James Montier: “I want to break free”

James Montier, former Societe Generale analyst now forking for GMO, put out couple days ago a nice report about risk, benchmarking and asset allocation. Whole report is worth reading, but let me high,light two things:

Problems with Policy Portfolios

Problem 1:  Risk isn’t volatility. To begin, we should ask ourselves why we are concerned with volatility as a measure of risk. Modern portfolio theory is so entrenched in the innermost workings of the world of finance that risk is typically defined as standard deviation or variance. 

However, risk isn’t a number. It is a concept. Ben Graham argued that we should focus on the danger of permanent loss of capital as a sensible measure of risk: What is the chance that I will see my capital permanently impaired by this investment? This strikes me as a much more sensible viewpoint than the mathematically elegant but ultimately distracting practice of assuming that risk is equivalent to standard deviation. 

Volatility creates opportunity, not risk. As John Maynard Keynes long ago opined, “It is largely the  uctuations which throw up the bargains and the uncertainty due to fluctuations which prevents other people from taking advantage of them.”  

For instance, let’s look at equity volatility. Exhibit 3 shows a measure of the volatility of the S&P 500. Were equities more risky in late 2007 or early 2009? If you follow the edicts of standard nance, then 2007 was a much less risky year than 2009. Now, tell me again that risk and volatility are the same thing! 

Problem 3:  Benchmarking alters behavior. The third problem is that benchmarking tends to alter investment Managers’ behavior along three important dimensions. First, managers motivated to compete against an index may lose sight of whether an investment is attractive or even sound in an absolute sense. They focus upon relative, not absolute, valuation.

Second, as soon as you give a manager an index, the measure of “risk” changes to tracking error: how far away from the benchmark are we? Sadly, in a benchmark-tracking-error, career–risk-dominated world, Keynes’ edict that “It is better for reputation to fail conventionally than to succeed unconventionally” governs the day. For benchmarked investors, the risk-free asset is no longer cash, but the index that they are compared against. 

This is evidenced by the drive to be fully invested at all points in time. After all, if the goal is to beat the market without falling significantly behind, it makes sense to remain 100% invested. (Witness Exhibit 6, which shows the relentless decline in cash levels in U.S. equity mutual funds.) Remaining fully invested at all times means that the investor simply chooses the best available investment. Relative attractiveness becomes the only investment yardstick. 

Finally, benchmarked managers start to think about return in a relative sense as well. I’ve always hated the idea of sitting in front of a client having lost money, but claiming good performance because I’d just lost less than an index. That very concept sticks in my craw as an investor.

The bottom line is that, effectively, everything becomes relative (risk, return, and valuation) in a benchmarked world. 

JM-I Want to Break Free