An investigation of investors’ estimates of returns earned and the effect of anchoring on these estimations

Abstract

This study examines overconfidence and anchoring in an investing environment to determine if there is evidence of the phenomenon amongst a sample of academics at participating universities.A survey was sent out to over 6,000 staff members at four South African universities assessing respondents’ estimations of their returns earned in unit trusts in which they were invested, as well as assessing whether they would adjust their estimate when presented with an anchor – the relevant Johannesburg Stock Exchange (JSE) All Share Index return. A total of 466 completed responses were obtained, of which 81 respondents indicated that they were invested directly in a South African equity unit trust to allow for statistical testing. The data obtained were analysed for evidence of overconfidence and anchoring by comparing respondents’ estimates of fund returns against historical returns and then checking whether they adjusted their estimate after being presented with an anchor.The study found that investors were under-confident rather than overconfident. Furthermore, it was found that older respondents were better able to estimate their past returns than younger respondents. The presence of an anchor appeared to have no effect on respondents’ estimates.