|
Article Excerpt
Abstract This paper examines the market for advice and the underlying perception that advice is useful and informative. We do this by first providing a theoretical examination of the informational content of advice and then by setting up a series of experimental markets where this advice is sold. In these markets we provide bidders with a demographic profile of the "experts" offering advice.
The results of our experiment generate several interesting findings. The raw bid data suggest that subjects bid significantly more for data than they do for advice. Second, in the market for advice there appears to be no consensus as to who are the best advisors although on average economists demand the highest mean price and women suffer a discount. In addition, we find that whether a subject suffers from a representativeness bias in the way he or she processes data has an impact on how he or she bids for advice and on his or her willingness to follow it once offered. Finally, we find that on average people impute a low level of informativeness onto advice, consistent with their bidding behavior for data versus advice.
Keywords Advice * Decision making * Risk aversion * Experiments
JEL Classification Numbers D81 * G11 * C91
1 Introduction
It is commonly thought that a picture is worth a thousand words. If that is so, one might ask how much data is a piece of advice worth. In other words, if advice is important then we should be able to measure it in two ways: how much data would a rational decision maker be willing to give up in order to receive a piece of advice from a person who has just engaged in the decision problem he or she is about to engage in or, alternatively, how much would that person be willing to pay for such advice from a person with a given set of characteristics.
The fact that we expect that people will bid different amounts for advice from different types of people implies that in the "market for advice" certain types of people are likely to fare better than others. Such markets for advice and influence might function in a number of ways. Under one scenario there may be a perception that certain people or types of people are worth listening to. These perceptions amount to broad stereotypes that may bestow huge rents on some of the agents in the market.
Such stereotypes, if they persist, can lead to what we will call "perception rents", i.e. amounts paid for the advice of agents in excess of the expected informational content contained in their opinion. If such rents are substantial, they present us with a potentially large inefficiency. An alternative to perception rents is what we will call the "chauvinistic bias". Here people tend to believe that advice from people like themselves is the best and hence tend to bid higher amounts for people with characteristics like theirs whether or not those types give the best advice.
A related question deals with the impact of what are sometimes called "representativeness" and "conservative" biases, relative to Bayesian updating, and their implications for the process of advice giving and following. For example, when updating one's beliefs, a rational Bayesian decision maker is expected to place a certain amount of weight on his previous prior (or the base rate) and a certain amount on new information (the sample) as it arrives. How much weight is placed on the new information depends on the strength of his or her prior. If a decision maker places more than the Bayes-optimal weight on the prior (or base rate) he or she is called "conservative" while if excessive weight is placed on the sample he or she is considered to be subject to the "representative" bias, thinking, in the limit, that the sample received is in some sense representative of the population from which it was drawn. Such people fail to take base rates or priors sufficiently into account.
These concerns have wide ranging implications for our research on advice giving and following. For example, if we could measure the degree to which a decision maker is subject to one of the biases discussed above, could we correlate that characteristic to the decision maker's willingness to pay for and follow advice. More precisely, if conservatives are reluctant to update their priors on the basis of new information, are they therefore less inclined to pay for advice and also follow it once it is given? Also who are more persuadable, conservatives or representatives? All of these concerns can be summarized under the name of "advice bias".
In this paper we study an experimental market for advice in an attempt to measure both the informational content of advice and the market for it. To do this we create a set of "experts" by having some of our subjects get experience playing a simple 2 x 2 "game against nature" a large number of times. These experts are surveyed to obtain information about their gender, GPA, major, and year in school. Advice is elicited from these experts and is then sold to a new set of subject "clients" who play the game once and only once. The prices generated by this market for advice furnish us with an opportunity to measure potential perception rents in the market. In addition, by observing the decisions made by the client subjects, we are able to measure whether these perception rents are wasteful or not. By observing the way subjects update their priors in the experiment, we are able to categorize them along the conservative-representative spectrum according to how much weight they put on their prior and how much they place on the data sample they observe. We are then able to correlate their degree of conservatism with their behavior in the advice-following experiment. For example, are economists or scientists more likely to be Bayesians and hence process data correctly? If so, does that explain their increased value in the market. Do subjects impute a degree of representativeness for their advisors that is greater than their own, etc? Finally, the data generated by our experiment allow us to calculate what the informational content of advice is by imputing how many observations a subject would be willing to give up in order to receive a piece of advice.
The results of our experiment generate several interesting findings. First we find that, in general, our client subjects bid for data amounts that, on average, are approximately equal to the expected values of the information they might expect to receive. The raw bid data suggest that subjects bid significantly more for data than they do for either advice or beliefs. We also find a little evidence for perception rents for economics majors and a certain amount of support for what we call the "chauvinistic bias," meaning that subjects tended to bid more for advice from people sharing the same major as themselves than for people of other majors.
In answering these questions we find that the way a person processes data, i.e. how much he suffers from a base-rate bias affects his or her behavior in the market for advice quite dramatically. For example, if subjects place more weight on data than they should under perfectly rational Bayesian updating they tend to bid more for information in the form of data, advice or beliefs than those who place less than a perfect Bayesian weight on data. Finally, we find that people tend to impute a higher degree of base rate bias to others, i.e. those who are giving them advice, than they themselves have. This tends to make advice worthwhile since it gives an advisee more insight into the sample of observations than he would get if his expert were, say, a person who processes data as he does. In other words, he would rather get advice from an expert who was different than himself than one who was the same.
The rest of the paper is organized as follows. In Section 2 we will describe the experiment run to investigate the questions raised above. In Section 3, present the theory of decision making with advice that will be used as our guide in analyzing the data. In Section 4 we present the results of our experiments. Section 5 contains a brief review of some related literature, and conclusions are contained in section 6.
2 Experimental design
2.1 Experimental overview
In this paper we report the results of an experiment that involves, among other things, subjects playing an "Investment" game, where in each period a subject must choose to invest or play it safe and keep his money in a secure asset. This game is first run on a set of subjects in order to create a pool of "experts" who will be used to give advice to subjects ("clients") who later play the game.
We then run a game where we create "clients" or "advisees." We run two variants of these. In the first, the Price-Elicitation game, we auction off (1) the different types of advice from different types of advisors (those with different demographics, gender, academic major, GPA, etc.). In the second, the Belief-Elicitation game, after a subject arrives in the lab we read them instructions and then elicit their prior belief about the state of nature...
|
