« Reviewing Richard Parker's Biography of Galbraith | Main | Joenertia »

March 12, 2005

Mankiw 0, Liberals 3

Greg Mankiw writes:

The New Republic Online: [T]here are three reasons [liberals oppose Bush's Social Security Plan]. The first is that... some Democrats will oppose anything he advances.... The second reason the left hates personal accounts is that... much of the left's rhetoric is a less elegant paraphrase of [Karl Marx's] worldview.... The third reason for the left's opposition to personal accounts is... Democrats are more averse to an economic system in which people play a larger role in taking care of themselves.... [This is] a valid concern... but this is not so much an argument against personal accounts as a reason why we need to get the details right. Any reform should include some restrictions to protect people from themselves. There should be limits on how much risk people can take in their portfolios, especially as they approach retirement. There should be requirements that people annuitize enough of their accumulation upon retirement to ensure they are kept out of poverty for the rest of their lives. 

Let me assure him that I know liberals, liberals are friends of mine, I am a liberal, and that Greg is at least two thirds wrong. Mankiw's claim that our view is a "less elegant" version of Marx's is 120% false, and reminds me of nothing so much as National Review's false claims in an earlier generation that there was no essential difference between John Maynard Keynes and Karl Marx. And seeking to make the country worse off for politial advantage--that's not a Democratic game, that's a game that was, as Mankiw notes, played in 1993 and 1994 by Republicans (like William Kristol, who wrote that the greatest danger was that the Clinton health care reform would pass and be a success). Two shots on goal, both of which miss. Mankiw 0, liberals 0.

As to his third point, Mankiw is correct that Bush's plan is a bad idea unless it "gets the details right." But are there any cases of complicated initiatives sponsored by the Bush administration that "get the details right"? Mankiw doesn't mention any. The farm bill? The corporate tax atrocity? The Medicare drug bill? Iraqi reconstruction? Bush's batting average on the important details is remarkably low. And that is not reassuring: here Mankiw kicks the ball into his own goal. 1-0, liberals ahead.

He scores similar goals against his own position with his claims that liberals mask their real concerns under two fake reasons: "two canards--one involving the deficit, the other involving risk."

The first--that Bush's plan "will require irresponsibly large increases in the budget deficit" is, Mankiw says, "mostly fatuous.... [T]he long-run impact of personal accounts on the government's finances is approximately zero.... [T]he government puts... $1,000 in his or her account.... The initial payment into the account requires $1,000 in extra government borrowing, but that debt is offset by a reduction in the government's liability to pay future Social Security benefits." So why, then, does Mankiw say that this worry is only "mostly" fatuous? Because there is a chance that it isn't fatuous at all. There is a chance that the plan will further reduce America's already too-low national savings rate if the reduction in future liabilities is seen as less salient than the increase in the current deficit. As Alan Greenspan--certainly no liberal--told the Congress:

The issue with respect to the financing is a difficult one to answer, because there are things we don't know.... First, we don't know the extent to which the financial markets at this stage, specifically those trading in long-term bonds, are discounting the $10 trillion contingent liability that we have.... If indeed the financial markets do not distinguish... then one would say, 'Well, if you wanted to go to a private system, you could go fully to a private system without any response in interest rates because, obviously, you're not changing the liabilities involved, you're just merely switching assets to the private sector'. But we don't know that. And if we were to go forward in a large way and we were wrong, it would be creating more difficulties than I would imagine.

To be sure, I think that Greg is probably right here: I think that the chances that the adverse budget and national savings impact of private accounts will create "more difficulties than we would imagine" are low: certainly less than 50%. 20%? That's my guess. But 20% is enough of a chance that there should be a Plan B in case that 20% probability comes to pass. And there isn't one.

A worry that Greg Mankiw calls "mostly fatuous" is one about which Alan Greenspan says "we don't know... [if] we were wrong, it would be creating more difficulties than I would imagine." Another own goal. 2-0, liberals ahead.

The second "canard," Mankiw says, is:

the claim that [Bush's plan] would leave retirees of the future facing too much risk. They argue that personal accounts would leave the safety net for the elderly with too many holes. The most obvious response is that the proposed personal accounts are voluntary. If you want to stay in a traditional defined-benefit plan, just don't opt in. And, even if you opt in but then want a low-risk retirement income, you can invest in inflation-indexed Treasury bonds. 

Here Mankiw commits a technical foul. Yes, those currently in the labor force can stay in the traditional defined-benefit Social Security system. But under the Bush plan, the traditional system gets smaller and smaller over time. Once the bend points are indexed to prices rather than wages, the share of your wages that are replaced by standard Social Security benefits falls and falls and falls. For my grandchildren born in the 2020s, the claim that they can not opt in and "stay in a traditional defined-benefit plan" that is anything like Social Security today is grossly misleading. The very long-run impact of the Bush plan is to drive the share of income replaced by traditional Social Security benefits to zero.

Moreover, here Mankiw commits yet another own goal. You see, Bush's private accounts are a good deal for the upper middle class and the rich: they effectively borrow from their defined-benefit Social Security account at 1.5% or 2% above inflation to invest their money, and they have enough other assets that they can effectively manage the risk. For poorer Americans for whom private accounts require effectively borrowing from their defined-benefit Social Security accounts at 3% plus inflation? That's a high interest rate. It's worth doing if expected asset returns are the 3% real for bonds and the 6.5% real for stocks that the SSA is projecting. But there are lots of reasons to fear that those projections are overoptimistic. And if they are, then private accounts are a lousy deal for those who are not relatively rich.

So an informed reading of Mankiw's New Republic piece comes up with three reasons to oppose Bush's private accounts plan:

  1. Bush has a track record, and a lousy track record: he gets the important details of policy wrong. He gets them wrong consistently.
  2. There is a serious worry that the Bush plan will reduce national savings yet again. There is no Plan B to guard against this.
  3. Realistic looks at risk and return make it probable that private accounts are a lousy deal for the non-rich.

And does Mankiw score any goals. Does he make any accurate and effective arguments for Bush's plan? He doesn't score any more own-goals--he doesn't claim that private accounts will solve Social Security's long-run funding plans. But he doesn't advance any effective positive arguments for his position either.

He does say that Harvard University offers its professors not a defined-benefit but a defined-contribution retirement program. But are there important reasons to think that the type of program appropriate for relatively rich Harvard professors with ample other assets is the type of program appropriate for a nationwide system covering the non-rich that is supposed to provide social insurance--a retirement income floor that you can count on no matter what? I would like to see such an argument made. But Mankiw doesn't even try the shot.

Mankiw 0, Liberals 3.

Posted by DeLong at March 12, 2005 09:59 AM