Nonparametric Evidence on the Effects of Financial Incentives on - - PowerPoint PPT Presentation

▶

Mar 21, 2024 32 likes •402 views

Nonparametric Evidence on the Effects of Financial Incentives on Retirement Decisions Day Manoli Andrea Weber UT Austin & NBER University of Mannheim June 2013 Manoli and Weber () Financial Incentives at Retirement June 2013 1 / 37

SLIDE 1

Nonparametric Evidence on the Effects of Financial Incentives on Retirement Decisions

Day Manoli Andrea Weber UT Austin & NBER University of Mannheim June 2013

Manoli and Weber () Financial Incentives at Retirement June 2013 1 / 37

SLIDE 2

Introduction

How do individuals adjust labor supply in response to wage variation? Answer important in many fields: macro, public finance, labor

◮ Design of equilibrium models of the labor market ◮ Normative tax policy analysis ◮ Modeling labor supply ◮ Dispute over magnitude of elasticities between micro and macro studies

Recent literature has highlighted the importance of distinguishing between the intensive and extensive margin (Heckman 1993, Saez 2002, Chetty et al. 2012) We focus on retirement decision in response to anticipated benefits Contribution: quasi-experimental estimates of extensive margin labor supply elasticities

Manoli and Weber () Financial Incentives at Retirement June 2013 2 / 37

SLIDE 3

Introduction

How do individuals adjust retirement entry to anticipated benefits?

◮ Relevant for design of pension systems, reforms, explaining retirement

patters based on financial incentives

◮ Complicated to disentangle incentives from various policies at

retirement

Research design based on a simple and salient incentive structure ’independent’ from public pension system Exploit discontinuities in financial incentives along a dimension other than age Reduced form concept of extensive margin intertemporal substitution elasticities Noparametric estimation method based on bunching estimators (Saez 2009, Chetty et. al 2012, Kleven and Waseem 2013)

Manoli and Weber () Financial Incentives at Retirement June 2013 3 / 37

SLIDE 4

Research Design

Retirement rule in Austria: individuals who complete 10 years of tenure by retirement qualify for lump-sum payment from employer

◮ Simple and salient rule ◮ Benefits are fully anticipated but small relative to lifetime wealth ◮ Focus on delay in retirement entry decisions

Examine retirement behavior around the thresholds

◮ Present graphical evidence on responses in retirement entry ◮ Examine heterogeneity across population groups

Examine magnitude of financial incentive

◮ Contrast legislative incentive with estimated incentives from the data

Elasticity relates retirement responses to financial incentives

Manoli and Weber () Financial Incentives at Retirement June 2013 4 / 37

SLIDE 5

Institutional Background I

Employer-provided retirement benefits within the severance pay system amount based on tenure at retirement mandated, lump-sum payments at retirement payments based on salary, not income / total compensation funds set aside by employers based on size of work force retirement is only voluntary separation that leads to payments severance payment is taxed at a constant, low rate of 6%

Manoli and Weber () Financial Incentives at Retirement June 2013 5 / 37

SLIDE 6

Severance Payments as Fraction of Annual Salary

.333 .5 .75 1 5 10 15 20 25 30 Years of Tenure at Retirement

Manoli and Weber () Financial Incentives at Retirement June 2013 6 / 37

SLIDE 7

Institutional Background II

Government-provided pensions: Normal (statutory) retirement ages: 65 (men) & 60 (women) Early retirement ages: 60 (men) & 55 (women) Ages 55-59: retirement through disability pensions Replacement rates ≈ 75% Even with bonuses for retirement at older ages, system is still actuarially unfair for most individuals

Manoli and Weber () Financial Incentives at Retirement June 2013 7 / 37

SLIDE 8

Exits from the Labor Force into Retirement

Manoli and Weber () Financial Incentives at Retirement June 2013 8 / 37

SLIDE 9

Data

Austrian Social Security Database 1972 - 2006

◮ matched employer-employee census of private sector ◮ complete earnings and employment histories ◮ some demographic information on workers and firms

