Nikolaj Moll Lund defends his PhD thesis at the Department of Economics

Candidate:

Nikolaj Moll Lund, Department of Economics, University of Copenhagen

Title:

Essays in Structural Demand Estimation

Supervisors:

• Bertel Schjerning, Professor, Department of Economics, University of Copenhagen
• Mogens Fosgerau, Professor, Department of Economics, University of Copenhagen

Assessment Committee:

• Thomas Høgholm Jørgensen, Associate Professor, Department of Economics, University of Copenhagen
• Pedro Mira, Professor, Department of Economics, CEMFI
• Helena Perrone, Associate Professor, Department of Economics and Finance, TBS Business School 

Summary:

This thesis consists of three self-contained chapters. The first chapter – co-authored with Anders Munk-Nielsen and Bertel Schjerning – proposes a new estimator, Direct Dynamic Regression, for single-agent dynamic discrete choice models – a workhorse model within economics when consumer demand has a dynamic component. The estimator relies on the fact that the logarithm of the denominator of the conditional choice probabilities and the integrated value functions are equivalent within the considered model class. The key contribution of this estimator is its simplicity and ease of implementation, since both structural parameters and integrated value functions can be estimated jointly by a least squares procedure at a modest efficiency loss. The chapter compares the proposed estimator against existing alternatives in a Monte Carlo simulation and on empirical data, highlighting benefits and limitations, and suggests extensions.

The second chapter – co-authored with Frederik Brandt and Jonas Slaathaug Hansen – considers the empirical properties of a newly proposed demand model, the Inverse Product Differentiation Logit model, against the current state-of-the-art models in the literature, by comparing them in an empirical application to the Danish car market for new cars. Our analysis shows that the Inverse Product Differentiation Logit provide flexible, but differing substitution patterns in comparison to competing models at great computational gains. We then apply these models to evaluate a major reform of the car taxation system in Denmark, demonstrating that these differences in substitution patterns lead to divergent counterfactual predictions.

The third chapter – co-authored with Maximilian Blesch, Kenneth Gillingham, Jonas Slaathaug Hansen, Fedor Iskhakov, Anders Munk-Nielsen, John Rust, and Bertel Schjerning – considers the interplay between the new and used car markets through a structural model incorporating both forward-looking consumers and endogenous equilibrium prices. By modeling how consumers interact across both markets, we explore how the price-setting of firms is affected by the presence of the secondary market. A key finding in this chapter is that the price sensitivity of consumers, that is, how strongly a consumer’s demand responds to a change in a given new car price, is closely connected to how the consumer responds to prices in the used car market. This directly affects the pricing power of car-producing firms, and therefore has implications for how such markets should be regulated.

An electronic copy of the dissertation can be requested here: lema@econ.ku.dk