Consensus and Dissensus Among Economic Science Academics in Mexico
This paper is closed for comments.
This paper reports results on consensus in economic opinions, scientific aspects of economics, as well as preferences and scientific activities, based on a survey applied to a defined population of economic science academics in Mexico. The results show that in the analyzed population there are areas of consensus, which, at least partially, is consistent with what was found in previous studies. The survey, which is broader than those used in other studies, identifies topics with significant dissensus whose determinants should be examined in future studies.
This is an excellent article, in fact I believe many readers would be able to understand the situation of economics in Mexico. My only suggestions for the current text would be to emphasize more the situation of economics in Mexico in relation to the international economic community, with possible especification to the readers of this journal (in other words, in relation to the fields of economic thought, methodology and political economy); and my second suggestion would be to incorporate two articles of the recent literature: Mark Horowitz & Robert Hughes. Political Identity and Economists’ Perceptions of Capitalist Crises. Review of Radical Political Economics, v. 50, n. 1, p. 173-193, 2018; and Henrik van Dalen. Values of Economists Matter in the Art and Science of Economics. Kyklos, v. 72, n. 3, p. 472-499, 2019. I know the reference list is already large, but these two also check current trends.
Rafael Galvão de Almeida
PhD. candidate at the Federal University of Minas Gerais
We thank to Rafael Galvão de Almeida for reading and making useful comments to our paper.
We will follow his suggestion on relating the situation of economics in Mexico and internationally. The papers mentioned by Galvão de Almeida serve this purpose. For example, van Dalen paper has some propositions that are equal or very similar to those we use in our work and that can be used to compare and discuss situations in the Netherlands and Mexico, specifically in the sections of economics as a science and economics assumptions. We also believe that other entries in our reference list can be used in the same way.
This paper provides results from an internet survey initially sent to 1,300 academic economists, soliciting their views on a number of economic issues. Many of these questions had been posed in prior research in similar internet surveys, so that the main objective is to see to what degree responses have changed. Throughout, the focus is not so much on the policy implications of specific responses but instead on the degree to which there is a consensus among economists on each of these questions.
One initial concern with use of an internet survey is that response rates tend to be very low, in this case only one in five. This raises a real question whether the subset who choose to respond is at all representative of academic economists as a whole. To the extent that the authors know any information about the underlying sample of 1,300 individuals, they should compare the respondents to the full sample in order to reassure readers that those who responded were largely representative of the underlying group. My guess is that younger individuals, those working part time, and retirees are more likely to respond, with the most successful individuals being unlikely to find the time to respond. If so, the respondents could be drawn more heavily from those whose knowledge of the academic literature is thin or dated. Response rates might also be higher among those with extreme views, who are looking for an outlet for their views.
In order to measure consensus, the paper makes use of an “entropy” measure that it claims is the one most used in past studies of this type. If the fraction of the sample choose response (e.g., agree), then the value of this metric equals divided by the maximum feasible value of this metric. One other choice the authors made was to ignore responses of “no opinion” or “don’t know”, reweighting the remaining responses.
Even though commonly used, I find this metric a questionable way to measure consensus in the profession. As an example, consider the following possible responses to the questionnaire:
Strongly agree Agree No opinion Disagree Strongly disagree
Statement 1 50% 50%
Statement 2 50% 50%
Statement 3 5% 90% 5%
All three of these statements generate the same value for this entropy measure: 1.0, which implies the minimum degree of consensus. Statement 1 certainly represents such a minimum degree of consensus. Statement 2, in contrast, should indicate a large degree of consensus, since all respondents agree with the statement, differing only in the reported strength of their agreement. Statement 3 also should show substantial consensus, but a consensus indicating the lack of any real evidence based on the past academic literature.
There are many alternative measures of the degree of consensus the authors might consider instead. The dominant fraction agreeing (or disagreeing) with a statement is one option. With the above three statements, this would be 50%, 100%, and 5%, providing a sharp distinction among these three possible patterns of response. A very similar measure would be the standard deviation of the response, assigning a value of 1 to agree or strongly agree, 0 to don’t know, and -1 to disagree or strongly disagree. Here the values for the above patterns of response would be 1, 0, and 0.016. These two approaches differ in the degree to which they highlight no opinion, since for example they would treat differently the assumed responses to statement 1 from those where half agreed and half had no opinion. A third approach would be the mean response, which would be 0, 1, and 0 for the above three statements.
I would also urge the authors to make a clearer distinction between distributional and efficiency issues. There is no reason for there to be consensus on distributional preferences among professional economists, even if there should be pressure towards consensus for statements where efficiency issues are dominant, at least to the degree that there is empirical evidence. To what degree is there less of a consensus on statements that largely raise distributional issues?
