Disclaimer

The views expressed by me on this blog are mine alone at the time of posting and do not necessarily reflect the views of any organization with which I am associated.

Tuesday, February 01, 2005

Krugman's Unhappy Returns

In his New York Times editorial today, Paul Krugman picks up on Dean Baker's theme of insisting that using historical rates of return on stocks alongside the lower-than-historical projections of economic growth in the Social Security Trustees Report is inconsistent. The argument, in a nutshell, has the following components:



1) In the long run, the corporate profits that generate stock returns can only grow as fast as the economy as a whole. Otherwise, the corporate sector eventually becomes larger than the whole economy, which is inconsistent. The long-term growth rate is 1.9 percent in the Trustees Report.



2) The rate of return on stocks, assumed to be 6.5 to 7.0 percent after inflation when the Office of the Chief Actuary evaluates reforms, must be partitioned into a dividend yield (including repurchases) and capital gains.



3) Since dividend yields are around 3.0 percent (Krugman's figure), that means that capital gains must be at least 3.5 percent.



4) With capital gains of 3.5 percent and profit growth of 1.9 percent, the Price-Earnings ratio will grow without bound. This is the inconsistency that Dean has been highlighting and that Krugman wrote about today.



This is the same argument that Dean posed to me (admittedly, a second time) while guestblogging over at MaxSpeak a while ago. Since that time, I have been working through Robert Gordon's Brookings Paper on productivity growth and some articles by Peter Diamond about what stock returns to expect for the future. I'm not done doing my homework, but here's an outline of a preliminary response:



The critical assumption in the Baker/Krugman example is that the dividend yield doesn't rise above a number like 3 percent, forcing the capital gains to cover the other 3.5 percent and be reinvested in the corporate sector. What if the payout ratio increased dramatically, so that capital gains accounted for only the same 1.9 percent return that matched the growth rate in profits and the economy as a whole? The inconsistency goes away, as the P/E ratio is stable. So one could rephrase the Baker/Krugman critique as, "Because of the low rate of economic growth, those holding to a 6.5 percent return are assuming an unrealistically high dividend yield."



But is a high payout ratio (e.g., 50% larger than what Krugman is positing) so unrealistic? I don't believe that economists as yet have a solid answer to this question, largely because they don't have robust models of what determines the dividend payout ratio. The most frequent answer that I have gotten when shopping this question around to better macroeconomists than I has been that the partition between capital gains and dividends is indeterminate in models used to study long-term growth. The Office of the Chief Actuary certainly doesn't generate its assumptions from a macroeconomic model that requires them to be consistent.



The most realistic impetus to drive the payout ratio higher is that the size of the elderly cohort will increase fairly dramatically relative to the size of the working-age cohort. Over the 75-year period, the projection is for there to be 80 percent more beneficiaries relative to workers. As more and more of the equity is held by the elderly, there will be a greater demand for firms to pay dividends (or repurchase equity) so that the elderly can consume their accumulated wealth.



According to this theory, the reason the economy doesn't grow faster and P/E ratios don't explode despite the solid return to capital is that firms don't reinvest their earnings. They pay out their earnings to satisfy the consumption demands of the large cohort of elderly. That the payout ratio exceeds historical highs is supported by the projection that the relative size of that elderly cohort also exceeds historical highs. Before I hitch my wagon too firmly to this horse, I'd like to know more about the extent to which elderly have a preference for dividends as opposed to capital gains. I would also like to hear more from macroeconomists about how to make internally consistent forecasts of the return to physical capital, the valuation of financial capital, and the role of demographics in both.



A couple of other issues:



First, as an indictment of a system of personal accounts, the Baker/Krugman argument carries more force if the return on equities stays permanently lower over a long period of time rather than being very low in the near term before resuming its historical rate. (Peter Diamond discusses this difference in the paper linked above.) If we set up personal accounts in 2006, and the market tanked in 2007, restoring the historical P/E ratio and thus making historical returns more feasible with a low payout ratio, then we lose money only on the one year of contributions. I should stress that Peter is not a fan of personal accounts--I am citing his paper only to point out where you can read more about the relationship of the Baker/Krugman argument to the timing of the low returns on the stock market.



Second, as I have noted in other posts, I am not wedded to the low economic growth rate assumed in the Trustees Report, but I also think that the mortality projections have life expectancy growing too slowly. If both of those parameters increased, Social Security's financial imbalance stays about the same but the argument I have laid out here becomes easier to make. The faster rate of economic growth makes it possible to have a lower payout ratio, and the greater share of elderly relative to workers enhances the attractiveness of dividends relative to capital gains.



