Why relatively poor people are not more supportive of redistribution

Redistribution of income from richer people to poorer people seems like something that poorer people should be more supportive of than are richer people. After all, they would benefit directly from such redistribution, and it would reduce inequality of income. However, as I've noted in posts before (see here and here), perhaps people don't prefer lower inequality, they prefer greater fairness. So, people's preferences for redistribution are somewhat more complicated than simply being based on their relative position in the income distribution.

A new article by Christopher Hoy (Australian National University) and Franziska Mager (Oxfam Great Britain), published in the American Economic Journal: Economic Policy (ungated earlier version here), provides some real food for thought on the topic of preferences for redistribution. They first note three assumptions that underlie the seminal theories of preferences for redistribution:

Firstly, most people are averse to large income differences in their country, and this is particularly the case for relatively poor people. Secondly, people who are more concerned about inequality tend to be more supportive of redistribution. Thirdly, poorer people should be more supportive of redistribution than richer people, as they benefit directly.

Hoy and Mager note that most people have a 'median bias' in their belief about their position in the income distribution. That is, poorer people tend to believe that they are richer than they really are, while richer people tend to believe that they are poorer than they really are. Hoy and Mager essentially tests what happens when you give people information about their actual position in the income distribution, on preferences for redistribution. As they explain:

Surveys across a range of high-income countries have shown that most people tend to think that they are positioned around the middle of the national income distribution regardless of whether they are rich or poor... This raises the question that if relatively poor people were made aware of their position in the national income distribution, would they be more concerned about inequality and supportive of redistribution?

We test how informing people that they are relatively poorer than they thought impacts their concern about inequality and support for redistribution through a randomized survey experiment with over 30,000 respondents in 10 countries (Australia, India, Mexico, Morocco, the Netherlands, Nigeria, South Africa, Spain, the United Kingdom, and the United States)... Respondents in each country were randomly allocated to receive either information about their position in the national income distribution (treatment group) or no information (control group). Prior to the treatment, respondents revealed their perception of the level of national inequality, their preferred level of national inequality, and their perceived place in the national income distribution... After the treatment, respondents were asked standard questions from the existing literature regarding their views about whether the gap between the rich and poor is too large in their country and whether they think the government is responsible for closing this gap...

Hoy and Mager note that:

Seminal theories of preferences for redistribution imply that informing people that they are relatively poorer than they thought would lead to greater concern about inequality and support for redistribution...

However, they find in their sample that:

...respondents in the poorest two quintiles of the national income distribution who were told they are relatively poorer than they thought are less concerned about the gap between the rich and poor in their country and are not any more supportive of the government closing this gap compared to respondents in the control group. This result occurs in seven countries (India, Mexico, Morocco, the Netherlands, Nigeria, South Africa, and Spain), and there was no effect caused by this information in the remaining three countries (Australia, the United Kingdom, and the United States).

In other words, this is the opposite of what the seminal theories would predict. Hoy and Mager then use a theoretical model to explain what might be causing these surprising results:

We illustrate that the likely channel causing the effect is people using their own standard of living as a “benchmark” for what they consider acceptable for others... People had perceived themselves to have an “average” living standard compared to other people in their country prior to the treatment even though they were actually relatively poor. Their previous assessment of their relative status implies that they thought there were similar shares of people poorer and richer than them in their country (this is as a result of placing oneself as being around the middle of the national income distribution). Upon receiving the treatment, this led people to realize two points. Firstly, there are fewer people in their country poorer than them than they had thought. Secondly, what they had considered to be an average living standard (i.e., their own standard of living) is actually relatively poor. In other words, relatively poor people consume more than they thought. Both of these points would suggest that the treatment provided to respondents would lead them to become less concerned about the living standard of poor people in their country.

Notice that this also accords with the idea that people prefer more fairness in the income distribution. Providing relatively poor people more accurate information about their position in the income distribution might make them believe that the distribution is fairer than they previously thought (after all, they are richer than they thought they were, relatively to others around them), and reduce their preferences for redistribution.

It is rare that a new study overturns some important seminal theory in such a convincing way as this study does. However, it would be easy to overstate what this study tells us about people's attitudes to inequality. In particular, it is important to note that these results are about preferences for redistribution, not aversion to inequality. As Hoy and Mager note:

...the results of our experiment illustrate that relatively poor people’s misperceptions of their position in the distribution do not appear to be lowering their concern about inequality. We provide evidence that the opposite is true. Relatively poor people would be even less concerned about inequality if they knew their true position in the national income distribution. In addition, we show that the elasticity of respondents’ preferences for redistribution to the treatment is substantially greater in middle-income countries than in high-income countries. This is a novel finding, as it suggests the literature on preferences for redistribution may well be qualitatively different if more research were to be conducted in these settings.