Income Tax Records 1994 - 2005

◮ employer reports to tax office at the end of the year ◮ annual salary plus withholdings of social security contributions and

income taxes

◮ separate category for severance payments

Sample restrictions:

◮ non-construction workers ◮ individuals still working at age 54 ◮ retiring within 6 months of last job ◮ with uncensored job-tenure at retirement ◮ retirements 1997-2005 matched to tax records

Sample: 89,426 individual retirements with 6-28 years of tenure

Manoli and Weber () Financial Incentives at Retirement June 2013 9 / 37

SLIDE 10

Summary Statistics

# Individuals 89,426 Mean Std.dev Female 0.51 Retirement Age 59.1 2.59 Tenure 15.66 6.10 Annual Earnings 29,327 12,949 Severance Pay 18,510 21,661 Implicit Tax Rate 0.81 0.21 Years of Employment 34.3 8.67 Years of Sick Leave 0.17 0.33 Fractions: Disability Pension 0.21 Early Retirement 0.57 Old Age Pensions 0.22

Manoli and Weber () Financial Incentives at Retirement June 2013 10 / 37

SLIDE 11

Outline

Empirical evidence on retirements Empirical evidence on severance payments Reduced form elasticity concept Estimation strategy Estimation results Discussion: interpretation and policy relevance

Manoli and Weber () Financial Incentives at Retirement June 2013 11 / 37

SLIDE 12

Distribution of Tenure on Retirement

200 400 600 800 1000 Individuals 10 15 20 25 Years of Tenure at Retirement

Manoli and Weber () Financial Incentives at Retirement June 2013 12 / 37

SLIDE 13

Adjusting for Coviariates

.02 .04 .06 .08 .1 .12 10 15 20 25 Years of Tenure

Manoli and Weber () Financial Incentives at Retirement June 2013 13 / 37

SLIDE 14

Distribution of Job Starts by Age

2000 4000 6000 8000 10000 12000 # of Individuals Starting New Jobs 40 45 50 55 Age Women Men

Manoli and Weber () Financial Incentives at Retirement June 2013 14 / 37

SLIDE 15

Hetergogeneity by Health Status

Manoli and Weber () Financial Incentives at Retirement June 2013 15 / 37

SLIDE 16

Hetergogeneity by Age and Gender

50 100 150 200 250 100 200 300 400 100 200 300 200 400 600 800 100 200 300 50 100 150 10 15 20 25 10 15 20 25 10 15 20 25

Male, Retire Age < 60 Male, Retire Age = 60 Male, Retire Age > 60 Female, Retire Age < 60 Female, Retire Age = 60 Female, Retire Age > 60

