EDN wrote:Valuethinker, we see things quite differently. I'm cautiously optimistic and realistic, you are pessimistic. That's OK. More specifically:
Thanks for all your data.
In an efficient market, an asset class cannot offer higher returns without higher risk. In the 20th century, that risk, if you managed (you couldn't generally, which is part of the trick) to be internationally diversified, you managed to grab the risk premium of equities. It's also arguable (Dimson and Marsh) that the 20th century was a surprise century in terms of how well equities did. That's only true if we pick the markets that didn't go to zero though due to political and economic disruption. So if we had the luxury of spending the 20th century investing in Australia and the US, say, and avoided for example UK Death Duties (which were punitive) as well as other personal taxation (80% tax rates at one point). The really big fortunes in the UK are held in land (the Duke of Westminster, the Grosvenor Estate, is a larger land owner than the Queen) for this reason - it generates an income with some correlation with inflation, and no UK government has ever really tried to confiscate land, probably won't unless we have a revolution. And if you don't sell it, no capital gains.
If we argue that there are a series of anomalies allowing excess return for equities- The equity return anomaly itself (hotly debated), small cap, value etc.-- then we have to have a theory why.
Otherwise what we have is a 100 year model run, and there was a set of outcomes where equities did worse than bonds. If that's not possible, then you are in a trap. Equities can never do worse than bonds, therefore they are less risky than bonds and they offer higher returns. Therefore the market is inefficient.
One theory is simply statistical. We have only one 100 year model run. Financial markets are a model, and we've run them exactly ONCE over that period. So past historical relationships are simply artefacts of the data.
That's more or less the efficient market theory.
Another theory is behavioural. For specific reasons, humans *in aggregate* make mistakes about expected returns and that makes them exploitable. Latest version of this is the low beta anomaly (that has been applied to understanding Warren Buffett's success). Also see David Laibson, Richard Thaler et al 'hyperbolic discounting' ie that humans are biased towards the short term, thus allowing long term investors to arbitrage that.
The trouble with behavioural explanations (that explain excess return without higher risk) are:
- you have to then argue why they are not easily arbitraged away
- they tend to be 'ad hoc' in the way they are applied-- so we have psychological explanations when it suits us, but we can't convincingly explain their general application eg ALL market actors make the same mistakes and underestimate the likely returns from equities long run
A balanced view on all this is that equities are *likely* to return more than bonds, by construction. However given that equities are starting a price to books of c. 2.0 times in developed markets, they cannot easily double that valuation (as they did in the 20th century, say). We can't rediscover what a great investment equities are. (this btw is more or less Stephen Wright and Andrew Smither's argument, and both the books they wrote are an excellent read-- perhaps not coincidentally, they are Brits writing about Wall Street
In the last 110 years we got lucky, We ran the model, and equities delivered high real returns. But the source of real returns that was rerating (an increase in PE or PB multiples) cannot reoccur except in some Panglossian universe. So then it comes down to corporate profits growth + dividends.
Well dividend yields are not high by historic standards (once upon a time, it was conventional wisdom that equity yields had to exceed bond yields-- that was true until about 1958). And there's no reason to think that corporate profit growth (except perhaps in emerging markets) in the 21st century is likely to exceed the 20th century *in aggregate*. In fact I'd argue the other way: profits relative to GDP in the US and other countries are at an almost all time high. In fact, what is more likely in the 21st century is that labour will start to claw back some proportion of the ground they have lost since the late 1970s. Partly due to labour shortages (demographics) and partly due to politics.
It's a guess based on past experience that real returns on stocks of between 3 and 5% are likely -- over a 30 year period. However the example of Japan has to loom behind us- -it's no dead certain thing, particularly not in a world of aging demographics (the Fed produced a paper showing the PE ratio of the stock market is well correlated with the size of the 35-45 year old cohort in the country, coincidence? We shall see).
Remember also Samuelson's paradox. (John Norstrad's website has a good explanation). Equities don't get any less risky the longer you hold them
(unless you have a fixed wealth target, then as you achieve that target you can switch into low risk assets-- of which TIPS (or rather Real Return Bonds) are the lowest risk asset, by design).
And that's why most of us are on ranges of 40-60% for equities. The 1970s weren't great for equities or bonds, but bonds were worse (and in fact TIPS, which didn't exist then, are perfectly designed for a rerun of the 1970s-- hence their attractions). The 80s and 90s were obviously great for equities, but the 2000s have not been.
There are people here who have been investing for a long time, and seen some pretty sickening lurches-- people who remember the 'tech' bubble of the early 60s, or the 'one decision' Nifty Fifty of the early 70s, or the 1966-1980 market that ended with 'Is Stock Investing Dead'? and c. -40% real returns over that period. Taylor Larimore's father was a stockbroker, so they nearly starved in the 1930s. I have friends who when they got into the stock market in 1979 (as brokers, or investment managers, or corporate financiers) they were the first new people hired in 10 years-- tells you what the 70s was like.
What we are fighting against is the tendency, usually from people younger than ourselves, to look at those graphs and say 'wow, if you hold onto equities you make so much more than anything else so 100% equities' or '100% small cap value' etc. We've been around long enough to see these 'sure things' disappointed.
We've also been around long enough to endure some really bad financial scares (and some bubbles). Of which I have to say the gap between 15 September 2008 and 13 October 2008 (between Lehmans bankruptcy and Gordon Brown's Reuters speech, which kicked off bailouts in the UK, US and Europe in quick succession) was the worst-- with the collapse of money markets, the world financial system stopped functioning-- liquidity dropped to zero for all meaningful purposes. It was a moment of sheer terror, when the authorities and participants seemed overwhelmed by events, that I hope never to again experience.
I mean 1998 (Long Term Capital Management) was bad, so was 2000-March 2003, but this one the world financial machine literally stopped. I had a former high school classmate, a broker at a major Canadian bank (some of the best capitalized banks in the world) tell me 'we didn't know if we'd be broke on Monday morning'.
The other thing about people here, and I am not counting myself in this, is that they have a high level of mathematical/ physics aptitude combined without the dead weight of years of finance training that crams everything into Gaussian normal distributions.
In other words, show them a stock market graph over any period of time, and they say 'fractal'. Meaning it's timescale independent.
That's a pretty profound insight, and it led me to read Benoit Mandelbrot 'The Misbehavio(u)r of Markets'. Which was a pretty extraordinary insight into financial markets returns. His metaphor of a blind archer firing arrows towards a wall (some fly parallel and never hit). And the longest series of daily financial data we have (Egyptian cotton prices since the 1500s) shows that.
The fractal nature of equity returns tells you what efficient markets tells you. Ie that yes, equities are likely to rise more than bonds, because they are riskier. But the data we have is one model run, and you cannot assume that future model runs will show equities finishing up higher than they started.