À propos de: Au Petit Bonheur la Science

Dans La Presse, Pierre Tircher produisait vendredi une réponse à mon article de mardi résumant les travaux d’Angus Deaton, Daniel Kahneman, Megan Teague et al. Je produis ici une contre-réponse.

La première affirmation avancée par M. Tircher veut j’ai choisi les éléments de l’article de Deaton et Kahneman qui m’intéresse. En conséquent, j’ai ignoré les autres éléments. Plus précisement, j’ai décidé d’ignorer les mesures de “bonheur émotionnel” afin de privilégier les mesures de “satisfaction de vie”. En effet, j’ai fait cet effort tout simplement parce que Deaton et Kahneman croient justement que les résultats sur la “satisfaction de vie” sont ceux qui sont pertinents.  Vers la fin de l’article, ils expriment clairement qu’ils croient que la “satisfaction” est une mesure “pure” du bonheur:

We speculate that the Cantril ladder of life is a purer measure of life evaluation than the life satisfaction question, which has an emotional aspect, and that the reports of the emotions of yesterday provide a purer measure of emotional well-being than the standard happiness question (…) The Cantril ladder is a serious contender for the best tool for measuring the degree to which individuals view themselves as achieving their goals, both material and other.

Et cet opinion est de plus en plus partagé parmi ceux qui étudient les relations entre “bonheur” et “revenu”. En effet, cet article dans Social Indicators Research conclut que la satisfaction de vie (basée sur l’échelle de Cantril) est nettement supérieur pour mesurer le bien-être des individus. (Voir aussi ces écrits de Deaton dans Journal of Public Economics qui réitèrent ce point)

Ensuite, il faut comprendre les types de données utilisées par Deaton et Kahneman qui les poussent à faire cette affirmation. Ils utilisent des données dites en “coupe transversale” c’est-à-dire des données sans variable de temps (par exemple, un recensement est un ensemble de données en coupe tranversale). À plusieurs reprises dans d’autres articles, ils ont souligné qu’il faut être prudent à l’égard de ce genre de données puisqu’il ne nous dit pas l’effet d’augmenter le revenu et/ou si le point de satiation entre bonheur (tant satisfaction de vie que bonheur émotionnel) se déplace vers le haut. D’ailleurs, M. Tircher se garde bien de souligner une phrase importante dans l’article de Deaton et Kahneman:

When interpreting our findings, it is essential to distinguish changes from differences. Our data speak only to differences; they do not imply that people will not be happy with a raise from $100,000 to $150,000, or that they will be indifferent to an equivalent drop in income. Changes of income in the high range certainly have emotional consequences

Dans un article du American Economic Review, Justin Wolfers et Betsey Stevenson ont remedié à cette limitation en utilisant des données d’enquête qu’on peut suivre dans le temps pour une multitude de pays. Voici le résumé qu’ils produisent de leur étude:

Many scholars have argued that once “basic needs” have been met, further rises in income are not associated with further increases in subjective well-being. We assess the validity of this claim in comparisons of both rich and poor countries, and also of rich and poor people within a country. Analyzing multiple datasets, multiple definitions of “basic needs” and multiple questions about well-being, we find no support for this claim. The relationship between well-being and income is roughly log-linear and does not diminish as incomes rise. If there is a satiation point, we are yet to reach it.

Il faut souligner que cette conclusion s’applique à tous les genres de mesure du bonheur (pas juste satisfaction de vie) qu’on peut générer. La conclusion qu’il n’existe aucun point de satiation entre bonheur et le revenu est relativement bien partagée depuis environ une dizaine d’années grâce à des avances importantes dans les méthodes statistiques et la qualité des données. Par exemple, le Journal of Applied Econometrics  a publié un article démontrant, utilisant une méthode encore plus sophistiqué que Wolfers et Stevenson qui conclut que:

Overall, we find no evidence of a satiation point, which is in support of Stevenson and Wolfers

En fait, ceci est cohérent avec une idée alternative provenant de certains résistants à la nouvelle conclusion qui se cimente depuis Stevenson et Wolfers (2013). Le point de satiation n’a pas besoin d’être stable et peut être repoussé. Dans un article du Journal of Economic Behavior & OrganizationDavid Clingingsmith trouve qu’il y a un point de satiation mais il se trouve à 200,000$ et non pas 75,000$ – un point qui se déplace.

Quant au dernier point de M. Tircher concernant ma lecture de Teague et al., il suffit de souligner que Rosemarie Fike et Virgil Storr (deux des auteurs) sont des amis personnels. J’ai parlé avec eux (particulièrement avec Rosemarie) du sujet de cet article dans Constitutional Political Economy. J’ai confirmé avec eux ma lecture de cet article et concernant le point de la corruption que M. Tircher fait. Ils soulignent que ma lecture est la bonne particulièrement parce que la mesure de la liberté économique qu’ils utilisent (dont je n’ai pas discuté dans mon article) est corrélée négativement avec la corruption et avec le revenu par personne. Ainsi, ils établissent un argument de co-dépendence entre liberté économique et positivement revenu (les deux se renforcent même si ils ont des effets séparés). La lecture de M. Tircher n’est donc pas celle des auteurs eux-mêmes.

