Within-Group Heterogeneity in a Multi-Ethnic Society

Nada es Gratis article by Miriam Artiles ’15 (Economics)

Photo by Adrian Dascal on Unsplash

Editor’s note: this article was originaly published in Spanish in the popular economics blog, Nada es Gratis, and is based on Miriam Artiles’s PhD from Universitat Pompeu Fabra. Her Job Market Paper was honorably mentioned in the third annual Nada es Gratis Job Market Paper awards.

Is ethnic diversity good or bad for economic development? When different languages, ethnicities or races coexist in the same society, there are challenges for the economy, but also opportunities. On one hand, if individuals within ethnic groups are homogeneous, and groups differ in preferences toward policies or public goods, then conflicting preferences can lead to inefficiencies in public good provision or to policy choices that may not benefit the entire society. Inter-group tensions can also result in civil conflicts or exacerbate mistrust and lack of cooperation. However, on the other hand, if ethnic groups differ in subsistence activities or skills, then complementary specializations can generate economic gains, stimulate innovation, and promote inter-group trade. Alesina and La Ferrara (2005) provide a review of this literature. While there is a general understanding that diversity brings opportunities and challenges, there is scarce evidence on which factors determine its positive or negative consequences. When is ethnic diversity good for economic development, and when is it bad?

I ask whether the effect of ethnic diversity on economic development depends on one characteristic of ethnic groups that has received little attention: the heterogeneity of individuals within ethnic groups. Underlying previous literature is the assumption that individuals within ethnic groups tend to be homogeneous. However, individuals may differ in many dimensions, including preferences, economic activities or skills, as well as cultural, genetic, and linguistic traits. I focus on having different economic specializations and skills within the same ethnic group, and I study whether ethnic groups with more heterogeneous individuals do better in multi-ethnic societies.

Consider two ethnic groups, A and B. The two groups differ in ethnicity. In turn, ethnic group A has individuals with diverse skills due to their different economic specializations, while ethnic group B is more homogeneous (individuals from ethnic group B have similar skills). The idea is that it may be easier for individuals of ethnicity A to live and to interact in a multi-ethnic society–they come from an ethnic group that is already highly heterogeneous. They will already be used to diverse environments. They will be more familiar with having to interact with heterogeneous individuals. If you come from an ethnic group that is highly heterogeneous, in terms of skills, you may be more willing to live and to interact with other ethnicities. In this case, positive interactions, mutually beneficial exchange, between ethnic groups will become more frequent.

The 16th Century resettlement of Peruvian ethnic groups

To study this, I collect new data on a natural experiment from Peru’s colonial history. I focus on highland Peru. There, Spanish colonizers resettled native populations in the 16th century. They forced together different ethnic groups in new villages, and this happened unintentionally. Importantly, in some ethnic groups, individuals had already been living in very different ecological zones of the Andes, at different altitudes, during the pre-colonial period, before the Spanish conquest. This creates within-group heterogeneity. In some cases, individuals from the same ethnic group were very different in terms of ecological specializations and skills – the types of lands and crops that they were used to cultivate. In other ethnic groups, everyone lived in the same climate zone, at the same altitude. I am asking: did the more heterogeneous ethnic groups do better once they were resettled in multi-ethnic villages?

Firstly, I use a map of the spatial distribution of ethnic groups at the time of the Spanish conquest. It allows me to compute the distance from each village to the closest ethnic frontier and use it as a source of quasi-random variation in ethnic diversity. During the pre-colonial period, individuals from the same ethnic group were distributed vertically, at different altitudes. This is the thesis of the anthropologist John Murra. He documents this vertical settlement pattern as a subsistence strategy in an environment in which differences in elevation create a variety of ecological zones and climates. At the time of the resettlement, the mountain environment of the Andes was new to Spanish colonizers – they were used to a flatter world. As a result, in villages that were created close to ethnic borders, they concentrated individuals from different ethnicities unintentionally (Pease 1978; Wachtel 1976). Secondly, I use spatial data on the distribution of ecological zones to compute a proxy for the heterogeneity of skills within each ethnic group prior to the conquest. It is important to note that ethnic groups with more heterogeneous skills may be different in other dimensions (e.g., group size, population density, etc). In the analysis, I use all the available data on the pre-colonial characteristics of ethnic groups to account for the main correlates of within-group heterogeneity.

