Friday, September 14, 2007

Professor Sir Clive Grainger

Last Wednesday night (12th September 2003), I went along to a free lecture given by one of the winners of the 2003 Nobel prize for Economics, the author of 12 books (one of which I used in my under-graduate degree) and over 250 articles and, a person who, by some accounts, is a bigger Welsh hero that King Arthur (I didn’t even know King Arthur was Welsh!).

The man in question is Prof. Sir Clive Granger, and the topic of his presentation on Wednesday night was "trends" (particularly in time series data), although there was quite a bit of discussion about forecasting as well. Trends and forecasting are both related to the art and science of pricing, so I thought I’d use this posting to talk about some of the key message I got out of the lecture…

Firstly, apparently no one has ever defined what a “trend” is, a term that first came into use in 1901. Interestingly, the word “trending” has been used since the 16th or 17th century. One commonly used definition is that “one end is different to the other end”, which would suggest that a flat line is not a trend;

There can be trends in levels (i.e. an upward trend), but there can also be trends in volatility – worth keeping in mind;


You can’t judge a trend – you need to know what’s going on behind it, what’s causing it. Prof. Granger suggested that the trend of Shanghai A-share index between 2002 – 2007 is a bubble, caused by the psychology of the investors


You can’t effectively forecast the economy more than three years in advance

What’s easy to forecast…things like population trends

What’s hard to forecast…exchange rates, stock market indices, anything related to an industry where policy issues matter, commodity prices and anything in a speculative market

What are we getting better at forecasting…demand for electricity (apparently)

No comments: