Pension Reforms and the Allocation of Retirement

Saving

Renata BottazziUniversity of Bologna, IFS and CHILD

Tullio JappelliUniversity of Naples “Federico II”, CSEF and CEPR

Mario PadulaUniversity “Ca’ Foscari” of Venice and CSEF

Prepared for the Annual Conference on Financial Security in Retirement18-19 September 2008

Motivation

Assess people awareness of retirement outcome innovations.

What do people know about their pensions? Provide the anatomy of the offset between

social security and private wealth. If social security wealth falls, do people increase more

financial or real wealth?

Study the demand for targeted retirement products.

Why are Italian pension funds still small?

The framework

We exploit a decade of pension reforms,

use data on subjective probabilities on

retirement outcomes,

and look at several components of private wealth, including financial (risky and safe) and real (business and housing) wealth.

Main results

Large revision of pension expectations, but many individuals have not completely updated their expectations yet.

Financial and real wealth have increased following the reforms, but the increase is more pronounced for real assets and housing in particular.

No effect on the propensity to hold targeted saving plans.

Offset and portfolio choice

Standard life-cycle framework: if social security wealth falls, private wealth should increase accordingly.

In a complete market world, the reduction of social security wealth should not affect portfolio rules.

With uninsurable income risk, borrowing (and short sale) constraints, portfolio rules become a function of age and wealth.

Background literature

Estimate social security by using current and projected legislation on pension eligibility:

Gale (1998), Gruber and Wise (1999), Attanasio and Brugiavini (2003), Attanasio and Rohwedder (2003)

using subjective expectations of retirement age and benefits: Bernheim (1990), Gustman and Steinmeier (2001), Bottazzi, Jappelli and Padula (2006)

The effect of the 1992-95 reform

Miniaci and Weber (1999): 1993 consumption drop partly due to the 1992 pension reform

Attanasio and Brugiavini (2002): offset between private saving and pension wealth (coefficient of -0.3)

The Italian pension reforms Three main reforms (1992, 1995, 1997)Features: retirement age and minimum years of contributions for

pension eligibility abolishment of seniority pensions (if started working after

1995) indexation of pension benefits to prices instead of wages less generous pension award formulas

Pension award formula:

Earnings model: 0.02 n w

Three groups:

Old ( 18n in 1995): Earnings model

Young ( 0n in 1995): Contribution model

Middle-aged (0 18n in 1995): Pro-rata model

1992 1995 1997

The Italian pension reforms

The eligibility rules and the pension award formula changeaccording to the years of contribution at the end of 1992

Three groups of workers:

Old, more than 18 years of contribution as of 31/12/1995

Middle–aged, less than 18 years of contribution as of 31/12/1995

Young, enter labor market in 1996

The Italian pension reforms: retirement age

Retirement age

Old age pensions Seniority pensions Minimum retirement age Minimum years of contributions

Private sector Public sector

Self-employed

Minimum years of

contributions Private sector

Public sector

Self-employed

Pre-1992 regime

All workers

60(55) 65(60) 65(60) 15 35 20 35

Old

Progressively rising to 65(60)

65(60) 65(60) Progressively rising to 20

40 before age 57 35 after age 57

40 before age 58; 35 after age 58

Middle-

aged

Progressively rising to 65(60)

65(60) 65(60) Progressively rising to 20

40 before age 57 35 after age 57

40 before age 58; 35 after age 58

Post-1997 regime

Young Subject to incentives: 57-65

5 Abolished

The Italian pension reforms: the pension award formula

Pension award formula

Private sector Public sector Self-employed

Pre-1992 regime

All workers Earnings model

2% years of contributions average of the last 5 years of earnings

2.33% years of contribution last year of earnings

2% years of contributions times average of the last 10 years of earnings

Old Earnings model

Gradually to 2% years of contribution average of last 10 years of earnings

Gradually to 2% years of contribution average of last 10 years of earnings

Gradually to 2% years of contribution average of last 15 years of earnings

Middle-aged Pro rata model

Earnings model before 1995, contribution model after 1995.

Post-1997 regime

Young Contribution model

Contributions (33% of gross wage for employees and 20% for self-employed) are capitalized on the basis of 5-years moving average of GDP growth. The capitalized sum is then multiplied by a coefficient that varies by retirement age, taking into account life expectancy.

