IMMIGRATION AND STATE INSTITUTIONS: DOES REGION OF ORIGIN MATTER?

• Author: Tuszynski, Meg

Some immigration opponents claim that immigrants import bad institutions and policies from their country of origin into their new home country. We argue just the opposite--namely, that immigrants are more likely to self-select into countries with better institutions than those in their home countries. Researchers have examined this issue in both a cross-country and within-country context. Their findings have been mixed. Although others have found small or nonexistent impacts of immigrants on state institutions, those papers assume that all immigrants are the same. Our approach is unique in that it divides immigrant populations in a variety of ways. We build on the previous literature that examines the relationship between immigration and institutions at the state level, where there are smaller inherent differences in economic institutions (compared to differences across countries). We do so by incorporating the regional diversity of immigrant populations, examining whether immigrants' countries of origin matter for economic outcomes in their new home country. Controlling for the diversity of immigrant populations in a way that previous researchers have not done improves our ability to assess immigration opponents' claims that immigrants from economically worse-off countries hurt U.S. economic institutions. We find virtually no evidence of an economically and statistically significant relationship between the levels of immigration we have experienced in recent decades and a decline in economic institutions in the United States, regardless of the region or economic conditions of recent immigrants' home countries. The limited statistically significant evidence we do find is mixed and small in magnitude. Thus, one of the key rationales used to call for immigration restrictions is not supported by our findings.

The economic impacts of immigration on native born populations have been studied extensively in the empirical literature. (This journal devoted its entire Fall 2017 issue to the topic.) For example, Akay, Constant, and Giuletti (2014) found a robust and positive effect of immigration on the subjective well-being of native populations, as measured by survey data from the German Socioeconomic Panel. Though there is some disagreement, the general consensus of the literature is that immigration is positively correlated with various measures of both subjective well-being and economic well-being (Leeson and Gochenour 2015; Kerr and Kerr 2011). Indeed, a recent meta-analysis of the literature by the National Academies of Sciences, Engineering, and Medicine (2016) found that the impact of immigrants on native wages and employment levels is small, and that immigrants have a largely positive effect on long-run U.S. economic growth. Furthermore, the report found that second- and third-generation immigrants tend to contribute more in taxes than either first-generation immigrants or native-born residents.

The link between immigration and recipient country institutions, however measured, has not been explored as extensively in the academic literature (Ashby 2007; Ashby 2010; Nejad and Young 2016; Clark et al. 2015; Powell, Clark, and Nowrasteh 2017; Padilla and Cachanosky 2018). Even assuming there are some (perhaps substantial) positive economic impacts of immigration, large levels of immigration might still be a net negative for destination countries if immigrants create significant negative impacts on host countiy institutions. Indeed, this is the concern of both Collier (2013) and Borjas (2014, 2015). According to Boijas, "For immigration to generate substantial global gains, it must be the case that billions of immigrants can move to the industrialized economies without importing the "bad" organizations, social models, and culture that led to poor economic conditions in the source countries in the first place" (2015: 968). Whether immigrants actually import their "bad" institutions, however, is ultimately an empirical question. That is the question we address in this paper.

This question is important because the concerns that Boijas and others raise--that immigrants from lower-GDP or developing countries will harm their new country's socioeconomic institutions--are often used to justify increased restrictions on immigration. If we fail to find evidence that immigrants actually do import the bad institutions from their home countries, then the argument for immigration restrictions is weakened. We will discuss what previous researchers have found on this topic and then explain how we will test their hypotheses, and ours, empirically. We provide extensive econometric results and then conclude by discussing their implications.

Immigration and Economic Efficiency

The global efficiency gains literature argues that the current level of international immigration is suboptimal due to various immigration prohibitions, and that immigrants' lives and broader macroeconomic environments would both be improved were these restrictions liberalized. Hamilton and Whalley (1984) argued that the reallocation of labor due to the removal of worldwide immigration restrictions could lead to as much as a 14.3 percent increase in world GDP. Since that time, numerous others have examined the quantitative welfare impacts of immigration. Docquier, Machado, and Sekkat (2015) have estimates that fall on the low end of the welfare gains spectrum, with efficiency gains of 7 to 18 percent of world GDP in the medium tenn, with a focal point around 12 percent. As Clemens (2011) details, the average estimates of welfare gains in the existing literature fall somewhere in the 50 to 150 percent of world GDP range. Though the precise estimates are fairly sensitive to particular modelling assumptions, even the more conservative models suggest staggeringly large efficiency gains as a result of the relaxation of mobility restrictions.

Still, there may be real reasons to be concerned about massive increases in international immigration. Clemens and Pritchett (2019) support a relaxation of current stringent immigration restrictions, but suggest that immigrants from poor countries might bring their low productivity with them to richer countries, thereby offsetting the substantial global efficiency gains found in Clemens (2011). They argue that there might therefore be a case for restricting this type of immigration, but clarify that current comprehensive restrictions are still inefficient. Though Collier (2013) does not run any empirical tests of his thesis, he posits that both migrants' origin and destination countries may be adversely impacted by their decision to move. He suggests that the increased diversity in the host country as a result of immigration might lower the existing level of social cohesion and mutual trust in that society.

Borjas (2015) takes Collier's argument a step further, and creates a model of the impacts of migration which incorporates these types of sociocultural externalities. In some simulations, the net gain to world GDP actually turns negative if enough people immigrate. According to Boijas (2015: 968), "For immigration to generate substantial global gains, it must be the case that billions of immigrants can move to the industrialized economies without importing the 'bad' organizations, social models, and culture that led to poor economic conditions in the source countries in the first place."

However, Borjas' argument hinges on the idea that immigrants will, in fact, bring these poor institutions with them. In many countries from which immigrants flee, those poor institutions were imposed by totalitarian dictators rather than chosen by the residents of those countries. So, it does not necessarily follow that immigrants would support implementing them in their new host countries. Indeed, the opposite seems more likely: that immigrants may selfselect countries with better institutions than those in their home countries. Still, since Borjas provides no empirical evidence of this dynamic, we use this paper to provide a limited test of his thesis.

Migrating Attitudes

The other literature to which this paper contributes is the migrating attitudes literature, which argues that immigrants bring with them many of the ideas and convictions of their home country when they move to a new country. The widespread use of worldwide public opinion surveys in recent years has facilitated the development of this literature. It is widely accepted that countries with higher levels of social capital (generally proxied using survey questions related to trust) tend to perform better on a variety of economic measures (see, e.g., Coleman 1988; Putnam 1993; Knack and Keefer 1997; and Platteau 2000).

The migrating attitudes literature builds on this idea. Algan and Cahuc (2010), for example, argue that immigrants from low-trust countries are more likely to exhibit the attitudes consistent with those low-trust societies. This is problematic because, as Aghion et al. (2010) show, individuals from more distrustful societies are more supportive of extensive government regulation. They suggest that this phenomenon results from the fact that in low-trust societies, people consider mechanisms of self-regulation to be lacking, so they favor external regulation to keep untrustworthy businesses and individuals in check.

In a similar vein, Alesina and Giuliano (2014) find that, not only do societies with more familistic values tend to have lower levels of social trust, but also that migrants bring these familistic values with them when they move. (1) Strikingly, they find that "family values inherited by children of immigrants whose forebears arrived in various European countries before 1940 are related to a lower quality of institutions and lower level of development today" (ibid.: 179). More recently, Bologna-Pavlik, Lujan-Padilla, and Powell (2019) found that higher levels of immigration were not associated with higher levels of corruption, and in some...