Economics, particularly financial economics, is not really a science. To be sure, it is methodical and relies heavily on math to give strength to intuition. While some things are trivially true and others, as in behavioural finance, can be made subject to experiment, this does not bridge the macro-micro gap that pervades all social sciences, from sociology to psychology to economics to psephology.

Psephology, of course, is polling. Pollsters attempt to infer from a sample of individuals the behaviour of a whole population. Sometimes they get it right.

Finance has had rather less success in explaining the irrationality of stock markets through the rationality of individual investors shepherded as if by an invisible hand. What’s missing is the data.

Indeed, William Sharpe, whose theories on capital asset pricing won him a Nobel Prize and for a long time underpinned index investing, once said it would take 1,000 years of data to determine the validity of the efficient markets hypothesis.

But we don’t have 1,000 years of data. At best we have 100. We have U.S. stock trading data going back to 1926. Elroy Dimson and his colleagues at the London Business School have international stock data that goes back to 1900 – but not for every market.

Which leads to a dilemma: does the data validate an investment hypothesis, or rather is it basically common sense that does the validation.

This becomes important as the data – such as it is – becomes more consolidated. FTSE is angling after the Russell Indexes. FTSE TMX is expanding its hold on U.K. and Canadian bond data to include European bond data. Bloomberg is buying up UBS’s commodities indexes. But no one has the complete 1,000-year dataset.

Instead, with ETFs, the common sense approach rules: low fees, a rational investment hypothesis, patience. Because there is no science to investing, at least not yet.

Scot Blythe is a Toronto-based financial writer.
Originally published on