Data

Survey of Household Income and Wealth (SHIW) – representative of Italian population

Subjective expectations on retirement age and replacement rate:

Retirement age: - all survey years - “When do you expect to retire? ” Replacement rate: - years 1989, 1991, 2004, 2006

“Consider the moment when you will retire. Setting your final monthly income before retirement equal to 100, what do you expect your first monthly pension to be? ”

1989 1991 2004 2006

Pre-reform transitional period

Post-reform

Expected pension wealth at retirement

Use subjective expectations on retirement age and replacement rate to construct the ratio of expected pension wealth at retirement to earnings (evaluated at each survey yeart):

( ) 11( | ) ( | )1 1

ttNN t

TNu N

t t t tNt

ggSSWSSWY P N t P Ny r r

expectedretirement

age

expectedreplacement

rate

P(•)=survival probability (by age and gender, before and after the reform)

g=growth rate r=interest rate

Social security wealth Expected Statutory

Pre-reform

Post- reform

Difference Pre-reform

Post-reform

Difference

Private Old 10.67 8.891 -1.779 9.356 9.052 -0.304 employees Middle-aged 8.3 6.925 -1.375 7.368 7.063 -0.305 Young 5.463 5.807 Public Old 10.75 10.09 -0.66 10.87 10.32 -0.55 employees Middle-aged 9.16 7.972 -1.188 9.032 7.578 -1.454 Young 6.17 6.097 Self Old 7.582 6.418 -1.164 7.687 7.907 0.22 employed Middle-aged 7.005 5.623 -1.382 6.568 4.97 -1.598 Young 5.039 3.624

Expectation error distribution of social security wealth before and after the reform

0.1

.2.3

-5 0 5 10

pre-reform post-reform

Trends in financial and real wealth

Financial

wealth Real

wealth

Pre-reform

Post- reform

Difference Pre-reform

Post-reform

Difference

Private Old 0.4978 0.6151 0.1173 2.697 4.705 2.008 employees Middle-aged 0.4247 0.4581 0.0334 1.796 4.325 2.529 Young 0.3922 2.951 Public Old 0.492 0.545 0.492 3.15 5.497 2.347 employees Middle-aged 0.4114 0.5405 0.4114 2.302 5.356 3.054 Young 0.3989 4.02 Self Old 0.6098 0.7034 0.6098 5.032 9.362 4.33 employed Middle-aged 0.4901 0.7414 0.4901 3.641 9.22 5.579 Young 0.5521 6.507

Offset between pension wealth and private wealth: the estimating

equation

tit it it itWY SSWY X

Financial (Real) wealth -to-income ratio

Social security wealth-to-income ratio

SSWY: - weighted sum of both partners’ SSWY - Gale adjustment factor

- Instrument with the statutory SSWY- Sample split: Informed (Expectation error less than the median) vs. Uninformed

Year, cohort, and employment dummies and their interaction, dummies, age and education of the hh, area of residence

Offset between pension wealth and private wealth: the results

Financial wealth

All Informed Uninformed -0.082 -0.128 -0.055

(0.011)*** (0.019)*** (0.013)***

Real wealth

All Informed Uninformed -0.478 -0.761 -0.318

(0.070)*** (0.080)*** (0.131)*

The anatomy of the offset

Risky financial wealth

Risky financial wealth

Riskless financial wealth Real estate

Business wealth

(inculding mutual funds)

(excluding mutual funds)

-0.006 -0.005 -0.076 -0.597 0.118 (0.003)* (0.002)* (0.011)*** (0.044)*** (0.052)*

The anatomy of the offset:Informed vs. Uninformed

Informed Uninformed -0.007 -0.009 Risky financial wealth

(inculding mutual funds) (0.005) (0.004)*

-0.007 -0.005 Risky financial wealth (excluding mutual funds) (0.003)* (0.003)

-0.120 -0.046 Riskless financial wealth

(0.018)*** (0.012)***

-0.676 -0.637 Real estate

(0.064)*** (0.071)***

-0.086 0.320 Business wealth

(0.045) (0.105)**

Summary

A reduction in social security wealth equivalent to 1 year’s income brings about an increase in financial wealth of just below 1 month’s income,

and an increase in real assets of about 6 times monthly income.

The effect is larger for the “Informed” and for housing wealth.

Other possible channels through which retirement saving increases

Targeted retirement saving products

Pension plans

Life-insurance policies

All Informed Uninformed All Informed Uninformed -0.007 -0.012 0.001 -0.017 -0.020 -0.022 (0.018) (0.026) (0.027) (0.015) (0.021) (0.025)

Conclusions

Large revision of expected social security wealth.

Larger effects of reform for real, smaller for financial wealth.

Information on pension outcomes is still important,

But the effect of pension reforms on the demand of targeted retirement saving products is small

Implications

Improving the dissemination of information about pension rights,

but increasing awareness of pension reforms might not be sufficient to prompt households to increase private wealth.

Pension reforms don’t seem to have diminished the propensity to invest in real estate.

Annuity markets are still at an infant stage.