Sometimes it seems as though all we ever discuss and write about is risk. In talking with investors who were burned during the most recent crisis and are now worried about the next one over the horizon, however, it's clearly never a bad time to rekindle some old flames of the discussion.
Unfortunately, the very definition of risk is elusive, as it can take on different shades of meaning depending on who's using the word. Here's a framework to think about it. These aren't widely accepted academic terms--just some labels to try and make the case.
True Risk
Sheesh. Where to even start? One of my colleagues invoked Justice Potter Stewart's famous declaration (about obscenity): "I shall not today attempt further to define ... [it]; and perhaps I could never succeed in intelligibly doing so. But I know it when I see it." Think of "True Risk" as the hypothetical, inherent risk in an investment--in other words, a Platonic ideal: The risk that exists, whether it's ever actually realised or not, and which any of us could only be certain of if we had perfect, godlike knowledge. Put another way, we're talking about risk that can never be truly and completely known or measured unless it rears its head and comes to fruition.
Some elements of True Risk are so ephemeral--such as whether a CEO will decide to perpetuate a fraud--that we're unlikely to ever have the capacity to measure them. Others are much more concrete and theoretically discoverable but can be very difficult to identify, especially in very complex or immature markets. It's not clear that the Financial Services Authority or other regulators would have acted differently before the financial crisis, for example; but there has been some reason to think that they probably did not have a full and complete picture of just how much financial liability there was in the system (not only actual debt, but also obligations from derivatives and other contracts). Even down at more granular levels, such as the dangers that developed among investment banks, their mortgage exposures, and their off-balance sheet exposures, it appears that regulators were unable to actually observe or understand certain risks. They were there, though--whether anyone could or couldn't see them.
If that's still not clear enough, think of it in another realm. If a thunderstorm passes over your office, there's a True Risk that lightning may strike its roof. You know that, but you probably can't say too much about its severity or likelihood of occurring. At this point, meanwhile, the scientific tools and resources necessary to perfectly describe that risk either don't exist or aren't practical to deploy.
A large portion of the work put into academic and applied finance, however, is devoted to coming up with ways to observe, predict, and measure investment risks. Those efforts fall mainly into the following two camps.
Fundamental Risk Estimates
This is arguably the closest we actually come to knowing True Risk (unless it's realised). It's what we estimate True Risk to be based on what most refer to as "fundamental" measures. Critically, though, even though many of these metrics can be calculated to measure things with great precision, they are ultimately measurements of those "things," and really are only estimates of True Risk.
When looking at the stocks of companies--or portfolios of stocks--for example, investors often look at accounting measures, such as the ratio of debt to shareholder equity, to gauge how highly leveraged they are. And while market capitalisation may be a little further removed, it, too, is the kind of basic factor that we often use to classify and think about how risky different stocks are.
For corporate bonds it might be debt coverage ratios that suggest how much cash or income an issuer will have available to pay its obligations. For mortgage bonds, one might look at any number of items such as homeowners' credit scores or their loan-to-value ratios. For bonds and portfolios of bonds (including mortgage bonds), we often use measures such as duration to predict how they will move, given changes in interest rates. The bottom line is that these are all different ways that we try to zero in on an investment's risk, based and built upon its fundamental characteristics.
Price Volatility as Risk
This awkward label is just an attempt to translate the über-jargony term "empirical volatility." Put another way, it's a reflection of how much an investment's price moves over time. It gets more complicated when academics layer on calculations, but most of modern finance is based on the following: Given that we can't really know True Risk, and that data for measuring Fundamental Risk can be really difficult to gather and can't usually be compared across different investments, the best proxy for "risk" is what we can observe through price fluctuations.
One can spend a lot of time picking and choosing time periods, but there is a limit to the usefulness of data gleaned from an investment's price movements. It can tell you what the "market" has perceived an investment's risk to be over time--which is certainly useful--but even that information can be muddied up by any number of factors.
Take U.S. Treasury bonds, for example. Putting aside speculation about Uncle Sam's economic situation and focusing just on the basic risk of whether or not bondholders will get paid, it's pretty well accepted that Treasury bonds are about the safest option around. By contrast, the likelihood that owners of high-yield corporate--or junk--bonds might fail to get some money back is generally a lot higher. Whatever precision we lack in being able to assess the True Risk of those sectors, we at least have good estimates of Fundamental Risk that tell us it's much greater for high yield.
The problem is, however, that for long stretches of time, broad swings in government bond prices can create the perception that they're a lot riskier than high-yield bonds. If you use a typical tool for measuring Price Volatility, such as the standard deviation, for Treasuries and high-yield bonds over the past 12 months, you'd find the former to be nearly twice as volatile as the latter. There are a few reasons for that (including the fact that Treasuries are much more liquid and actively traded), but the fact is that during unremarkable times for the market, high-yield bonds often just chug along, with their prices changing just a little bit every day, even if they're marching upward over time. In other words, unless serious risk actually materialises in high-yield prices--that is, bad things actually happen--then the statistics can actually send the wrong message.
The only way that "Price Observed Risk" would help you see the bigger True Risk in the high-yield bond sector is if you were to include a period during which high-yield bonds actually "blew up," as they did in 2008. Before that, though, you would have had to go all the way back to 2002. That means that high-yield risk remained relatively well hidden for a roughly five-year stretch (2003 through 2007). During that period high-yield bonds looked less risky as determined by Price Volatility than U.S. Treasuries.
The point here isn't to completely dismiss the value of academic research or the use of statistical tools to try and measure risk. Each method has some advantages and utility, and Morningstar makes liberal use of both in order to help investors understand risks, both hidden and visible. In fact, the only way to get really close to understanding True Risk, as we've labelled it here, is to combine many different indicators of risk. Of course, sifting among them and using the results to make wise and profitable investment decisions is an exercise even the best among us spend a lifetime trying to perfect.
