Long-run Connections in the Distribution of Income

Capital Shares and Income Inequality: Evidence from the Long Run

by Erik Bengtsson (Lund) and Daniel Waldenström (Paris School of Economics and CEPR)

Abstract – This article studies the long-run relationship between the capital share in national income and top personal income shares. Using a newly constructed historical cross-country database on capital shares and top income data, we find evidence on a strong, positive link that has grown stronger over the past century. The connection is stronger in Anglo-Saxon countries, in the very top of the distribution, when top capital incomes predominate, when using distributed top national income shares, and when considering gross of depreciation capital shares. Out of-sample predictions of top shares using capital shares indicates several cases of over- or underestimation.

Freely available for a limited time at: The Journal of Economic History, 78(3): 712-743
DOI: https://doi.org/10.1017/S0022050718000347

Review by Leandro Prados de la Escosura (Universidad Carlos III, Groningen, and CEPR)

Until recently, many academic economists would react sceptically to the idea of income inequality. It is absolute poverty what matters, they would argue, and sustained growth is the answer. It would be misled, however, to conclude the only lately has inequality become part of the economists’ agenda. Actually inequality has always been present in economists’ preoccupations. Its symptoms have varied, as social sensitivity to inequality has changed over time. Classical economists identified personal income distribution with the functional income distribution as inequality largely depended on the gap between average incomes of capital and labour. This is clearly exposed in David Ricardo’s famous passage in which he noted that the explaining the distribution of “the produce of earth (…) among (…) the proprietor of the land, the owner of the stock or capital (…) and the labourers” is the main purpose of political economy. In the early 20th century, as the share of skilled workers in the labour force was increasing, Simon Kuznets noted the dispersion of labour incomes and highlighted its role as a driving force of income inequality. At the turn of the century, the increase in the concentration of income at the top of the distribution has renewed the concern about inequality as the contributions of the late Sir Tony Atkinson, Thomas Piketty, and their collaborators evidence. It is in this context where Eric Bengtsson and Daniel Waldeström’s important contribution should be read.

The purpose of Bengtsson and Waldeström was to the test the existence of a long run connection between the functional and the personal distribution of income. As economic historians they find no reason to assume the relationship would be stable over time and across countries since many other dimensions (technology, institutions, personal incomes composition, …) will impinge on it.

Their analytical framework is the decomposition of income inequality, using the coefficient of variation, into wages and capital income dispersion, factor shares, and the correlation between capital and labour income, from which a link between capital share and income inequality is predicated. More specifically, a positive association between the two metrics is hypothesised: as capital returns are more unevenly distributed than those accruing to labour, a rise in the share of capital in national income would result in an increase in personal income inequality.

Once their hypothesis is defined, they deal with the data. On the basis of capital returns and GDP, mostly derived from previous scholarly work, they put together a reasonably homogeneous dataset on capital shares, that is, the ratio of capital incomes (interest, profits, dividends, and realized capital gains) to national income for 21 countries (mostly present-day OECD countries) since 1900. Capital returns are computed form the income side of historical national accounts, which raises the challenge of how to distribute mixed incomes (those of self-employed) between capital and labour. The so called labour method that attributed to the self-employed a labour income equal to that of the average employee in each specific sector of economic activity is the usual approach. Then, they choose income concentration at the top of the distribution as a measure of personal income distribution, since alternative metrics such as the Gini coefficient are not available on yearly basis for their country sample and time span. The dataset on the share of income accruing to the 10, 1, and 0.1 top per cent derives from the World Inequality Database (https://wid.world/wid-world/).

The approach to test the association between personal and functional income distribution is panel regression analysis, in which a log-linear relationship is predicated between top income shares, as dependent variable, and the net (gross) capital share, so the latter’s parameter represents the elasticity of income concentration (inequality) with respect to the capital share. In addition to the baseline equation, they also compute more complex models which include control variables (level of development, proxied by GDP per head; structural change, by the agricultural labour share; relevance of private capital by stock marker capitalisation; and size of the public sector, by the government spending to GDP ratio) country fixed effects, and a linear time trend. The overall view of all countries over the 20th century is complemented by a breakdown by epoch (pre World War II, 1950-80, and 1980-2015) and type of countries (three clubs, the Anglo-Saxon, the Scandinavian, the Western European).

