Experimental Economics by Pablo Branas-Garza & Antonio Cabrales

Author:Pablo Branas-Garza & Antonio Cabrales
Language: eng
Format: epub
Publisher: Palgrave Macmillan

Figure 5.2 Bayesian updates in markets with fixed and flexible prices

Note: Diamonds indicate the asset’s unconditional price. Squares represent the update price after receiving the good/bad signal.

In the fixed price case, if at some moment the number of purchases is higher than the number of sales by two, in spite of receiving a negative signal, the price updated à la Bayes will be higher than the price established by the market maker (50). This motivates the agent to buy, affecting by his behavior all those who act after him. Using the same argument, a trade imbalance of –2 makes agents, who still have not made their choice, sell – independently of their private information. When the price is flexible, the market price after two purchases is 84.48. If the third agent receives a positive signal, he will buy because by applying Bayes’ rule with his private information it will lead him to calculate an expected value of 92.7 for the assets (as above-mentioned, the third agent will sell if his signal is negative because his updated value will be 70, which is lower than the market price offered by the market creator, 84.48).

Experimental evidence of herds in financial markets: fixed and flexible prices

Empirically, it is difficult to test herd models, because there is no data on the private information of agents. In addition, it is difficult to know if agents react differently depending on whether the exchanged asset’s price is fixed or flexible. This can be done in the laboratory, where subjects can receive private information on the value of the asset and observe the history of exchanges.

Cipriani and Guarino (2005) used 216 subjects to test differences in herd behavior depending on prices. Their experiment consisted of ten rounds, in each of which 12 subjects had to make their choices sequentially. Before every round, an experimenter would toss a coin to determine whether the value of the asset was going to be zero or 100. The participants did not know the result of the coin toss, but they received an informative signal. There were two bags in the experiment. One contained 30 blue tokens and 70 white tokens; the other contained the opposite. The bags were identical. Each subject, when asked to make his choice, had to draw a token from the bags. If the coin toss resulted in heads (tails), the participant will draw a token from the first (second) bag, so that the color of the token was an informative signal on the asset’s value in that round.

Each participant, after privately observing the color of the token, was to return the token to the sack and to announce out loud if he wanted to buy, to sell or to not do anything with the assets. The price of exchange was determined by the asset’s price, fixed on the board. Also, the decisions and prices that had taken place so far were registered on the board, so that participants were not only receiving their private signal, but they were also obtaining information about the history of transactions and prices.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.