Not long ago Capital Chronicle was cited on another website as being Asia-ignorant. There's truth in that statement, but we've been trying to educate ourselves. A part of our study is in the exhibit below:

Exhibit A: Real GDP growth rates in selected Asian countries, 1986-2003

The table sketches a risk profile of the mean-average GDP growth rate. The variance to the mean measure of standard deviation is widely known: skew and kurtosis complement this by describing distribution and probability of the data. How might these measures be interpreted in Exhibit A?

Israeli psychologists' Kahneman and Tversky experiments have become a standard part of behavioural finance teaching. The link above provides background, the third paragraph being all that's needed to follow this article. Consider the choices yourselves to get the idea.

The implication of that third paragraph is that we are risk-averse when it comes to financial gains, but risk-takers when it comes to losses. This latter point suggests the degree of emotional distress suffered does not increase that much more for a big loss versus a smaller one: knowing a loss will happen reduces (up to a point) sensitivity to its magnitude.

If that's the case (and it's contested), investments concentrating more returns above their long term mean are highly appealing. Such a distribution of returns is negatively skewed (a picture example is represented at Exhibit B), or below the perfect "zero" a normal distribution has. That makes the growth patterns of Korea, Indonesia and Malaysia stand out.

Exhibit B: negatively skewed distribution
The downside of these negatively-skewed distributions is that the sub-mean performance, when it comes, is bigger than the more frequent gains. But hey, a loss is a loss - remember?

Adding kurtosis to the mix changes everything. Kurtosis is a measure of the return to be expected above or below what a normal distribution would give. In other words, it's the likelihood of surprise results.

A kurtosis higher than the "3" of a normal distribution means more results around the mean and the tails, with less in between. Lower, and there are more results between the two tails and the mean with less outside these zones. Graphically, low and high kurtosis look like Exhibit C.

Exhibit C: Examples of low and high kurtosis
In the case of Korea, Indonesia and, to a lesser extent, Malaysia this suggests the effects of a negatively skewed distribution will be turbocharged.

That's terrific if the behavioural finance explanation convinces - just hang on for the occasional supermodel-like mood swings.

But find time in the portfolio too for the low kurtosis, kiss-your-sister attractions of lower growth, negatively-skewed but a whole lot less variant patterns of economies such as the UK and the USA.

* China GDP data from the Asia-Pacific Research Center at Stanford's Institute for International Studies (APARC Dispatches, October 2004)
* Kahneman & Tversky information from Peter Bernstein's book Against The Gods
* World Bank data via
* Asian Development Bank data via their Key Indicators 2004 publication (

Bookmark and Share


Related Posts with Thumbnails