From the persecuting to the protective state? Jewish expulsions and weather shocks from 1100 to 1800

Warren Anderson, Noel Johnson, Mark Koyama, 18 February 2013



Throughout history, religious and ethnic minorities have been the victims of persecution or expulsion. Persecutions and expulsions were frequent in the pre-industrial world, particularly in medieval and early modern Europe. Such events are rare in the developed world today, but they continue to occur in the developing world. Notably examples include the expulsion of Asians from Uganda under Idi Amin, the persecution of the Laotians in Vietnam in the 1970s, and the treatment of south Asians in Kenya in recent years, as well as the genocide that occurred in Rwanda in the 1990s (for more on the economic aspects of ethnic conflict, see Spolaore and Wacziarg 2009; Campos 2009).

New research

When and why did this transition from the persecuting state to the protective state in Europe take place? In our recent work, we focus on the persecution of the Jews in medieval and early modern Europe, one of the most numerous and best documented minorities throughout European history (Anderson et al. 2013). We obtain two key results:

  • Using data on climatic variation, we identity the effect that negative economic shocks had on minority rights in the pre-industrial period.
  • We show that the relationship between climatic shocks and the expulsion or persecution of Jewish communities was strongest in 15th and 16th centuries and disappeared after 1600.

Our identification strategy exploits the variation in city-level temperature to test whether expulsions were associated with colder temperature during the growing seasons (that is, April to September). A decrease by one standard deviation – or about one third of a degree – in average growing-season temperature in the 15th and 16th centuries was associated with a 1-2 percentage point increase in the likelihood that a Jewish community would be expelled in any given five-year period. This effect is significantly stronger for cities with poorer quality soil.

Our analysis allows us to quantify the decline of the persecuting state and the rise of religious tolerance in the period before democratisation or the onset of modern economic growth. After 1600 there is no longer a discernible effect of weather on expulsions.

Data and empirical strategy

Agrarian economies were vulnerable to weather shocks. Our hypothesis is that negative economic shocks should have led to the scapegoating of minorities and to political unrest. We therefore expect periods of cold weather to be associated with a higher probability of expulsion. To test this, we construct a panel data-set comprising 785 city-level expulsions of Jewish people from 933 European cities which are recorded as having Jewish populations between the years 1100 and 1800. We also collect data on 614 incidents of violent persecution that fell short of full expulsion. All of our data comes from the 26-volume Encyclopedia Judaica (2007).

Good-quality data for output or national income does not exist for the middle ages. To obtain a measure of economic stress we therefore use weather data.

Information on yearly growing-season temperature comes from Guiot et al. (2010) who assemble information from proxy sources including 95 tree-ring series, 16 indexed climatic series based on historical documents, ice-core isotopic series, and pollen-based series to construct a 32-point grid of reconstructed temperature during the growing season for all of Europe between the year 900 and the present day. We use geospatial software to interpolate the temperature for the area between the grid points so that we have a smooth map for each year and extract the yearly temperatures for each of our cities. Figure 1 plots the cities in our database and the weather grid and corresponding heat map for the year 1100.

Figure 1. The distribution of Jewish cities overlaid with the temperature grid and the corresponding heat map of average temperature during the growing season in 1100

In order to refine our hypothesis we use the suitability for wheat cultivation of the area around each city as a measure of how vulnerable a city was to a weather shock. This data comes from the UN Food and Agriculture Organization. To control for market development we use data from Bosker et al. (forthcoming) as a measure of urban population.

Figure 2. The effect of a negative deviation in temperature on the probability of expulsion. Fixed-effect regression for cities with poor soil quality

Our econometric results indicate that cold weather has a strong effect on the probability of expulsion in the 15th and 16th centuries. The baseline probability of an expulsion in our sample during any given five-year period in the 15th or 16th centuries was about 2.5%. A change of one standard deviation in average temperature is associated with a 1-2 percentage point increase in the likelihood of expulsion in a given five-year period. This means, for example, that a Jewish individual from the 15th or 16th century, who lived to 50 years old, faced roughly an 18% chance of being expelled during their lifetime. This risk was almost twice as great during a cold year.

