, Departamento de Economia. c/Madrid 126, 28903 Getafe (Madrid). Tel. +34 916245744. Fax +34 91 6249875. E-mail: [email protected] The author wishes to thank Pedro Albarran, Rachel Carrasco, Marcel Jansen, Barbara Petrongolo and seminar participants at ESPE and EALE Conferences 2002 for helpful comments. Financial support from a Marie Curie Fellowship of the European Community program Improving Human Potential under contract number HPMF-CT-2000-00504 is gratefully acknowledged. The data used in this paper are available from the Data Archive of the University of Essex.
1

1 Introduction
Notwithstanding the pension reforms implemented in the last three decades, portability of occupational pension rights is still a matter of public policy concern in the United Kingdom. While provisions such as the reduction of the vesting period, inflation indexation of deferred rights and the introduction of transfer options, have improved the position of early leavers from occupational pension plans, there is still concern that a lack of pension portability could be detrimental to labour market efficiency. Little empirical work has been produced to support the policy debate. The few available empirical studies are based on cross sectional data with retrospective questions or on short panel data and/or do not take into account workers’ selection into pension arrangements. The ”selection” issue is particularly relevant in the UK context, where workers have the right to choose among different pension arrangements provided to supplement the basic State pension. This institutional framework is the outcome of a number of reforms aiming to stimulate labour market flexibility through more extensive pension choices.
We estimate discrete time hazard rate models of job to job mobility where transitions are defined according to exit reasons (voluntary/involuntary) and destination states (quit to pension/non pension jobs). Our objective is to analyze the impact of
1

occupational and personal pension arrangements on voluntary job mobility. If pension participation status reflects workers’ choices, neglecting this endogeneity will introduce bias in the estimates. We deal with this problem using an instrumental variables (IV) approach. In particular, to identify the effect of pension choices on job mobility hazards we exploit the exogenous variation provided by a derived instrumental variable representing the occupational pension offer rate by industry, union coverage and firm size. The occupational pension offer rate is expected to bear a positive and significant correlation with occupational pension coverage and participation while being unrelated to job turnover. Such a variable is included in the pension choice probit equations while being excluded from the job to job transition hazards.
We find that the effects of occupational pension schemes on job mobility hazards change significantly while moving from the "simple" hazard model to the IV hazard model. In the former, workers either offered and participating or offered but not participating to an occupational pension plan have significantly lower quit rates. However, once endogeneity is accounted for the occupational pension effects are no more significant. The validity of our instrument is confirmed by its statistical significance in explaining pension choices as well as by an Hausman endogeneity test. These results confirm the importance of accounting for the endogeneity of pension choices, an issue
2

that has been disregarded by most of the empirical literature. The additional finding that workers participating to an occupational pension plan are significantly less likely to quit towards a non pension job can be interpreted as indirect evidence that they are in ”good” jobs.
The remaining of the paper is organized as follows. The next section describes the structure of the UK pension system, focusing on its second pillar and on occupational pensions portability regulation. Section 3 reviews the empirical literature. Section 4 introduces the empirical modelling framework. Section 5 describes the data. Section 6 illustrates the results. Section 7 concludes.
2 Pension Choices and Portability Issues in the United Kingdom
The current UK pension system has a three tiered structure. The first tier is public, and consists of a basic flat-rate pension as well as of a means-tested benefit. The second tier is mandatory and pension provision is split between the State - in the form of the State Earnings-Related Pension Scheme (SERPS) - and private companies - in the form of occupational - employer sponsored - and personal pension schemes. Finally, there is a third tier of voluntary private retirement saving.
The wide variety of retirement pension plans currently offered in the UK is the result
3

of a number of reforms undertaken over the last 25 years, whose main aim was to offer individuals a wider choice of retirement pension vehicles. Central to this strategy was the ”contracting-out” mechanism, introduced originally in 1978 as a means of integrating existing occupational DB pension schemes into the SERPS.1 While employees with earnings in excess of a ”lower earning limit” were automatically enrolled into SERPS, initially they were also given the option of contracting-out into an approved defined benefit (DB) occupational pension scheme.2 The 1986 Social Security Act extended the contracting-out option to approved defined contribution (DC) plans3 and to approved personal pensions.4 Moreover, any contractual membership requirement to an employer’s scheme was abolished, while employees were allowed to ”opt out” from an occupational plan to subscribe a personal pension.
Employers are not mandated to sponsor an occupational pension plan nor an approved occupational pension plan5, while employees can always decide to remain into SERPS or to later reenter it. Finally, individuals can eventually top up their occupational or personal pension with additional voluntary contributions or free-standing additional voluntary contributions (up to the limits permitted by the Island Revenue).6
Tables 1 and 2 report figures elaborated from the Occupational Pensions Schemes Survey collected over the 90s by the Government Actuary.7 Table 1 indicates that there
4

