 What are the causes of herd behaviour?
Economists have carried out a significant amount of theoretical and empirical research on herd behaviour in recent times. Herd behaviour is when some decide to make an investment, then observe that others are not making the investment, and in consequence change their minds. Or they decide against an investment and then change their minds and carry out the investment when they see others investing.
The reasons why investors are influenced by others’ decisions fall into two broad categories.
Investors, as individuals, often have incomplete information about financial markets and the countries in which they operate. Individuals may copy the behaviour of other investors or groups of investors in the belief that this group knows something they do not.
Institutions and their managers have a fear of being different. This fear may be underpinned by incentives and bonuses. Fund managers may be penalised for below-market returns rather than be rewarded for above-market returns. They will aim to meet the average return for all funds. In these circumstances, ‘word of mouth’ will play an important role in herding. Investors will aim for conformity with other investment professionals by engaging in consultation and informal networking.
Empirical evidence suggests that there is a greater tendency to herd, and hence greater volatility, in emerging financial markets than in the markets of developed
countries. Investors often have incomplete information about newly liberalised emerging markets. In such circumstances investors assume (wrongly perhaps) that others have inside knowledge. They tend to follow those with a reputationfor picking winners. Weak reporting requirements and lax accounting standards add to the problem because they make it difficult to take decisions on the basis of ‘fundamentals’.
‘Herding’ as discussed above may be unfortunate (and even reprehensible from the standpoint of its likely negative impact on pension funds), but it is rational. Eventually, however, herd behaviour may lead to irrationalexuberance, producing a bubble which eventually bursts. Asset prices begin to rise to heights which cannot be rationalised by reference to fundamentals. Investors start to buy stocks which are already rising in price at an unprecedented rate. The positive feedback effect rein- forces price rises and increases volatility. In the end, the bubble bursts and financial crisis ensues.
 Is it possible to predict a crisis? Can countries be identified in terms of their vulnera- bility to capital reversals? Using past data, economists have proposed models which could, in principle, act as early warning systems. There are a large number of models, but they can be broken down into two categories:
Models which use a signals approach. Indicators have been isolated from past data. They include such things as the differential between foreign and domestic interest rates on deposits and indices of equity prices. If these indica- tors exhibit unusual behaviour over a two-year period, it is a signal of a risk of crisis.
Models which use regression analysis to predict likely currency crises on the basis of past data. Variables indicating the probabilities of crises include:
high foreign investment rates;
high domestic credit growth;
overvalued exchange rates;
large current account deficits.
Regression models in current use were mainly formulated prior to the 1997 crisis in Asia. When used (retrospectively) to ‘predict’ (statistically) the 1997 crisis (Berg and Patillo, 1999) the results have been mixed. Some models give very poor results: no better than would have been achieved by guesswork, and significantly worse than would have been achieved by an informed observer such as an econo- mist or even a financial journalist! It is highly likely that the models perform so badly because they do not incorporate important variables relating to portfolio behaviour, which is acknowledged to have been critical in determining the currency crises in Asia.
The problem seems to be that many of the important variables relating to financial markets and portfolio behaviour, which can trigger a currency crisis and capital outflow, are difficult to model statistically. Weak banking systems, political insta- bility, social unrest and loss of investor confidence appear to have been significant causal factors but cannot easily be modelled. Of the variables which canbe modelled statistically, three emerge as significant in regressions for ‘predicting’ the capital outflows and crises of the second half of the 1990s:
high domestic credit growth;
overvalued exchange rate;
large current account deficit.
Of course these three are simply evidence of the economic ‘mismanagement’ from which many countries suffer from time to time. Translating them into early warning systems is another matter.
If, as is suspected, recent crises – especially the one in Asia – were triggered by investors withdrawing from shaky companies, it may be that the debt/equity ratios of companies are significant explanatory variables. In the Asian countries, prior to the crisis, the debt/equity ratios in the corporate sector were very high. Debt had to be honoured, irrespective of company profits. A high debt/equity ratio can be a problem. Equity finance, related to profitability, is a better bet for a company in a downturn. The corporate balance-sheet positions left companies in Asia very exposed. Asian economies were vulnerable because their corporate sectors were highly leveraged. Debt/equity ratios were high. There is emerging research in this area which suggests that adding corporate balance-sheet positions to second-generation models substan- tially improves the explanatory power of these models in the case of the Asian crisis. This is what we would expect when investor behaviour is a significant influence on capital flows.
 The literature on early warning systems tends to concentrate on predicting the likeli- hood of capital outflows and currency crises. There is an alternative approach to dealing with the downside of portfolio flows which is to try and establish the funda- mentals of an economy in which crises never occur. This is assumed to constitute a safe or near-safe environment for capital inflows.
Not only does this approach offer a positive guide for policy-makers in countries which wish to attract significant inflows of funds, but it also helps investors by providing a check-list of information needed in order to manage risk. It might even, in this role, help to deter the herding instinct which contributes to volatility.
The fundamentals which go to make up a safe or near-safe environment for capital inflows have been estimated from the positive experience of emerging markets from 1985 to 1998 (Osband and Rijckeghem, 2000). These are those markets which managed
over a 12-month time horizon to avoid a currency crisis. The fundamentals can be listed as follows:
prudentgovernmentborrowingandspending;
highinternationalreserves, measured either as a ratio to short-term debt or as a ratio to imports (the traditional measure);
lowrateofgrowthofprivatedomesticcredit, used as a proxy for the strength of the banking system;
competitive exportsand a healthy currentaccounton the balance of payments;
a competitive,non-overvaluedexchangerate;
highgrowthrateofrealGDP, and of industrialproductionin particular;
a diversifiedexportbasewhich helps the country withstand problems with the terms of trade and other external shocks.
Based on the above, it is possible to identify environments with such strong funda- mentals that they face little or no risk of a currency crash in a 12-month time horizon. Levels that exceed the thresholds are deemed safe or near-safe. The researchers applied the filtering technique to a sample of emerging markets between 1995 and 1998, and were able to classify 47 per cent of the observed environments as safe or near-safe, which seems like good news.
But it means of course that over half of the environments were ‘unsafe’ in terms of selected fundamentals. This may be a fair reflection of the real world or it may be that the thresholds are set too high. Or they could even be too low. For example, if the global financial environment becomes more volatile in the future, the need for foreign exchange reserves will rise. On this basis, an environment which was safe or near- safe in the 1990s may no longer be so. Likewise, if trade begins to falter globally, then an even stronger and more diversified export base may be necessary to create a safe or near-safe environment for foreign investment.
|
Скачать 142.55 Kb.