The first result in the paper documents the direct effect of ethnic diversity, which I benchmark against previous results in the literature. I find that ethnic diversity is robustly associated with lower living standards in the long run. Specifically, I explore a variety of outcomes that capture contemporary living standards. As proxies for local economic activity, I use light intensity per capita (2000-2003) and a measure of non-subsistence agriculture from the agricultural census of 1994. For access to public infrastructure, I use data from the 1993 population census on access to public sanitation and the public network of water supply. This result is in line with the literature on the costs of ethnic diversity, though it also highlights the persistent consequences of forced diversity at the local level. When examining the effect of ethnic diversity and within-group heterogeneity, I find the following pattern:

The figure shows the average effect size of ethnic diversity as a function of within-group heterogeneity. I find a robust pattern: the more heterogeneous an ethnic group was prior to resettlement, the lower the cost of ethnic diversity. On average, where ethnic groups have more heterogeneous individuals in terms of skills, the negative effect of ethnic diversity is reduced, and ethnic diversity may even become an advantage for economic development. To understand the evolution of these long-term effects, I use data from the 1876 population census on occupations and literacy rates, showing that the documented pattern persists over time.

Why is this happening? When exploring potential channels, I find evidence consistent with cultural transmission. Individuals from more heterogeneous ethnic groups in terms of skills are more likely to interact with other ethnicities. Using data from colonial records, I find evidence suggesting cooperative behavior and more open attitudes when interacting with other ethnic groups. Overall, understanding whether individuals from more heterogeneous ethnic groups are better able to integrate in a multi-ethnic society is a relevant question, not only in an increasingly globalized world, but also in the context of forced displacements and migrations, like in the case of refugees.


Alesina, A., and La Ferrara, E. (2005). “Ethnic Diversity and Economic Performance.” Journal of Economic Literature, 43 (3), pp. 762-800.

Murra, J. V. (1975). Formaciones económicas y políticas del mundo andino. Instituto de Estudios Peruanos.

Pease, F. G. Y. (1978). Del Tawantinsuyu a la historia del Perú, Instituto de Estudios Peruanos.

Wachtel, N. (1976). Los vencidos: los indios del Perú frente a la conquista española (1530-1570). Alianza.

Connect with the author

Miriam Artiles ’15 is a PhD candidate in Economics at Universitat Pompeu Fabra and will soon start as an Assistant Professor at Pontificia Universidad Católica de Chile. She is an alum of the Barcelona GSE Master’s in Economics.

This post was edited by Ashok Manandhar ’21 (Economics).

Environmental Investment Tax Incentive Reform in Spain: a Lost Opportunity?

Nada es Gratis article by Kinga Tchórzewska ’15 (Economics)

Photo by Clayton Cardinalli on Unsplash

Editor’s note: this article was originaly published in Spanish in the popular economics blog, Nada es Gratis, and is based on Kinga Tchórzewska’s PhD from the University of Barcelona. Her Job Market Paper was honorably mentioned in the third annual Nada es Gratis Job Market Paper awards.

Firms are often reluctant to invest in green technology. As for the reason why – they point to high fixed costs and the resulting capital market failure. However, instruments that could possibly address this problem, such as environmental investment tax incentives, are not very popular among regulators – even though they may offer an interesting alternative to environmental taxation or even investment subsidies, since tax incentives are easier to implement at the administrative level. Could Environmental Investment (EI) tax incentives be successful at encouraging green investment? And how do firms react to the modifications in existing EI tax credits with respect to employment and innovation decisions? I try to tackle those questions using the EI tax credit reform in Spain.

Spanish Environmental Investment (EI) Tax Credit

Spain is a very interesting country in which to study a large-scale national tax incentive program because the EI tax credit went through some unusual transformations over the years of its existence. The specific EI tax incentive analyzed in this paper was first introduced in 1996 at 10% of the firm’s level of investment and survived in such form until 2006, when its slow phase-out was announced. The phase-out was implemented as the then government believed that the tax incentive was mostly financing end-of-pipe technologies (which do not affect the production process but purely reduce the pollution level at the end of the production line e.g. filters and sulphur scrubbers) rather than cleaner production technologies, very often required by law already. The phase-out was then successfully continued until tax credit’s complete elimination in January 2011. Unexpectedly, in March of 2011, this tax credit was re-introduced for 4 more years at the stable rate of 8% investment level. It was possibly done to mediate the effect of the financial crises on the industrial sectors. Figure 1. shows the chronology of events and the expected versus actual rates of the tax credit.

