Wednesday, April 22, 2009

Chuck Blahous: Social Security myths are vanishing along with the surplus

Chuck Blahous, senior fellow at the Hudson Institute and the Bush administration's main man on Social Security reform, corrects some long-standing Social Security myths in the new issue of National Review. Well worth a read.

'When something is unsustainable," so goes the aphorism, "it tends to stop." Usually invoked about tangible phenomena, the old saw is equally true about unsustainable arguments. If a conception is unsustainable in the face of overwhelming evidence, it must die. Sometimes it dies with a whimper, sometimes with a bang.

In recent years a fashionable myth took hold on the left end of the American political spectrum: the myth of overly conservative Social Security projections. Social Security didn't really face a shortfall, it was said in these quarters. The supposed threats to the system amounted to a "manufactured crisis," based on faulty projections of the Social Security trustees, hyped by conservatives (and their enablers in the scorekeeping agencies) who wanted to destroy the treasured program.

The myth is exploding into pieces before our eyes, done in by evidence that can no longer be brushed aside. Yet such was its power that it continued to be invoked even as Social Security's projected near-term surpluses were disappearing. While some were renewing their efforts to deny the coming Social Security shortfall, the Congressional Budget Office was quietly providing updated projections for congressional staff. The new numbers showed that the FY2010 Social Security surplus would be almost wholly eliminated: a mere $3 billion, a pale shadow of the trustees' projection, last year, of $88 billion.

To the extent that things are turning out far worse than the trustees previously envisioned (to say nothing of the persistently more optimistic CBO, whose projections we will discuss later), this was primarily due to factors that no one could precisely have foreseen. Social Security this year faced the programmatic equivalent of a perfect storm: the recession depressed payroll-tax revenues at the same time that Baby Boomer benefit claims were surging, and the program was paying out its largest cost-of-living increase (5.8 percent) since 1982.

Even before the recent downturn, however, the myth was groundless and should never have gained traction. Those who bothered to look could see that there was never a plausible chance that the projected shortfall would vanish. The myth was sustained by various fallacies. Among them were:

Conflation of aggregate and per capita growth.
The "National Jobs for All Coalition," in a typical example of this mistake, recently stated that the Social Security projections "assume that in years to come real GDP will grow much more slowly than it has over the past century or more, when it has averaged around 3.2 percent per year." The implication — indeed, often the overt statement — of this line of argument is that without this inexplicable projection of a growth slowdown, the problem would be much smaller.

But total growth depends on a number of factors, and some of those factors are changing. The Social Security projections were not arbitrarily assuming a decline in productivity growth per worker — they were taking into account a very realistic projection of a decline in the number of workers added to the labor force every year, as the Baby Boomers head into retirement. From 1963 to 1990, annual labor-force growth was never lower than 1.2 percent, and reached as high as 3.3 percent. Now, net labor-force growth is expected to drop to 0.5 percent by the end of the next decade and stay there.

If the workforce grows by less than one-third the rate it used to, we can quite reasonably expect the economy to grow more slowly as well. This demographic reality was glossed over by some, to create the misimpression of arbitrarily conservative future economic assumptions.

Overstatement of the relative impact of economic growth.
In a closely related example, a recent MarketWatch column titled "Why Social Security Isn't Going Broke" stated: "The actuaries' own low cost projection assumes an average annual growth rate of 2.9 percent between now and 2085. . . . Guess what? Under the actuaries' low cost projection, the Social Security system never runs out of money!"

This rather makes it sound as though faster growth by itself will eradicate the problem, doesn't it? But this "low-cost" projection of the trustees isn't only about faster economic growth. Rather, it's a hypothetical illustration of what could happen if virtually every variable improbably breaks in the direction of a smaller Social Security deficit. The "low-cost projection" assumes that fertility rates permanently return to levels not seen since the 1970s. It assumes that longevity grows less over the next 75 years than it has over the past 30. It assumes that inflation after 2010 never rises above 1.8 percent. And so on. The scenario is a compendium of every rosy outcome (from a fiscal perspective) that one can dream up. It is, in short, not plausible.

Misrepresentation of the projection record of the Social Security trustees.
Another longstanding myth is that the trustees' past projections have proven too conservative. In a typical statement from a 2005 budget hearing: "They [the trustees] have been wrong because they have consistently understated economic growth. I believe, in all likelihood, they are wrong again."

This allegation has developed in direct contradiction of the facts. As I pointed out in a 2007 paper, the trustees' projection record since the last major reforms is one of impressive accuracy, although they have been slightly too aggressive — that is, they have, to date, somewhat overstated the program's fiscal health.

The myth of excess past conservatism has been fed by cherry-picked references to the few, unrepresentative trustees' reports that were in fact too conservative: those of the mid-1990s, made just before an unexpected surge in economic growth caught all government forecasters (not only the trustees) by surprise.

