Page 1 sur 67

Pale in Comparison

The Economic Ecology of the Jews

as a Rural Service Minority∗

Yannay Spitzer†

The Hebrew University of Jerusalem

February 9, 2017

Abstract

The five million Jews who lived in the Pale of Settlement at the turn of the century were

overwhelmingly over-represented in towns and in cities. They specialized in seemingly urban

occupations, were relatively literate, and were almost absent in agriculture. This pattern per- sisted overseas where one third of them would eventually immigrate. Hence, Jews were typically

characterized as an urban minority. I argue that the opposite was true. The economic ecology

of the Jews, the patterns of choices of occupation and location, are described in a model in

which Jews were countryside workers with a comparative advantage in rural commerce, com- plementing agricultural workers, and without comparative advantage in denser urban settings.

Using data from the 1897 census, I show that the cross-sectional patterns across districts and

localities were consistent with all the predictions of this model. When the share of Jews in the

population grew, Jews spilled across two margins—occupational, as manufacturing workers, and

geographic, as rural frontier men. Non-Jews were imperfect substitute for Jews, rendering the

latter indispensable to the countryside economy. No evidence of urban advantage is evident in

the data. Turn of the century Pale of Settlement Jews ought to be understood as rural workers,

in and of the countryside. In this light, the patterns exhibited in the US after immigration

appear as a sharp break from, rather than a continuation of, old country economic tradition.

∗

I thank my committee members Joel Mokyr (chair), Igal Hendel, and Joseph Ferrie for their endless support and

advice. I am grateful for comments from seminar participants at Bar-Ilan University, Brown University, Northwestern

University, European Historical Economics Association Meeting, Pisa, Israel Economic History Association, Haifa,

Markets and States in History Conference, Warwick, and for discussions with and suggestions from Rapha ̈el Franck,

Oded Galor, Marlous van Waijenburg, David Weil, and Ariell Zimran. Data used in this paper was constructed with

Gennady Polonetsky.

†Department of Economics, The Hebrew University of Jerusalem. Email: yannay.spitzer@huji.ac.il, personal

website: yannayspitzer.net.

Page 2 sur 67

1 Introduction

Jews are typically considered to have been a quintessentially urban minority. Since the Early Middle

Ages, in almost every country and period in which they developed significant communities, Jews

clustered in urban environments, specialized in urban occupations, and were only rarely directly

employed in agricultural labor.1

In particular, this was the case at the end of the nineteenth century

in the Pale of Settlement—the western provinces of the Russian Empire in which Jewish residence

was generally allowed—that was home to five million Jews, more than half of world Jewry at that

time. In the words of Jacob Lestschinsky, arguably the most prolific scholar to have ever worked

on the demographics of the Jews: “There is general consensus that Jews are an ‘urban’ people, and

East European Jewry was no exception.” (Lestschinsky 1961, p. 72).

I propose to modify this consensus, by asking whether Jews were really urban or rural agents.

From a narrow economic perspective, I suggest to view Pale of Settlement Jews as a rural service

minority. I show that explaining the patterns of settlement and occupation—the economic ecology—

practiced by Jews in the Pale of Settlement during the late imperial period, requires thinking of

them as a rural, not an urban, organism, the countryside its habitat.2 This economic ecology had

been qualitatively different from the metropolitan nature the Jewish communities soon assumed

when they migrated to the New World and to western Europe. Whereas after their migration, Jews

overwhelmingly clustered in metropolitan centers and were largely absent in the countryside, in the

old country Jews tended to disperse uniformly across space, and though clustered in small urban

localities, they were present in virtually every rural district and showed no particular preference

for larger urban environments. I explain the economic ecology of Pale of Settlement Jews using

a model of an ethnically segregated labor market in a partial spatial equilibrium. This model

considers Jews as complementary labor force to the non-Jewish largely agricultural labor within

a local rural economy. The model predicts a number of empirical patterns that characterize the

Jewish economic ecology; some of these patterns were hitherto unknown, unnoticed, or merely

speculated due to lack of data. I use newly coded data from the 1897 Russian Census to motivate

and justify the assumptions of the model, and to test its predictions.

In this model, two ethnicities, Jews and non-Jews, live in a rural economy that uses labor of two

types of occupations: agriculture and commerce. Agricultural labor and commerce labor are im- perfect substitutes. The primary underlying assumption is that Jews on average have comparative

advantage in commerce and comparative disadvantage in agriculture, such that as a result, Jewish

and non-Jewish workers are effectively themselves imperfect substitutes. Within each ethnicity,

there is heterogeneity in the individual degree of comparative advantage in commerce, and in equi- librium these distributions of comparative efficiency between the two occupations determine the

labor supply.

1 On the historical process that brought about this specialization, see Botticini and Eckstein (2012).

2 To be clear, the urban-rural distinction is used here in a restricted economic-demographic sense, and is not

meant to bear any cultural meaning.

2

Page 3 sur 67

Labor of both occupations exhibits decreasing returns to scale. Workers choose an occupation

freely within a district-level labor market, according to where their marginal productivity of labor

is higher. Jews can migrate across districts, but migration is costly. Thus Jews face two margins: an

occupational frontier and a geographical one. When the marginal productivity of commerce labor

declines, Jews are increasingly likely to opt out of commerce, or to migrate to another district. The

model describes a state of partial equilibrium, in the sense that each worker chooses the occupation

that maximizes his income within the district, but differences in wages across districts may exist.

Migration across districts is not formally incorporated into the model, but it is useful to think of

broadly-defined costs of migration that enable differences in wages to persist across districts. Some

migration does take place gradually over time, flowing from districts of lower wages to districts

of higher wages. For Jewish workers, this would mainly imply migration from districts in which

Jewish density is high, and thus commerce wages are low, to districts in which Jewish density is

low and commerce wages are higher. Non-Jews also face the occupational margin; despite being

disadvantaged in commerce, when the marginal productivity in commerce is sufficiently high, a

larger minority of non-Jews opts for commerce.

This simple model of Jews as a rural service minority makes several non-trivial testable predictions.

First, Jews would avoid clustering at the district level. They may form a majority at the level of

the locality, but unlike other ethnic minorities, never at the level of the district. Obversely, they

will tend to distribute evenly across space, and would spill over to new geographic frontiers when

the option becomes available. A Jewish vacuum at the level of the district will not be a viable

state—if open for Jewish immigration for sufficient amount of time, and to the extent that the

costs of migration are not prohibitively high, higher wages in commerce will attract Jews, such

that all districts will have a sizable Jewish minority. Cases of relatively low Jewish density will

occur in regions in which the option to settle is new, and will reflect a state of partial equilibrium,

with the density gap gradually offset by immigration from high density regions. Everything else

equal, in districts with higher Jewish density there will be more commerce workers overall, but

the share of commerce workers among both Jews and non-Jews will be lower. That is, a larger

proportion of Jews crowds both Jews and non-Jews out of commerce.

The model does not include cities as separate labor markets. This has two important implications,

that are at odds with a view of Jews as a metropolitan minority, or one that places Jews in a town- level labor market, rather than a regional one. Thus, it is agnostic on whether Jews would choose to

cluster in large cities. A competing model, one that would describe Jews as a metropolitan minority

with comparative advantage in greater urban centers, would have predicted that Jews would be

increasingly over-represented in larger cities. Furthermore, since in the model I propose the labor

markets are at the level of the district, it does not predict patterns reflecting Jewish congestion at

the level of the locality. The correlation between Jewish density and the total, Jewish, and non- Jewish shares of commerce workers will not exist at the level of the town, conditional on the degree

of Jewish density in the district. The correlations will be found only at the level of the district.

3