You've left Luskin an opening to hitch a rescue/theory around the rate of depreciation.

A grumbly thought. Could low long term interest rates be in part an anticipation of a secular decline in the economic growth rate?

Something Robert Solow wrote in 1956 keeps ringing in my head. Actually, it is odd that we worried a lot about capital shallowing over the past decade given the drop in the national savings rate (which this White House seems to desire to lower even more) and now we are talking about a fall in the growth rate of effective labor leading to the prospect of capital deepening. But you are right - how can the CEA right something so utterly inconsistent with Solow's growth model?

r = a(y + d)/s

A powerful, indeed, robust, explanation of the recent 18-year bull market in the U.S. stock market.

Thank the Lord, I'd never heard of it in 1982, wot?

Appreciate the analysis.

It's a bit discomforting to know that these guys shape government thinking and policy.

What's the real problem, anyway? Why do they make such sweeping statements if they haven't applied correct economic formulas and models to support their conclusions?

I find this discussion a bit strange. Isn't the fact that U.S. real equity returns have averaged 6 or 7 percent more than those on bonds known as the equity premium *puzzle*? Stochastic versions of the standard growth model that Professor DeLong and the CEA staffers are throwing around could only rationalize such high returns with very high levels of risk aversion, levels of risk aversion that are inconsistent with other macro facts.

If we are trying to think about whether high equity returns are likely to persist, shouldn't we have a model that explains those high returns? If we don't have such a model, shouldn't we be a bit more cautious in tying our fortunes to the continuation of high returns?

Perhaps some would say that while there is not a good theory of why equity returns have been so high, we can take an empirical approach and just note that since they have been high for 60 years or longer, we can probably count on them being high for a good while longer. The trouble with that argument is that if we view the 6% gap between equity and bond returns as a point estimate of an underlying mean difference, the corresponding confidence interval would be something like 6 percent plus or minus 5 percent. Even if we think the premium is unchanging over time, there is even after 60 or 100 years a lot of uncertainty about its value.

It is not obvious to me that there is an equity premium unconnected to the government inflating away the return on bonds prior to 1980, and legislating away the corporate income tax subsequent to 1980.
However, I am not a political economist.
After correcting for changes in government currency and tax policy since, say, 1905, is there truly an equity premium?

The equity premium has not been what it was for quite some time :) The last 10 years the Vanguard S&P Index was up 11.4%, while the long term bond index was up 9.7%. Starting date changes the margin, of course. Since 1973 the S&P Index was up about 11.4%, and the long term bond index was up about 9.2%.

There is no reason to expect the equity premium to be more than 2 percentage point in future, but that is a significant margin. Also, tax law favors dividend income and capital gains over interest.

Please do argue, but I can find no reason to fear any Vanguard bond portfolio if an investor is willing to be patient during a period of rising interest rates. Putting together a portfolio of bonds other than Treasuries that will match Vanguard's, would be more than just difficult.

"Perhaps some would say that while there is not a good theory of why equity returns have been so high, we can take an empirical approach and just note that since they have been high for 60 years or longer, we can probably count on them being high for a good while longer."

Substitute the words "productivity growth" for "equity returns" and you have just made the case that we will beat the numbers required under Low Cost to totally fill the "gap". Because they rely on future growth no higher than those of the "60 years or longer" time frame you are appealing to here for prediciting returns on equity.

The desperate attempts to have it both ways, to maintain the truly dismal economic numbers needed to maintain "crisis" under Intermediate Cost and still promote an equities based solution are getting more strained by the day. More and more I am feeling like a Borg Commander:

"Embrace the Low Cost Alternative: You will be Assimilated"

2004 Report: Economic Assumptions http://www.ssa.gov/OACT/TR/TR04/V_economic.html#wp159107
2004 Report: Trust Fund Ratios http://www.ssa.gov/OACT/TR/TR04/II_project.html#wp106217

Anne -- be careful in your bond analysis.

We have been in a 20 year bull market as
yields fell from historic highs in early 1980s back near historic norms. It is the flip side of the rise in yields from 2% in 1950 to almost 20% in 1980. We still can get back to 2%-- but it will not be a straight line.