Amazingly, if the CBO's projections are off by a mere $3 billion and the program enters cash deficits next year, the deficit date will have arrived sooner than predicted in every single trustees' report since the 1983 reforms. In other words, not a single projection over all that time will have been conservative enough. If the projections' critics haven't been 180 degrees off of empirical reality, they've been 179.9.

The acceleration of Social Security's difficulties is leaving egg on many a face — not only those of the most vocal problem-deniers, but on those of other forecasters as well. Just last year, CBO, on the eve of this sudden downturn, inexplicably revised the long-term Social Security assumptions to be more optimistic. CBO's projection changes were mysterious at the time, and look startlingly ill-timed now. One assumption they made was that 2001/03 tax policies would expire, along with relief from the AMT, and that federal income-tax collections would be allowed to grow to exceed permanently the historic highs. However defensible as a literal application of "current law" in the near term, no responsible agency should have adopted such an implausible basis for lowballing, in its communications to legislators, the estimated size of the long-term shortfall.

Even stranger was CBO's sudden decision to increase its long-term real-wage-growth assumptions to 1.4 percent annually — more than 50 percent higher than the historical average over the last 40 years. This put CBO's latest estimate near the faulty 1983 assumption of 1.5 percent, which proved to be far too aggressive. CBO's change was never adequately explained, and its implementation on the eve of current economic difficulties has turned out to be a masterpiece of mistiming.

Whatever the explanations for these various missteps, one thing is now clear: Not only have the trustees' projections not been too conservative, we now face a bigger problem even than previously projected. The refutation of the mythmakers has arrived in the form of an economic crisis that will exact a heavy price: from all of us, including those who saw the problem coming, those who denied it, and those who were innocently oblivious. People from every camp will need to work together to fix the problem that now faces us. Reality can sometimes be an unforgiving teacher. Let's hope that enough of her pupils learn and respond.

Mr. Blahous, a senior fellow at the Hudson Institute, formerly served as deputy director of the National Economic Council and executive director of the President's Social Security Commission. He is currently writing a book titled Social Security: The Unfinished Work.


Arne said...

The folks doing projections for 1991 totally missed because of that recession; they were too optomistic. By 1995 they seemed to have built in the idea that the downturn was permanent, and now they were too pessimistic. It is hard to fault their failure to see the recession. (Using my hindsight ability to look at years worth of reports) it is more obvious that the downturn was long but temporary. In 1995 employment numbers were still rising to to prior levels.

Recessions lately have recovered much more slowly than they hit. I would expect projections to have a similar non-uniform behaviour.

Bruce Webb said...

It would help is Blahous pointed out that the dip to $3 billion in 2010 is not permanent (as the piece implies) and that the over date for Primary Deficit has been moved up by about two years to 2017 and not in fact to 2009 (February is an structural anomaly on three different counts, something covered in my most recent AB post).

It would also help if Blahous explained why policy makers would respond to a drop in labor growth with an aggressive immigration policy that would keep both legal and illegal immigration permanently below their recent highs.

And I still maintain that ignoring interest effects is special pleading. That Blahous apparently doesn't want to pay back the money borrowed from workers which ended up flowing disproportionately to capital (most workers not having much invested in defense stocks or holding large portfolios or Treasuries) is the problem of the people borrowing the money and not of the people lending it (and no they don't entirely overlap).

People like me called the Trustees pessimistic starting in 1997 because on a year over year basis they were proved too pessimistic. And the response was to ratchet future numbers down.

If you plugged in the growth numbers projected by the tax cutters then Social Security would have ended up fully funded. Greenspan spent the late nineties insisting that productivity would hold up over the long run even as the OACT was producing numbers suggesting the exact opposite.

Trust Fund Depletion was projected for 2029 in 1997. Seven years later it was projected to be in 2042. The reason that someone was able to make the following claim:
"In a typical statement from a 2005 budget hearing: "They [the trustees] have been wrong because they have consistently understated economic growth. I believe, in all likelihood, they are wrong again."

Is that based on the best information available AT THE TIME it was precisely true. And given that the Bush Administration was insisting that the 2003 tax cuts were just about to demonstrate their magic growth effects it was not unreasonable to agree.

Blahous is just using 20/20 hindsight here. Was there anyone in a position of authority actually predicting the productivity crash of Q4 2005?

I don't think it is me that is cherry picking here.

Anonymous said...

I am glad that Mr Blahous' sponsors, colleagues and fellow travellers destroyed the US economy just so they cold explode this myth. Thanks guys. Heckuva job.

Bruce, near term projections will always seem conservative in a time of unanticpated (or unsustainable) good times. Projections that are perfectly accurate over then long term average will almost certainly go through period in which they are both over and underestimating the short term results.