Social Influences on Predictions in Simulated Financial Markets

Sammanfattning: During the last decades the interest in financial markets has increased enormously around the world. Since investors are economically dependent on their investments, it is essential fo them to make accurate investment predictions. One observed behaviour among stock investors in actual markets is that they tend to choose the same stocks. Such behaviour is referred to as herding when investors are imitating each other. This thesis examines whether different bonus systems or consistency among investors has an impact on the level of herding. In Study I undergraduates who were recruited as participants were in two experiments requested to predict an “upmarket” or “downmarket” conditional on diagnostic information and on the predictions ostensibly made by three others. Economic incentives for making the same predictions as either the majority or the minority of the others were introduced. In Study II other undergraduates who participated in two experiments predicted the future price of a stock given the current price and the predictions made by five fictitious others. Consistency was varied as variance within and between the others’ predictions. The existence of herding was supported in both studies. Financial rewards for following a majority or minority led however to asymmetrical results in Study I: the participants in the majority-bonus condition more frequently followed the majority, but no such effect was observed in the condition with a bonus for following the minority. The results of Study II showed that the level of following the herd increased with the number of the others’ predictions that were correlated