Depending on conventional theory to predict markets.
“Modern Portfolio Theory (MPT) is a misnomer since MPT is almost six decades old. The idea is even older - trying to offset risks without reducing returns. Every business in the same economy has common exposures to macro-economic risks like policy risk and national fundamentals (deficits, consumption and investment patterns).
There are also industry-specific risks. The fortunes of the finance industry is dependent on interest rates, cars on the cost of metals, power on fuel supplies and so on. Finally, every business has company-specific risks. The CEO could fall ill, for example, as in Apple.
Industry-specific risks may be opposed or non-correlated. IT and autos have non-correlated industry risks. Coal producers and power generators have opposed risks with negative correlations, which is a good reason for integrating energy value chains. Similar businesses tend to have highly correlated risks, with the differences due to company-specific factors. Correlations and covariances are reflected in price. If a set of assets with low or negative correlations are held, overall risk is reduced, without lowering overall returns. The effectiveness of such diversification depends on the respective weights and the covariances of assets.
Keynes instinctively used the concept in managing the King's College portfolio. John Templeton took it further by diversifying across national economies.
“Keynes and Templeton diversified via common sense and empirical inspection. In the 1950s and 1960s, researchers such as Markowitz and Sharpe worked out methods of calculating what is called the “Efficient Frontier” for portfolios.” A given portfolio includes assets held in varied weights. A given weight may be zero, or negative if it's short-sold. The expected return for an expected risk can be worked out by juggling weights. The Efficient Frontier is the line along which the maximum expected return can be derived for given levels of expected risk. As expected risk rises, the expected return should rise.
MPT uses historical returns and it's based on many idealised assumptions. One is that future returns and variance of returns will remain close to historical returns. A second assumption is that historical correlations between assets won't change much.
Originally, risk was represented by simple standard deviation (SD) of averaged return. A high SD was considered risky. SD doesn't distinguish between upside or downside. So, a higher-than-average return is treated to be as undesirable as a lower-than-average return.
There's been much tweaking and many different ways of massaging variances have developed. Nobody takes MPT and Efficient Frontier calculations as gospel. But it's an useful way of looking at assets.
Some generalised insights can arise. If all assets in an economy show high correlation, the economy is likely to be over-regulated. If, at a given time, the Efficient Frontier is reached only by making the portfolio highly overweight in very few assets, or many short-sales are required to reach the Efficient Frontier, the economy is not healthy.
India has always had high correlations between assets. This reflects the License Raj. There are few negatively correlated assets. A simple example. In most economies, primary energy producers (oil, gas and coal) yield good returns during periods of high inflation. They tend to be negatively correlated to the overall economy. The negative correlation offers a hedge against inflation.
India's petro-subsidy policy destroys that hedge (while failing to combat inflation). Similar situations arise in commodities like sugar and cement.If and when further deregulation occurs, some correlations should drop and eventually, negative correlations should become visible. Overweight situations can indicate bubbles. The efficient Frontier line in 1999-2000 could be hit only by over-weighting of ICE (IT, Communication, Entertainment-media). In 2007-08, the Efficient Frontier line showed the overweighting of real estate and infrastructure. Both times, the overweight assets were in bubbles and underperformed subsequently.
If the Efficient Frontier line also implies the short-selling of many assets, the diagnosis is a well-developed bear market. Similarly a period of reversal out of a bear market into a more bullish phase seems to be indicated by short-sells disappearing from the Efficient Frontier line. This pattern of implied short-selling showed up in 2002-03 and gradually disappeared in 2004-05. It was visible again in 2008 and it's visible now. Such an analysis is theoretical and it involves fiddling with lots of data and making unrealistic assumptions like widespread short-selling, which cannot translate into actual portfolio construction. But there is some value to it. Even where negative correlations are not available, low correlations provide defensive strength and reduce risks. The indicators for bubbles and cyclical change can also certainly help any investor.