Around the blogosphere:



Max takes credit (appropriately) for introducing this argument on his blog a while ago.



Brad links to the Krugman article, and then takes Luskin to task (appropriately) for the false assertion that higher productivity growth would not improve Social Security's finances. (It would, but it is not clear that sufficiently higher growth to restore solvency completely is realistic. See this earlier post of mine.)



Angry Bear likes the Krugman article, but then wonders about what is happening to national saving. That's close in spirit to the argument I've made in this post, though I am not asserting that AB would agree with anything I have written here.



Atrios links to the article and wonders what rate of growth would plug the hole in the finances. Again, see this earlier post.



Other blogs commenting on this post



26 comments:

PGL said...

Baby Angrybear (PGL) agrees with most of this Andrew and your dividend yield comment is something I 100% agree with. I was working through the financial economics and may have something more on this. But you scooped me. Well done! And thanks for mentioning our blog.

Anonymous said...

There are two issues I'd like to bring up:

1. Isn't the dividend yield going up equivalent to assuming a crash in the price of stocks? I thought average current P/E ratio of the market was something like 25-30 - so that (at most, assuming *no* retained earnings) the current pile of earnings could support dividend yields of 3-4%. Throw in your earnings growth of 1.9% and it sounds like the upper bound for the (long-run) real return of stocks today is 5.9% (and probably less given that the assumption of 0% net retained earnings across the market seems unrealistic - unless we expect some companies to borrow money to pay dividends). Expecting dividend yields to keep up the long-run return of stocks seems to imply that the market is currently overvalued significantly - which, of course, highlights one of the shortcomings people see in a private accounts plan.

2. Assuming historical rates of return are what to expect for stocks and bonds builds in the assumption that the "equity premium" is immutable. And that sounds incorrect given even a simplistic supply and demand analysis. Particularly since the plan on the table is to borrow money to finance private accounts, it sounds to me like the price of bonds should go down (more are being sold) and the price of stocks should go up (more are being bought). This, of course, would close the equity premium and substantially reduce the potential financial upside of private account. Consider the alternative - if the equity premium is really immutable, why bother with taxes at all? The government should just borrow money to invest and the extra return will (in the long run) finance continuing operations.

Cent21 said...

If I were arguing the conservative's line, I'd say that globalization might keep big global corporations growing faster than the US economy.

But as a progressive advocate for my kids and grandkids, I'd suggest to the kids that they should be more concerned about their lifetime wage earnings than their return on investment on savings of a few percent of their earnings. They'd wind up with more if they can improve their wage earnings by 1% per year rather than improving their rate of return on savings by 1% per year. Those two might be correlated, but aren't necessarily correlated.

It's one thing for economists to argue against protectionism; it's quite another to ask a couple generations of workers to bank both their job security and their retirement security on returns on foreign equity.

JG said...

"the Baker/Krugman argument carries more force if the return on equities stays permanently lower over a long period of time rather than being very low in the near term before resuming its historical rate."

When Krugman three columns ago endorsed the Kinsley argument that stock yields must fall and bond yields rise to close the gap between them , and (sort of) quoted Jeremy Siegel as supporting the idea that stock yields will fall, he neglected to mention how Siegel also said in the same statement that he expected bond yields to be 0% over the next several years.

So apparently Krugman takes your point above much more readily in regard to bonds than stocks.

Patrick Sullivan said...

Yes, there seems to be quite a bit of confusion about the difference between capital's share of product, and the growth of capital's share of product. Suppose Congress passed a law confiscating 100% of capital gains; wouldn't that raise dividend payouts?

Other points that Krugman isn't seeing, is that returns on investment (stocks, bonds, money market instruments) are going to be higher than future returns on workers P/R taxes. And, that workers can invest worldwide, not just inside the U.S.

Anonymous said...

A number of commenters have mentioned the possibility of investing abroad to get out of the box imposed by GDP growth and reasonable P/E ratios. That is certainly a theoretical possibiliity, but in a nation with a great and growing current account deficit, relying on net foreign investment seems to be practically unwise. If we were (or plan to) invest abroad to fund our retirements, we would need a great and growing current account surplus instead of our current deficit. Otherwise, when we get to the far future, whatever income we will be getting from foreign assets will be dwarfed by the share of GDP devoted to servicing our external claims (leaving nothing to fund our retirement). The foreign investment option might be interesting if the proposal on the table plausibly increased national savings (since that would be a step towards turning around our current account deficit). Unfortunately, the proposal on the table finances the accounts with newly issued government debt, leaving the overall savings picture unchanged.