Of course, attitudes towards many statements include both efficiency and distributional considerations. As an example, take attitudes towards use of tariffs. The theoretical literature argues that there would be an efficiency loss from use of tariffs when the country is a price-taker in the world market for a good, providing an efficiency argument against use of tariffs on the grounds that Mexico is a “small” country. But even for a small country, there could be a few goods where the country is not a price taker in the world market, providing some efficiency case for tariffs on these goods. In this case, the response from a profession economist would depend on which goods might face tariffs. Even in this context, there would likely be an efficiency gain from signing a trade agreement preventing all signatory countries from making (much) use of tariffs, given the joint efficiency loss from use of tariffs. Tariffs can also provide an efficiency gain, though, to the extent that tariffs offset other distortions, e.g. reestablish a level playing field between domestic and foreign producers in the retail market when there are differential taxes across domestic firms.
Regardless there are always distributional effects across industries, skill groups, and consumers with differing consumption bundles, from use of tariffs, providing an alternative rationale for tariffs. However, the political decision-making process on tariffs favors concentrated interests. Trade agreements are often argued for as a way to tie a government’s hands to avoid this prisoners’ dilemma outcome on efficiency grounds.
Many responses then are possible to such a statement about tariffs even among those fully knowledgeable about the academic literature, given alternative distributional preferences as well as alternative inferences about the policy context of the question. (The academic literature on many of the other statements has comparable complexity.) Disagreement among professional economists due to different inferences about the policy context, though, would reflect ambiguity of the statement rather than any underlying lack of consensus.
To what degree, though, is the lack of consensus a reflection of differing distributional preferences (or political leanings)? To this extent, the stated opinions of economists do not carry the hoped for weight of professional expertise. Do patterns of response, for example, divide into liberal vs. conservative political leanings, rather than reflect a common base of evidence from the academic literature? Here, one would expect to see the pattern of responses among the set of statements breaking into two (or more) “camps”. Since the authors have acquired some information about what “camp” each of these respondents identifies with, one approach would be to look at the degree of consensus within each camp compared to that in the whole sample. More consensus within a camp rather than between camps would then provide caution about how to interpret the stated views of economists.
We appreciate the very thoughtful comments to our paper by Dr. Robert H. Gordon. In order to answer Prof. Gordon’s criticism in a thorough manner, we have classified his comments into four topics.
First. About the concerns on the statistical representativeness of our results, at the beginning of our research we set out to obtain answers from the largest possible number of faculty members of Mexico’s economics schools (or departments). Since the information we collected from most of the departments’ published on the internet was incomplete, outdated or even non-existent; we decided to request this information directly from department heads. Tipically we received lists of the names and emails, so making an exhaustive comparison between our sample and the full sample was not possible.
An exercise (not included in the submitted draft) consists of taking the 1,315 individuals that appeared in the full sample and assigning them a gender according to their given names. There are a few unisex names, but we are truly sure that the percentage of female in the full sample would be 30-32%, which nearly coincides with 30.9% of the sample of the respondents. We could try to reduce the number of ambiguous names in order to add the datum in our final draft. Also, we know that 9.1% of our entire sample was working in private universities, which is close to 10.2% of the sample of respondents.
Prof. Gordon thinks that younger individuals, those working part time, and retirees would be overrepresented in our results. We believe this is not the case because:
1. As indicated in our paper 89.1% of respondents are full-time faculty members.
2. We requested information of active faculty at the time that our research was conducted. If we recieved information of retirees and they answered our survey, this is something we did not know.
3. We know that 14.5% of respondents were 35-year-olds or younger. We do not have the age distribution of the full sample, but we believe that such percentage does not imply an overrepresentation of such population.
Second. In our paper we present general results for most of the entire survey and, when it is posible, compare with others works. Because for our consensus results the straightforward comparison is with those of Urzúa (2007), who only used the entropy index and did not report “don’t know” responses, in order to do so in a reasonable way we had limit our consensus measure to the entropy index and to extract the “no opinion” or “don’t know” responses.
After this article was sent to “Economic Thought”, two of us published a Spanish-language paper in which we propose a more comprehensive method to measure consensus (Andere & Canché-Escamilla 2019). In line to Fuller & Geide-Stevenson (2003, 2014), in this paper we use three measures to obtain a global consensus index, which are:
1. Entropy index.
2. Chi-square test of goodness of fit to a uniform distribution of responses.
3. Percentages of respondents answering to some degree of agreement with the statement.
This method already appears in Andere & Canché-Escamilla (2019), but I think we can discuss some of those results in this paper.
Third. Prof. Gordon’s comment on distinguishing between efficiency and distributional issues is pertinent. In fact, questionnaire’s statements on economic opinions could be classified by efficiency/distributional considerations or by a positive/normative distinction in the sense of Friedman (1953). One option we see is to identify the degree of consensus (measured by the entropy index) between the distributional and efficiency issues that Prof. Gordon refers to.
