When Measurement Gets Politicized: The Case of Public Pension Liabilities
Filed Under (Retirement Policy, U.S. Fiscal Policy) by Jeffrey Brown on Apr 13, 2011
Many academic economists, including me, often weigh-in on public policy issues. One of the things we quickly learn is that academic discourse and political debate can be quite different. One example of this is that academics are quite good at isolating specific questions (e.g., “holding all else constant”), while political debates often combine issues in an attempt to “spin” the discussion for or against a certain idea.
The public pension debate is a prime example. There are at least four very important – but conceptually distinct – issues that often get discussed. Three of these are areas where there are legitimate grounds for disagreement. The fourth, however, is a pure issue of measurement over which there is virtually no disagreement among academic economists (regardless of ideology). But others have succeeded in politicizing the issue, and the implications of this are important and unfortunate.
What are the four issues?
Question 1: Should public sector workers continue to be offered Defined Benefit plans, or should they be offered Defined Contribution plans instead?
Question 2: If we continue to offer DB plans, should we fully pre-fund them?
Question 3: Assuming we do at least some pre-funding, how should the assets be invested?
Question 4: What is the value – in today’s dollars – of the future pension benefits that we owe?
These are all distinct questions. Two people could completely disagree on whether public workers should be offered DB or DC plans, but they might still agree that if a DB is offered, it ought to be fully funded. Or they could agree that they both like DB plans, but then disagree on the optimal portfolio allocation. Indeed, for each of the first three questions, there are a number of intellectually defensible answers, and smart, well-educated, good-intentioned individuals can disagree simply because they place different weights on different factors. Fair enough.
But question 4 is unlike the other three. Question 4 is not a question about values or weights or the perception of pros and cons. Question 4 is a measurement issue, pure and simple. Financial economic theory – and centuries of experience with financial markets – provide clear principles on the right way to discount future pension liabilities. Namely, you pick a discount rate that reflects the risk of the liabilities themselves. Every academic financial economist I have ever asked (and there are many, including several Nobel Laureates) agrees on this point (and this is true regardless of their personal political ideology). Furthermore, they agree that the right answer to this question is *completely* unrelated to how a plan invests its assets (question 3), or whether the plan pre-funds (question 2), or whether the individual prefers a DB or a DC plan (question 1). They agree that it is a simple measurement issue. Just like 1+1=2, and this is true for both liberals and conservatives.
Unfortunately, a large number of non-academics – ranging from the Government Accounting Standards Board to some plan administrators to some ideologically-motivated “think tanks” – have managed to turn a clear measurement issue into a muddled ideological and political issue. In essence, they have begun to argue that 1+1 is actually equal to 1.5, not 2. And they further imply that those who say 1+1 is equal to 2 are just out to destroy DB plans.
They do this by saying that those who would discount public pension liabilities the correct way (using a risk-adjusted discount rate -which results in an estimate of about $3 trillion of under-funding in public plans – rather than the intellectually vacuous but “official” estimates of about $1 trillion) are just out to make DB plans look “more expensive.” They accuse scholars of trying to inflate the costs of DB pensions for some political reason, such as a desire to privatize the system.
All of this is nonsense. Many of these same economists disagree on the answers to questions 1, 2 and 3, but we all agree that we ought to at least start with an accurate measurement of the size of the pension liability. Whether one believes DB plans are the greatest human invention of all time, or the worst sin ever committed, should have no bearing whatsoever on how we calculate the present value of our future pension liabilities. It is also true that how we invest our assets has no bearing on the size of the liability (after all, a dollar invested in stocks today is still worth the same as a dollar invested in bonds today).
Unfortunately, this politicization of a fundamental economic principle is not merely an intellectual frustration to academic financial economists. Understating the true economic costs of future pension promises has real consequences. It distorts decision-making. It artificially stacks the debate in favor of some reform options and against others. It promotes excessive risk-taking. And, perhaps worst of all, it disguises the true cost of government to current taxpayers.