The theory of minimum variance

With five central banks meeting this week in the region and none had eased policy or expected to do so – Bank of Thailand met yesterday, Bank of Korea, Bank Negara Malaysia, Bank Indonesia and Bank of Japan meeting today, how is it that we are getting a risk rally in the region? Especially given no fresh insights from the Fed’s FOMC minutes released last night.

Well, on days like this, let’s go back to basics – theory. Let me bring up the theory of minimum variance. It’s interesting to read about it in an energy product brochure I came across recently which mentioned the application of the theory to animal psychology and from there, relate back to us, humans, and thereby, the behaviour of us as market participants and our guardian angels, the central banks. So what does the theory suggests for animal psychology? Here’s what it is in a nutshell.

Minimum variance theory on animal psychology

“If you toss a frog into boiling water, it will instinctively jump out because the change is so drastic and painful. However, if you placed that same frog into a pot of lukewarm water, and then gradually increase the temperature of the water one degree at a time, the frog will boil to death without moving a muscle. Basically, animals including human beings are terrible at detecting subtle changes in their environment.”

So that I believe is what Ben Bernanke and folks are doing to us in the markets. Are we in a slow cooker, boiling to death while the Fed gradually turns on the heat and siphons out the QE?

Now think of the opposite case – PBOC’s sudden clamp down last month. Did we not all feel the sudden pain and quickly abandoned all ships and swam for our lives? So did more survive in this case and are raring to jump back on the bandwagon? So that may be what’s  happening to us in the Asian markets right now. Back to our happy days as our central banks are all in a Mexican stand-off with the Fed.


The theory of minimum variance unbiased estimation (MVUE) was first introduced in 1963 as a statistical technique. Given the condition that the theory can be applied only to unbiased samples, its practicality was thought to be limited and soon we have other techniques such as the Maximum Likelihood (MLE) method. The often used Markowitz model is probably the most well-known use of the minimum variance portfolio theory in the world of finance.

In my little spare time this morning, I also discovered the minimum variance theory being applied in geography – how river meanders – and neuroscience to explain eye and arm movements. For those interested, I’ve extracted the introduction on the latter since us in the markets, we’re probably more interested in our eye and arm movements than the meandering of rivers in the outback of USA.

“One of the most puzzling issues in neuroscience is why neurons employ stochastic rather than deterministic signals to process information. There are many different approaches to tackle the problem such as stochastic resonance, correlated computation and others. Recently Harris and Wolpert proposed a theory, the minimum-variance theory, as an example of optimal control models of movement, aiming to provide a plausible explanation of ‘noise calculation’ in the cortex. The idea of the theory is quite straightforward: to find a noise signal which minimizes the variance of certain quantities, say the arm movement trajectory.”

Alright, I’m back to investigating a messy case of biomass mess. Have fun, everyone!