Artificial Software Agents on Thin Double AuctionMarkets – A Human Trader Experiment

This paper studies how software agents influence the market behavior ofhuman traders. Programmed traders with a passive arbitrage seekingstrategy are introduced in a double auction market experiment with humansubjects in the laboratory. As a treatment variable, the influence ofinformation on the existence of software agents is investigated. We foundthat common knowledge about the presence of software agents triggersmore efficient market prices in the presence of the programmed strategywhereas an effect of the information condition on behavioral variablescould not be observed. Surprisingly, the introduction of software agentsresults in lower market efficiency in the no information treatment whencompared to the baseline treatment without software agents.