Their main finding is a robust association between top income shares and capital shares over the long run, that results in a high elasticity of income inequality with respect to the capital share that declines when specific periods (1950-1980 and 1980-2015) are considered. Thus, a 10 per cent rise in the net capital share is corresponded by an almost similar increase in income concentration at the top. It is worth noting, however, the lower elasticity –and statistical significance- found for the pre-World War II era. When more complex models, including covariates, are used the fit of the regressions increases but reduces the coefficient for the capital share that, nonetheless, still holds up. The association becomes stronger as top incomes shares are restricted to the 0.1 per cent and a possible explanation is that these mainly correspond to capital earners.

The authors also explore the extent to which capital incomes at the top of the distribution account for the association between top income shares and the capital share. They confirm previous findings suggesting that capital incomes predominate in the incomes of the top earners and increase within the income top. However, the authors find that high-paid salaried employees have been replacing capital earners at the very top of the distribution. In any case, the association between top income share and the capital share is stronger when capital rather than wage or total top incomes are considered.

1913: The wealthy man profits from the sweat of the worker. The cartoon is titled ‘The Reflection’ and the capitalist exclaims ‘I’m a self-made man, look at me’. (Photo by Hulton Archive/Getty Images)

The paper concludes with the authors addressing the crux of the matter, is the capital share a good predictor of inequality? In order to do it, they re-run panel regressions leaving aside the country whose inequality is to be predicted. The results are positive in general, but the authors acknowledge that capital shares are not perfect predictors of concentration at the top of the distribution, to which one may add that top income shares are neither a precise measure of personal income inequality.
This thought-provoking paper raises some reflections. Would not it more intuitive the framework proposed by Milanovic (2005) than the authors’ breakdown of the coefficient of variation? Milanovic decomposes inequality into between-group and within-group inequality. In a similar scenario, with only two groups: proprietors, to whom capital (and land) incomes accrue, and workers who receive the returns to labour, personal income inequality results from both the gap between average incomes of capital and labour (that is, a metric that captures the functional distribution of income) and the dispersion of incomes within both capital owners and workers.

A reference to top income shares shortcomings is missing. Uncritically assuming it is a good proxy of personal income distribution neglects that fact that although top income share satisfies axioms of income scale independence, principle of population, and anonymity only weakly does the Pigou-Dalton transfer principle. Moreover, it is silent on how inequality evolves at the bottom of the distribution.

Some of the results could have been explored more. For example, their finding of a lower elasticity of income inequality with respect to the capital share in the pre-World War II era (that is also found when the Ginis is used as dependent variable) deserves further examination. On the one hand, it seems at odds with the inference of a much higher association between income inequality and the capital share in earlier phases of economic development, as between-group inequality (the average capital incomes to labour incomes ratio) has a larger weight in total personal inequality (the dispersion of capital and labour incomes is presumably lower). On the other, a potential explanation would be the increasing dispersion of factor returns –wages, in particular- in the early 20th century, a finding consistent with Kuznets’ focus on the rise of skilled labour and rural-urban migration.

Another neglected issue (somehow a paradox given the authors’ background) is why income concentration at the top and the net capital share are not associated in the Nordic countries since 1980.

The avenues for research the paper opens deserved more detailed consideration. It is true that in the last section the authors address the issue of how good a prediction of personal inequality the capital share is. However, an obvious extension would be to use the capital share as a proxy for income inequality prior to the early 20th century, when hardly any data on top income shares are available and other measures of inequality such as the Gini are rare.

The paper is worth reading as it represents an ambitious and successful project to deal with income inequality over space and time. Moreover, a most valuable online appendix that complements the freely accessible dataset provided by the authors’ webpage accompanies the paper.

References

Milanovic, B. (2005), Worlds Apart: Measuring International and Global Inequality, Princeton: Princeton University Press.

