Studies of framing, judgment and choice

Abstract: This dissertation consists of four essays. Essay 1: Studies of Framing, Judgement and Choice. Essays 2-4, the main essays in this dissertation, present studies within the research domain of economic psychology. The focus in these three essays is to investigate how people form their judgements and their choices in different contexts. In order to position these three essays in their research context an introduction to economic psychology as a research field is presented in Essay 1. In this essay the experiment as a research technique and its advantage and disadvantage are also discussed. Essay 2: The Utility and the Curse of Winning an Auction – An Experimental Analysis. A general assumption in auction theory is that overbidding in auctions occurs because bidders do not adjust their bids to the information asymmetry in the auction market. Two auction experiments were conducted in order to investigate whether other factors, such as the subjects’ desire to win an item, or prevailing market conditions can explain overbidding. The experiments were designed as an emulation of a Treasury securities auction. Students in business administration and professional actors in the Swedish money market participated in the experiments. Furthermore, by offering an award to the best end-of auction performer(s) the subjects were also encouraged to fulfill their task to the best of their ability. In one experiment the bidders were offered a notably high award to the best performer. In another experiment bidders were offered a moderate award. The bids were on average submitted close to to the calculated expected value of the items. The mean bids in the two auction types were not significantly different indicating no effect on the general bidding due to the size of the award. Two main strategies were distinguished in this study indicating that the auction institution attracted bidders differently. The bidders who purchased no items in the auctions accommodated early to the strategy close to the items’ calculated expected values thus allowed them to optimize their financial outcome from the auctions. The bidders who purchased one or more items used a strategy that allowed for some temporay ‘controlled’ gambling. In the competitive auction context the bidders who intended to buy an item consistently had to submit bids in excess of the calculated value. Otherwise their bidding was moderate and in line with that of the nonbuyers. There was a tendency among the buyers to express a higher interest in financial risk taking and arousal-stimulating activities than among nonbuyers in the auction experiments. In real-world auctions risk taking may be less controllable and manageable and the effects from excessive bidding may be less beneficial to the winning bidders. Essay 3: Who is Cursed in a Tresury Securities Auction? – An Empirical Analysis of the Swedish 6-month Treasury Bill Auctions from 1991 to 1994. This study investigated whether information is perfectly distributed in the Swedish 6-month Treasury bill auctions or whether any private auction information systematically affected the price system and thereby the efficiency in the distribution of the issued securities in the auction market. It was furthermore investigated whether the prevailing auction rules generated strategic trading among bidders. The study covers a period from January, 1991, to June, 1994. With a time series technique, the auction variables were each divided into an anticipated and an unanticipated component from the variable’s past history as derived from an established time series model. The anticipated component shows the publicly known information about the value of a security while the unanticipated component shows information that is new to the market for instance information about recent deals and liquidity needs. Regression analyses showed that the winning bidders succeded in purchasing the Treasury bills in the auctions at prices that were, on average, lower than that of an identical bill in the secondary market. Furthermore, results from the regressions indicate that some privileged bidders traded strategically in the auction market by withholding private information and utilized their information so as to gain a more advantageous position in the auctions. The stable bidding environment provided the bidders in the Swedish 6-month Treasury bill auctions regular and correct feedback. This in turn may explain the bidders’ efficiency in selecting and implementing their bidding startegies and to act rationally and adaptively. In summary, signs of price manipulation indicates that the Swedish 6-month Treasury bill auctions are not fully efficient. Essay 4: Is Confidence in One’s Belief a Matter of Knowledge or of Cognitive Illusion. In many studies on judgement and choice it has been found that people have a tendency to be overconfident in their decisions. Overconfidence in experiments is defined as the general tendency to overestimate the likelihood of having selected the right alternative among a set of avilable answers. In order to investigate wether the selection method is the key to creating as well as eliminating overconfidence two types of two answer-alternative questions were employed in the present study, namely informally selected and representatively selected items. Moreover, to investigate the extent to which perceived difficulty and familiarity could explain confidence level the subjects were also asked to report whether they perceived an item as difficult or familiar. 102 university students participated in two confidence experiments. It was found that a representative selection of items decreased the subjects’ overconfidence but at the cost of poor resolution. Moreover, the results showed increased ‘realism’ when the subjects were asked to make confidence assessments. Also, a reversal of the hard-easy effect was found for perceived difficulty. The results showed that the confidence level could be explained to a significant extent by the subjects’ perception of a question as either difficult or familiar. The subjects tended to display decreased confidence when items were perceived as difficult or very unfamiliar and they displayed increased confidence in the opposite condition. The results in the present study demonstrate the importance of making a distinction between objective and perceived difficulty as determinants of the subjects’ confidence level.