Measure for measure

Buttonwood

Aug 30th 2005
From The Economist Global Agenda

Investors are mad for risk-appetite indices. What do they actually tell us?

IT CAN now be revealed that Buttonwood has not always been a fan of cricket. Brought up in the robust and marginally faster-moving world of baseball, she has long found the appeal of five-day contests involving leg-spinners and silly mid-offs inexplicable. Until this brilliant holiday weekend, that is, when The Economist trounced the Spectator and, in a slightly more publicised match, England edged past Australia.

In much the same way, America’s stockmarket edged past news on Monday August 29th that Hurricane Katrina was set to become the world’s costliest storm ever for insurers, helping to push oil above $70 a barrel: share prices moved up within half an hour of the market’s opening, though they fell again the next day. It was certainly not the first time that investors had bought stocks when the news was bad, like a pitcher shaking off his catcher’s signals. Were investors cleverly re-assessing economic fundamentals (basically solid growth and still-good prospects for corporate earnings) and upping their rational valuation of shares? Or were they just irrepressible exuberants on a tear?

This question, it turns out, is at the centre of one of the big divides in financial theory. Those who believe in efficient markets think that prices reflect all known information about future returns. If an investor happens to feel irrationally feisty when he rolls out of bed in the morning, no matter: smart money will move in and arbitrage away the difference between the price he paid for the share and the price that correctly reflects the discounted value of expected cash flows.

Advocates of behavioural finance, on the other hand, reckon markets are imperfect. Sophisticated investors are often unable or unwilling to offset sentiment traders entirely, especially when there are constraints to arbitrage (it is not always possible to short a stock, for example). For many, “investor sentiment” does matter.

Which is just as well, as there are an awful lot of people measuring it. Take last week. On August 22nd, the UBS/Gallup Index of Investor Optimism said the slow but steady rebound of investor confidence since April had continued in August. On August 23rd, State Street Global Markets’ Investor Confidence Index also showed improvement since the recent low point in May. CSFB’s Risk-Appetite Index, which only in March was reflecting “euphoria” before the big plunge in April and May, shows another sharp fall in risk appetite over the past two weeks or so.

The divergences among these measures are perhaps less surprising than their congruencies, for they are constructed along different lines with different variables. UBS polls qualifying American investors about their expectations for financial markets. CSFB looks not at what people say but at what prices suggest they are actually doing, evaluating returns on 65 asset classes around the world to see whether riskier ones are outperforming safer ones. State Street concentrates on quantity rather than price. Drawing information from the $9.5 trillion in global assets that it holds in custody, the bank analyses periodic changes in this pool to determine whether institutional investors are assuming more or less risk.

How well do the different indices capture investor sentiment? The authors of an article in the Bank of Canada’s June Financial System Review put ten of them under the microscope. They chose five recent episodes in which markets reacted extremely and indices might have been expected to reflect unambiguous investor sentiment: the Russian debt default in 1998, the peak of the dotcom-fuelled bull market, the beginning of the 2000-2002 bear market, the terrorist attacks of September 11th 2001 and the beginning of the 2003 bull market. The chart below shows the results, with CSFB’s index best capturing the moment.

There are literally dozens—maybe hundreds—of other measures of investor sentiment, ranging from official versions, such as those the International Monetary Fund and the Bank of England have for surveillance purposes, to private-sector models, many but not all of which purport to give a trading advantage. They are branching out fast. Yale’s School of Management, which already has a much-watched index based on a poll of institutional and retail investors, and tracks investor sentiment in Japan, is launching similar indices in a number of new countries, beginning with India next week. And State Street is planning to break down its global index to show risk appetite specifically in North America, Europe and Asia.

Assuming that these indices do actually reflect what they are supposed to, what do they tell us that is of practical use? Do they predict future returns? If not, why are investors prepared to pay an arm and a leg for some of them? After all, the Chicago Board Options Exchange's VIX contract, which tracks implied volatility in the S&P 500 share index, is known as the “investor fear gauge”—though it has come unglued of late.

Academic journals are stuffed with articles about how to measure investor confidence and what to do with it once you have measured it. Most suggest that sentiment is not a reliable predictor of long-term future returns in general. But it may help to predict returns on certain kinds of shares, says Malcolm Baker of the Harvard Business School, whose research with Jeffrey Wurgler, of New York University’s Stern School of Business, will be published soon in the Journal of Finance. Shares in new, small, non-dividend-paying firms—basically, those without much of a track record and whose shares are hard to arbitrage—are more likely to outperform when investor sentiment is pessimistic at the start of the period being studied.

“I love to see those academic studies saying that sentiment doesn’t matter,” says Jason Goepfert, chief executive of Sundial Capital Research in Minnesota. “That means it’ll work all the better. We’ve looked at all sorts of indicators—fundamental, technical. It’s not that investor sentiment doesn’t fail, but it fails less often.”

Sundial’s sentimenTrader.com offers a variety of confidence indices to subscribers, both institutional (including hedge funds) and individual. Mr Goepfert points to one real-money gauge in particular: the Rydex fund family, one of the few to make its asset data readily available and in the past a good indicator of sentiment. Allocations are close to their all-time most bearish now, with loads parked in the money markets and a leveraged short position that is greater even than last April, when most sentiment indicators dived. This suggests that returns are set to rise over the coming weeks.

Investor sentiment seems to be most predictive when it is near historical extremes, and then only in the short term. In the longer run it has less to say, for returns should reflect the business cycle and corporate earnings. The big question now is what the impact of sharply higher oil prices will be. If they act as a brake on global economic growth, as seems likely, it will knock investor sentiment and returns, both short-term and long-term, for six.

Send comments on this article to Buttonwood (Please state whether you are happy for your comments to be published)

Read more Buttonwood columns at www.economist.com/buttonwood

Copyright © 2005 The Economist Newspaper and The Economist Group. All rights reserved.