Reconstructing the B-School

Clio in the Business School: Historical Approaches in Strategy, International Business and Entrepreneurship

by Andrew Perchard (Stirling), Niall Mackenzie (Strathclyde), Stephanie Decker (Aston) and Giovanni Favero (Venice)

On the back of recent and significant new debates on the use of history within business and management studies, we consider the perception of historians as being anti-theory and of having methodological shortcomings; and business and management scholars displaying insufficient attention to historical context and privileging of certain social science methods over others. These are explored through an examination of three subjects: strategy, international business and entrepreneurship. We propose a framework for advancing the use of history within business and management studies more generally through greater understanding of historical perspectives and methodologies.

Keywords: History, strategy, international business, entrepreneurship, methodology

Freely available for a limited time at: Business History, 59(6): 904-27

Review by Mitchell J. Larson (University of Central Lancashire)

Recently Martin Parker (Bristol) has taken to the airwaves promoting the idea of bulldozing the business school. In sharp contrast, Andrew Perchard, Niall McKenzie, Stephanie Decker, and Giovanni Favero make a compelling case for certain disciplines in the management sciences to open themselves to alternative methodological and epistemological approaches. They argue that the fields of strategy, international business, and entrepreneurship have not embraced historically-oriented research to the same extent as other fields within business and management studies. The authors also admit that many scholars conducting historical business research have not made a sufficiently solid case about the robustness of their historical methodology(s) or data to convince other social scientists about the validity of their claims. Drawing upon an impressive range of previous works to develop their discussion, the paper attempts to reconcile these discrepancies to highlight how a more explicit articulation of the historian’s process could overcome the concerns of ‘mainstream’ management scholars regarding theorization and methodology in these three fields specifically and in management studies generally.

One major concern held by non-historians is that historical work illustrates an alleged a-theoretical or even anti-theoretical nature of scholarly writing (Duara, 1998). A second major concern is that historical methods (i.e. of data collection) are not sufficient grounds upon which to base management theory. The authors demonstrate the complexities of these issues with respect to existing historical work in business and management studies, such as the ‘cherry-picking’ of outlier events to support a more general point – especially by scholars in other fields applying historical methods rather casually – and place responsibility upon business and management historians to make their process(es) more transparent and explain themselves and their work better to other social scientists. The article claims that the “continuing distinctions drawn between the primary data created by social science research…with the collection of ‘secondary’ documentary evidence in archives…are misleading.” (p. 915). Whereas social science researchers will be aware explicitly about potential sources for bias in their data and often include discussions about this in their work, the historical process ‘internalizes’ these judgements and thus appears to hide them from the reader. The discipline of history, so accustomed to the individual historian’s assessment of the materials being examined, assumes that with satisfactory preparation the historian’s assessment will be reasonable based on her (or his) knowledge of the historical context, the actors involved, and assumptions about the rationality and practicality of the various decisions that might have been made at any particular point in the timeline. But it is this internalization of decision-making and assessment which so troubles non-historians and why the authors call for business and management historians to “more clearly articulate the methodologies adopted by historians to show the value of history to business and management studies…” That there is value to be realized is shown through the acceptance of historical approaches by other branches of the management studies arena, and their point is that these three sub-fields have been slower to warm to their use than others.

The major difficulty here lies in the way data are encountered: the social scientist generates new ‘primary’ data through his or her interaction with respondents whether actively (through interviews or questionnaires) or passively (through observation). Given the nature of historical work, of course this style of primary data generation is seldom possible: all the protagonists are gone and even the organization(s) to which they were affiliated may have disappeared or transformed beyond anything the historical actors could have imagined. Indeed even the labels of what constitutes ‘primary’ and ‘secondary’ data differ between historians and other groups of social scientists. What the historian then faces is piecing together traces of the past much like an archaeologist might do when exploring new ruins. The main difference is that the business historian deals with written records while the archaeologist deals with physical remains, but in both cases often as much (or more) remains hidden as is brought to light in the process of discovery. This process, and the gaps in data continuity that it allows, appear to bother social scientists whose epistemological approach is steeped in the rationalist arguments of the physical sciences and applies only to the data they have actively sought to collect. That other elements can be discarded as irrelevant to the analysis likewise troubles historians for whom contingency and context are vitally important pieces of the story.