Agreed. The bull market in bonds began in December 1981, and who would have imagined it going till now? Please do continue in such thinking.

pi-

I'll keep on posting this data in a quixotic attempt to destroy the pernicious meme fostered by the 1926-present Ibbotson's/CRSP data everyone uses as a reference point. Post-1945 data produces an even more misguided sense of confidence that stocks are vastly better capital appreciation vehicles than bonds.

Before Inflation

Stocks Bonds

1801-1900 6.51% 4.99%

1901-2000 9.89% 4.85%

After Inflation

1801-1900 6.76% 5.23%

1901-2000 6.45% 1.57%

There is an equity premium, but much of the excess return stocks delivered over bonds in the 20th century was due to persistent inflation. Bonds, returning a fixed coupon, are much more vulnerable to excessive/surprising inflation than stocks, whose return is linked to corporate profits and expectations thereof.

http://www.efficientfrontier.com/ef/402/2cent.htm

Some further part of the excess stock over bond return post-1945 is probably due to investors ratcheting down the equity risk premium. In 1945, most people thought stocks were for rich speculators. In 2000, most people thought stocks were the best way to generate long-term capital appreciation.

http://www.jeremysiegel.com/view_article.asp?p=127

Particularly given current valuations, it is extremely unlikely that domestic stocks will return 6-7% more than bonds after inflation during the time horizon of those now alive in the United States. That 6-7% figure is a stockbroker's estimate--or of a most euphemistically-titled "financial advisor."

Also, I note that Brad's theoretical model linking stock returns and GDP is empirically supported by virtually all the data we have on economies and equities.

Mycroft
Thank you. I was wondering about that, myself. I did know that inflation transferred wealth from bondholders to stockholders, but I didn't have any real numbers.
I am expecting inflation in the near future. I expect the equity premium to be much higher in the near future. I could be wrong.

There well may be a rise in long term interest rates as traders and investors react to Alan greenspan's remarks. The 10 year Treasury note has risen from about 3.95% to 4.25%.

we should also note, anne, that there are some early signs that foreign (for which we can read chinese, in particular) central bankers are losing a little of their taste for the 10-year US paper. If they lose a little more of their taste, we will be looking at 5% long rates very soon, and my long-awaited stagflation-lite will really rear its ugly head....

"You've left Luskin an opening to hitch a rescue/theory around the rate of depreciation."

That's right! Because thanks to the Bush supply side tax cuts, the rate of depreciation is going to turn into the rate of APPRECIATION! The value of every capital good in the country will increase by . . . well, by whatever the president says it should.

Only you miserable ant men, who despise the pure genius of the creative giant, could fail to understand this. You mock our beloved president, just as you mocked Howard Roark. But you'll all see -- once we have The Precious, all will bow before us! We'll be master then!

Besides, Larry Kudlow agrees with me.

Mycroft's efficient destruction of the "it always has been" even leaves out the spanner in the work: it wasn't until 1967 that the Federal Government was permitted to issue bonds with higher than a 4.00% coupon.

So all that post-WW II growth of the US economy--when you could literally talk about "only safe haven"--was racing against bonds tethered to a prison ball.

Which is why, pace Ellen1910, one can look at GDP and the stock market from 1967-1983 and reasonably predict that the market must rise significantly.

Stock dividends alone were higher than bond yields 50 years ago. The conditions that produced so large an equity premium through the last century, would seem unlikely to recur. Remember too the bond prices are high, but stocks too selling at price earning ratios that are higher than the long historical average.

Just wondering... is anyone else getting a little tired of this Social Security debate? Should this really be the top item on anyone's agenda? Aren't there other problems we should be focusing our resources on?

Focusing on the Social Security debate has taught me more than I might have thought a few months ago. Think about the depth and nuances of just this post by Brad. Think of how these issues are relevant across countries. Time is not wasted, I am not the least tired of the debate though worried about the consequences of attacks on our social benefit structure.

something else not to forget about equity returns: about half of the appreciation in the major indices post '81 is just from investors willingness to pay twice as much for a dollar of earnings today as they did then. Earinings growth has been much less impressive over that time period.

