Common Sense on Social Security
A Centrist Strategy for Social Security Reform
Social Security Reform: Breaking the Stalemate
Section 6. Removing the Stock Market's Rose-Colored Glasses
Much of the discussion on Social Security reform has been predicated on the stock market's assumed ability to deliver superior returns. If only Social Security could tap into the stock market, it is argued, the task of achieving solvency would be ever so much easier. It is further asserted that anyone investing in an index fund based on the Standard & Poors' 500 would, over the past seven decades, have earned average real returns of 7%. It is therefore concluded that Social Security should have no problem earning 7% returns on the stock market in years to come, either in PRA's, in the Trust Fund, or both.
Real returns are a composite of two primary factors, capital gains and dividend yields, and two secondary factors, portfolio management fees and capital gains taxes. This analysis will focus on capital gains and dividend yields. Capital gains on the S&P; 500 index have averaged about 2.3% a year over the past several decades, if calculated on an inflation-adjusted basis. Dividend yields in the S&P; 500, meanwhile, averaged 4.6%. If one assumes that all dividends are reinvested, the overall compound growth rate on the S&P; 500 turns out to have been 7% a year (1.023 x 1.046 = 1.07), with dividend yields generating two-thirds of the growth.
Now the question is whether the S&P; 500's historic track record can be used as a reliable guide to the future. The evidence says no.
Capital Gains. A growing GDP provides the framework within which corporations and their related stock values also grow. Whatever the GDP growth rate turns out to be over time, the total value of the stock market is likely to grow at roughly the same rate.
Over the past several decades, GDP has grown at an average rate of 3.3% a year, adjusted for inflation. In the future, however, it wouldn't be surprising to see the GDP growth rate soften somewhat, thanks to an emerging slowdown in the population growth rate. Social Security demographers expect population growth rates of only 0.4% a year in the decades to come, nearly a full percentage point less than the historic average. And they expect the workforce growth rate to slow even further, to 0.1% a year, in the latter decades of the 21st century. As these slowdowns occur, they are almost sure to be reflected in the nation's GDP growth rate. By 2050, a GDP growth rate of 2% a year may well be regarded as fairly vigorous. (On the other hand, Social Security's GDP growth forecast of only 1% a year seems much too low.)
One must also take note of the portfolio lag factor, the differential between long-run GDP growth rates and long-run portfolio growth rates. Averaging 2.3% a year, capital gains in the S&P; 500 lagged GDP's 3.3% growth average by a considerable margin. While the overall value of the stock market is likely to grow at roughly the same rate as GDP, empirical evidence indicates that a stock portfolio based on a market-wide index won't grow quite as rapidly, a point that Wharton's Jeremy Siegel has noted. When new companies are listed on the stock market, the total value of the stock market reflects the change. But it's not a change that registers in the value of an index fund portfolio. Whenever an existing company raises funds by selling new stock, the total value of the stock market rises. Meanwhile, the value of the index fund portfolio stays put. As a result, the growth rate of the total stock market regularly outpaces the capital gains growth rate of an index fund portfolio.
The empirical factors all point in the same direction. Capital gains growth rates aren't likely to be as strong in the future as they were in the past. Over the next several decades, with the population growth rate slowing substantially, the real capital gains growth rate of a typical index fund portfolio could be as low as 1.5% a year.
Dividend Yields. Dividend yield refers to the value of the dividend paid, divided by the value of the stock. If the dividend is $1 and the stock price is $25, the dividend yield is 4%. Should the payout stay at $1 while the price rises to $50, the dividend yield falls to 2%. To estimate future dividend yields, one must be able to estimate future movements in both the numerator and the denominator. What's likely to happen to corporate dividend payouts (the numerator)? What's likely to happen to stock prices (the denominator)?
Corporate dividend payouts - the numerator - are fairly stable, if measured in cents per share terms. Stockholders don't approve of dividend payments that bounce around from one quarter to the next, and corporations are loath to anger their shareholders. Moreover, if corporate dividend payouts are measured in GDP terms, the same stability can be observed. Over the past two decades, dividends distributed by publicly listed corporations have averaged 2 to 2.5% of GDP. Dividend payouts appear to be unaffected by changing fashions in the P/E ratio. As a result, it is not so difficult to forecast future trends in the dividend yield numerator.
Forecasting stock prices is a different matter, and forecasting the stock market's total capitalization - the denominator - is only slightly less difficult. From 1925 to 1995, with a great deal of cyclicality, total stock market capitalization averaged a moderate 65% of GDP. Given the capitalization level, it's not surprising that dividend yields on the S&P; 500 averaged about 4.6% a year over the same time period.