There are a number of significant factors here which the article discusses at some length, but what is striking about the discussion is that there is, perhaps ironically, seemingly little consideration for how these disciplines arose and evolved over time and whether these differences in development might be at the root of the issue. History as an activity reaches back to antiquity but the modern discipline of history received fresh articulation in the early nineteenth century. In contrast, the fields one might ascribe to the ‘social science’ area relevant to business and management (anthropology, communications, economics, geography, sociology, and psychology, for example) tend as a group to be newer and as part of their growth had to justify space in the academic environment for themselves. The process of doing so led these fields to ally themselves with the methods and approaches of the physical sciences to gain scientific credibility in a way that the traditional subject of history never did. The discipline of history, and by extension business and management history, is now playing a catch-up game to find ways to articulate and justify its value as a discipline in the face of criticism from practitioners in other fields. Perchard et. al. try to move this process forward by explaining to historians how their work could or should be explained differently (not necessarily done differently) to assist non-historians in assessing and appreciating its value. Here they remind us of the work of Andrews and Burke (2007) whose ‘five Cs’ (change over time, causality, context, complexity, and contingency) provide a useful guide to help non-specialists appreciate the aspects that historians are likely to fix upon as explanatory variables. The authors also point to the work of Jones and Khanna (2006) and Maclean, Harvey, and Clegg (2016) as helpful in making historical work relevant to mainstream business and management studies.

The article is a valuable contribution to the on-going effort to bring management and business historians closer to those studying and theorizing about management and business activity. Its relevance touches on a number of critical issues both in the academic field of study and related to the career development of those engaged in this kind of research.

References

Andrews, Thomas and Burke, Flannery (2007), “What Does it Mean to Think Historically?”, Perspectives in History, available at: https://www.historians.org/publications-and-directories/perspectives-on-history/january-2007/what-does-it-mean-to-think-historically

Duara, Prasenjit (1998), “Why is History Anti-theoretical?”, Modern China, 24(2): 106.

Jones, Geoffrey and Khanna, Tarun (2006), “Bringing History (Back) into International Business,” Journal of International Business Studies, 37(4): 453-68.

Maclean, Mairi, Harvey, Charles and Clegg, Stewart (2016), “Conceptualizing Historical Organization Studies,” Academy of Management Review, 41(4): 609-32.

Palm Oil, Rubber, and Colonialism

The Emergence of an Export Cluster: Traders and Palm Oil in Early Twentieth-century Southeast Asia

by Valeria Giacomin (Harvard-Newcomen Fellow in Business History during 2017/2018)

Abstract: Malaysia and Indonesia account for 90 percent of global exports of palm oil, forming one of the largest agricultural clusters in the world. This article uses archival sources to trace how this cluster emerged from the rubber business in the era of British and Dutch colonialism. Specifically, the rise of palm oil in this region was due to three interrelated factors: (1) the institutional environment of the existing rubber cluster; (2) an established community of foreign traders; and (3) a trading hub in Singapore that offered a multitude of advanced services. This analysis stresses the historical dimension of clusters, which has been neglected in the previous management and strategy works, by connecting cluster emergence to the business history of trading firms. The article also extends the current literature on cluster emergence by showing that the rise of this cluster occurred parallel, and intimately related, to the product specialization within international trading houses.

Freely available at Enterprise and Society, Volume 19, Issue 2, June 2018, pp. 272-308

Review by Helena Varkkey (University of Malaya)

In this article, Giacomin presents an archive-based historical analysis of how palm oil became one of the most important traded commodities from Southeast Asia to the world in the early- to mid-1900s. She uses the cluster (defined as a “geographically proximate group of interconnected companies and associated institutions in a particular field, linked by commonalities and complementarities”) approach to explain how the organizational structure of the pre-existing rubber cluster in the Malay peninsula (at the time a British colony) and Sumatra (under the Dutch) formed the basis for an emerging palm oil cluster in the same geographical region.

