Consumer preferences, Bayesian statistics, and contract design: three essays in applied microeconomics

Thesis

Chapter 1

In this chapter, I execute and evaluate a novel elicitation technique for consumer preferences, using data from a survey I conducted in Bangladesh. Prompting binary comparisons of two bundles of various types of goods at a time, I illustrate that the preferences over 51 different goods expressed by a population of microfinance recipients can be approximated with the output of a low number of salient survey prompts. Processing the resulting rich dataset, I present easily applicable reduced-form analysis that verifies the empirical viability of bundle comparison surveys for preference elicitation and I provide actionable guidance on how to achieve optimal results through such surveys.

Chapter 2

Elicitation of consumer preferences is of significant importance in many applied microeconomics settings but it is challenging to execute empirically without relying on strong assumptions or incurring prohibitive surveying cost. In this chapter, I present an alternative method, driven by Gaussian Processes, that allows preferences elicitation across a high-dimensional universe of 51 goods. First, I show estimation approach on a illustrative simulation – I also showcase performance in comparison to conventional MLE approaches. Then, I present estimation outcomes based on empirical data collected through preference data of 500 Bangladeshi respondents. I conclude by highlighting familiar Cobb-Douglas preference patterns which this method was able to reconstruct from survey data alone.

Chapter 3

I compare debt-based and equity-based microfinance contracts in an incentivised lab-in-the-field experiment with Bangladeshi freelancers. Building on the literature that stresses the need for understanding agent heterogeneity to predict microfinance performance, I show that equity-based contracts overperform debt-based instruments in an environment that requires partner matching with noisy ability signals. Using structural estimation, I demonstrate that microequity contracts enabled participants to respond to exogenously added signal noise by adjusting the rules of thumb they used towards behaviour more in line with a rational expectations framework – while conventional microfinance contracts did not.

Authors: Schaffner, N

• DOI: 10.5287/ora-ppovnobkm
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English