Fragility of Credit Markets

Project Start:07/2014
Researchers:Sascha Baghestanian, Baptiste Massenot
Category: Macro Finance, Experiment Center
Funded by:LOEWE

Topic and Objectives

Can the recent extreme events in financial markets be explained by individual irrational behavior? Economists do not have a definite answer to this question. Some economists argue that behavioral biases play an important role in financial booms and busts. Others say that competitive markets offset these biases. Irrational investors should, indeed, either be driven out of the market as their wealth evaporates or be outweighed by the more rational participants.

In order to study whether markets are successful in offsetting individual behavioral biases, we compare the outcomes of experimental financial markets in a standard market economy to a more artificial “island” economy. In the market economy, subjects compete against each other for credit which they can use to invest in a risky project. The equilibrium interest rate ensures that all available credit is allocated. In the island economy, the environment is identical except for market interactions. Instead, subjects report their demand for credit and only receive some if their bid is higher than the realization of a random variable. The allocation mechanism is thus independent from the other participants on the island while it is based on the decisions of the other participants in the market. The comparison of aggregate outcomes between these two environments tells us about the causal impact of an essential feature of competitive markets.

Key Findings

  • There are substantial differences between market and island outcomes. Market prices display a bubble pattern. They first increase above the fundamental value of the project and then typically crash towards the end of the session. By contrast, average prices across islands closely track the (constant) fundamental value of the project. These results do not support the claim that competitive markets correct individual biases, rather the contrary.
  • We also find that gambling for resurrection is an important factor to explain these patterns. Empirically, average losses are positively correlated with subsequent prices to a significant extent.
  • Motivated by this evidence, we show that a simple model of investment in which investors suffer from a desire to gamble for resurrection leads to implications consistent with the price patterns we observe. When subjects suffer initial losses on average, a desire to make up for these losses induces them to take even more risk which further increases prices. As a result, further losses accumulate and so does the desire to resurrect. As their wealth evaporates, subjects become more and more constrained and have to decrease their demand at some point. Prices start decreasing as a result.
  • Gambling for resurrection also nuances the claim that irrational investors will be driven out of the market as they accumulate losses. This claim implicitly assumes that individuals are either always rational or always irrational. However, while investors may have different fixed preferences, the behavioral biases we uncover depend on prior outcomes that are independently distributed across subjects. This implies that biases are to some extent random and that the same individuals might be more or less rational in different periods. Thus, it may take some time before irrational behavior completely disappears from the market, if ever.

Related Working Papers

No.Author/sTitleYearProgram AreaKeywords
104Sascha Baghestanian, Baptiste MassenotCredit Cycles: Experimental Evidence2015 Macro Finance, Experiment Center -