Anonymous said...

Jim writes:

"With more than 4 billion mostly young people in economies in Asia, South America and Eastern Europe (and dare we hope Africa?) that expect to be growing *a lot* faster than 2% year -- and we'd better hope they do, for the world's sake -- is it unreasonable to imagine profits earned abroad will grow in proportion to total profits over the next 50 years, as US economic growth slows?...Are profits earned abroad on this scale in the future entirely implausible in light of the likely course of future world economic development?"

Short answer: yes. That's why no economist has been willing to write down numbers showing this. They realize it would make them look silly to their peers.

Interestingly, when Baker has made his own calculations of likely stock gains in the future, he has factored in increased profits earned overseas. Just not crazy levels. See http://www.cepr.net/Social_Security/letter_to_feldstein2.htm

Moreover, even if profits earned overseas DO rise at some implausible rate, then we could add additional revenue to Social Security via a tax on them. There would be the same gain to retirees, without the individual risk.

JG said...

"A number of commenters have mentioned the possibility of investing abroad to get out of the box imposed by GDP growth and reasonable P/E ratios. That is certainly a theoretical possibiliity, but in a nation with a great and growing current account deficit, relying on net foreign investment seems to be practically unwise. If we were (or plan to) invest abroad to fund our retirements, we would need a great and growing current account surplus instead of our current deficit."

Well, I'm just talking about the level of corporate profits and where they can come from -- to get that trophy. I'm not talking of the entire economy.

The government is dissaving big time while individuals aren't saving nearly enough and are sure to incur a haircut of tax increases and benefit cuts in the foreseeable future, which will have bad consequences.

But the level of corporate savings is fine, and as US economic growth slows while growth picks up elsewhere it's logical that more of those savings will be invested abroard to get higher returns (especially if domestic comsumption declines due to that haircut) so the foreign contribution to profits will increase.

That's *hardly* a panacea for funding our retirements. The world would be a much better place if we'd all put aside real savings not only for our retirements but also our future health costs (and if the government had funded the SS and HI trust funds with real assets instead of spending them and leaving IOUs).

But with the government promising to pick everything up for everybody, who was going to do that?

One interesting point about international flows is that our big trading partners -- Japan and Western Europe -- are all in the same boat as us as far as unfinanced retireee commitments go, them worse than us actually.

What sort of international equilibrium is that going to create when everyone tries to make up for their too-little savings at once?

PGL said...

My agreement and simple modeling of your point is up over at Angrybear. Enjoy - PGL

Anonymous said...

[Max Sawicky; I really should register, I come here enough.] I provoked a response to this from Dean:

http://maxspeak.org/mt/archives/001111.html

Anonymous said...

If you need to get 6.5% real returns on invested capital, ask an investor how to do it - don't ask a bunch of economists. You can find many different long term investments that yield 6.5% - just ask a few of the old guys at Sutter Trust or John Hancock how to put the money to work. You're all wound up trying to read the tea leaves in the share markets. Ask Warren Buffet if it is possible to get a better rate of return than the <2% available from SSI. You have to manage the money every day and not try to place the investment for 75 years like somekind of experiment. Of course the government will F-It-U and thats why the idea of private accounts is attractive. Anyways, as the old saying goes, no mans property is safe as long as the legislature is in session.

Anonymous said...

Jim,

Do you have a source for the figure "the S&P 500 over the past 55 years to 1/1/05 has appreciated 0.9% a year on average faster than GDP"? Somebody I was talking with on another board keeps saying that since stocks returned at 7%, and the economy at 3-4%, the stocks must be outpacing the rest of the economy by 3-4%. Is the error simply that he's failed to account for dividends, and only the capital gains return should be counted for considering the actual increase in valuation of the index?

Patrick Sullivan said...

"in a nation with a great and growing current account deficit"

Today. But that can change pretty quickly.

I fail to see why that has anything relevant to say to individuals searching for places to invest over 40 or 50 years.

Certainly we can't tax the labor forces of those foreign countries to provide for Americans' retirement.

Patrick Sullivan said...

Thanks to Tom Maguire, I now see that I've got Brookings on my side too:

http://www.csis.org/gai/Graying/speeches/bosworth.html

JG said...