The first part of her paper focuses on how the regional rubber cluster structure developed. While the literature commonly credits unique local factors in the development of such clusters, Giacomin instead looks at nonlocal factors: (1) mainland Chinese traders that controlled regional trading routes from major Southeast Asian ports and brought in low-skilled tappers and harvesters from surrounding territories, and (2) Western traders that brought in capital inputs (seeds, machinery and finance) and highly-skilled human resources (estate managers and engineers) from Europe and other parts of the Empire and established headquarters in major European trading ports, allowing them to access crucial market information on demand. Both of these foreign merchant communities congregated in the emerging trading hub of Singapore, strategically located in between British Malaya and Dutch Sumatra, and developed a mutual dependency: Chinese contacts were vital for Western traders wanting to run a business in the Eastern colonies, while the Chinese needed Western traders to scale up their region-based commercial activity to a global scope.

The second part of the article explains how palm oil became the “spin-off” crop of the rubber cluster in the region. During the natural rubber boom in the early 1900s, the Malaya-Sumatra rubber cluster became over-dependent on this export and thus over-specialised in terms of existing practises, agronomic knowledge (through R&D agencies like the Rubber Research Institute), and coordinating institutions (eg. the Rubber Growers’ Association). When the advent of synthetic rubber in the 1920s caused natural rubber prices to fall, companies desperately looked to diversify their production to recoup and replace their losses. However, this over-specialisation meant that they could consider only a limited range of crops similar to rubber for diversification. As it happened, the rubber estate structure could be conveniently repurposed for into oil palm estates. Furthermore, the oil palm flourished in a much narrower latitude span compared to rubber, giving confidence to companies that the demand for palm oil would be more sustained since supply would be more limited.
Giacomin concludes that even though the literature often regards over-specialisation as fatal, in the case of the Southeast Asian rubber cluster, this serendipitously led to the emergence of one of the most enduring regional clusters serving the global economy. Today, Indonesia and Malaysia account for over 90% of global palm oil exports.

While the significance of the rubber sector in paving the way for palm oil in Malaysia and Indonesia is well known, this paper remains an important and significant addition to the current literature, not only on general business management and strategy, but also more specifically in terms of (1) palm oil expansion and development and (2) agricultural systems (estate vs. smallholdings).

Firstly, the specific role of nonlocal (especially Chinese) entrepreneurs in connecting the production areas in this region to the consumption areas in the West was previously not well understood. In the context of the “global north” and “global south”, palm oil can be considered a uniquely “southern” vegetable oil. Compared to other oils like sunflower, rapeseed, and soya bean, the production, major business players, beneficiaries and direct impacts of palm oil is situated more comparatively in the global south. Giacomin alludes to this in reference to the narrow latitude where oil palm can be grown. This northern/southern framing has coloured much of the recent debates and controversies over palm oil today. This paper’s analysis on the historical role of Chinese merchants is especially useful in further informing the idea of palm oil as a “southern” oil, while at the same time, the equally important role of the Western merchants that Giacomin highlights may be useful in moderating certain northern arguments in this ongoing debate.

Secondly, the historical nature of Giacomin’s analysis of this sector is especially timely in the current period where other regions, like West Africa and Latin America, are looking to increase their global trading share of palm oil. Giacomin mentions that even though the oil palm originated from and was first produced commercially in West Africa during colonial times, Western African territories were unable to effectively penetrate global markets because they did not display the same institutional cohesion across neighbouring territories, something that Southeast Asia managed to do through the pre-existing rubber cluster. This “cluster” model may thus provide an exemplar to be used by emerging palm oil production regions and companies as an effective way to possibly break the current oligopoly (Indonesian and Malaysian firms) which is the palm oil industry. Especially for West Africa, which is considered the current “greenfield” area for palm oil outside Southeast Asia, current strategies can be developed to avoid past mistakes.