has been a downward trend in private sector occupational pension schemes’ membership over the decade. Alternatively, the distribution of active members by type of plan has remained relatively stable over time, with more than 80 percent of plan participants belonging to DB plans. Table 2 indicates that the level of contracting-out also has been fairly stable over time, with more than 80 percent of private sector workers participating to contracted-out plans. Most workers participate to contracted-out DB plans - Contracted-Out Salary Related (COSR) schemes - while a minority of workers participate to contracted-out DC plans - Contracted-Out Money Purchase (COMP) schemes. A minority of workers participate to Contracted In Salary Related (CISR) or Money Purchase (CIMP) schemes.
Pension regulation usually defines the standard portability options available to a worker leaving an occupational pension plan before retirement age. According to this general framework, pension portability rules in a pension plan define the rights of early leavers. Typically, these rules provide that an individual is entitled to join a pension plan only upon satisfaction of some eligibility condition (related to service, age or employment status). Once eligible, workers joining an occupational pension plan are usually required to complete a further vesting period before being entitled to any pension rights’ accrual. Eligibility conditions and vesting periods apply to occupational
5

pension plans independently of their DB/DC nature. While vesting periods are usually short - e.g. 2 years in the UK - a more relevant portability issue is typically arising in DB plans, because of their implicitly backloaded pension rights’ accrual structure. The typical DB plan promises workers to pay a retirement pension annuity related to the length of service and to the final salary. In case a separation occurs before retirement, vested workers are entitled to a deferred retirement pension determined on the basis of earnings received upon leaving the firm. If deferred benefits are not indexed to inflation and to productivity growth even workers moving betweens firms offering identical DB plans and wage profiles will accumulate lower total pension benefits than workers remaining in the same firm throughout their career. The shortfall of actual retirement benefits from those that would have been paid if there had been no change in scheme membership as a consequence of job separations during the worker career represents the opportunity cost of leaving the current pension plan8.
In contrast, workers covered by DC arrangements have a legal claim on the individual pension account in which all pension contributions have been invested. If the funds remain in the account after the worker leaves the firm, the account will continue to grow by the accumulated returns on invested assets. Alternatively, the funds can be transferred to a different occupational or personal pension plan. In either case, DC
6

plans can be defined as fully portable, given that a job changer retains the full value of his/her pension account.
While in the UK there are not specific legislative provision regarding eligibility conditions to an occupational plan9, a number of legislative changes have contributed to improve the situation of early leavers over the last 25 years. Before 1975, early leavers had no legal right to transfer their accrued pension entitlements to a new scheme or even to receive deferred benefits from their old scheme. Under the current rules, the vesting period is set at two years of pension plan membership. In particular, vested early leavers from DB plans can have their accrued rights preserved in the pension scheme as deferred benefits, to be revalued until retirement guaranteeing a minimum ”limited price indexation” in line with the Retail Price Index, up to a maximum of 5 percent. Alternatively they can take a tax free transfer value to a different occupational pension scheme (either DB or DC) or to an approved personal pension or to purchase a retirement annuity.
3 Literature
The impact of employer provided pension plans on individual job mobility choices has been widely investigated in the US pension literature. While early empirical studies10 documented a significant negative correlation between plan participation and job mo-
7

bility, more recent literature has tried to explain this stylized fact using three main arguments: pension portability losses arising to early leavers from DB plans11, compensation premiums accruing to pension covered workers12 or the ”self-selection” of ”low discounters” into pension jobs13.
Evidence for the United Kingdom is essentially limited to five previous studies. McCormick and Hughes (1984) use the 1974 General Household Survey (GHS) to estimate ”intentions to quit” and turnover logit equations. While they explicitly derive pension portability losses, their empirical specification only contains a pension participation dummy and its interaction with job tenure. They find that pension workers are significantly less likely to move, while the negative and significant size of the interaction term suggests that tenure matters only in pension jobs. Henley, Disney and Carruth (1994) use the 1985 GHS to estimate hazard rates of exit from jobs without distinguishing between exit routes. They find that occupational pension scheme membership significantly decreases the hazard, while transferability of pension rights increases it. They include a quadratic pension-tenure interaction term which turns out to be negative and significant, confirming the McCormick-Hughes proposition that the pension loss function is time dependent and possibly non-linear. Mealli and Pudney (1996) is the only study that takes into account the potential en-
8

dogeneity of pension participation choices, although the use of the retrospective sample design provided by the 1988-1989 Retirement Survey could undermine the accuracy of their results. They analyze the role of unobserved individual specific characteristics in explaining the relationship between pension coverage status and labor force transitions. Estimating a random-effects competing risks model they provide evidence of a strong positive association between the length of job tenure and pension participation status, while no role is found for unobservables. These findings run against the self-selection hypothesis; however, the role of pension portability losses is not explicitly modelled.
In a recent paper14, we use data from the European Community Household Panel (ECHP) survey to analyze the effect of occupational pensions and pension portability losses on interfirm job mobility within a switching regression framework. We find that UK workers participating to an occupational pension plan are significantly less likely to move, while pension portability losses do not seem to act as a significant mobility impediment. Rather, the finding of positive wage premiums accruing to occupational pension workers is consistent with the view that the latters are less likely to leave because they hold ”good” jobs.
A more recent study is provided by Disney and Emerson (2002) using BHPS data. They find that not only those workers who join an occupational pension plan offered
9
WorkersPension PlanPension ChoicesLeaversPension Portability Losses