Individuals Years of Tenure at Retirement

Manoli and Weber () Financial Incentives at Retirement June 2013 16 / 37

SLIDE 17

Distribution of Severance Payments, 10 Year Threshold

5 10 15 20 Percent .25 .33 .5 .75 1

A1. Tenure 9 years

5 10 15 20 Percent .25 .33 .5 .75 1

A2. Tenure 10 years

Manoli and Weber () Financial Incentives at Retirement June 2013 17 / 37

SLIDE 18

Distribution of Severance Payments, 15 Year Threshold

5 10 15 20 Percent .25 .33 .5 .75 1

B1. Tenure 14 years

5 10 15 20 Percent .25 .33 .5 .75 1

B2. Tenure 15 years

Manoli and Weber () Financial Incentives at Retirement June 2013 18 / 37

SLIDE 19

Distribution of Severance Payments, 20 Year Threshold

5 10 15 Percent .25 .33 .5 .75 1

C1. Tenure 19 years

5 10 15 Percent .25 .33 .5 .75 1

C2. Tenure 20 years

Manoli and Weber () Financial Incentives at Retirement June 2013 19 / 37

SLIDE 20

Distribution of Severance Payments, 25 Year Threshold

5 10 15 Percent .25 .33 .5 .75 1

D1. Tenure 24 years

5 10 15 Percent .25 .33 .5 .75 1

D2. Tenure 25 years

Manoli and Weber () Financial Incentives at Retirement June 2013 20 / 37

SLIDE 21

Distribution of Severance Payments

.2 .4 .6 .8 1 Sev Pay Fraction 5 10 15 20 25 30 Years of Tenure at Retirement

A. 15th Percentile

.2 .4 .6 .8 1 Sev Pay Fraction 5 10 15 20 25 30 Years of Tenure at Retirement

B. 25th Percentile

.2 .4 .6 .8 1 Sev Pay Fraction 5 10 15 20 25 30 Years of Tenure at Retirement

C. 50th Percentile

.2 .4 .6 .8 1 Sev Pay Fraction 5 10 15 20 25 30 Years of Tenure at Retirement

D. 75th Percentile

Manoli and Weber () Financial Incentives at Retirement June 2013 21 / 37

SLIDE 22

Definition of Responder Sample, 10 Year Threshold

200 400 600 800 Individuals 7 8 9 10 11 12 13 Years of Tenure at Retirement

A1. 10 Year Threshold, Responders

200 400 600 800 Individuals 7 8 9 10 11 12 13 Years of Tenure at Retirement

A2. 10 Year Threshold, Non-Responders

Manoli and Weber () Financial Incentives at Retirement June 2013 22 / 37

SLIDE 23

Definition of Responder Sample, 15 Year Threshold

200 400 600 800 Individuals 12 13 14 15 16 17 18 Years of Tenure at Retirement

B1. 15 Year Threshold, Responders

200 400 600 800 Individuals 12 13 14 15 16 17 18 Years of Tenure at Retirement

B2. 15 Year Threshold, Non-Responders

Manoli and Weber () Financial Incentives at Retirement June 2013 23 / 37

SLIDE 24

Definition of Responder Sample, 20 Year Threshold

200 400 600 Individuals 17 18 19 20 21 22 23 Years of Tenure at Retirement

C1. 20 Year Threshold, Responders

200 400 600 Individuals 17 18 19 20 21 22 23 Years of Tenure at Retirement

C2. 20 Year Threshold, Non-Responders

Manoli and Weber () Financial Incentives at Retirement June 2013 24 / 37

SLIDE 25

Definition of Responder Sample, 25 Year Threshold

200 400 600 Individuals 22 23 24 25 26 27 28 Years of Tenure at Retirement

D1. 25 Year Threshold, Responders

200 400 600 Individuals 22 23 24 25 26 27 28 Years of Tenure at Retirement

D2. 25 Year Threshold, Non-Responders

Manoli and Weber () Financial Incentives at Retirement June 2013 25 / 37

SLIDE 26

Mean Severance Payments, Responder Sample

.15 .2 .25 .3 .35 Severance Pay Fraction 6 7 8 9 10 11 12 13 Years of Tenure at Retirement

A. 10 Year Threshold

Manoli and Weber () Financial Incentives at Retirement June 2013 26 / 37

SLIDE 27

Mean Severance Payments, Responder Sample

.3 .35 .4 .45 .5 Severance Pay Fraction 12 13 14 15 16 17 18 Years of Tenure at Retirement

B. 15 Year Threshold

Manoli and Weber () Financial Incentives at Retirement June 2013 27 / 37

SLIDE 28

Mean Severance Payments, Responder Sample

.55 .6 .65 .7 .75 Severance Pay Fraction 17 18 19 20 21 22 23 Years of Tenure at Retirement

C. 20 Year Threshold

Manoli and Weber () Financial Incentives at Retirement June 2013 28 / 37

SLIDE 29

Mean Severance Payments, Responder Sample

.8 .85 .9 .95 Severance Pay Fraction 22 23 24 25 26 27 28 Years of Tenure at Retirement

D. 25 Year Threshold

Manoli and Weber () Financial Incentives at Retirement June 2013 29 / 37

SLIDE 30

Participation Elasticity

Relate retirement responses to financial incentives ε = ∆p/p −∆(1 − τ)/(1 − τ) Retirement response: relative increase in retirements at the tenure thresholds ∆p/p Financial incentive: increase in the implicit tax rate ∆(1 − τ)/(1 − τ) = SP ∗ (1 − tsev) y ∗ (1 − τ)

