This book falls into the ‘economics of happiness’ category of literature. It uses the method of perception survey data to measure human well-being in Latin American countries. When qualitative and quantitative data are compared, some paradoxes appear, such as ‘unhappy growth’ and ‘happy peasants and frustrated achievers’. Some policy implications about these paradoxes are also discussed. The new field of happiness economics is intended to develop complementary measures of human well-being, providing analysis of the self-reported subjective well-being of individuals (‘happiness’ or ‘satisfaction’ about life). These data are supposed to complement objective data on human well-being. If objective and subjective data are found inconsistent, a paradox is said to exist.

The book relies on the Gallup World Poll for its measurement of the self-reported subjective well-being of individuals. This poll has been carried out in 23 Latin American countries and in more than 130 countries of the world since 2006. The crucial measure of happiness/satisfaction is derived from the question, ‘Generally speaking, how happy are you with your life?’, or ‘How satisfied are you with your life?’, using a scale of zero to ten for the response. The book presents seven studies, written by 14 authors, in which happiness is associated with objective variables such as income, health, education and job security.

Regarding income, the study finds that levels of happiness are positively correlated with income within countries; that is, the self-reported average level of happiness of the rich is higher than that of the poor. Across countries, the study finds that ‘countries with higher GDP per capita have, on average, higher levels of happiness … Yet controlling for levels of per capita GDP … individuals in countries with positive growth rates have lower happiness levels – the paradox of unhappy growth’ (p. 9). Another paradox is that, within countries, the wealthy are more likely to be less satisfied with public services than the poor, despite the lower performance of the services that the poor have access to. This is part of a larger paradox known as ‘the happy peasant and the frustrated achiever’.

The book derives implications for public policies from the paradoxes that are empirically identified. Policymakers are given several pieces of advice to interpret these paradoxes correctly and use them in a positive way. The expected contribution of the book is clearly stated as: ‘[I]t is precisely in identifying these gaps between subjective and objective measures of welfare, and seeking an explanation for them, that this book provides a contribution to our understanding of quality of life in Latin America and beyond’ (p. 10). It is fair to say that, given the methodology of the economics of happiness, the book indeed presents the results and their policy implications in a well-balanced manner. The book is written in a very professional way – there are no misleading conclusions or implications. The chapters show the empirical results in a direct and comprehensible fashion and the introductory chapter by Graham and Behrman summarises very well the content of the subsequent chapters.

What is debatable, however, is precisely the given methodology of the economics of happiness. Is economics making progress as a science with this methodology? ‘Science is measurement’ is an accepted scientific principle; do self-reporting data on happiness or satisfaction of individuals constitute scientific measurement?

The distinctions between measurable variables made by the utility theoretician Nicholas Georgescu-Roegen in his book The Entropy Law and the Economic Process (1971, chap. 4) include cardinal, weak cardinal and ordinal. These are defined as follows: cardinal is when absolute zero exists and differences are measurable, as in flows of output such as tons of steel per day; weak cardinal applies when only differences are measurable because the origin is arbitrary, as in temperature (Celsius/Fahrenheit) or chronological time (BC/AD); and ordinal (qualitative) is when only ranking or order is measurable, as in ‘friendly’ (high/moderate/low), ‘inequality’ (high/low) or ‘democracy’ (delegacy/participatory). Happiness and satisfaction belong to the ordinal category, and aggregation is not allowed in this type of variable – yet the studies shown in this book aggregate these values and use average values, which are inserted as the dependent variable in the regression analyses. This procedure has no logical justification. One finds in the book the statement that ‘Happiness surveys are ordinal in nature and do not attach cardinal weights to the answers’ (p. 17), but, all the same, the studies in the book proceed to the aggregation of individual values of happiness.

The methodology of happiness economics assumes that happiness of individuals can be measured by self-reporting surveys, just as self-reporting monthly nominal salaries is a good way to measure income. Some limitations of measuring subjective well-being by the self-reporting method are pointed out in the book. The introductory chapter indicates the shortcomings of this type of information, although in just one paragraph, and referring to its framing only: ‘How questions are asked, and where they are placed in the survey can make a big difference in the responses’ (p. 3). Is that all?

One can think of many other limitations, starting with the meaning of the variable ‘happiness’ or ‘satisfaction’, which may vary across cultures; then there would be the problem of misperception; and finally, even if the previous issues are absent, there is the problem of distortion of the true answer, which may be replaced by the supposedly politically or socially correct answer, or by a simple lie. There is no way to control for these problems when using opinion surveys.

On the options for using methods other than self-reported surveys to measure happiness, Georgescu-Roegen has asked himself the following question: ‘Are there ordinally measurable attributes that could not possibly be measured by a pointer-reading instrument? Any definitive answer to this question implies at least a definite epistemological, if not also a metaphysical, position’ (1971, p. 100). Many authors have expected that a ‘hedonimeter’ (a term used by F. Y. Edgeworth) would eventually be invented. Surprisingly, according to the book, such a device currently exists: ‘Research by psychologists, meanwhile, demonstrates that answers to surveys are in large part validated by psychological and neurological measures of happiness and well-being’ (p. 2). No reference is given in the book to this crucial point. The problem is also discussed in a box (box 4.1, chap. 4), but no epistemological position is found there either. Until the hedonimeter is indeed discovered, the contribution of subjective studies based on self-reporting surveys of individuals, such as those presented in this book, will be very problematic.

The other lesson one can draw from reading this book is that similar measurement problems also arise when incomes, expenditures or employment situations of individuals, the ‘objective’ measures, are obtained from household surveys, which mostly rely on self-reported information. (Even the Gallup poll asks for happiness and income in the same survey!) Cardinal variables do not imply that they are measured correctly. Maybe some of the paradoxes pointed out in the book do not exist because the corresponding objective variables are not measured correctly; or maybe there are more paradoxes than those reported in the book.

In fact, according to some philosophers of science, such as Rosenberg in his 2008 work Philosophy of Social Science, one of the difficulties for the progress of economics as compared to physics is that in economics, empirical measurement of relevant variables is not precise enough. The lack of good instruments of measurement is a general problem in economics and thus applies not only to ordinal variables but to cardinal and weak cardinal variables as well.