Chapter 6

The risk of equity markets

Understanding the risk you take to get returns

It seems that every pre-bubble period is characterised by an abundance of changing paradigm stories or that ‘this time it’s different’, only for history to repeat itself and markets falling. What follows are the inevitable stories about people who saw it coming and those who predict further gloom.

While nobody really knows what will happen to the stock markets, we can make some observations about the risks we take in investing in them. Figure 6.1 shows world and US equity market risk over the past 30 years, illustrated as the trailing 12-month standard deviation (SD) of the returns (explained below).

What you immediately notice is that the risk moves around a lot – it won’t surprise you that the markets moved around a lot during the 2008–09 financial crisis (notice the spike in 2009 where the standard deviation was over 40% for world equities), while market returns were far less volatile in 2007, right before the crisis. What you also notice is how closely tied the world and US risks appear. This is not a surprise as the US market is the largest component of the world market, but also because the world is far more interconnected than it used to be. But what we can also see from Figure 6.1 is that expecting a standard deviation of the equity markets of 20% on the basis of how it has been in the past is not a terrible guess.¹

Figure 6.1 Risk of equity markets (trailing 12 months standard deviation)

The standard deviation is important as it is meant to give you an idea of how much returns may vary. It assumes that returns are distributed around an expected average return of all the many potential outcomes, and the standard deviation tells you how different from the average return future outcomes may be. A higher standard deviation means that more outcomes are very different from the average outcome, while a low standard deviation suggests that most outcomes are clustered around the average outcome. While we don’t know what the future outcome will actually be (unless you have a crystal ball), the standard deviation helps us understand how great a variation there may be in actual future outcomes. Please see the Appendix for a more in-depth discussion of the standard deviation, including the common critique that standard deviation measure massively underestimates the frequency and magnitude of large declines in equity markets.

You can lose a lot!

Most investors who invest in equity markets worry less about small fluctuations in the value of our investments – we worry about the risk of losing a lot of money quickly.

With an increase from around 40 at its inception in 1896 to trading at around 20,000 at the time of writing, the Dow is a manifestation of the equity market’s staggering success story and a neat illustration of the riches that are to be gained from investing in stock markets. But even this shining example of the success of equity markets has periodic massive declines in value, most notably around the Great Depression. Figure 6.2 shows the worst losses the Dow has incurred since its inception.

Figure 6.2 Biggest losses in the Dow Jones since 1900

It is obvious that investing in the stock markets can lose you a lot of money.

Figure 6.2 is a great way to put things in perspective. A super-simple and non-mathematical way to use the chart would be to say, ‘If there is a crisis that is as bad as the 1930s depression, at least we know that then the stock market declined by over 80%.’ Or, ‘If the 2008 financial crisis was to be repeated we know that in that case the markets declined by 50%.’

While future events that lead to losses in the market may not look like past ones, this kind of simple comparison is a useful way to frame risk.

Recovering from the magnitude of losses like those in Figure 6.2 can take a long time, even if you were able to endure the painful losses and hold on to your equity position. Keep in mind that had you invested $100 in the Dow before the 1930 declines started you would have been left with only about $15 at the bottom, challenging even the most patient investor.

Figure 6.3 shows how long it took in years for the Dow to recover from the declines shown in Figure 6.2.

The point is that even once you have hit the bottom it may take a long time to be dug out of that hole, and in many cases longer than most investors’ time horizons.

Figure 6.3 Time (in years) for the Dow Jones to recover from losses

Don’t assume that markets always bounce back

Of course the charts above are based entirely on the US stock market and specifically the Dow, a small subset of that market. People often use the US stock market for data analysis as it is not only the place with the most comprehensive data sets but it has, at least historically, dominated financial academic circles. Until only a couple of decades ago it was not as simple to get international data seamlessly and even if you did get the data it was not easy to analyse.

One of the problems with using US-based data is the large selection bias that is introduced. The twentieth century was the American century and the stock market reflected this success. But just because we use data from a very successful century in a very successful geographical area does not mean that things will be like that in future.

Imagine if you were an investor in the Russian stock markets or government bonds just before the 1917 revolution. You would have lost everything without hope or recourse. Likewise, there have been many instances of large-scale and irrevocable losses for investors. When studying the charts above there may be a dangerous tendency to believe that ‘it may take a while, but equity markets will always come back’. They may not.

The disadvantage with the view that markets will always bounce back is that some investors will want to ‘average in’ to falling markets, i.e. buy on weaknesses, dips, etc. After all, if markets always bounce back, the reasoning goes, you will eventually be fine. That thinking is akin to a gambler going to a casino, betting on red and keeping on doubling down whenever he loses. That strategy works really well until the one time when it really, really doesn’t work and you run out of money.