"Somebody I was talking with on another board keeps saying that since stocks returned at 7%, and the economy at 3-4%, the stocks must be outpacing the rest of the economy by 3-4%.

"Is the error simply that he's failed to account for dividends, and only the capital gains return should be counted for considering the actual increase in valuation of the index?"

It's the difference between simple and compound interest. Over that period S&P 500 returns were 7% annually, comprised of 4% appreciation in value and 3% dividends and other distributions. The appreciaton compounds, the dividends are simple interest -- like interest on a bond. (Although you can reinvest it to compound it, of course)

You can look up the numbers for yourself easily enough, they're hardly secret.

Go back in time futher and dividends made up more of the 7%.

"SIEGEL: From 1871 through 1980 the average dividend yield on stocks was 5%, with firms paying out 75% to 80% of their earnings as dividends." -- http://www.forbes.com/forbes/2004/0419/096_print.html

That's 5% dividends + 2% appreciation = 7%.

Until surprisingly recently it was accepted wisdom that because stocks are riskier than bonds they *had* to pay a dividend higher than the interest rate on bonds, or nobody would buy them.

Times change. So does what everybody knows to be true.

JG said...

"Baker has made his own calculations of likely stock gains in the future, he has factored in increased profits earned overseas. Just not crazy levels. See http://www.cepr.net/Social_Security/letter_to_feldstein2.htm"

From which...

"this assumption implies that 20 percent of the profits of U.S. corporations will be coming from developing nations by 2050"

Which seems a rather modest definition of crazy levels of growth for 50 years, being that foreign profits are 17% of profits now.

"Moreover, even if profits earned overseas DO rise at some implausible rate, then we could add additional revenue to Social Security via a tax on them."

Not even close. Remember that we are going to be increasing all income taxes, including those on profits, by a good 35% by 2030 just to cover the operations of the trust funds, as is.

"There would be the same gain to retirees, without the individual risk."

Ha. With a 20-1 multiple of stock price to after-tax profits you are going to cut $20 of capital income to retirees to give them a $1 tax transfer and call it "same gain, without risk".

PGL said...

Andrew: I'm sort of saddened by the normally reasonable and smart liberal commentators over at Angrybear. They are frustrated at your blog even if I've been saying you have a solid point. I'm as liberal as it gets but when a conservative has a point - especially one as nonpartisan and honest as yourself - I say either accept it or challenge it on principle. But maybe I'm not be as clear as can be. Which is to say: HELP, by being yourself: smart, opinated and yet very honest. But I need to work on my writing skills. Cheers!

Anonymous said...

Maybe I'm missing something, but at first glance I don't see how "shifting" the assumed returns from capital gains to dividends changes the picture AT ALL. Sure, you can increase dividend growth and keep PE ratios from going to infinity...but now you've got dividends growing faster than profits, forever, and dividend-yields going to infinity.

Isn't the bottom line very simple: TOTAL returns to stocks CANNOT grow faster than profits in the long run, regardless of the form the returns take?

Anonymous said...

For some reason, Blogger doesn't remember me. So I'll post this via "anonymous" and sign my name below.

There are significant pressures to increase dividends that go beyond the demographics of baby boomer retirement. Two come to mind immediately: first, a more favorable tax regime for dividend income and second, pressures on corporate governance.

There has been a slow buildup of pressure on corporations by shareholders to make profits more liquid to them. That is classically done most directly through dividends. It's reasonable to assume that that pressure will continue in an environment in which many more people are investing directly. The latest tax changes encourage this as well.

No time to post on the other issues now - great thread!

Robin Burk
http://www.windsofchange.net/

exguru said...

