Most workers in history and a substantial share of workers today are coerced into accepting employment and its terms through force or the threat of force. What drives labor coercion, and what are its consequences?

This paper provides the first estimate of the causal effect of a major hypothesized determinant of coercion, labor scarcity. Theoretically and observational, however, the effects of labor scarcity are ambiguous. For example, Western Europe saw a decline in coercion following the ravaging of the Black Death, while coercion intensified in Eastern Europe. I focus on Estonia before and after the Great Northern War plague (1710-12), drawing on extensive, hand-transcribed data on tens of thousands of serf households and their labor contracts. I show that the plague generated immense spatial variation in labor scarcity that is uncorrelated with a host of characteristics. Exploiting this locally quasi-random variation, I find that labor scarcity substantially raises coercion, as measured by serfs’ work obligation. Investigating mechanisms, I find that this effect is enhanced by the lack of outside options and increased labor monopsony power, in accordance with the prediction of historians and theoretical models. Taken together, these findings give important insights into the circumstances under which coercion increases and provide suggestive evidence why it did not in other cases (e.g., post-Black Death Western Europe).

Turning to the consequences of labor coercion, I exploit the uncovered scarcity-coercion relationship and instrument coercion intensity with plague deaths. I find that (instrumented) labor coercion decreases historic literacy and present-day levels of trust. These findings contribute to a literature that reports mixed consequences of labor coercion and introduce migration as a new mechanism.

How do historical events change individuals' values, and when do they enter collective memory? A fundamental problem in the field of cultural economics is the difficulty of measuring these changes in values when targeted surveys are lacking. To tackle this problem, a promising recent literature uses first names as proxies for values, however, we argue that the scope and systematization of this approach can be augmented.

To demonstrate this, we develop new and more formalized methods for estimating values from first names. Importantly, we expand the breadth of values that can be measured to, e.g., militarism, entrepreneurship, and scientism. This is achieved by leveraging the fact that parents often name their children for namesakes in professions they admire, e.g., military generals, entrepreneurs, and scientists. We determine what first names characterize each of those professions by 1) relying on occupation information in the full-count US census or 2) by drawing on an extensive global list of famous individuals in different professions (e.g., George Washington, John D. Rockefeller, or Isaac Newton). With this, we attach measures of values to each individual in a population, such as the US Census, based on their first name, after also applying a novel machine learning method of first name choice that determines the effect of class, ethnicity, and other identities.

We then turn to applying and validating our first name-based value measure. First, we show that the positive effect of early Irish and Scottish migration on contemporary crime rates in the US South, as reported in Grosjean (2014), disappears once we control for differences in militaristic values. In a second application, we find that individuals who were exposed to American Civil War battles display collective memory in the form of decreased militarism (in names) in the following decades.

In sum, this paper provides scholars with a tractable approach to extracting measures of various values from first names. Our approach is applicable to many settings, given that in its most simple version, relying on famous individuals as namesakes, the only setting-specific data required are first names in the target population. We also provide researchers with US county-level ideology measures for all census years, which we hope will inform future persistence studies and mitigate the ‘black box’ problem caused by missing data in the intermediate period.

The political consequences of refugees for receiving countries have received much attention in recent years and have sparked a burgeoning literature. However, evidence on the long-run consequences of refugees is lacking.

The expulsion of 8 million Germans (so-called expellees) from Eastern Europe to post-WWII West Germany serves as a natural experiment that allows us to estimate the long-run effects of refugees on far-right voting, violence, and immigration sentiments.

Using administrative data, we show the (instrumented) historic settlement of expellees increases far-right voting and violence today, particularly in locations where many of them initially settled and where their characteristics were more mismatched with those of the receiving locality.

We turn to survey data to investigate whether this effect is driven by expellees themselves, their offspring, or by non-expelled ‘natives’ who were exposed to them. Creating new expellee and expellee offspring identifiers in the German Socio-Economic Panel (SOEP), we find that first-generation expellees view migrants and refugees significantly more favorably than other Germans, while their offspring hold more hostile views. Natives are only significantly more likely to vote for the far-right and have anti-immigrant/refugee sentiments in locations where expellees experienced higher unemployment in the past.

Taken together, our findings suggest that past displacement can have lasting negative effects on attitudes toward migrants through contact, particularly when the displaced were in more dire conditions and were more dissimilar to natives. This effect arises both from the attitudes of refugee offspring and natives.