Personal Accounts, Risk, and Replacement Rates

[National Dialogue Moderator <moderator>]

Personal Accounts, Risk, and Replacement Rates--by Carolyn L. Weaver

With a system of personal accounts, workers would be able to tailor their 
investment portfolios to achieve their desired level of "income security."  A 
worker could, if he or she wished, invest entirely in bonds, even price 
indexed bonds.  Less investment risk and less expected return, but more 
predictability.

With social security, workers have no control over the risks they bear.  And 
the risks are real.  Social security has already undergone two major 
"reforms," in 1977 and 1983.  We are bracing for the next.  Just what 
benefits will be--either in terms of level, relationship to prior earnings 
(the replacement rate), or return on taxes--is not known.  Retirees are no 
longer immune to the financial pressures of an aging population and an aging 
social security system.  Proposals to reduce cost-of-living adjustments and 
to increase benefit taxation abound.  Centralized investment would expose 
workers and retirees to new unforeseen risks.

The Burtless data reported by Bob Reischauer is, as noted by one of the 
participants, quite misleading.  It is based on two key assumptions.  First, 
workers are assumed to invest entirely in stocks for their entire working 
lives.  In fact, workers are observed to hold a mix of stocks and bonds and 
to increase their holdings of fixed-income securities as they get older.  
Second, on reaching 62, workers are assumed to go into the marketplace and to 
convert the full value of their investment funds into life 
annuities--regardless of how favorable or unfavorable market conditions are 
that day.  There are many good ways to manage the withdrawal of retirement 
funds and this is not one of them.  A lot of unnecessary risk (thus 
volatility in replacement rates) is imbedded in the Burtless study on both 
the investment and distribution side.  

Having said this, trying to compare a pay-as-you-go social security system 
with a system of personal accounts by reference to replacement rates (or 
benefits relative to pre-retirement earnings) is inherently misleading.  The 
reason is that the "investment" in --or implicit taxes or contributions 
to--the two kinds of systems are not the same.  Social security taxes have 
risen over the years and, to keep replacement rates constant in the face of 
the economic and demographic changes ahead, would have to keep on rising in 
the future.  With a system of personal accounts,  a contribution rate can be 
fixed in the law and left there indefinitely, much as companies establish a 
contribution rate for their 401(k) plans.  Comparisons between systems where 
the "investments" are not the same must be interpreted with great care.  (As 
an aside, replacement rates under social security were far from stable in the 
1970s.)

While Bob Reischauer concedes that the data are extreme, he goes on to 
suggest that we can infer something about the political (un)sustainability of 
a system with volatile replacement rates.  I disagree.  With a system of 
personal accounts, the size of the investment fund an individual accumulates 
would depend on the contributions and investment choices he or she makes and 
on market conditions.  Accumulations would differ from person to person.  In 
a decentralized system of private accounts, though, who would know and what 
would it matter?  TIAA-CREF manages much of my investments for retirement.  I 
have no idea what my colleague's investment accumulation at TIAA-CREF is or 
how it stacks up against mine, or how our annuity payments would differ one 
day, should we decide to purchase one of the various kinds of annuities 
offered.  If I want to increase my expected retirement income, I can do so.  
If he wants to increase the predictability of his retirement income, he can 
do so.  Differences in the "replacement rates" we ultimately achieve, from 
our various sources of income, would be little cause for alarm.  

A widespread decline in expected retirement incomes, owing to generally 
depressed market conditions, is a different matter altogether, one that 
raises problems for investment-based retirement systems and pay-as-you-go 
retirement systems alike.

It seems to me that volatile replacement rates are more properly a concern 
with a public, one-size-fits-all system like social security today.  Two 
workers who have the same lifetime earnings and pay roughly the same taxes 
expect to get the same benefits.  (In reality, their benefits will differ 
depending on whether they are married or have children, when they choose to 
retire, their birth cohort, and other factors.)  When replacement rates 
differ significantly--as they did for the "notch babies" and those born 
before them--we all hear about it!

In the end, a system of personal retirement accounts offers workers the 
prospect of generating larger and more secure retirement incomes.  It also 
offers workers the prospect of accumulating real financial wealth that they 
own, something that social security, as currently designed, can not offer.  
Financial risk is part-and-parcel of an investment-based system and, with 
modern financial markets, there are many good financial institutions and good 
techniques to manage that risk.

Carolyn Weaver