Christoffer Jessen Weissert forsvarer sin ph.d.-afhandling

Christoffer Jessen Weissert forsvarer sin ph.d.-afhandling: "Essays in Macroeconomics: Inflation Inequality, Consumption Behavior and Impulse Responses".

Ph.d.-forsvaret foregår i CSS 26.2.21.

En elektronisk kopi af afhandlingen kan fås ved henvendelse til: charlotte.jespersen@econ.ku.dk

Bedømmelsesudvalg

  • Lektor Tomas Høgholm Jørgensen , Økonomisk Institut, Københavns Universitet, Danmark (formand)
  • Assistant Professor Jessie Handbury, University of Pennsylvania, US
  • Assistant Professor Mikkel Plagborg-Møller, Princeton University, US

Abstract

The dissertation consists of three self-contained chapters on topics within macroeconomics. All chapters share an orientation towards empirical analysis of macroeconomic questions. New methods for empirical analysis are presented in the first and third chapter. The method presented in the first chapter contributes to analyses of inflation inequality. The method presented in the third chapter contributes to analyses on how the economy responds to shocks. The inflation inequality discussed in the first chapter relates to households' consumption behavior. In the first and second chapter specific empirical analyses on US households' consumption behavior is conducted.
Chapter 1 / A Nonhomothetic Price Index and Inflation Heterogeneity
with Phillip Hochmuth & Markus Pettersson
In this chapter, we look at a classic and central topic in macroeconomics: inflation. Specifically, we look at inflation at the household level. We derive a microfounded, nonhomothetic generalization of all known superlative price indices, including the Fisher, the Törnqvist, and the Sato-Vartia indices. The index largely avoids the need for estimation, aggregates consistently across heterogeneous households, and admits different index weights across the expenditure distribution. The latter property rationalizes the methods used in most previous measurements of inflation inequality. In an empirical application to the United States using CEX-CPI data for the period 1995 to 2020, we find: (i) poor and rich households experience on average the same inflation rate; but (ii) inflation for the poorest decile is more than 2.5 times as volatile as that of the richest decile; and (iii) this higher volatility primarily stems from a larger exposure to price changes in food, gas and utilities. In these findings, substitution between goods as prices change plays only a second-order role. Instead, almost all differences come from mechanical changes in the cost of different base-period reference baskets.
Chapter 2 / Quality and Consumption Basket Heterogeneity
with Rasmus B. Larsen
In this chapter, we look at another core topic within macroeconomics: consumption behavior. We investigate what households put into their shopping carts and study in detail the differences across the income distribution. We study how the quality of households' consumption baskets varies with income using detailed household-level panel data on purchases. By exploiting the randomized disbursement timing of the Economic Stimulus Payments of 2008, we show that households increased spending when receiving the payment and spent more money on goods of higher quality. While the spending effects are concentrated among low-income households, middle-income households drive the quality effects. These findings support the theory of nonhomothetic demand. To model this, we embed nonhomothetic preferences over quantity and quality in an otherwise standard buffer-stock model. Contrary to the standard model, the nonhomothetic model can be used to match that the marginal propensity to spend is decreasing in income. Moreover, the calibrated model implies that households trade up in the quality of consumption when receiving a transitory income payment. Compared to the standard model, our nonhomothetic model also generates a more unequal wealth distribution, which is closer to the data.
Chapter 3 / Local projections or VARs? A data-driven selection rule for finite-sample estimation of impulse responses
with Anders F. Kronborg
In this chapter we look at how one in empirical analyses estimates impulse responses. We derive a data-driven rule for when to choose local projection, vector autoregression (VAR) or a mix of these two methods’ estimates of impulse responses in finite samples. We show that local projections and VARs are linked in finite samples: local projections can be derived from a first-step estimate of the VAR impulse responses and a second-step linear correction of the forecast errors of the VAR. The sum of these two estimates yields the local projection estimate and the second-step estimate therefore captures the difference between local projections and VARs. We coin the difference between local projections and VARs the local projections contribution to VARs and since this difference is estimated we can perform inference on it. The selection rule is based on this inference and depends on a simple key statistic for the local projection contribution: the coefficient of variation. When the coefficient of variation is large, the local projection contribution is imprecisely estimated and more weight is put on the VAR estimate. When the coefficient of variation is small more weight is put on the local projection estimate. We show that the selection rule performs well in a host of Monte Carlo studies.