« Extraordinary Complicity? | Main | McKinnon: "China is Not a Currency Manipulator" »

Thursday, April 20, 2006

Defining the Social Security Privatization Debate

This article from Economists' Voice distinguishes seven features to keep in mind when the debate over Social Security privatization begins again. It would be useful to have a similar analysis of proposals for Health Savings Accounts in order to avoid the types of ambiguities and muddled debate noted in the article:

The Many Definitions of Social Security Privatization, by Don Fullerton and Michael Geruso, Economist's Voice: In the past decade, debate over social security privatization has exploded in the United States. The Bush Administration, think tanks, and academic economists have all pushed cases for privatizing the United States public pension system. When you look under the hood, though, "privatization" can mean a different thing to almost every advocate. ...

This ambiguity of the term privatization muddles the public and scholarly debate over changes to social security. Some proponents of privatizing Social Security desire a strengthening of our public old age insurance system, while other proponents of privatization desire the end of it. Probably worse, it is common for an advocate of reform to bundle together several aspects of privatization and then justify the package by pointing to benefits that could be achieved with much less radical change.

Table 1
Direct Comparison of Seven Attributes
  Current Social Security   Typical Private Pension
  PAYGO Funding   Full Funding
  Defined Benefit   Personal Defined Contr.  Accts.
  Mandatory Contributions   Voluntary Contributions
  Government Management   Private-Sector Management
  No Investment Choice   Investment Choice
  Annuity Only   Lump-Sum Payout Option
  Redistribution   Actuarial Fairness

Decomposing privatization

Seven vectors capture the salient features of most privatization plans. As shown in Table 1, the current system is pay-as-you-go (PAYGO), meaning that funds from current workers are used to pay the benefits of current retirees. It is a defined benefit plan, which means that each retiree’s benefits are prescribed in a way that does not depend on funds in a personal account. Workers and employers are required to make contributions to the system through Old Age Insurance payroll taxes, and any receipts in excess of expenditures are held in an aggregate fund managed and invested by the federal government. Benefits are paid only as an annuity (on a monthly basis throughout retirement with no lump-sum payout option). The design of tax and benefit schedules and other rules mean that the system redistributes across demographic groups.

In contrast, looking down the right column of Table 1, a purely private pension would be characterized by full funding and by personal accounts of a defined contribution plan. It would have voluntary levels of contribution, private-sector management, individual investment choice, a lump-sum payout option, and actuarial fairness.

A reform to privatize Social Security might change any one of these seven rows, or attributes, without necessarily affecting any of the others. Complete repeal of Social Security would include all seven. Ignoring potential dependencies, 127 possible plans could be formed from this list (by including at least one, and possibly all seven attributes)! ...

Most actual privatization proposals would reform more than one but not all of the seven attributes. For example, in 2000 Laurence Kotlikoff ... proposed a plan endorsed by a long list of academicians that has a fully-funded system under which workers make mandatory contributions into personal accounts managed by the private sector. These accounts would offer only a single investment portfolio, and contributions would be matched on a progressive basis to add redistribution. Upon retirement, account holders would have to annuitize their benefits. That plan includes only three of the seven possible attributes of privatization.

Of course, describing privatization as a list of seven binary choices is a massive simplification. Since each attribute is actually a vector, we mean to allow for degrees of adoption. For example, workers may not be given full freedom in choosing investments, but they may be allowed to choose among a constrained set of funds managed by competing private-sector entities...

The Problem with Ambiguity

The definition of privatization matters. The term’s ambiguity has led to imprecise analysis and commentary on changes to Social Security. The most striking problems occur when the analysis of just one attribute of privatization is used to justify a much broader schedule of changes. Consider a 1998 Washington Post article by José Piñera, Chilé’s former minister of labor and social security, who presided over that country’s transition to a private pension system. ... In this article, Piñera invokes Martin Feldstein’s analysis of the benefits of funding to argue that the United States should follow Chilé’s example. But in doing so, he twists together several attributes of privatization that are completely separable. ... The desire to achieve benefits derived from just one feature of privatization is thus conflated with the need to adopt a whole array of changes.

Definitions also differ among academic economists. In his 1996 analysis, Laurence Kotlikoff considers a privatization plan that sounds like repeal of social security, and he measures efficiency gains from improved labor supply and saving incentives. Yet these gains do not require repeal. His results depend on funding, but they do not require changes in any other row of Table 1, even that assets be held in personal accounts! Martin Feldstein’s (1998) definition of privatization also calls for funding, but he would have mandatory contributions to individually managed accounts. In contrast, a 1998 paper by John Geanakoplos, Olivia Mitchell, and Steve Zeldes characterizes privatization as the creation of a system of individual accounts, but this definition does not require funding, nor that individuals choose investments. Thus, even though each author has a consistent definition of privatization, the same term is used to denote incompatible concepts. ...

In general, anyone who aims to show gains to economic efficiency from privatization should specify which attributes of Table 1 are key to the results. In many economic models, Social Security’s full privatization will increase national savings. But the gains to national savings might have nothing to do with putting investment decisions in the hands of individuals and everything to do with fully-funding the system. ..

We may have good reasons for funding our nation’s Social Security system. And we may have good reasons for keeping contributions to the system mandatory, or keeping investments under government management, or keeping benefits redistributive. Fortunately, the choice to fund Social Security does not constrain our choices along any of these other vectors.

Likewise, if we fear the prospect of retired Americans ending up penniless and desperate in old age, then perhaps giving workers full control over their retirement savings presents a risk too great to bear. But the desire to preserve this form of social insurance no more rules out Social Security’s privatization than full-funding requires Social Security’s repeal. Like Singapore’s Central Provident Fund, a privatized United States pension system could restrict retirement investments to relatively safe options such as government bonds and government-managed savings accounts. Thus, before Wall Street brokerage firms become too enamored of the idea of privatization, let’s ask whether Wall Street should be directly involved in a privatized system at all. ...

    Posted by on Thursday, April 20, 2006 at 12:43 AM in Economics, Policy, Social Security | Permalink  TrackBack (0)  Comments (44)



    TrackBack URL for this entry:

    Listed below are links to weblogs that reference Defining the Social Security Privatization Debate:


    Feed You can follow this conversation by subscribing to the comment feed for this post.