The US is a really, really weird country:
IF THERE is an article of faith among investors, it is that equities are the best investment over the long run, far better than government bonds. But research from Elroy Dimson, Paul Marsh and Mike Staunton of the London Business School into returns since 1900, published this week in the “Credit Suisse Global Investment Returns Yearbook”, suggests that this belief is misleading. Their data show, for example, that global bonds have delivered a better return than equities since the start of 1980. Thirty-three years is a long time by most people’s reckoning.
Add to that the problem of survivorship bias. Most investment research has focused on America, where there are a lot of finance professors. America was the great winner of the 20th century, both militarily and economically. Although its success seems obvious now, it was not the only great power a century ago nor was it the most-favoured market of early-20th-century investors.
The chart shows that in the 50 years after the end of the American civil war, the Russian stockmarket easily outperformed Wall Street. Russia, with its vast territory and industrialising workforce, was seen as the exciting growth opportunity for the 20th century. (Argentina was another favourite bullish bet.)
. . .
Austria has been another great historical disappointment. In the early years of the 20th century Austria-Hungary was still one of the great powers of the world, with an empire spanning much of south-eastern Europe. Defeat in two world wars, the break-up of the empire and two periods of hyperinflation meant that Austria had the worst real return of all 20 countries in the London Business School data, not just for equities but for government bonds and bills as well. An American investor who placed $1m in Austrian government bills in 1900 would now have just $100 left.
Austria shows that equities do not always pay off over the long term. Between 1900 and 2012 an investor in Austrian equities would have endured a period of 97 consecutive years of real losses. Investors in Italy and Belgium suffered real losses over periods lasting more than 70 years.
The inclusion of Russia and Austria in the database (plus China, where investors also suffered a 100% loss in 1949) is one reason why the professors show a lower historical global real annual return from equities (5% versus 5.4%) than they did in the 2012 edition of the yearbook.
. . .
Indeed, American investors may not realise how lucky they were. The real return from American equities between 1900 and 2012 averaged 6.3% a year; the return from the rest of the world was just 4.4%.
Lots of people have suggested that America’s amazing and unexpected economic success might explain our equity premium. I wonder if it also explains why Robert Shiller has not done particularly well in forecasting the US stock market in recent years, despite having what I view as the best anti-EMH model (and one I’ve used successfully on occasion.)
Suppose there were periods where Americans thought we were about to become another Argentina, but then each time we stepped back from the brink. Because these fears never panned out, the US market would appear (ex post) much too volatile. Obviously 1932 was one such period. But even in 1980 I can recall a sense that the dollar was becoming a joke, and America was becoming a banana republic. Stocks were quite depressed in real terms. As it became apparent we would avoid the worst case of currency debasement, stocks rallied sharply. And of course the period around 9/11 and 2009 were also periods of great pessimism.
Shiller seems much better at telling people to sell than to buy. Contrary to widespread impression his 1996 “irrational exuberance” call was not particularly accurate. And he missed the great 2009-13 bull market in US stocks. In my view historical P/E ratios are misleading, as the 21st century will see persistently low real interest rates on T-bonds. That doesn’t mean I have any magic formula to predict stocks, I’m just trying to give a sense of why the market might think high stock prices are justified, whereas Shiller’s model (which does fit past data pretty well) says stocks are much too high.
Please don’t take this as a green light to go out and buy stocks—the market has just doubled, for God’s sake. Even if I don’t agree with Shiller, it’s hard to ignore what’s happened in the past after stocks have soared. The real purpose of this post is to argue that finance studies relying on an outlier like the US might be very misleading. Yes, US stocks might be overvalued—but don’t rely on studies using past US data when reaching that conclusion.
PS. I wrote this 2 weeks ago—I should have posted it right away.