We live in an instant-gratification society, especially if you are five years old. “But I need it now!” is a phrase I’m becoming increasingly familiar with as Jackson becomes ever more able to express himself. The substitution of “need” for “want” is a wrinkle that initially surprised me, but clearly it comes from parental questions about whether he really “needs” that new toy. The adventure continues.
I was thinking of this, now versus later and need versus want, in light of the recent employment data. I wrote yesterday about ”snowdown” versus slowdown, and today’s stats emphasize my points. Employment gains were up, despite a snowstorm in early February, and this time, the establishment survey did better than the household, narrowing that gap. The unemployment rate rose slightly because more people were looking for work, which is a good thing. The slowdown fears arose from an excessive focus on short-term data.
One of the key points about the recent slowdown fears is that the trended data, measured over a period of months, continued to improve even as the individual data points bounced around. You got a much better picture of the reality of the situation by looking at the bigger, longer-term picture, rather than focusing on the immediate results.
You can also see this in the stock market. For many technical trading methods—I exclude the high-frequency traders, who are doing different things—longer-term metrics provide better signals than shorter-term. Based on my research, you get better results, for example, with a 10-month moving average signal than a 2-month. You get cleaner signals, and fewer false signals, with a 20-month as opposed to a 10-month. There are reasons to use both, but the signal is cleaner the longer the period you use.
Which brings us to measuring risk for individual portfolios. I read an excellent book recently, The Little Book That Still Beats the Market, that’s worth a look for a number of reasons. In particular, it points out the notion—beats you over the head with it, really—that the approach it recommends can underperform for months or years, but it doesn’t matter. Long-term results are what matter.
This is a great framework to define risk in our portfolios, and to determine what matters. Consider the most commonly used risk measure, standard deviation—will it matter in the long term? I would argue that it doesn’t: no one really cares how windy the road is when they arrive, unless the road was so windy as to be unsafe. Windiness is not a good measure of whether you crash or not, though.
What is a good measure of crash risk is drawdown. A severe enough drawdown, or loss, is exactly like a crash on a journey. A drawdown impairs your ability to continue, in a way that moderate standard deviation does not.
You can actually see this mathematically. A 50-percent drawdown requires a 100-percent gain to get back to even. In fact, any drawdown requires a larger gain to get back to even. Drawdown therefore directly affects how long it will take to start making money again, which is directly relevant to reaching the final destination.
I know I’m shamelessly mixing metaphors here, and there are problems with drawdown as a metric as well. I plan to spend some more time looking at this topic, but this is the broad context where my own research is focused right now.