The accompanying chart shows the Obama administration's predictions for what the unemployment rate would be if (a) the stimulus package was implemented, (b) what it would be if it wasn't implemented, and (c) what it actually has been. The abject failure to be accurate in these predictions has resulted in strong criticism from the president's foes (and even some friends). But Obama isn't the first president to make predictions (or for that matter, promises) that didn't hold (recall, for example, "read my lips: no new taxes!").
Qualifying predictions is important. I recall a few years ago when a couple sports pundits didn't bother to offer their predictions for the first round of the baseball league playoffs, as it was "a given" as to what teams would prevail in the first round. Well, they were wrong. And yet, their pomposity and arrogance seemed to give their predictions undeserved credibility.
I have voiced my concerns with ex ante risk measures, but recognize that clients and managers still want to see them, which is fine, as long as they recognize the underlying assumptions and the qualifications of these predictions. To make bold statements about what will happen will usually lead to errors of one sort or another.
.png)
Your points are well taken regarding the inability of anyone to "predict" the future. While that word only speaks of "telling" what will happen before it does happen, the more sesquipedalian version of the word is more ridiculous in both its sound and its intent: prognostication - implying that these "predictors" actually "know" what will happen in the future. We all know that "we can't time the markets" and yet everyone has a "forecast" of everything from interest rates to the level of markets, currencies, commodities and anything else that is eligible for investing (and let's not forget the real culprit of all this activity: generating broker fees and other administrative profits for Wall Street.)
ReplyDeleteBut... and it's a reasonable point to be made:
We invest for the future, and we need some reasonable basis as stewards and fiduciaries for the course of action we intend to take. This leaves us with only two options:
a) Close our ears to the "Talking Heads" (rather like the sailors lashing themselves to the mast to avoid the calls of the Harpies to dash their boats upon the rocks) and simply invest in a well diversified portfolio, making sure we have adequate liquidity for short term cash needs, and leave the plan in place - rebalancing the portfolio in spite of the levels of the markets. That's what is advised by many advisors and all academics.
b) Make well-reasoned statements of your expectations for the markets and make modest adjustments to your strategic asset allocation, As an example, in the current environment, many are investing more heavily in large U.S. companies and certain growth sectors that are expected to weather the expected slow-growth economy that we expect to continue, while allocating less money to bonds in this zero interest rate environment. You'll notice that this DOES NOT "bet the ranch" by making huge swings in the asset allocation plan (i.e. "Get out of bonds" or "Go to cash" or "Go gold.")
As a performance professional, you frequently analyze the behavior of active managers using methods that involve an 'allocation effect' and a 'selection' effect. What is that allocation effect besides an active decision to overweight and underweight certain sectors of the economy as the result of "predicting" what will happen in those sectors before it does? In effect, every active decision is the result of an expectation or a "prediction" of future conditions and the response of investments that should generate a relative profit to simply showing up the overall market. Why else invest actively?
So, is there really a difference? You've explained the difference in terms of the forcefulness of the statements, or the confidence of the ones making those predictions. That sounds reasonable, but then again, if you expect to influence the behavior of others, you must believe in what you are saying. Yes, we COULD be wrong, but we believe that we are right. We are not making guarantees, but we are willing to put our money (or more realistically "other people's money) behind these expectations. We call it "conviction" but perhaps it's nothing more than the behavioral bias of "overconfidence." The difference is probably simply whether you were right or wrong.
An obvious addition that should be made to the typical risk prediction is a confidence interval. A likely reason why confidence intervals are rare is because they are quite wide for risk estimates. Danielsson and Macrae have more thoughts along these lines.
ReplyDeletePat, thanks for your input. We see confidence levels with VaR, and too often folks ignore the other side. We hear this with weather predictions, right? (e.g., 90% chance of rain) Confidence intervals at least reinforce that we're not 100% confident with our prediction.
ReplyDelete