Finally, Giacomin’s analysis of early smallholders is useful to inform current discussions on the ideal agricultural systems for oil palm. Her paper argues that in the mid-20th century, the fact that the palm oil was an estate crop (involving high costs and favouring large-scale production) provided a solution to the problem previously faced by rubber companies that were facing competition from and losing market share to rubber smallholdings. While this might have been the case historically, oil palm today has been successfully adapted to the smallholder model in both Indonesia and Malaysia, with a significant share of both countries’ production (about 40% each) coming from either organised or independent smallholders. Giacomin’s analysis stops at the early decolonialisation period, before the newly independent nations began to formulate oil palm smallholder schemes as a strategic tool for rural development and poverty eradication for both countries. Her analysis however can serve as a useful starting point in the ongoing debates on if and how both the estate and smallholder systems can co-exist efficiently and in harmony.

Overall, this paper is a valuable piece of business history that helps to further shed light on a controversial agro-economic sector often shrouded in secrecy. The fact that palm oil continues to be a hot topic worldwide today underlines the relevance and importance of such forays into history to inform the present.

From VoxEU – Wellbeing inequality in retrospect

A must read – Leandro Prados de la Escosura (databases and open access book on inequality)

The Long Run

Rising trends in GDP per capita are often interpreted as reflecting rising levels of general wellbeing. But GDP per capita is at best a crude proxy for wellbeing, neglecting important qualitative dimensions. 36 more words

via Wellbeing inequality in retrospect — VoxEU.org: Recent Articles

To elaborate further on the topic, Prof. Leandro de la Escosura has made available several databases on inequality, accessible here, as well as a book on long-term Spanish economic growth, available as open source here

View original post

Evaluating the Distinctive Economic Impact of Historical Female Migration in the United States

A Woman’s Touch? Female Migration and Economic Development in the United States

By Viola von Berlepsch (London School of Economics), Andrés Rodríguez-Pose (London School of Economics) and Neil Lee (London School of Economics),

Abstract: Does the economic effect of immigrant women differ from that of immigrants in general? This paper examines if gender has influenced the short- and long-term economic impact of mass migration to the US, using Census microdata from 1880 and 1910. By means of ordinary least squares and instrumental variable estimations, the analysis shows that a greater concentration of immigrant women is significantly associated with lower levels of economic development in US counties. However, immigrant women also shaped economic development positively, albeit indirectly via their children. Communities with more children born to foreign mothers and that successfully managed to integrate female immigrants experienced greater economic growth than those dominated by children of foreign-born fathers or American-born parents.

URL: https://econpapers.repec.org/paper/cprceprdp/12878.htm

Circulated by NEP-HIS on 2018-05-08

Review by Fernando Arteaga (George Mason University)

Summary

What is the economic impact of female migration? The authors seek to answer the inquiry by using the United States in the late 19^th and early 20^th century as their study case. The goal of the paper is to highlight the distinctiveness of women immigration (compared to that of men), both in the processes that led women to migrate, the characteristics they had, and the places where they finally settled. The main thesis of paper stresses the long-lasting effect women have had; through their family role, as mothers, they facilitated the formation and transmission of social capital, which had a pervasive positive effect on income.

Female migrants in early America tended to settle mainly on urbanized areas in the Northeastern coast – compared to that of male immigrants (who settled mainly in the South and West). The migration levels of women, in absolute terms, were lower, and their marriage rates higher. More importantly, their labor participation rates were low: women tended to stay and work in domestic chores rather than find occupations in the market. These characteristics make female migration distinct to that of men, and motivate the goal of the paper in trying to assess their particular relevance.

 

 

Figure 1: Immigrant Women in the United States, 1880. By the nature of the variable, counties with a larger share of immigrant women imply a lower share of immigrant men.

 