What makes this EI tax credit reform especially interesting is that it generated a lot of confusion until the very last moment and while introduced in March 2011 – it was done specifically with the intention to favor cleaner production over end-of-pipe technologies. In the analysis, I focus on industrial firms as the main beneficiaries of the program and consider the time period between 2008 and 2014. In the first part, I compare firms’ behavior before and after the change in this policy instrument using difference-in-difference analysis. This will show if the modification of the tax credit discouraged end-of-pipe technologies as well as how the policy reform affected green employment. In the second part of the analysis, by using instrumental variable approach with difference-in-difference, I examine the proportionate effect of an increase in the amount of the tax credit. I study its proportional effect on firms’ investment, employment and R&D outcomes. Thus, I perform the first quasi-experimental econometric analysis of the effectiveness of EI tax credit at encouraging adoption of green technologies directly, but also indirect green employment and green R&D effects.


I find evidence that firms did in fact decrease their investment through the tax credit in the end-of-pipe technologies as a result of the policy change. This also includes the technologies specifically reducing air-pollution alone such as filters/sulphur scrubbers. We can, therefore, conclude that the modification was implemented quite successfully. That being said, there is no evidence to support the claim that this policy change led to an increase in the investment in cleaner production technologies. Unfortunately, the policy change also had a few unexpected indirect effects. It appears that firms reduced the number of green employees as well as the expenditure associated with the salaries of green employees, as can be seen in Figures 2a and 2b.

After performing the heterogeneous analyses, it is also clear that firms responded differently depending on their size – Figures 3a and 3b. More specifically, small firms seem to have benefited the most from the policy change, by considerably increasing their investment in cleaner production technologies. The opposite has happened to the large firms, who decreased their investment in the cleaner production technologies through the modified tax incentive.

By studying the proportional effect of the EI tax credit on investment outcomes it becomes apparent that Spanish environmental investment tax incentive was generally successful at inducing all types of green investment. This means that even during times of financial crises tax credit was drove firms’ green investment. However, they favored air-pollution-reducing over energy-efficient technologies, not necessarily end-of-pipe over cleaner production technologies, as per the concern of the government at the time. Additionally, I find further evidence that the increase in the amount of environmental investment tax credit results in a proportionate increase in the number of green employees and even private environmental R&D. Those indirect effects are quite hopeful, showing that a successful EI tax credit can also drive employment and create positive externalities through R&D.


This analysis provides a multitude of important implications for policy makers. Firstly, it encourages the usage of EI tax credits, which is also in agreement with previous literature, especially the work done by Ohrn (2019). However, this is in stark contrast to the decision of the Spanish government to eliminate this fiscal incentive from the Spanish Corporate Income Tax completely. This analysis supports its continuous use and perhaps even further modification, rather than a complete phase-out.

What we can learn from this green tax incentive is quite straightforward – adopting green depreciation incentives leads to increased business incentives and green employment outcomes, even during times of economic downturn. Additionally, the government can be successful at modifying the existing tax incentives, such that they discourage those technology choices that the central government considers undesirable. While the results clearly indicate that the tax credit should have been redefined even further so as to also encourage more investment in cleaner production technologies, this empirical work does not justify its complete phase-out. The fact that there is an increased investment in cleaner production technologies for smaller firms is also very important, as those are exactly the companies frequently faced with capital market failure – especially in the time of financial recession such as this one.

Of course, more research is needed to assess whether these types of incentives are the most efficient way to improve firms’ economic outcomes, and how the tax credit also affected their employees over the short- and long-run – especially after the complete elimination of the tax credit in 2015. Lastly, even given the financial burden that tax deductions and subsidies entail, they might still be economically justified in some cases. For instance, when positive externalities appear, such as increased green private R&D, which is the case here.


Ohrn, E. (2019). The effect of tax incentives on US manufacturing: Evidence from state accelerated depreciation policies. Journal of Public Economics, 180, 104084.

