Complexity and Sophistication

Add to Calendar
Date and Time: 
Thursday, June 11, 2015
12:00 pm – 1:00 pm
Speaker: 
Abstract: 

The growing recognition that individuals often make poor choices that hurt their economic circumstances has initiated an important public debate about how to tackle the problem. Policy proposals have focused on either equipping individuals with better decision-making skills (i.e., financial education) or using nudges and defaults designed to point individuals in the right direction. A complementary approach that has received less attention is reducing the complexity of decision-making problems. Simplifying may reduce decision-making mistakes, but the remedy may come with a cost: Reducing the number of options, for example, may remove one’s optimal choice or exacerbate overconfidence. To figure out whether simplification is an approach worth pursuing, the first step is to better understand how complexity affects decision-making.   
 
To answer this question, we randomly assigned subjects to simple or complex investment problems. The problems are designed such that any portfolio in the simple condition can be re-created in the complex condition and vice-versa. This feature of the design permits isolating the effects of complexity. We find that complexity leads participants to select investment portfolios with lower return and lower risk and there is suggestive evidence that complexity deteriorates the quality of decision-making. To investigate the effects of complexity on choice avoidance, randomly selected participants could choose between investing and an outside option. Complexity leads to decision-making avoidance among participants with low decision-making abilities, but these participants end up worse off when they have the outside option than whey they are forced to make active investment decisions. Our results suggest that complexity does lead to decision-making mistakes, but that providing a default option as an alternative to avoid complex decision-making may make things even worse.