Now, however, the nature of the game has changed considerably. Stock market capitalization has climbed to 180% of GDP, nearly three times the market's historic average. Pause for a moment and do the dividend yield math. The Numerator: Corporate dividend payouts equaling 2% to 2.5% of GDP. The Denominator: Market capitalization equaling 180% of GDP. Divide corporate payouts by market capitalization, divide 2%-plus by 180%, and you'll have calculated the market's current dividend yield. About 1.2% or 1.3%. Not very high. Dividend yields are sharply below their historic average.
Now take a look at the reasoning process needed to estimate future dividend yields. The numerator - dividend payouts relative to GDP - isn't likely to change very much. But what about the denominator? What about stock prices and total market capitalization? Is tomorrow's stock market capable of doubling or tripling in value, as stock market analysts James Glassman and Kevin Hassett believe that it is? (Atlantic Monthly, September 1999) Conversely, is the market riding a bubble, on the verge of a devastating plunge? Or is it simply nearing a natural crest, ready to stabilize at a new plateau? No one can forecast tomorrow's dividend yields without taking a stab at future capitalization levels.
Those who reason from a conventional perspective expect the worst. It is a long-standing tenet of corporate finance that the typical stock market investor demands both a risk-free return of 3% and a risk premium of 4%, for a total return of 7%. Once the Dow levels off and the current joyride ends, the investor community will awaken to an unpleasant morning-after moment of truth. Long-run real returns of 7% are not to be had in a stock market with highly inflated prices and severely depressed dividend yields. Once the market plateaus, capital gains growth rates will drop below 2.5% a year. Dividend yields won't quite reach 1.5%. Even if stock buybacks provide an extra boost, investors could well find themselves facing real returns of 4.5% or less. In frustration, many will bail out. P/E multiples and stock prices will go into a long downward slide until investors once again see opportunities to earn real returns of 7%.
Is this a realistic scenario? Probably not. For a 7% return rate to be restored in a stock market that's crested, dividend yields must rise to 4.5%. For dividend yields to triple, total market capitalization must sink to a third of its current level. The Dow would have to fall to 4000. If the Dow refuses to plummet to that level, future dividend yields and real return rates are certain to remain well below their historic levels.
Quite a different paradigm has been put forward by Glassman and Hassett. Those who hold stocks for the long run, say these authors, no longer require a risk premium. The riskiness of a buy-and-hold strategy is now understood to be so trivial that conventional risk-premium arithmetic has become pass�. Sophisticated investors may well be satisfied by stock market returns as low as 3%.
If the Glassman-Hassett hypothesis is correct, one can use their algorithm to estimate the stock market's future. If capital gains (and stock dividend payouts) can be expected to grow at 2.5% a year, the sophisticated investor may be willing to accept dividend yields as low as 0.5%. (2.5% capital gains + 0.5% dividend yields = 3% real returns.) Since the Dow would have to more than double before dividend yields decline to half a percent, the authors wouldn't be surprised to see current stock prices double or even triple before the Dow finally crests and levels off.
Were I called upon to prepare Social Security's dividend yield forecast, I'd come down somewhere in the middle. The conventional risk arithmetic of
At the same time, the notion that the risk premium has actually fallen to zero seems somewhat improbable as well. Once the Dow crests and stock market growth falls in step with GDP growth, the market's steady state dynamics will re-materialize for the first time since the market's 1974 - 1984 trough phase. In a flattened market, paced only by the GDP growth rate, investors may not be satisfied with steady state real returns of 4.5% or less. If they're not, they'll allow P/E ratios to drift slowly downward until they finally start receiving a steady state return rate that gives them the risk premium they're looking for. I doubt they'll be satisfied until steady state returns rise to at least 5%, implying a new risk premium of 2%.
My bottom line? Capital gains of roughly 2% over the next three or four decades, shrinking to 1.5% later in the century as population growth rates slow even further. Dividend yields a little above 2%, based on long-run market capitalization in the 120% of GDP range (down from its current 170% to 180% range). Hidden dividends from share repurchasing worth half a percent, perhaps a full percent. Total real returns of 5% a year for the next four decades, shrinking to 4.5% as zero population growth rates slowly become the norm. If this forecast is off, as it almost certainly will be, I may well have erred on the side of optimism.
Now, having sifted through the evidence bearing on growth rates and dividend yields, what is absolutely extraordinary is the widespread spirit of denial. On the right, real stock market returns of 7% are vital to the success of the Archer-Shaw proposal. On the left, the Nadler proposal is equally dependent upon 7% real returns. From the Cato Institute to the Brookings Institute, Social Security specialists in every camp (with the initial exception of Dean Baker, and the recent addition of Peter Diamond) have shown themselves willing to bet the farm on the stock market's ability to deliver 7% returns. No one dissects the historical record; no one evaluates its significance for Social Security's future. The rosy notion of 7% returns has become little more than an empty mantra.
A Non-Profit Successor to the Collaborative Democracy Project
Page Version 1.02
Revision Date April 13, 2006