Connect with the author

Kinga B. Tchórzewska ’15 is a Postdoctoral Researcher at the ZEW – Leibniz Centre for European Economic Research. She is an alum of the Barcelona GSE Master’s in Economics.

This post was edited by Ashok Manandhar ’21 (Economics).

The importance of norms for development

Essay by Oguz Korkut Keles ’20 (Economics)

Photo by Anthony Garand on Unsplash

The relationship between institutions and development is a long-standing topic in economic research. However, economists have tended to only evaluate formal institutions (such as laws and property rights), neglecting the informal (like conventions and norms). This overspecialisation precludes the analysis of ideas and ideologies. Without considering these abstract drivers of development, the space for ethically and politically dangerous explanations of success appears (such as for genetic reasons).

Contrary to recent literature, I argue that informal constraints are actually the basis of institutions and therefore the real generators of growth and development. I show this by examining revolutions – the cauldrons where new systems, ideas, and conventions begin, and old ones end.

The illusion of separation

Scholars of comparative development have noted the increasing divergence between developed and developing countries: the gap between Northern and Southern Europe and the underdevelopment of the Middle East and sub-Saharan Africa being major examples. Several theories attempt to explain this divergence, considering possible factors such as geographical characteristics and institutional differences. Notably, comparative development has even been attributed to levels of genetic diversity (Ashraf & Galor, 2013).

In particular, the crucial historical link between institutions and development is well known. Famous examples include the advantage of limited royal power (Acemoglu, 2005); reformed constitutional arrangements and strengthened property rights (North & Weingast, 1989); and the balance of power between merchants and princes (De Long, 1993). Yet, these studies put their emphasis on formal rules, neglecting the norms, ideas and ideologies that underwrite them.

The latter are fundamental elements of institutions as they influence formal rules. In a seminal contribution to institutional economics, North (1994) distinguishes between two forms of institutions: formal rules (constitutions, laws, property rights etc.) and informal constraints (norms of behaviour, conventions, self-imposed codes of conduct etc.). In a later work, he argued that institutions evolve incrementally and successively over time (North, 1991). When those two forms are approached as two separated sources of institutions, the role of informal constraints in institutional evolution will be missed, throwing a veil over a core aspect of institutions and leading us to fallacious conclusions about the key determinants of growth and development.

Similarly, Karl Popper (1945) distinguishes an open society from a closed society based on whether a distinction exists between normative laws and natural laws. Where there is none – what Popper calls a closed “tribal society”– taboos and conventions act as if they were natural law. This gives them a powerful role in society and a fundamental role in development. By creating formal laws, societies recognise the distinction between norms and natural laws, weakening the effect of conventions (although, as we will see, they still act through both formal and informal laws).

Both North and Popper agree on this chronological development of institutions meaning a better understanding of causation is needed. Myrdal (1978) convincingly argues that the mechanisms of social systems are determined by an endogenous cycle of causation that affects the distribution of power in a society and economic, social and political stratification). This means that a change in informal constraints will alter formal rules, which will then return to affect the former. Therefore the scaffolding of institutions consists of norms of behaviour, conventions and self-imposed codes of conduct.

The revolutionary crevasse

Just as a crevasse provides a glimpse deep into the ice, revolutions open a window to the creation and destruction of social systems. Revolutions are beloved by social scientists (especially in institutional economics) as they provide natural experiments to investigate causal effects. They can shed some light on the importance of norms and convention, as well as their relationship with ideas, ideologies and leaders.

In the literature, for example, Acemoglu et al. (2008) and North & Weingast (1989) have respectively examined the impact of the French Revolution on development and the Glorious Revolution on institutional structure. However, these types of studies have focused only on the secondary changes (in laws and property rights) instead of the initial causes of change (norms and conventions). In this regard, a re-evaluation of revolutions and their characteristics is necessary to observe the initial changes.

Let us first consider which elements prepare amenable conditions for the emergence of revolutions. Gottschalk (1944) identifies three broad factors:

  1. demand for change stemming from (a) personal discontent and (b) social dissatisfaction
  2. hopefulness derived from (a) popular programs of reform and (b) a leader
  3. weakness of the conservative forces – perhaps the most important.

