The Role of Public Employment in Stabilizing the Macroeconomy: Theory and Evidence

在20世纪下半叶，西方世界的许多国家建立了广泛的政府支出计划。在商业周期研究中，一个被低估的事实是，在许多发达经济体，政府现在雇佣了很大一部分劳动力：英国和美国为16%，欧盟为20%（经合组织，2019年）。因此，在经济衰退期间，政府的行为对于直接决定失业率的变化至关重要：政府是否也在解雇工人，从而加剧了失业率的上升，还是在雇佣工人，从而可能弥补一些空缺？尽管政府就业对劳动力市场有这种“直接”影响，但大多数关于政府周期性作用的研究都是从政府就业中抽象出来的。相反，大量的文献解决了“政府支出乘数”到底有多大这个古老的问题。从经验上看，这衡量的是典型的抽象的政府支出增加对经济的影响。从理论上讲，这是在问，如果政府通过从私营部门购买更多商品来增加支出，从而对劳动力市场产生的任何影响都是“间接的”，那么经济将如何反应。在本文中，我的目标是通过实证和理论两方面的研究来填补这一文献空白：1)政府就业的变化如何在商业周期中“直接”导致失业；2)政府在经济衰退期间暂时增加就业以应对失业是否是一个好主意；以及这一政策与传统财政刺激或转移支付（如失业救济金）等更深入探索的工具相比如何。为了回答这些问题，我建议写一篇三章的论文，将以宏观经济学几个领域的最新进展为基础。我的第一章将研究公共就业在经济周期中的作用。这样做需要一个现实的劳动力市场，为此，我使用了Coles和Kelishomi（2018）在搜索和匹配文献中的最新进展。他们认为，与自由进入的标准搜索和匹配模型不同，私营企业在为失业工人创造新的就业机会方面往往行动迟缓，并估计就业机会的创造在数据中是非弹性的。在这一章中，我的主要观点是，如果私营部门创造的就业机会在这种方式上缺乏弹性，那么在衰退期间改变公共就业就会对失业率产生强大的影响。我将建立一个定量模型来估计公共就业对经济的影响，包括公共就业“乘数”，并估计英国政府就业变化对失业的历史贡献。在我的第二章中，我将扩展模型，以研究异质性以及公共就业如何影响风险和不平等。我将首先扩展双主体新凯恩斯（TANK）模型（Cantore and Freund 2020; Ravn and Sterk, 2020），将第一章中的公共就业和劳动力市场模型包括在内。这在很大程度上建立在我的MRes论文的基础上。然后，我将这个模型扩展到一个成熟的异构代理新凯恩斯主义（HANK）模型（Auclert et al. 2018; 2020; Kaplan et al. 2014; 2018; Hagedorn et al. 2019）。据我所知，这将是第一个使用TANK或HANK模型研究公共就业的模型，因此我将能够使用最具创新性的商业周期模型来研究政府就业、不平等和总需求之间的相互作用。在我的第三章中，受COVID大流行的启发，我将研究大流行对就业的影响，以及公共就业如何减轻封锁的影响。为此，我将把流行病学SIR模型与第一章中的模型结合起来，以检验大流行期间公共就业在支持健康和失业方面的共同作用。

During the second half of the 20th century extensive government spending programs have been established by many countries in the Western world. It is an underappreciated fact in business cycle research that governments now employ large fractions of the workforce in many developed economies: 16% in the UK and the US, and 20% in the EU (OECD, 2019). The behaviour of the government during a recession is therefore crucial in directly determining what happens to unemployment: Is the government also firing workers, and therefore adding to the rise in unemployment, or are they hiring workers, and potentially picking up some of the slack?Despite this "direct" effect of government employment on the labour market, most research on the cyclical role of the government abstracts from government employment. Instead, a voluminous literature tackles the age-old question of how large the "government spending multiplier" is. Empirically, this measures the effect on the economy of a typically abstract increase in government spending. Theoretically, this asks how the economy responds if the government increases spending by purchasing more goods from the private sector, and therefore any effects on the labour market are instead "indirect".In this dissertation, I aim to fill this gap in the literature, by investigating both empirically and theoretically 1) how changes in government employment "directly" contribute to unemployment over the business cycle, and 2) whether it is a good idea for the government to temporarily increase employment during recessions to fight unemployment, and how this policy compares to the more-explored tools of traditional fiscal stimulus, or transfer payments such as unemployment benefits.In order to answer these questions I propose to write a three-chapter dissertation, which will build on recent advances in several fields of macroeconomics.My first chapter will study the role of public employment over the business cycle. Doing so requires a realistic labour market, and for this I use recent advances in the search and matching literature by Coles and Kelishomi (2018). They argue that private sector firms tend to be sluggish at creating new jobs for unemployed workers, unlike standard search and matching models with free entry, and estimate that job creation is very inelastic in the data. My key insight in this chapter is that if private sector job creation is inelastic in this way, changing public employment during recessions has powerful effects on unemployment. I will build a quantitative model to estimate the effect of public employment on the economy, including a public employment "multiplier", and estimate the historical contribution of changes in government employment to unemployment in the UK. 11111111111111111In my second chapter, I will extend the model in order to study heterogeneity and how public employment affects risk and inequality. I will first extend a Two-Agent New Keynesian (TANK) model (Cantore and Freund 2020; Ravn and Sterk, 2020) to include public employment and the labour market model from chapter 1. This builds heavily on the work from my MRes thesis. I will then extend this model into a full-blown Heterogeneous-Agent New Keynesian (HANK) model (Auclert et al. 2018; 2020; Kaplan et al. 2014; 2018; Hagedorn et al. 2019). To the best of my knowledge, this will be first model studying public employment using either TANK or HANK models, so I will be able to study the interaction between government employment, inequality, and aggregate demand using the most innovative business cycle models.In my third chapter, motivated by the COVID pandemic, I will study the effect of pandemics on employment and how public employment can mitigate a lockdown's effects. To do so I will combine an epidemiological SIR model with the model in chapter 1 to examine the joint role of the public employment in supporting health and unemployment during the pandemic.