Fourth. There is a body of literature wich uses information from surveys similar to ours to identify the respondents’ ideological leanings by creating a “liberalism index” (i.e. Klein et al. 2013, Klein & Stern 2007, Stastny 2010). We agree with Prof. Gordon that relating liberal or conservative leanings to the degree of consensus would be very important issue to investigate, but we sincerely believe that it exceeds the scope of our paper.
Andere, J. L., & J. L. Canché-Escamilla (2019). Entendiendo el consenso en la profesión económica en México. Análisis Económico, 34(86), 9-34.
Friedman, M. (1953): The methodology of positive economics, en Essays in Positive Economics, Chicago, University of Chicago Press.
Fuller, D. A. y D. Geide-Stevenson (2003). Consensus among economists: Revisited, Journal of Economic Education, vol. 34, No. 4, p. 369-387.
Fuller, D. A., & D. Geide-Stevenson (2014). Consensus among economists–An update, Journal of Economic Education, vol. 45, No. 2, p. 369-387.
Klein, D. B., W. L. Davis, & D. Hedengren (2013). Economics Professors’ Voting, Policy Views, Favorite Economists, and Frequent Lack of Consensus. Econ Journal Watch 10(1): 116-125.
Klein, D. B., & C. Stern (2007). Is There a Free-Market Economist in the House?. American Journal of Economics and Sociology, 66(2).
Stastny, D. (2010). Czech Economists on Economic Policy: A Survey. Econ Journal Watch 7(3): 275-287.
Urzúa, C. M. (2007). Consensos y disensos entre los economistas mexicanos, Revista de la Cepal, 91: 153-165.
The economy like practically all the disciplines of the social sciences, since its theoretical foundation by Adam Smith, has been the subject of great debate and controversy in relation to its aims, its objectives, its instruments, as well as in relation to the role that the economists and in this context if their role should be positive or normative, and it is, both in schools, public and private faculties, as well as in economic research centers.
In 1929 in Mexico, the first school of economics was founded at the Autonomous National University of Mexico, from that year onwards, schools have been established, in such a way that currently in the whole Mexican Republic there is at least one school or faculty.
Although the initial orientation of the new public schools and then the private ones, was based on the curriculum of the faculty of economics of the UNAM of orthodox orientation, but with Marxist dyes, each of them subsequently took their own theoretical path.
In the field of applied economics, development models have historically varied considerably in the country, based on the classical school, Keynesian and post-Keynesian approaches: while, in Latin America, the influence has come from the structuralist and neo-structuralist model. During the last 40 years, neoliberal models of neoclassical origin have prevailed. Before the arrival of a new government in Mexico, the theory that is intended to be applied is that oriented towards the welfare of the population, but in practice it contemplates orthodox orientations.
In this sense, the study on consensus and dissent among economic science scholars is relevant. The study addresses eight aspects of great importance ranging from opinions related to various topics of economic policy to the situation of the Mexican economy over the next five years.
The topics related to teaching in Mexican universities seem relevant to me. This is because since many years ago, the National Association of Teaching and Economic Research Institutions, constituted by schools, faculties and public and private research centers in Mexico, has been conducting studies to determine which are the most appropriate and relevant subjects for the teaching of economics in Mexico, as well as its theoretical orientations.
I consider that an important document to guide this association is the study that I am commenting on, however, it seems pertinent to deepen the analysis in the pedagogical orientations that should guide the teaching of economics.
Currently, the government carries out a basic knowledge test called EGEL-ECONO, whose approval allows obtaining the degree. The exam consists of the following areas: business economics, financial economics, public economics, economic growth and development, and international economics.
The analysis of these areas, as a reason for controversy, it seems to me a very fruitful case to advance in the study with the same methodology, but taking as subjects of study the graduates under this modality.
I also believe that progress must be made in the analysis of the new theoretical paradigms that have been present for several years, such as social capital, the new institutionalism, the knowledge economy, competitiveness and innovation models just to name a few.
An important conclusion of these studies is that there is no universal pattern of thought that guides economic knowledge.
The debate between the different ways of thinking is not over, nor would it have to be done because the economy, like every concept, is necessarily relative and historical.
We thank to Alberto Quintal Palomo for reading and commenting our paper. His historical and institutional remarks on the education of economics in Mexico are valuable for us.
We agree with him about the diversified, controversial and changing nature of the economic thought. In our paper’s perspective, economic thought can be conceived broadly as composed for qualified opinions, theories, methodologies, but also for the way that it is taught and researched. Thus our paper displays both levels of consensus and disention as the way they vary in time and space, for a number of topics.
Perhaps by displaying the former, we are contributing to the open debate referred by Prof. Quintal Palomo.