Demand comes from widespread provocations (corruption, taxation, poor infrastructure etc.) which generate social dissatisfaction. Yet, demand by itself is not sufficient for the revolution. Some hope of success is also needed. This comes from programs of reform, as provided by the Voltaires and Rousseaus, the Lockes and Ademses, and the Marxes and Kropotkins (Gottschalk, 1994). However, tuneless emphasis on widespread provocations that are based on the formal rules underestimates the phycological mechanisms that are mainly based on informal constraints.

Personal discontent (arising for idiosyncratic reasons) only appears at the individual level yet plays an essential role in generating the leaders of revolutions. These leaders then support the new-born ideas and ideologies based on the program of reform which has a multiplier effect by coherently spreading revolutionary sentiment. This is crucial once we think of revolutions as risky events over which individuals have varying valuations of the possible outcomes.  Gneezy et al. (2006) show that individuals, faced with a complex choice, may choose to stay in the old system if they value the risky benefits of revolution less than the worst outcomes of rebelling. However, once the revolutionary “lottery” is based on intellectuals’ programs of reforms and explained by leaders it becomes easier to code.

In this way, agents facing complex task (in this case revolution), might act following the leader through many of the channels identified by behavioural economics such as Tversky and Kahneman’s simple heuristics, Walker and Wooldridge’s conventions and Shiller’s narratives. These share common features which affect the majority’s decision-making processes – especially when tasks are complex.

This process is essential to notice the importance of informal constraints and how they become formal rules since leaders are the symbol of ideologies and ideas. As Axelrod indicates norms precede laws and laws strengthen norms. After the success of the revolution laws strengthen the norms through formalization. And after social conventions are entrenched, they become thoughtlessly accepted by individuals (Epstein, 2001).

As with the example of revolutions, before a change in the formal rules, an ideological revolution has occurred when intellectuals provide the programs of reforms. The ideas become conventions during the revolution, changing societal expectations. Notions of equality and liberty – in the case of the French Revolution – became the convention as the system was upended. The relationships between ideas, ideologies, norms and leaders encourage us to take them into account when evaluating growth and development.  


I have argued that ideas, norms and ideologies are the initial drivers of development and have had an immense effect on our civilizations. However, traditional political economy’s overemphasis of formal rules fails to capture this. The insularity of this approach is highlighted by examining Revolutions, which provide evidence in favour of more inclusive definitions of institutions and the importance of ideas, ideologies and leaders in creating social systems. Therefore, I contend that a more holistic approach to analysing development is required otherwise alternative and ill-founded explanations of growth with remain.


Acemoglu, D., Cantoni, D., Johnson, S., & Robinson, J. A. (2008). From ancien regime to capitalism: the French Revolution as a natural experiment. Natural Experiments…, op. cit, 221-256.

Acemoglu, D., Johnson, S., & Robinson, J. A. (2005). The rise of Europe: Atlantic trade, institutional change, and economic growth. American Economic Review95(3), 546-579.

Ashraf, Q., & Galor, O. (2013). The ‘Out of Africa’ hypothesis, human genetic diversity, and comparative economic development. American Economic Review, 103(1), 1-46.

Axelrod, R. (1986). An evolutionary approach to norms. The American Political Science Review, 1095-1111.

De Long, J. B., & Shleifer, A. (1993). Princes and merchants: European city growth before the industrial revolution. The Journal of Law and Economics36(2), 671-702.

Epstein, J. M. (2001). Learning to be thoughtless: Social norms and individual computation. Computational economics18(1), 9-24.

North, D. C. (1991). Institutions. Journal of Economic Perspectives5(1), 97-112.

North, D. C. (1994). Economic performance through time. The American Economic Review84(3), 359-368.

North, D. C., & Weingast, B. R. (1989). Constitutions and commitment: the evolution of institutions governing public choice in seventeenth-century England. The Journal of Economic History, 49(4), 803-832.

Popper, K. R. (1945). The open society and its enemies. Routledge, London.

Myrdal, G. (1978). Institutional economics. Journal of Economic Issues12(4), 771-783.

Gottschalk, L. (1944). Causes of revolution. American Journal of Sociology50(1), 1-8.

Connect with the author

Oguz Kortut Keles ’20 is an alum of the Barcelona GSE Master’s in Economics.

This post was edited by Ashok Manandhar ’21 (Economics).