key: cord-0782894-4f6glbvn
authors: Lee, Sang Yoon (Tim)
title: The Political Economy of Early COVID-19 Interventions in U.S. States: Comment
date: 2022-01-15
journal: J Econ Dyn Control
DOI: 10.1016/j.jedc.2022.104304
sha: 20e47d443f7bf82c234af2ffa60f8a4d07fbf4d7
doc_id: 782894
cord_uid: 4f6glbvn

nan

investigates whether the stringency of lockdown measures across U.S. states were affected by a state governor's party affiliation and re-election concerns. Importantly, the outcome they are concerned with is the policy itself, and little attention is paid to whether a policy in fact succeeded in reducing deaths or economic losses. So in some sense, Gonzalez-Eiras and Niepelt (2021) is less about Covid-19 than understanding how political concerns, rather than economic or public health concerns, lead to different kinds of policies being implemented.

Unsurprisingly to anyone familiar with American politics, they find causal evidence that Republican governors implemented shorter and less stringent lockdowns. More interesting is their finding that governors facing re-election implemented longer and more stringent lockdowns, even controlling for several other factors including a governor's party affiliation. R and D stand for Republican and Democrat, respectively. The second bar also includes the governor of Puerto Rico, who belongs to neither party. "Reelection" are the subsample of governors who were seeking re-election in an early election to be held in 2020 or 2021, who also imposed some form of a lockdown. "Early election" are governors who were not seeking re-election in an upcoming early election, who also imposed a lockdown. The "control" group are governors not facing early election among those who imposed a lockdown.

These are both interesting and important findings in and of themselves. They also find that other usual suspects, such as divided governments or the gender of the governor, matter little. The rest of the paper focuses on a particular interpretation of their results, which are one of the three caveats I discuss here.

The main outcome variables of interest are whether or not a lockdown was announced, the duration of a lockdown, and a stringency index derived from the "Oxford Covid-19 Government Response Tracker" (Hale et al., 2021) . The authors do a careful job checking that their results are consistent across all measures, and their extended analyses are very convincing.

Still, robustness of the outcome variable as well as potential omitted variable bias is a real concern, especially given the rather small sample size. In Figure 1 , we see that re-election concerns are identified by less than 10 out of 52 observations (45 if states which did not implement a lockdown are dropped). So potentially, the results can be sensitive to using a different outcome variable (e.g. the "Stay-At-Home" index developed in Baek et al. (2021) ) 1 , or the inclusion of additional independent variables (economic structure as in Aum et al. (2020) ; Lee et al. (2021) , public health variable such as share of population with health insurance, or something com-pletely unrelated such as whether the state allows capital punishment).

The small sample size restricts further, potentially interesting analyses: Given the large gap between Republican and Democrat governors' responses, we may also expect them to behave differently in the face of re-election. While possible in theory, unfortunately it would be hard to ascertain statistical power by comparing less than 5 individuals to each other. (2021) present a model that features a not-necessarily-benevolent policymaker who realizes that economic activity can lead to more infections according to a standard SIR model, and needs to decide on how long and severe economic restrictions should be. The framework is simple and transparent, and motivates the empirical regressions that follow. The framework also makes it clear how environmental variables such as the electorate's ideology and urbanization should enter the decision problem.

However, the analysis stops short of making predictions on how the policymaker's decisions should depend on any of the underlying variables. So while the theory is useful for shaping how we should think independent variables affect a policymaker's decision, a theoretical explanation for why Republicans avoid restrictions and why re-election concerns lead to stronger restrictions is absent. So in some sense, how we interpret the coefficients is left to the reader.

The authors invoke existing political economy theories-not modeled in their own theoryto give an interpretation, but the link is not obvious. In fact the results do not necessarily imply political motivations at all: For example, to the extent that the chosen policy deviates from a benevolent planner's, it could be that the policymaker is indeed benevolent, but is misinformed of the underlying parameters of the economy and epidemiological environment (e.g., Republicans or governors facing re-election may have a better understanding of the environment).

The authors interpret the fact that re-election concerns led to stronger restrictions as evidence that these governors cared more about signaling competence than generating campaign contributions. This is based on the presumptions that i) voters care more about public health than the economy when evaluating competence, and ii) less restrictions lead to more campaign contributions.

But there is no clear theoretical reason nor empirical evidence for either assumption. Not imposing restrictions might as well be viewed as competent for some (more Republican) electorates, and there may well be wealthy (presumably Democrat) donors who desired more restrictions and would thus raise their campaign contributions toward politicians who implemented stricter measures.

At a more fundamental level, there is little reason to believe that imposing stronger restrictions implies a concern for lives (public health) over livelihoods (the economy). Since people voluntarily adjust their activity even in the absence of restrictions, less restrictions could reflect a concern for non-Covid medical emergencies, and more plausibly, early restrictions can be better for the economy later on (Aum et al., 2021) . It could even have nothing to do with lives or livelihoods, in fact, and can be purely ideological (e.g. https://www.psa.ac.uk/psa/ne ws/freedom-or-self-interest-motivations-ideology-and-visual-symbolsuniting-anti-lockdown).

Inequality of fear and self-quarantine: Is there a trade-off between GDP and public health?

Unemployment Effects of Stay-at-Home Orders: Evidence from High-Frequency Claims Data

The Political Economy of Early COVID-19 Interventions in US States

A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker)

Hit Harder, Recover Slower? Unequal Employment Effects of the Covid-19 Shock