And that points to the importance of doing more research on this topic, in a wider range of countries. It also suggests that, if people in developed countries are given accurate information about where they are placed in the global income distribution, they may become more favourably disposed towards global redistribution. This assumes that the effect on relatively richer people works in exactly the opposite manner to the effect that Hoy and Mager find for poorer people (they don't show the effects on richer people in their study, but something similar was found in this study). That would suggest a way for advocacy groups to increase the support in developed countries for foreign aid, or perhaps to increase donations from 'ordinary people' in developed countries targeted at poverty reduction in developing countries. Changing preferences for global redistribution could have substantial global-welfare-improving implications.

Read more:

• Who cares about income inequality?
• Changes over time in attitudes towards redistribution in the U.S.
• Inequality, redistribution, and fairness
• People want fairness, not equality

Does the CORE textbook not go far enough?

As I noted in a post last year, Samuel Bowles (Santa Fe Institute) and Wendy Carlin (University College London) wrote an article that looked across modern economics principles textbooks, and argued for the CORE textbook The Economy (which is the book we use in my ECONS101 class) as an improvement. In their view, CORE 'teaches economics as if the last 30 years happened', rather than simply using the standard models that had been taught for decades prior. CORE represents a big shift in the focus of economics texts (which is part of the reason we use it in ECONS101). But does it go far enough?

In a recent article published in the journal Challenge (sorry I don't see an ungated version), John Komlos (University of Munich) argues that CORE stops short of the fulsome education he would like to see in economics principles. Komlos offers an exhaustive list of 49 additional topics or insights that he would like to see included. There are far too many for me to list or explain here, but some of the key ones include more coverage of political and economic power, more consideration of the sources and consequences of inequality, more on business practices (especially those that disadvantage the consumer) and oligopoly, more behavioural economics, more finance, more economic geography, more ecological economics, and more on the economics of information.

I have some sympathy for Komlos' views. However, there is one aspect of economics that he appears to be forgetting - scarcity. Economics principles is just one course in a student's degree. The textbook they encounter is just one textbook. There is limited lecturing and tutorial time that can be devoted to the range of topics that we might cover in a principles course. There is limited attention and study time that students can devote to reading a textbook. It is not reasonable to expect an economics principles text, or an economics principles course, to cover literally everything. Choices need to be made. If a course attempted to include all of the things that Komlos is arguing it should, then we have to ask, what is the opportunity cost of including those topics? That is, what content would be given up in order to spend more time extensively exploring the overlap between economics and political science, sociology, or geography? That isn't to say that those topics are not important, but the real question is, are those topics so important that they should squeeze something out of an already-packed curriculum?

It is easily possible to include some of Komlos' pet topics into introductory economics. My ECONS101 class includes a lot of the business economics that Komlos advocates. My ECONS102 class includes behavioural economics, the economics of information, inequality and redistribution. However, including the laundry list of additions simply wouldn't be possible across two courses, let alone a single principles course. I think what Komlos is really arguing for is a reconsideration of what is taught across the entire economics major, rather than just in principles. That would be more feasible.

However, there is one additional issue with what Komlos suggests. A lot of the topics he advocates for overlap with other social sciences, including political science, sociology, and geography. My students will know that I have no respect for disciplinary boundaries in my teaching, but adopting Komlos' approach would represent a clear move into economic imperialism, which may not be appreciated by our colleagues in the other social sciences. Komlos could well consider that economics students should receive some of these other perspectives from academics of those disciplines, but he doesn't make that clear. And students in PPE (Philosophy, Politics and Economics) programmes, to the extent that those programmes still exist, will be getting a lot of that content already.

So, what do we take away from this? The CORE textbook The Economy is an improvement on the existing textbooks. Komlos argues that it represents only half of a paradigm shift, and:

...is a further step in the evolution of - and toward a revolution in - economic thought.

I probably wouldn't go that far. But the textbook is at least one tool that we can use to ensure that students receive an economics education with modern features.

[HT: John Komlos]

The minimum wage and poverty and inequality in Brazil

One of the purposes of a minimum wage is to reduce poverty, by ensuring that the lowest paid workers are paid more than they would be without the minimum wage (at the cost of potentially reducing employment among low-wage workers, as I most recently noted here). The minimum wage might also reduce inequality, if it raises the income of low-income groups, while not affecting higher-income groups, and provided any disemployment effects are not too great. It is an open question how effective the minimum wage is at reducing poverty and inequality.

Most of the empirical evidence on the minimum wage comes from developed countries (and a lot of it from the US, where the federal minimum wage is very low and less likely to be binding than in other countries). So, I was interested to see this recent article by Orlando Sotomayor (University of Puerto Rico), published in the journal World Development (sorry, I don't see an ungated version), because it looks at poverty and inequality, and examines the case of Brazil. Sotomayor uses data drawn from the Monthly Employment Surveys conducted by the Brazilian Institute of Geography and Statistics, and a difference-in-differences analysis.

The choice of a difference-in-differences analysis is interesting. Although the minimum wage is specified at the national level in Brazil, differences in the income distribution will mean that the minimum wage is more binding in some states than in other states. For example, Sotomayor notes that earnings in Rio de Janeiro are lower than in São Paulo, so comparing changes in poverty (or inequality) in Rio de Janeiro with those in São Paulo after an increase in the minimum wage will provide some idea of its impact on poverty (or inequality). Sotomayor also makes use of a synthetic control method as well (which I've written about before, for example here).