Some may view the rebound from the 2008 financial crisis as evidence that markets do always bounce back, and that historical instances of complete and ‘un-rebounded’ losses are exactly that: history. To those I would say, look at Japan.

In Figure 6.4 the main Nikkei index is trading at an approximate 50% discount to its peak in the early to mid-1990s, despite rebounding more than 100% from the worst point (so 20 years after investing, some investors were still down 75% in Japan). When I was studying economics at university in the early 1990s, I remember how often we considered the Japanese miracle. We were told how special characteristics like lifetime employment and superior production techniques had led to Japan’s spectacular rise and the implication was clear: Japan would continue to prosper. But already as I headed to business school in 1996–98 there were few mentions of the Japanese economic miracle, other than the odd case on auto production techniques.

Figure 6.4 The Nikkei index since its peak

Even if the Nikkei does recovery fully it is entirely possible that this will be so far into the future as to be irrelevant for many current readers’ financial lives. From the vantage point of 1990 many would have predicted the chart shown in Figure 6.4 to apply to leading European or US markets, not Japan. So don’t exclude the possibility that this can happen in your local equity markets or even the wider world equity markets.

We don’t know what is to come, but it’s dangerous to extrapolate too much from historical data alone. We don’t know what we don’t know, and it is hard to incorporate this factor. Over the next year, decade or century we could be blessed with peace and prosperity, or some completely unpredictable calamity. Looking at how the US stock market reacted to the dire events in the twentieth century may give us an indication of how the world stock markets will react in future, but it’s also quite possible that future losses will look very different.

Diversification and the false sense of security

Although correlations between various national stock markets have generally increased over the past decades, the benefits from diversification are still clear. Local companies, in general, are more dependent on the global markets than they were decades ago, but there are still unique characteristics to national economies. The latter operate in different legal and political climates and while they may be dependent on local factors such as access to commodities, skilled labour, tourism, etc. they are also influenced by natural disasters (as in Japan) or political upheaval (as in the Middle East, Brazil, etc.). Diversifying away from such exposures in one region or country makes a lot of sense.

The main problem with higher correlation in down markets is that we are not always afforded the protection of diversification when we really need it. We are probably less concerned that the volatility of our monthly returns is slightly higher because correlations are higher than expected if this happens in a market where we are up 10% a year (see Figure 6.5). But if we are down 40% because of high correlations between our different investments then we care a lot. If we believe that the correlation between investments is more or less a constant number we would have understated our portfolio risk in bad markets and be more exposed to losses than we thought. Increased correlation was a factor that affected a lot of investors in the 2008 crash, when not only did various international stock markets correlate, but also several supposedly uncorrelated asset classes did so as well. (Certain government bonds were a notable exception, being the safety asset in the turbulence.)

A major selling point of the US sub-prime investment proposals was that there had never been a case where all housing markets in the US had declined at the same time. This diversification was supposed to provide great investment security and was a major driver of the high ratings and attractiveness. In simple terms, investors did not believe that the housing markets in Miami, Las Vegas and Dallas would all collapse at the same time and were therefore willing to provide more debt at a low cost. During the crash, correlations between the various housing markets shot up to the point at which they acted like one market instead of providing diversification against a location-specific decline. The resulting eggs of embarrassment on the face of the financial community and huge monetary losses became all too obvious.

Figure 6.5 World equity returns and correlation between markets (illustration only)

Some people look back at historical market declines that used to be more geographically contained and wish they had had the chance then to diversify internationally. Now they can diversify, but correlations are up. You can’t have it all. Over time, as cross-border capital flows to foreign equity markets increase and the world generally becomes more interconnected because of trade or information flows, correlations will probably increase even further. Higher correlation means that we would be fooling ourselves if we think diversification alone protected us against bad things and we accepted higher risk in our portfolio as a result.

Risk rethought

To some, standard deviations or skew might sound like an archaic finance-related term that will get us all in trouble. What I’m suggesting is not to use this as an exact science, but rather as a tool that gives a general idea of what the perceived risk of investing in equities will be in the future. The next step is to figure out how to use that in understanding the risk of our portfolio.

Other than basing things on a gut feel or what we glean from newspapers, we can get a more accurate sense of where things stand. A future market standard deviation of 20% is clearly not the same as one of 40%: the market is expecting far less risk. Similarly, you can’t just assume that future returns are distributed like the neat bell curve of the standard deviation – they are not. Understanding that there is skew helps to explain this; it makes more sense that really bad events can and probably will happen on occasion despite the standard deviation suggesting that it is as unlikely as being hit by a meteor.

This chapter has been less about giving answers than about informing the gut feel that investors probably already have about the equity markets. In the long run we can expect good returns from equities, but this is not without risks, and those risks can be unpredictable and severe. We’d better plan for that.