I'm with Joe Deely.

If you believe the Trust Fund has real assets (loans made by low-moderate earners to high income earners to be paid back by the latter), the "debate" is over.

Note: I invest my IRA/401(k) in broad indexes -- 60% equity 40% debt ratio. You can choose your own. But that has nothing to do with Social Security.

John Bogle's research shows that mutual fund investors have over the last 25 years made significantly lower returns on stocks than the S&P Index returns. The problem of course is investment cost. Mutual funds are expensive; too darned expensive.

Well, the equity premium has been lower than through the century these last 10 and 20 and 25 years. With mutual fund costs that are even 1% the equity premium is cut in half. With 2% costs, the equity premium is about gone.

The price earning ratio for the S&P Index in 1979 was about 10. In 1989 the p/e was about 15. In 1999 the p/e was about 30. Lately, back to 19.

This is probably a really dumb question, but here
goes anyway: when every talks about "average
returns" from stocks or bonds, are they using
the arithmetic mean of the percentage growth,
or the geometric mean of the growth factor ?

It's fairly obvious that anyone honest should be
using the geometric mean of the growth factor -
if I gain 50% and lose 50% next year, then I have
a 25% loss overall, which is not the average of
+50 and -50 - but there's enough blatantly
dishonest debate going on these days that I'd
like to be sure.

I belive "average returns" are essentially treating the entire mass of publicly traded companies as a unit. (Or a "representative subset" such as the S&P or a standard index fund that tracks same.)

You look at the collective price, and divide by the collective earnings, and there's the average P/E. Take the total dividends, divide by total price, and there's the dividend yield. Look at total buybacks, divide by price, and that's the buyback yield. Look at the year-to-year rise in total price (minus the amount from buybacks), and that's the average capital gain.

Incidentally, one thing that's commonly elided by anti-privatizers is that if you own the RIGHT subset of the market, you do have the potential to make outlandishly good returns. But it's mathematically impossible for EVERYONE to beat the market as a whole.

Richard-

Geometric mean, yup. There is no such thing as a dumb question--these days especially. We need more questions.

The whole world of finance/investing mostly exists in a narrow and one might argue rather pedestrian band of numbers between 0 and 2. When you get over 2, you break out the Veuve Clicquot.

Returns 25% 15% -5% 5%

1.25 * 1.15 * .95 * 1.05

= 1.43

-1 => 43% cumulative return

1.43

take to the 1/4 power (1 over time periods), and

1.094

-1 => 9.4% annualized return

Goodness knows if the shills are using arithmetic means. I'm certainly not.

Auros,

Bush is trying to sell privatization on the basis that we all live in Lake Woebegone and everyone's returns will be above average. They won't.

Brad two points. Your analysis is a steady state analysis. In particular you assume that Y/K will be fixed. This fits the facts (like rK/Y) is stable, but should be stated.

More to the point, in the Solow model, capital can be freely converted to and from consumption goods so Q must be 1. One (like Brad DeLong) can (and has) made sense of the CEA argument by noting that equity returns, unlike GDP grwoth, have the initial price of equities in the deonominator. If the S&P 500 cost $1 there would be plenty of room for 6.5% real returns. The missing link in the CEA argument is that, if GDP growth is 1.9% and investors require 6.5% real return on equity then the P/E ratio is X. X is, I think something on the order of 7 (must be lower than average post WWII P/E to give the same return with lower GDP growth. The naive observer might think the S&P p.e. is 19, but the CEA has just proven that it must be lower, because we know (somehow) that investors require 6.5% real returns and we know (somehow) that they are rational.

Of course the President also argues that investors are irrational and that the equity premium is a free lunch not a risk premium.

On bonds and inflation I don't see how surprise inflation could possible be positive on average without actual inflation increasing. The inflation rate now is about what it was in 1900. How could people have underpredicted it on average ?
The 4% coupon rule doesn't explain why anyone was willing to buy such bonds, nor why corporate bonds paid similar returns.

[comment spam]