Sotomayor finds that:

...within three months of a minimum wage hike, poverty and income inequality decline, on average, by 2.8% and 2.4%, respectively. Effects fade over time, particularly with respect to bottom-sensitive distribution measures, a process that is consistent with resulting job loses that fall more heavily among poorer households.

In this study, the minimum wage is associated with small improvements in poverty and inequality through raising incomes of the poor, but these improvements are reduced or eliminated by the disemployment effects in the longer run. This adds to the literature from more developed countries such as the US and suggests that, in the longer run, there may be more effective ways to reduce poverty and inequality than relying on the minimum wage. Redistribution through a more expansive social security system, or a wage subsidy for low-wage employers, are two options that deserve more consideration, as they would not be associated with the same disemployment effects as the minimum wage.

Strategic citing

One common piece of folk wisdom among academics is always to make sure that you cite articles in the journal to which you are submitting. Taken a step further, if this strategy is successful, then citing journal editors or potential reviewers should be the next obvious step. At the margin, this sort of 'strategic citation' must lead to some papers (and journals, and authors) receiving more citations than the quality of their work warrants, creating inefficiency in the research system.

How much of this 'strategic citation' goes on? That seems difficult to assess, but this recent article by Amir Rubin (Simon Fraser University) and Eran Rubin (California State University), published in the Journal of Political Economy (ungated earlier version here), used a natural experiment to find out. Specifically, Rubin and Rubin look at the case of the Journal of Business (JB), which was discontinued in 2006, despite being one of the top five international journals in finance. If you're as surprised as I was to learn that a seriously highly ranked journal would simply cease publication, Rubin and Rubin explain in one of the footnotes that:

We thank Douglas Diamond, the editor of the Journal of Business for 13 years, who told us that the main reason for the discontinuation was the difficulty in finding an editor from within Booth’s faculty.

Wow. Anyway, this natural experiment allows Rubin and Rubin to explore aspects of strategic citation, because:

In broad terms, rather than just serving their intended objective of referencing relevant work, citations of articles published in top-tier journals may be driven by agency considerations because authors focus on achieving professional goals with respect to these journals. These personal goals include, most obviously, the desire to obtain publications in top-tier journals but can also include the desire to get invited to a conference sponsored by such journals, become a referee in those journals, or receive reference letters from scholars associated with the journals. To facilitate this, authors may cite top-tier journals as a way to enhance their relationships with those journals...

And once the Journal of Business ceased publication, the incentives for authors changed, and:

...authors may reduce their tendency to reference low-relevance JB articles, and there may be increased negligence in citing relevant JB articles. 

Rubin and Rubin collected data from the top five finance journals (Journal of Business, plus the Journal of Finance, the Journal of Financial Economics, the Review of Financial Studies, and the Journal of Financial and Quantitative Analysis) covering the period 1996 to 2016 (i.e. ten years either side of the discontinuation). Matching articles from the Journal of Business with similar articles (in terms of publication year and early citation trajectory) from the other top journals, and then comparing subsequent citation measures, they find:

...strong evidence that compared to the matched articles of the other four journals, the citations of JB articles that were published prior to 2006 were significantly negatively affected after 2006. Furthermore, we do not find any change in the citation count difference in the years prior to the discontinuation decision, which implies that the change in citing practices occurred during the time when publishing in JB was no longer possible...

This is well illustrated in their Figure 2, which compares the citation trajectory of articles published in the Journal of Business with the same trajectory for the other four top finance journals:

That suggests that articles in the other four journals received about 20 percent more citations after the discontinuation of the Journal of Business than did similar articles published in the Journal of Business. Taking their analysis a bit further, Rubin and Rubin importantly find that:

...articles that do not cite JB articles (conditional on them being expected to cite JB articles) in the post-2006 period are more likely not to be cited after their publication, which is consistent with the idea that strategic citations are associated with reduced research quality.

In other words, authors who engage in strategic citation are often doing so as a way of attempting to signal that their own paper is of higher quality than it actually is. So, what should the research system do about this? Rubin and Rubin suggest that:

...if authors of academic studies were to include more information on references cited (as done in patent applications), it could potentially benefit academic research and help reduce adverse citing practices.

Closer monitoring is one of the solutions to agency problems, as we note in my ECONS102 class. However, closer monitoring is costly, both to the principal (the journal editor and reviewers would need to review the inclusion of all, or perhaps just a sample, of the included references in every reviewed article), and to the agent (because authors know their citations are going to be closely monitored, they will spend more time ensuring that they summarise them more accurately). However, it isn't clear that such a practice would eliminate strategic citing. Articles are often broad enough in their coverage that they can be cited, even when they may not be the optimal citation to make a particular point.

Strategic citing is clearly a problem, but I don't think that this article provides us with a workable solution.

[HT: Marginal Revolution]