New working paper: State Capacity and the Post Office: Evidence from 19th Century Quebec

Earlier this week, Michael Makovi and I submitted the following paper for publication. It concerns the theory of state capacity and its measurement. We argue that one of the proposed channels from which strong and capable states are able to foster economic development is not generalizable. That channel, the Post Office’s role as market coordinator, has been used frequently especially in economic history. We use the insights proposed by others to test it in Quebec between 1831 and 1861 (an area for which there is excellent data). The abstract is below and the paper is available here on SSRN: 

The theory of state capacity predicts that states with powerful abilities—as long as they are constrained—can promote economic growth, because these abilities can serve to support market development. A difficulty in the study of state capacity relates to its measurement. Some scholars, especially in economic history, argue that post offices area proxy for government effectiveness and state capacity (Chong et al., 2014; Acemoglu et al., 2016; Rogowski et al., 2017; Jensen and Ramey, 2019), because a well-functioning communication network helps markets operate. We test whether this claim is generalizable by using data from nineteenth century Quebec. We use a difference-in-difference method to estimate the effect of gaining, or losing, a post office on the value of agricultural output per acre between 1831 and 1861. We find no treatment effect, implying that post offices had no relationship with agricultural productivity. This may suggest that post offices are not a valid proxy for state capacity or, at the very least, that they do not constitute generalizable evidence

Divergence, convergence, and the history-augmented Solow model

A few weeks ago, Vadim Kufenko, Klaus Prettner and I received news that Structural Change and Economic Dynamics accepted our paper “Divergence, convergence and the history-augmented Solow model“. The paper is available here at the journal’s webpage and the abstract is below:

We test the recently proposed history-augmented Solow model with respect to its predictions on the evolution of cross-country income inequality between nowadays industrialized countries. Using a broad range of deterministic and stochastic simulations, we illustrate that the model predicts the following pattern. There is low cross-country income inequality before the Industrial Revolution (during the period of “Malthusian Stagnation”), strongly increasing cross-country income inequality afterwards (the period of “The Great Divergence”), and finally declining cross-country income inequality toward a level that is higher than the level before the Industrial Revolution (the period of “Club Convergence”). Tests on the structural break of the observable time series of income dispersion and segmented regressions show that this development is fully consistent with the empirical evolution of cross-country income inequality since the late 19th century. Additional tests using quadratic polynomials, fractional polynomials, polynomial ridge regressions, and polynomial LASSO regressions confirm this finding.

Forthcoming – How poor were Quebec and Canada during the 1840s?

Gonzalo Macera (Texas Tech University) and I heard back from Social Science Quarterly. Our paper harnessing wages from the 1842 censuses of Lower and Upper Canada have been accepted. The older version of the paper is available here on SSRN, the abstract is below and the online appendix is here :

This paper uses the censuses of 1842 of Canada East (modern-day Quebec) and Canada West (modern-day Ontario) to help explain the historical differences in living standards between Canada and the United States. The wage and price data contained in the censuses suggest a gap of 42 percent between Canada East and Canada West. We argue that Canada East was substantially poorer than the rest of Canada and, as it represented such a large proportion of the total population of the initial four Canadian provinces (over 35 percent), that relative poverty weighed heavily in determining the extent of differences in living standards between Canada and the United States. These findings change the perspective on the roots of the differences between the two countries. We propose that any research agenda trying to explain those differences should focus heavily on Quebec.

Forthcoming – State Capacity and Economic Development: Causal Mechanism or Correlative Filter?

Alexander Salter and I just received news that the Journal of Economic Behavior & Organization has accepted our paper on state capacity. The paper can be found here on SSRN and the abstract is below.

In this paper, we explore why there are no examples of societies with low state capacity and high economic development. We argue that such an outcome is unlikely because of the nature of investments in state capacity. Societies that become rich in the absence of a strong state invite predation by societies that develop such states. Thus societies invest in state capacity, in part, to plunder other societies’ wealth. Those investments are a form of rent-seeking. Potentially preyed-upon societies are forced to invest in state capacity in turn so as to deter potential attackers. This entails that as soon as a rent seeker enters the game, the likelihood of a low-capacity, high-development society surviving falls. This explains the historical lack of such societies. We thus interpret state capacity not as a causal condition for widespread economic prosperity, but a survivability condition for enjoying this prosperity.

Forthcoming: Multilingualism and the decline of French in Quebec

Yesterday, I received news from the Journal of Multilingual and Multicultural Development will publish my article (co-authored with PhD student Alex Arsenault Morin – economics at Queen’s University) on why the vitality of the French language in Quebec is underestimated. In the last decade, fears of a “French crisis” have surged (with the belief that the French population would grow increasingly marginalized). Skeptics point to the fact that French-usage in the public sphere is rising while believers point to a decline of French-usage at home. We show that the rise of multiple language use within households (and across spheres of life) lead to this debate. We propose a new metric that resolves this problem. We show that there is no decline (nor are there any improvements).

The paper is here on SSRN. 

Forthcoming: Improving deflators for estimating Canadian economic growth, 1870–1900

I received news yesterday that a revised version of my paper with Michael Hinton on the improvement of price deflators for Canadian GNP figures between 1870 and 1900 will be published by Research in Economic HistoryThe old version of the paper is here on SSRN. The key result from our paper is that Canada’s rate of growth is heavily underestimated during the country’s first three decades after “independence” (for lack of a better term). Instead of being a less-than-average performer, Canada is the fastest growing economy of the western world.

The data, once the paper is published, will be shared in full detail on the Macro History of Canada section of my website with read-me files (i.e. methodology, details, explanations of choices made).