The Economist this week has a great article about an academic study looking at the financial performance of companies that appear on the front page of three leading business magazines: Business Week, Fortune and Forbes. The study looked at 549 cover pages, categorising the stories as very positive, very negative, or neutral. The research found that there were many more positive stories than negative, perhaps for fear of legal ramifications, or even because of fear about losses in advertising revenue?
The study found that in the 500 days before the story appeared, companies with a positive story had over-performed the index by 43%, whilst those with a negative story had under-performed by 35% in the 500 days leading up to the date of the story.
However, when the researchers looked at the following 500 days they found those with a negative story over-performed by 12%, and those with a positive story over-performed by 4%. The difference between the two numbers was not significant. What was interesting, however, was that the period of over-performance and under-performance was largely over by the time the cover story was printed.
In a similar vein the article reported analysis presented by Bill Miller, a fund manager at Legg Mason. He showed that the performance of ‘uncommon stock’ tipped by Lehman Brothers sometimes beat the market and sometimes lost, in the 12 months after they were tipped, but almost always beat the market in the 12 months before they were tipped.
The Economist highlighted the problem as ‘recency bias’, the tendency of people to assume that recent trends will continue in the future. This bias is ever present in market research situations. When we scope the research, when we create the questionnaire, when we analyse the results we need to guard against this temptation, particularly in a world where change is happening faster and faster.
The academic research referenced by the Economist was “Are Cover Stories Effective Contrarian Indicators?”, by tom Arnold, John Earl and David North in Financial Analysts Journal, Volume 53, No.2.