Traditionally, the stock market is not the place one would seek evidence of superstition, whatever about other forms of irrational behaviour. One form of unusual repetitive behaviour that academics and practitioners have observed since the 1930s is the general tendency for stock markets to rise on Fridays. However, an article by respected US academics Kolb & Rodriguez in 1987 cast doubt on the immunity of markets from superstition, showing that the US market, examined from the 1930s to the 1980s, seemed to decline on Friday 13ths.

Other work, by equally credible academics and portfolio managers, cast some doubt on this, but were unable to discredit the possibility completely.

Of course, the world is not simply confined to the US, and work on UK markets in 1998 by Mr Andrew Coutts of Sheffield Business School confirmed what other international authors had found for their markets - rather than fall on Friday 13th, stock indices tended to rise, abnormally so.

In the academic debate around these results, it was shown that what was found depended on the precise nature of the statistical tests used, and also on the type of index being investigated. As these differed from study to study no definitive conclusion could be drawn.

This issue may well seem trivial, but the dominant theory of how stock prices move, the so-called "efficient markets hypothesis" - a theory that underlies the decisions of the pension fund managers working for your long-term financial well-being, the asset allocation decisions of the new national pension fund, and the financial mavens who toil over exotic financial products such as options and derivatives to ensure that your bank can offer long-term fixed rates on your mortgage - has no room for and cannot accommodate such regularities.

After all, the markets know when Friday 13th occurs and any long-term regularities (such as stocks showing a tendency to decline on this day) should be taken into account by traders. Eventually, when all recognise its existence and incorporate this into their buying and selling prices, the regularity should, it is argued, be eliminated.

In a set of papers to be published in two journals, the Journal of Economics and Finance in the US and Applied Economic Letters in Britain, I examine world stock market indices in the 1980s and 1990s. Using sophisticated statistical techniques, and on a set of data designed by the Financial Times to allow international comparisons to be made between stock markets, I can show that stock markets have typically risen on Friday 13th, a rise that is greater than the typical Friday rise observed internationally and substantially greater than the average daily rise in stock markets.

The implications of this for fund managers and academics are potentially highly significant. In the papers, I suggest that this regularity, when juxtaposed with other well-known regularities, casts significant doubt on the efficient markets hypothesis. While other theoretical frameworks have been suggested as a basis for financial market activities, these also cannot easily accommodate such empirical regularities. Nothing in existing finance theory can explain what mechanism might drive stock markets to show such rises, especially across national and indeed cultural boundaries.

My conclusion is that human psychology, culture and irrational behaviour - all aspects of reality that finance and economic theories have struggled to incorporate successfully in their models - perhaps offer a better guide to the markets than abstract theories that are based on the non-existence of these issues.

Brian Lucey works in the School of Business Studies at Trinity College Dublin.