Abstract
Inequality has risen to the top of the policy agenda throughout the world. There is a burgeoning academic literature on inequality (including my own work), and it has become a focal point for research and policy proposals by the World Economic Forum and the World Bank. There is a mini-industry in measuring inequality, including the United Nations University-WIDER database as well as cross-national estimates by Branko Milanović and Thomas Piketty and Emmanuel Saez. Piketty’s dense 700-page tome Capital in the Twenty-First Century even became a best-seller in the United States and Europe.
I have followed the literature in the field and examined how inequality has led to less generalized trust over time in the United States and across nations and to lower levels of corruption, both directly and through social policy (universal education). So I was anxious to see what new contributions Steven Hitlin and Sarah Harkness were making. I struggled my way through Unequal Foundations: Inequality, Morality, and Emotions across Cultures and, alas, I still don’t see a new contribution.
This book is only marginally about inequality. The first chapter is a summary of some recent research on unequal distribution. But coverage is rather sparse. The authors pay no attention to race, to ethnicity, to immigrant status, to residential segregation, or even to public opinion on inequality. After a cursory discussion of trends in inequality, they briefly discuss some consequences but largely wave away inequality as the major factor underlying a wide range of social problems. They dismiss the impact of inequality on driving down generalized trust (without citing my work supporting the linkage) because their “conception of the relationship between structure, inequality, and culture is ‘stickier’ than that of other theorists tackling the issue of how inequality affects outcomes” (p. 115). They do not recognize that I have argued that both inequality and trust are “sticky” (I use that very language in laying out the “inequality trap” in Corruption, Inequality, and the Rule of Law [2008:26–27]).
Much of the book is not devoted to inequality at all, but rather how to study morality. I don’t see what this discussion adds to the study of inequality, especially since the authors do not discuss issues such as whether inequality really is a problem (some say that it isn’t), whether it is a worse problem now than in the past (Milanović’s estimates suggest that inequality has declined in much of the world), what the best way is to measure inequality (including or excluding benefits people receive), the difference between inequality of results and inequality of opportunity, or what people think of inequality. This is especially surprising since Timothy Smeeding (together with Lars Osberg) has devised a novel way of measuring preferences for levels of inequality by using the International Social Survey Program surveys to compare what people in more than 25 nations believe people in various professions do earn compared to what respondents think they should earn.
But Hitlin and Harkness are not devotees of survey research. Instead they argue that small samples are preferable to large-scale surveys (p. 136) but never justify their preference. They also have data on language usage (I don’t understand their methodology—their discussion is difficult to follow, but they use a technique designed to measure language or events [“Affect Control Theory” using EPA dictionaries—not sure what they are] from various universities in the United States, China, Germany, Japan, and Canada). I don’t understand how this technique can measure preferences for equality. Nor do I see how these five countries—four of which are advanced western democracies—can tell us much about preferences on equality, especially when the authors admit at the very beginning of the book that estimates of inequality for China are problematic (p. 19). How are measures of language from countries that are more alike than different based on student samples (when there are large-scale surveys available for public download) reliable? Then Hitlin and Harkness use logit analysis to predict moral emotions (again not clear what this means or how it is related to inequality) from a set of independent variables that are not listed. But these logit analyses are based on samples that are anything but random, and the authors’ use of tests of significance—much less two-tailed tests—is not justified.
It would help if the authors were to discuss what these “moral emotions” are and how they relate to the larger theme of the book. Instead, I kept staring at the wordclouds in Figures 9-1 through 9-4 wondering what they mean and how they were derived, what the simulation program Interact did, and how they handled the data. Table 9-2 contains the “Average Actor-Behavior Object Events Producing Self-Sanctioning Moral Emotion by Country.” For the United States, the actor from the Male Dictionary is “stepsister.” For Canada, it is “egghead.” For the female dictionary, the object is “Lesbian/Defendant” and for China it is “Debtor/Divorcee.” They then divide the emotions into positive/self-transcendent and negative/sanctioning (how?).
In sum, we have a book that does not address the most common issues in inequality research, is not based on the literature on inequality, and employs a methodology that at best needs elaboration. In the end, “sanctioning is predicted to be more likely in the United States and China, while feelings of compassion and especially praise are more likely in the less hierarchical countries of Germany, Japan, and Canada” (p. 180). Finally, I found something that made sense to me, since we would expect more positive (“compassionate”) outcomes in countries that are less polarized by income—or so I argued in The Moral Foundations of Trust (2002). But then the authors call “these findings” “counterintuitive.” I was at a loss for words.