Figure 2 plots the coefficient from our preferred specifications. It shows that during the 15th and 16th centuries a 1°C decrease in average temperature increases the likelihood of an expulsion occurring by between four and seven percentage points. We obtain similar results when we control for events such as the Black Death and the expulsion of Jews from Spain in 1492. We also use data on persecutions of Jewish communities and obtain results that are consistent but with a larger magnitude.

The decline in expulsions and persecutions

The relationship between weather shocks and expulsion breakdowns after 1600 saw expulsions still occurring, though mostly in eastern Europe. Over time, they were less frequent and uncorrelated with supply shocks.

Sevreral explanations are consistent with the historical evidence:

  • First, a diminution in religious intensity and prejudice after the end of the Thirty Years War;

This proposition is controversial among historians (see Kaplan 2007) and there is plenty of evidence that anti-Semitic attitudes can persist for centuries (see Voightländer and Voth 2012). Nevertheless, we cannot fully exclude this on the basis of the available evidence.

  • Second, the rise of more powerful states capable of resisting popular anti-Semitism;

Medieval states were vulnerable to economic shocks and popular unrest, and though they depended on minority groups such as Jewish people for financial expertise and as a source of revenue, they were willing to sacrifice the rights of those very same minority groups in order to sate popular anger. The states that emerged after 1600, in contrast, succeeded in building fiscal and legal capacity; they enjoyed greater political stability and ceased to be responsive to popular unrest and antisemitism. This argument is consistent with the large literature that identifies an improvement in economic and political institutions in Europe from the seventeenth century onwards (e.g. Acemoglu, Johnson and Robinson 2005).

  • Finally, the rise of markets and changing attitudes towards commerce and particularly moneylending and banking may have played a role in reducing the likelihood that popular unrest would lead to the scapegoating of Jewish minorities.

Conclusions and implications

This column examines the effect of negative supply shocks on the treatment of religious or ethnic minorities. We exploit the fact that the economies of medieval and early modern Europe were predominantly agrarian, and use exogenous variation in temperature during the growing season to identity the effect of supply shocks on the probability of a Jewish community suffering an expulsion or persecution.

Our results quantify the decline in persecutions and expulsions that occurred in Europe before the onset of modern economic growth or the rise of democracy. They are consistent with the argument that an improvement in political and economic institutions and state capacity occurred in the period 1600-1800. Finally, our findings point to the cost of economic volatility in weak states today.


Acemoglu, Daron, Simon Johnson and James Robinson (2005), “The rise of Europe: Atlantic trade, institutional change, and economic growth”, The American Economic Review 95(3), 546–579.

Anderson, Warren Robert, Noel D Johnson, and Mark Koyama (2013), “From the Persecuting to the Protective State? Jewish Expulsions and Weather Shocks from 1100 to 1800”, memo.

Bosker, Maarten, Eltjo Buringh and Jan Luiten van Zanden (forthcoming), “From Baghdad to London: unravelling urban development in Europe and the Arab world 800-1800”, Review of Economics and Statistics.

Campos, Nauro F (2009), “The growth effects of changes in ethnic fractionalisation”,, 22 December.

Encyclopedia Judaica (2007), 2nd edition, Macmillan Reference.

Guiot, Joel and Christopher Corona (2010), “Growing season temperature in Europe and climate forcings over the past 1400 years”, Plos One 5(4), 1–15.

Kaplan, Benjamin (2007), Divided by Faith, Harvard University Press, Cambridge, M.A.

Spolaore, Enrico and Romain Wacziarg (2009), “Kinship and conflict”,, 7 July.

Voigtälnder, Nico and Hans-Joachim Voth (2012), “Persecution perpetuated: The medieval origins of Anti-Semitic violence in Nazi Germany”, Quarterly Journal of Economics 127(3), 1–54, also for, “The geography of hate: How anti-Semitism in interwar Germany was influenced by the medieval mass murder of Jews”, 22 May 2011.

Topics: Economic history, Frontiers of economic research
Tags: Agriculture, expulsion, history, Jews, Judaism, weather

Assistant Professor of Economics, University of Michigan-Dearborn

Assistant Professor of Economics, George Mason University

Assistant Professor in Economics, George Mason University