Culture

There has been a long academic and public interest in the contribution of culture and institutions. One famous theory is Max Weber (1905)'s The Protestant Ethic and the Spirit of Capitalism, in which he argues that 'institutions are the main reasons for entrepreneurship and economic development, while “Protestant Ethic” (culture) is just something extra.' On the contrary, Martin Wiener (1981) and David Landes (1998) claim that culture, rather than institutions, is the main reason for economic development. However, these arguments are all based on the people’s behaviour under different institutions. It is technically difficult to distinguish the effects of culture and institutions because there is no good measurement to quantify culture or institutions. In this research project, we attempt to tackle this difficulty in answering the following questions:

  1. How to distinguish the effects of culture from institutions on entrepreneurship?

  2. What are the qualitative and quantitative effects of culture on entrepreneurship?

  3. Are these effects persistent over time?

  4. Are these effects different for male and female?

The novelty of our research design is to use the US as a 'natural experiment' to isolate the two effects. The US is a country made up of immigrants from various cultures. It provides an ideal laboratory because:

  • Immigrants with various cultures face the same institutions.

  • We can distinguish the “entrepreneurs” by inheriting the family business from the true entrepreneurs by setting up new business.

  • Even though the immigrants are not the random sample from their cultures, the comparison is still valid because we are comparing like with like.

We use US census microdata in 1910 and 2000 to evaluate the effect of culture on entrepreneurship and the persistence of the effect. We exclude the US born natives because they are more likely to inherit a family business, so we only use the first-generation migrants to the US. Arguably, all first-generation migrants take their original culture with them to a common institutional environment. This way, we can isolate the effect due to institutions from the effect due to culture.

Let's look at the crude entrepreneurial ratios by a selection of countries, representing the main cultures. Overall, the chances of entrepreneurship decline from 5.1% in 1910 to 3.2% in 2000. In line with Weber's argument, Catholic group in 1910 are on average lowest, and Protestant group are relatively higher. In particular, in 1910 the Greece and Turkey entrepreneurial ratios are quite similar, which can be interpreted as consistent with common ethnicity back then. Mexico, Cuba, Wales, Netherlands, Turkey and Japan show persistence of entrepreneurial culture. It is not surprising that most of the chances are not persistent over the century because the immigrants from the same country might have different cultures.

Note that the crude ratios in the table do not control for the differences in individual characteristics (the supply side of the entrepreneurial market) and industry attributes (the demand side of the entrepreneurial market). If we do, the entrepreneurial behaviour differs purely because of the cultural difference. Therefore, we use the logit model to characterise the entrepreneurial decision behaviour, controlling for gender, marital status, residence duration, education, English ability, age, wealth, industry, and country of origin. The marginal effects of the countries of origin (indicating cultures) are:

Surprisingly, after purging off the influences of other factors, the result shows contradiction against Weber’s conclusion that those with “Protestant Ethics” tend to be more entrepreneurial than those with “Catholicism”. There are negative marginal effects for all the Protestants, while positive marginal effects for Italian (Catholic). Nevertheless, the results shows support to the findings that Jewish culture still maintains highly entrepreneurial. The Jewish migration in 1910 was mainly from the Russia/Poland due to persecution, but in 2000 was mainly from Israel.

A statistical test of persistence shows that Cuba, Italy, England, and the Netherlands have strictly persistent entrepreneurial coefficients over time. The Jews from Russia/Poland (1910) and from Israel (2000) also show persistent entrepreneurial effects. The Greece and Turkey show close correlation in variations across time. These persistence results from the relative internal stability of the social and economic structure. However, many other cultures seem to have different effects across the time, e.g. China and Japan. It might be due to two reasons: the two countries underwent dramatic changes in culture over the century: cultural revolution and westernised; immigration control was introduced against the Chinese immigration and naturalization since 1960s.

Furthermore, based on the same methodology and data, we have also exploit the 'damping effect' of gender on the entrepreneurial culture. We argue that the within-culture variation is mainly attributed to gender, which dampens the cross-culture variation. The principal hypothesis is that there is a greater chance of females in 'unentrepreneurial' cultures being relatively entrepreneurial compared to males.

The publications out of this research project include: