The easiest to get this is to look at Hussman's predictions and the total return of the S&P500 (with dividends reinvested). IIRC, he started to get bearish at around 1000 points in the S&P. I'm sure he will eventually get his pullback and he will say 'see, I told you that these valuations were not sustainable', yet he is being off by so much (in terms of time) that the pullback will still likely to lead to a level of dollar return (total return with dividends reinvested) that will be magnitudes higher than the 1000 points level where he was sounding the red alarm. He shorted a convex function with his whole overvaluation/mean reversion thesis and destroyed his career as a result of that. Its far more important to understand these sorts of risks and balance your exposures (that is, protect yourself from being short convexity, seeking to be long of it) then to look at these garbage 'models' that are short compounding and are super error prone
The thing is when I say short convexity/compounding I'm not talking about the regular 'short' convexity exposure one gets when someone moves from stocks to bonds in normal times. Like say someone gets out of the stock market (which is expected to deliver 5% a year) to bonds (expected to deliver 3% a year). That person is giving up 2% a year on the hopes of seeing a decline in stocks so they can invest in a stock market expected to deliver 7-8% a year. This is ok (specially if there is a reentry strategy), that's not what I'm referring to I'm referring when these mean reversion aholes see some economic progress (or the chance of economic progress) and they go all 'mean reversion' on people. There is an extra dose of convexity there (one that is not obvious). Like in that example where ROEs were 10% and they go to 15%, if it turns that such performance will go longer than usual, the mean reverter will face a BIGGER stock market return than he expected, so the cost of being in bonds vs stocks will be much higher than expected. And this extra performance will compound yearly for god knows how long. So a higher % is being given up for an uncertain amount of time, this is an usual amount of convexity, its 'market timing short convexity' on steroids. Same thing happens when there is the chance that ROEs will increase or when growth rates will increase (or when a bubble might be starting) Market timing has its time and place, but its NOT when there is economic progress going on (which historically has lasted a certain amount of 'average length' and one wants to fade that) or when there is a macro regime shift (2009 economic reversal, 2016 Trump election). The funny thing is that when people have REAL short positions, in some level they understand all of this. They will close that short position as they start to lose money and they hear about the bull case. In some level they know they are exposed to a HUGE loss even if they think the scenario is unlikely. They go ahead and close 'just in case'. Well, that's what bulls do, they buy it 'just in case'. If its rational in one instance why it is not in the other? If there is a small chance of a giant payoff, why is it wrong to speculate on such scenario? If equities are priced to deliver 5% but there is a chance that something will happen and they will deliver 7-8% for an uncertain amount of time, people will buy 'just in case' because they know they don't have to be right very often in their speculation to make it the right decision. That's the power of that extra compounding/convexity So this is why now I prefer to go outright short SPY/ES when I'm bearish (and keep my long stock positions). Because I know I will treat that short position with the care that it needs, I wont let it ride to infinity. But if I go to cash/bonds, I might very well do that
So a rule can be derived from this: Avoid shorting (outright, by being in cash/bonds or through bearish comments) when the upside is unusually uncertain and certain catalysts could be increasing that upside. Err on the side of being long instead. If shorting is so risk, knowing how much risk one is taking is crucial. In situations where that is hard to asses it makes no sense to short. If its so bad to be short, being long is likely to be a better bet
This rule can be applied to situations where a stock markets get hammered hard but then puts some kind of big bottom and then a positive news catalyst hits. In that case, shorting is so ridiculously bad (market upside is uncertain and catalysts could be increasing it), a long starts to make sense
I guess another way of saying this is that the stock market is quite forgiving of optimism that turns out to be mistaken. The stock market, however, is brutal to pessimism that turns out to be mistaken. All that to the convexity there which is amplified compared to most other markets. Therefore, as a rule, most things should be interpreted as being bullish even if you got reasons to think that they are not. That is, the threshold for wanting to sell should be very high. Especially in relient economies/markets like the United States.
Here is an article talking about valuation and stocks in a different perspective. Very good read http://www.philosophicaleconomics.com/2017/02/asset-markets-as-banks/
So to recap, when someone sells/shorts/warns people about equity valuations in periods where business are or might do especially well, they are short so many things -They are exposed to that good business period going longer than it did in the past (with exponential/convex consequences) -They are exposed to the fact that the market might be under apreciating the chance that bussiness will be doing well -They are exposed to valuations going to a level (or avg level) that is higher than it did historically (with exponential/convex consequences, in Japan the CAPE ratio reached 80). When they come back down, they still hang around in a area that is much higher than it did historically -They are exposed to the fact the business might do even better than they are doing now, even though they are already doing better than average (in other words, they are short progress). With exponential consequences -They are short risk premia (as most bearish/defensive instances are) -Other exposures that I haven't thought about And what do they get in return? A chance to buy the market cheaper (at higher implied returns). But this 'benefit' is limited, it can only be so much, meanwhile they can lose a run that is massive. So its a limited gain massive risk 'trade'. When one considers that people (in their working years) are short valuations (so they are already naturally exposed to this trade), increasing it even further is just flat out irresponsible, if not lunacy
Jim Chanos has a saying that says 'I have seen a lot of stocks go to $0 but I have never seen any stock go to infinity'. But he is wrong IMO, he needs to look at stocks that go up by 200-300-500-1000-5000%+ and weight those losses against the gains of a short that go to $0. I have seen countless stocks rise 1000% (over many years) and the consequences to a short when that happens is huge. People seem to miss that because when a short works out, usually the decline is very fast (50%+ in a matter of months) but when a stock works out, usually it will grind 10% a year for decades. So people miss out on that asymetry The same idea can be applied to the index. A bearish instance needs a much higher conviction and % chance of playing out than a bullish one. A bull will be bailed out by the natural assymetry of stocks, a bear will get hurt by it. That's why as a rule, most things are or should be interpreted as bullish (or should be ignored). Going bearish is almost like a nuclear button, it should be saved for rare situations. Especially in resilient economies
But this doesn't mean that I think equities are so wonderful that one should have 100% of their assets in it at all times (like Buffett recommends). The data shows that people don't have that kind of risk tolerance (I certaintly don't, not unless I'm seeing HUGE bargains). One needs to figure out how much equities they can tolerate (assuming they will drop 50% in the next 6 months) and hold on to that allocation (which usually will be 10-30%) like a pitbull. Not try to 'time' it, not listen to valuation assholes and not get scared out by headlines. They should own that bucket pretty much forever, even if the individual securities in the bucket change
Along with this logic I just added a 1% position to SNAP (I know, sounds nuts right) at $26. Stop at $23. If US equities will be in this bull market for years as I expect, SNAP might go on a monster GPRO/FB type tear so there is a huge convex exposure there. And my risk is limited to the IPO low. I'm adding this to my 'convexity' stock basket (VRX, FNMAS, GREK). All I need is just 1 of these to work to pay for all the losses and perhaps give me a profit. 1% position with a close by stop turns this into a super low risk trade, but it might double or triple from here on pure speculation