Manoli and Weber () Financial Incentives at Retirement June 2013 30 / 37

SLIDE 31

Estimating Changes in Retirement

Manoli and Weber () Financial Incentives at Retirement June 2013 31 / 37

SLIDE 32

Estimating Changes in Severance Pay

.15 .2 .25 .3 .35 Severance Pay Fraction 6 7 8 9 10 11 12 13 Years of Tenure at Retirement

A. 10 Year Threshold

Manoli and Weber () Financial Incentives at Retirement June 2013 32 / 37

SLIDE 33

Participation Elasticities

Threshold 10 Year 15 Year 20 Year 25 Year Average N=21,729 N=19,724 N=15,588 N=18,461 Change in Retirement Probabilities 0.1414 0.2424 0.3777 0.2123 0.2434 (0.0233) (0.0277) (0.0350) (0.0251) (0.0157) Change in Sev Pay Fraction 0.0620 0.1056 0.1202 0.0514 0.0848 (0.0046) (0.0058) (0.0049) (0.0070) (0.0028) Change in Net-of-Tax Rate 0.2916 0.4963 0.5651 0.2415 0.3986 (0.0215) (0.0275) (0.0229) (0.0331) (0.0131) Elasticity 0.4848 0.4883 0.6684 0.8790 0.6301 (0.0892) (0.0622) (0.0683) (0.1668) (0.0559)

Manoli and Weber () Financial Incentives at Retirement June 2013 33 / 37

SLIDE 34

Participation Elasticities by Gender

Threshold 10 Year 15 Year 20 Year 25 Year Average Men Change in Retirement Probabilities 0.0975 0.1616 0.3889 0.1926 0.2102 (0.0364) (0.0397) (0.0560) (0.0322) (0.0227) Elasticity 0.3881 0.3815 0.7904 0.8455 0.6014 (0.1613) (0.1024) (0.1298) (0.2475) (0.0877) Women Change in Retirement Probabilities 0.1729 0.2999 0.3713 0.2375 0.2704 (0.0288) (0.0375) (0.0421) (0.0415) (0.0203) Elasticity 0.5380 0.5502 0.5990 0.9106 0.6495 (0.1051) (0.0762) (0.0737) (0.2691) (0.0795)

Manoli and Weber () Financial Incentives at Retirement June 2013 34 / 37

SLIDE 35

Discussion

Reduced from elasticity not derived as a structural model parameter Independent from parametric assumptions and specific model context Robust to scaling assumptions that arise in a static model framework a la Saez (2010) Elasticity can be used for calibration in standard macro model , i.e. Rogerson and Wallenius (2009) But: under more general model assumptions extensive margin intertemporal labor supply elasticity is an ambiguous concept (Attanasio, 2012) Policy experiment

◮ Government provides lump-sum bonus for retiring at age 61 or later ◮ Implies increase in average implicit tax rate from 0.8 to 1.00 ◮ Increase in fraction retiring at age 61 by ˆ

e ∗ dln(1 − τ)

Manoli and Weber () Financial Incentives at Retirement June 2013 35 / 37

SLIDE 36

Labor Supply Responses to Retirement Bonus at Age 61

10,000 20,000 30,000 Individuals 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 Baseline - No Bonus with Bonus at 61+, e=0.25 with Bonus at 61+, e=0.65

Manoli and Weber () Financial Incentives at Retirement June 2013 36 / 37

SLIDE 37

Conclusion

How much are individuals adjusting retirement decisions in response to anticipated benefits? Exploit policy discontinuities in severance pay at retirement in Austria Clear evidence of retirement responses: Pattern of spikes and dips in retirements by tenure Analyse relevance of financial incentives in the data Reduced form elasticity concept relating retirement responses to financial incentives Results: Moderate elasticity of about 0.6, roughly three times larger than estimates based on legislative incentives Little evidence of heterogeneity across population of workers later in life-cycle Limited response to financial incentives in retirement decisions

Manoli and Weber () Financial Incentives at Retirement June 2013 37 / 37