The notion that GDP is an imperfect guide to well-being is arguably as old as the concept itself. Its limitations were, for example, already discussed in detail by Simon Kuznets, and they were part of the debate in the 1940s between Richard Stone, Milton Gilbert, and Kuznets about the exact measurement of national income. There has since been an undercurrent of literature trying to compensate for the flaws of the official System of National Accounts, for example, by incorporating inequality, or environmental problems, or health and education into the measured concepts. The Human Development Index, inspired by the welfare-theoretical work by Amartya Sen and developed by the United Nations Development Programme (UNDP), is the best-known example of this new approach. Since the financial crisis of 2008 and the report by Stiglitz, Sen, and Fitoussi (2018), the well-being debate has both broadened and deepened. The core ideas are simple: firstly, the ultimate goal of economic development is not just an increased command over resources (as is measured by GDP), but the growth of the wellbeing of the relevant population. Secondly, wellbeing is multidimensional—we do not live by bread alone—, and health, political rights, education, a clean environment, inequality, etc. are also important dimensions that have to be taken into account when the “achievements” of economies are assessed. However, different dimensions with different scales cannot simply be compared. One way to deal with this is to aggregate the different dimensions into a single, composite index. But for a concept that has many dimensions, this is not an easy task. That we are interested in something multidimensional but that we would like to measure it by one single index, has led to the proliferation of attempts to solve this puzzle.

Economic historians have traditionally struggled with these issues as well. The British Standard of Living Debate, the American “Ante-bellum growth puzzle,” and the literature on the increase in wellbeing during the Interwar period are examples of debates in which various indices of wellbeing and growth developed differently over time, questioning the use of “traditional” measures such as real wages or real income. A more recent debate focuses on the period after 1980, when economic growth continued (albeit at a lower rate than before), but trends in inequality and health suggest a stagnation in the increase of wellbeing.

Leandro Prados de la Escosura (LP), one of the leading economic historians of his generation who has published many innovative papers on European (and global) economic growth, has produced an ambitious contribution to this debate. It offers, based on a rich dataset, a reinterpretation of the evolution of wellbeing in the world since 1870. He builds his new view on data about four dimensions of wellbeing: life expectancy, education, GDP per capita, and liberal democracy, the latter being an index of the quality of political institutions (and therefore the degree of political freedom). Together, they form the Augmented Human Development Index (AHDI). Its analysis leads to other insights into long-term development of wellbeing. An important example of this new view is that between 1920 and 1950 wellbeing increased much more than the increase in GDP per capita suggests, because literacy and education continued to grow at a robust rate.

A researcher wishing to go beyond GDP and measure multidimensional wellbeing in the past, has had to make a number of choices to get to a composite wellbeing index that charts the long-term trend. Firstly, how many and which dimensions to include. Secondly, how to transform the series of the individual dimensions to make them comparable with each other—how to standardize series with very different scales. And thirdly, how should the various series be aggregated; what weights can be used? In contrast to recent attempts to measure the 11 dimensions of the How’s Life? (OECD 2020) framework (by some considered the best international standard), LP focuses on four dimensions only, which all show significant progress over time. The “dark side” of economic growth—pollution, inequality, biodiversity decline—is not taken on board. He then transforms and standardizes the four series in such a way that, whereas in the real world, economic growth is in absolute terms much bigger than the increase in the other dimensions, the standardized series shows the opposite pattern. As a result, his wellbeing index is dominated by life expectancy (which accounts for 37 percent of the total variation) and educational attainment (32 percent), with the remaining 31 percent accounted for by the other two underlying series, GDP per capita and political rights. Rather than concluding that this is perhaps a bit unbalanced and that his estimates may be biased, he takes this for a fact and argues that economic growth was not a driver of the increase in wellbeing. His conclusion to Chapter 2 illustrates this: when listing the causes of human development progress he mentions “life expectancy was the main contributor,” “Education … was a steady contributor,” and “political and civil liberties…added substantially” (p. 64), but economic growth is not mentioned.

The early 1930s clearly demonstrate where this led to. Between 1929 and 1933, when the world economy collapses, mass unemployment peaks, democracy is on the defensive, and Hitler seizes power—in short, during our worst nightmare—the AHDI still shows an increase in wellbeing, thanks to the increase in life expectancy and education, which overpower the declines in GDP per capita and liberal democracy (pp. 32–34). By this standard, the world’s population was still somewhat better off in 1933 than in 1929, a result that is, to say the least, challenging established views. This happy growth of wellbeing continues in the rest of the thirties (next data point is 1938), and between 1938 and 1950, the next year for which estimates are available. It explains the “superior” development of wellbeing compared with GDP in the first decades of the twentieth century, which is one of the main conclusions LP draws from the reconstruction.

These results are also caused by the specific transformation of the two dominant series, life expectancy and education, that LP applies (drawn from the work by Kakwani Reference Kakwani1993). This gives a lot of weight to changes at the upper tail of the distribution; for example, five years of basic education has the same impact on wellbeing as one year of university education. For more details see van Zanden (Reference van Zanden2022), which also includes a response by LP.

Properly measured multidimensional wellbeing can be an important guide to the ex post assessment of the outcomes of economic development, including the effects of policies to counterbalance negative externalities. Growth can be balanced, where all dimensions of wellbeing increase in a comparable way, or very unbalanced, with large negative externalities and growing inequality. The measurement of wellbeing addresses such issues. What causes, in the long run, the increase in wellbeing, which drivers can be identified, is an entirely different story. But this is sometimes mixed up, and the rise in life expectancy and education that is statistically “causing” the rise of the wellbeing index developed by LP, is also identified as the final cause of the gains. I think, to make a long story short, that productivity growth made possible by the accumulation of ideas is still the best candidate for the most important driver of development and increase in wellbeing in the long run. LP probably disagrees. The debate continues.