Not enough attention has been focused on exactly where all the ss money will be invested. The federal employees plan offers several choices--enough to make your eyes glaze even when you understand what all of them are about--and they also have a "default option," for federals who fail to make any selection. I'm told the default option puts them in a conservative balanced fund with something like 40% bonds and the rest in big cap stocks. But one factor I can't find out is what percentage of participants came in via the default option. I suspect it is a very high percentage, since federal workers (and other people) tend to trust the government, and, as I said, many have little aptitude and less confidence about doing it themselves. One can see if the same experience, or higher, proves normal for the monster body of the new ss accounts, which it probably will, the man who gets to define the default package will hold the keys to the kingdom... Another thing. There can never be enough supply of large cap shares to feed this insatiable monster, month after month, so of necessity much of the ss money will be going right back into Treasury bonds. (Don't worry, enough will go into stocks that the P/Es will rise remarkably.) This will increase demand for government bonds right when the Treasury is borrowing to finance the transition. Still, one suspects their prices should rise a bit because of the huge demand, and yields should fall. Remember, however, most of these people are a long ways from retirement, and those bonds will be inert in these accounts, and not come on the market again for many years--just as the securities now in the "trust fund." This should also help ease the transition. Rightly or wrongly, most people think of Treasury bonds as "the safest" investment, and they will opt for the safest vehicle with any savings, especially retirement money. It may be that 75% or 80% of this money which Democrats think will be running through the oily fingers of stock brokers, will actually go immediately into Treasury bonds and even federally insured CD's. While that isn't very sexy, the owners will still do three times as well as under the present system... Even if it did not materially increase retirement benefits, of course, the individual account system would be worthwhile, because it would change ss from a hugh Ponzi with Gargantuan unfunded liabilities into a system which was fully funded--and allows spouses or heirs to collect in the event of premature death.

Anonymous said...

Sam begins by saying:
"In his New York Times editorial today, Paul Krugman picks up on Dean Baker's theme of insisting that using historical rates of return on stocks alongside the lower-than-historical projections of economic growth in the Social Security Trustees Report is inconsistent."

I say:
Maybe I'm being thick here but how does anything else Sam says apply to this paragraph. Government projections about Social Security bankruptcy assume one rate of growth while government projections about stock performance assumes a different rate of growth.

It's nice to be able to attempt to reconcile these different assumptions by speculating about external and as yet unseen factors such as improved management and foreign investment but to do so invites equally speculative counter-hypotheses. Since examples of the latter (e.g. the projections in the Trustees Report are implausibly pessimistic) may be more realistic than former (e.g. Walmart stock could outperform the U.S. economy by opening stores in a soon-to-be-middle-class Africa.)

Disappointing for someone either way (Krugman or Bush) but we're talking about guidance for the entire American economy, not some Mars probe. There's no point finessing outcomes when half the President's team inputs inches and the other half inputs centimeters.

If Trustees Report assumptions were reconciled upwards the Social Security crisis would resolve itself, while if returns on stock are reconciled downward towards the Trustees Report them privatization will fail. Therefore if the President is to prevail politically (if not mathematically) he has to equivocate. Sympathizing with his dilemma, however, can't excuse it.

David Innes, www.8020.org

Anonymous said...

I do not share the confidence in the industry of our intrepid correspondent who recommended we stop talking to Paul Krugman and start with Warren Buffet, but he's absolutely right. Only an armchair economist could possibly ask why we couldn't just crank up the dividend payout ratio. An investor, at least an old-style kick-the-tires-and-read-the-10K guy, wouldn't even be able to answer you. He'd be too busy trying and failing to shut his jaw. A hedge-fund trader would just be staring at you, wondering what positions you had on and how he could get on the other side of them.

And please, don't stop investing your own money. I'm beginning to understand why I have been able to beat the market for the last decade or so.

fling93 said...

As more and more of the equity is held by the elderly, there will be a greater demand for firms to pay dividends (or repurchase equity) so that the elderly can consume their accumulated wealth.I'm still clueless on how the stock market fits into the whole economic situation. Can somebody please explain the above to me? How are companies be incentivized by demand for shares of their stock? I can only think of narrow cases and indirect motivations, like not wanting to be easily acquired, the board and management not wanting their personal holdings to decline in value, and being able to raise more money if you issue new shares.

brad said...

I would say that it is clear that sufficiently higher growth to restore long-run solvency is *not* realistic. It would help, significantly, but it would be more likely than not that a significant long-run financing problem would remain--especially since I think Andrew is correct in his assessment that the SSA is currently lowballing likely life expectancy.

David said...

I've been mulling on some similar thoughts. Seems to me that, if the growth rate goes down, the dividend payout ratio will go up. Earnings are retained (as opposed to paid out in dividends) to the extent that management expects funds for capital investment to be required. Less growth, less capex need.

The other factor that must be considered here is the fact that US corporations do substantial business outside the US, some of it in countries which will arguably have long-run growth rates which are considerably higher than the US rate.

photoncourier.blogspot.com

David said...

On the other hand, where does he get a 3% dividend yield? I believe the actual yield on the S&P500 is more like 1.6%. Yes, one also needs to consider the impact of buybacks...but these are partly negated by various dilution factors.

photoncourier.blogspot.com