The text relies on standard econometric analyses, based on intuition and on the literature of migration and culture transmission. The main data sources are the historical censuses of 1880 and 1910, which capture the amount of, male and female foreign-born population residents in each US county (among other data).  The paper presents two base regressions that aim to assess the direct and indirect economic impact of female migration, both in the short and the long term. The first model regresses economic income (GDP per capita by county) on female migration (foreign-born women as a share of the total population in the county). They find that the variables are negatively correlated: lack of labor market participation hindered the female contribution to income. The authors also found that this has had a long term negative effect (income today is also negatively correlated with female migration in late 19^th and early 20^th century) [1]. To correct for potential biases and to establish a causality linkage that goes strictly from migration to income, and not the other way around, the authors use three different instruments: a) the percentage of married persons; b) the number of persons living in a household; c) the urbanization rate of the county being examined. The first instrument accounts for the fact that female migrants tended to be married in larger shares than the rest of the population. The second accounts for the idea that migrants, especially women, tended to stay with members of their families through their lifetime. The last one maintains that female migrants favored settlement in urbanized zones. The validity of an instrument (marriage percentage, household size and urbanization) hinges upon it being correlated with the dependent variable (income) only by it causing the highlighted mechanism (female migration). The authors do several post-hoc statistical tests to evaluate the instrument’s validity and conclude that it is indeed a valid and strong one. In any case, the instrument variable outcomes do not change the results of the baseline ordinary least squares scenario, they just allow a more robust interpretation of them: it can be said that female migration did have a negative impact on income.

The second model emphasizes the indirect impact of migrant women. Maybe women themselves did not positively contribute to the economic wellbeing of their communities, but they could have done so through other means. The authors refer to the literature that stresses how mothers influence their children behavior and thus have an important role as social capital transmitters (which could positively affect economic wellbeing).  They regress economic income today on the share of children (in 1880 and 1910) born from: 1) a migrant mother and an American-born father; 2) a migrant father and an American mother; 3) both American parents. The standard base of comparison is the share of children that had both parents as immigrants [2].  By definition, the model can only capture the long-term effect of female migrants. The authors find that US counties with an historical larger share of children with migrant mothers are correlated with larger incomes today – in comparison to the other explanatory variables; having American parents is negatively correlated with income today; having a migrant father, and American mother, has a non-significant and null effect on economic outcomes today. The argument, again, rests on the case of social capital transmission: women, as mothers, matter very much.  To corroborate their OLS results they also use an instrumental variable. The authors assume that American-born women that had migrant mothers followed the cultural transmission pattern established by their forebears. They call this the “supply-push” component, which they estimate and use as their instrument. Just as the first model, the instrumental variable inclusion does not modify the basic results, it only permits to talk about causality from migrants in the past to better economic outcomes today.

In conclusion, the paper finds that female immigration, while having a negative direct short-term impact on economic income, has a long-lasting positive effect through the “cultural carrier” channel.

Comment

The paper is a very interesting one, being one of the few studies that aims to disentangle the impact of women as migrants compared to that of men. The results the authors present make intuitive sense. I would like to make just small technical comments based on the variables they use and how they use them.

First, related to the semantics of the concept of “migration.” Migration is normally thought as a flow variable, but here it is used as a stock variable. Given the data they use (measuring migrants as people classified as foreign born in two censuses) the authors cannot measure the impact of migration as a flow, only the impact of it in broad terms. This is not a problem. I just would have liked to see a minor explanation on the paper that clarified the interpretations that we could get out of this. In fact, I think it could explain why they find a negative impact of migrant women in income (if the variables were flows, through migration rates and economic growth, the results may be different).

Second, on a more technical note, I’m skeptic of the instruments being used. Even though the authors argue that they are valid and strong, I remain unconvinced. The authors show that all four of them are correlated with the dependent variable and uncorrelated with the error terms, yet there is almost no explanation, backed up by a narrative, of how exactly these instruments impact on income only through female migration. For each one of the instruments used I could think of other alternate channels by which they could impact income. For example, the use of percentage of marriage by county could indeed be correlated with female migration, but is that the only potential channel? Could it not be that maybe poverty or religion could be impacting income as well?

Lastly, I wish the narrative part could be explained in larger detail. For example, how exactly female migrants in 1880 have a direct impact on income in 2010. Or how exactly children of foreign mothers in 1880 and 1910 could affect income today. It is one thing to say that culture matters, it is another different thing to point how exactly it does. In fact, even though they do mention the pervasiveness of cultural traits through time, they fail to mention that this pervasiveness does not imply ipso facto a good outcome is assured. Sometimes, social capital is also correlated with bad outcomes.

[1] The authors do not provide a concise explanation of why this could be happening: how could a century year old female migration pattern directly impact economic wellbeing today?

[2] All the interpretations of results are in comparison to that baseline.