California money manager Robert D. Arnott may own 11 vintage motorcycles, but he's no easy rider. In a series of articles, he has challenged the conventional investment wisdom that equities always produce a real return over bonds, attacked the idea that reinvesting corporate earnings leads to higher growth, and blasted the notion that stock buybacks enable earnings to grow faster than the economy. Now, he warns that workers will have to work until their early 70s. A four-time winner of the prestigious Graham & Dodd Scroll for best article on investment theory, Mr. Arnott just has been named to a three-year term as editor of the Financial Analysts Journal. He remains chairman of First Quadrant Corp., and chief executive of a new enterprise, Research Associates LLC. Research Associates has a subadvisory deal with PIMCO Funds, and also has lined up projects doing work in asset-liability modeling and software for tax-advantaged investing. Mr. Arnott talked to Chief of Bureaus Joel Chernoff about motorcycles and his controversial research.
Q Why did you visit Jay Leno's house?
A I've been to see his collection (of cars and motorcycles) three times. He's a very relaxed, down-to-earth fun guy. I went on the Love Ride last year, which is a motorcycle fund-raiser for muscular dystrophy, with him and his group. My biggest concern about Leno is that his taste in motorcycles and mine are almost identical. So if I'm interested in a particular bike, and I have any concerns about him being interested in the same bike, he can outbid me hands down.
Q What is your outlook for the stock market?
A Near term, kind of neutral. The bear market has wrung out a lot of the excesses. Long term, I'm still a bear. And, adjusted for inflation, stock prices won't be higher 15 years from now than they are today.
I see four elements of risk. One, the risk premium is too low (roughly 40 basis points). Two, simple valuation measures are still too high. If you normalize earnings by taking a 10-year average, prices relative to 10-year earnings, you're looking at a multiple of 25 times. That's too high. Historically, (prices) got above 25 times 10-year average earnings only in 1901 and in 1929. In both cases, p/e ratios were 70% lower 20 years later.
I don't think the actual outcome will be that gloomy. But it's important to note that a secular bear market doesn't mean the market cycle has been repealed. You still have bull and bear markets. It's just that the bull markets are anemic and the bear markets are worrisomely damaging. We saw this in 1965-'82. Adjusted for inflation, stocks fell 70% in 17 years.
The third element of risk is quality of earnings. Quality of earnings is still rubbish. Pension fund rate assumptions are still four percentage points above long-bond yields; that's worth between $8 and $10 per S&P share. S&P earnings were $54 in 2000. Stock options weren't being expensed. That imputed value for 2000 was between $6 and $8 a share; today, it's about half that. You take known fraud - WorldCom, Enron - you add it up, and it was $3 a share for 2000. Now you're down to $36 a share, and that was a peak year.
So I think normalized earnings for the S&P are probably not higher than $35. It's an extreme view; most of Wall Street still thinks it's over $50. But if I'm right on this, and I'm pretty confident that I am, the market is at 25 times normalized earnings.
The fourth and potentially biggest issue is demographics. In 20 years, we have more retirees than ever before in history, selling assets to fund their retirement needs to a proportionately smaller roster of buyers, the workforce, than ever before in history. That's terrible for asset values.
Q You've challenged the notion that higher reinvestment of earnings leads to faster future earnings growth. What did you find?
A One of the most common arguments out there is don't worry about low dividend yields because companies are retaining earnings more than ever before, they are reinvesting in wonderful new ideas that will lead to faster earnings growth than ever before. But you only need to look at history to find that quite the opposite holds true. The more companies pay out in dividends, the faster their subsequent earnings growth.
If management is retaining a little bit of the earnings, they're investing in the most important project or projects for their long-term survival or competitiveness. If they retain more, they start to go down their wish list of additional projects. And there's some real hubris if management thinks their fifth-best idea (out of a limited universe) could possibly be better than their shareholders' best idea out of a limitless universe.
So what we find is when payout ratios are low, subsequent earnings growth is atrocious. And that's exactly what we saw, yet again, at the peak of the bubble. The retained earnings become a piggybank for pet projects and empire building.
Q What about stock buybacks allowing earnings to grow faster than the GDP?
A If earnings are growing as fast as the economy, and per-share earnings are growing 2% a year slower, the implication is that there's a 2% dilution going on - new share issuance exceeding buybacks by 2% a year. If the S&P rises 10% and S&P capitalization rises 15%, then you've got 5% more shares. So if we track the growth of the S&P index over the last 50 or 100 years and then track the growth of S&P capitalization over that same span, and take the ratio of the two, we have a very direct, very simple measure of dilution. And it mirrors what we found in terms of earnings and dividend growth being slower than GDP growth by essentially an identical margin. In the U.S., the shortfall in both cases was 2.3% per annum over the last century.
Q Can you explain your assertion that Social Security is facing a demographic crisis, not a financial one?
A This was the most disturbing paper I've ever worked on. When my generation retires, there's going to be proportionately more of us than ever before in history supported by proportionately less workers than ever before in history. In order for that support ratio to not rise, my generation has to retire at age 72 or 73.
And that boost in retirement age takes place between now and 2030, with most of the increase taking place between 2010 and 2020.
There have been a lot of arguments raised to suggest ways out. How many (immigrants) do we need to keep dependency ratios where they are? We need 118 million immigrants in the next 20 years - and they'd better not bring their parents with them.
What about imports? The goods and services that retirees consume are mostly domestic goods and services. You can't import a golf course. You can't import a nurse to take care of your health needs.
People say, if productivity rises 2% a year, the support ratio can rise 2% a year. That holds true if the people who are producing more goods are perfectly content to hand 100% of that to the retirees. That's not going to happen. The people who produce goods and services more productively want to enjoy the benefits of rising productivity.
Robert D. Arnott, chairman, First Quadrant Corp., Pasadena, Calif.
Assets under management: $12 billion
Performance (for periods ended Sept. 30):
1 year -4.0% (-8.2%)
3 year -4.6% (-4.5%)
u.s. market neutral:
1 year +7.6% (+1.8%)
3 year +4.5% (+4.1%)
u.s. core equity:
1 year -13.9% (-20.5%)
3 year -10.0% (-12.9%)
Benchmarks (respectively): customized; treasury bills; s&p 500