Global | What do we mean when we compare productivity across countries?

Following the Draghi Report, the debate over productivity and Europe’s ability to converge with the United States has been reignited. However, confusing diagnoses often arise when countries are compared using different definitions of productivity. It is therefore essential to clearly define the metrics used and to understand the mechanisms underlying each definition.

In economic debate, three distinct concepts are commonly used: productivity per employed person, productivity per hour worked, and total factor productivity (TFP). Mixing them can introduce inaccuracies into comparative analyses and create biases in economic policy.

Productivity per employed person is defined as the GDP generated, on average, by each worker, while productivity per hour measures that output relative to the total number of hours worked. The relationship between the two is determined by the average number of working hours per employee: an economy may have high productivity per hour and lower productivity per employed person if its workers work fewer hours per year.

It is worth recalling that the ultimate purpose of economic growth is to increase well-being. To that end, productivity per employed person must grow on a sustained basis, as it is the main real determinant of wages per worker in the long run, as well as of consumption and leisure. When output per employee stagnates, raising wages or reducing working hours without equivalent efficiency gains tends to increase unit labor costs and harm investment, competitiveness, and employment.

Historical evidence shows a clear regularity: annual hours worked per employed person decline as productivity rises. This phenomenon reflects the income effect. Technological progress makes it possible to produce more with fewer hours, raises disposable income, and allows households to consume more, but also to demand more leisure. In this sense, working fewer hours is a voluntary choice compatible with higher well-being. What is difficult to reconcile with workers’ revealed preferences, as reflected in collective bargaining agreements, is a voluntary reduction in hours when productivity per employed person and real wages remain stagnant. In that case, more leisure is not accompanied by more consumption. If, in addition, the reduction in working hours keeps monthly compensation unchanged, it ceases to be a natural consequence of progress and becomes an increase in unit labor costs. The relevant question is not whether an advanced economy should aspire to work fewer hours, but whether it does so as a result of higher productivity or as a substitute for it.

The third metric, TFP, requires additional caution. It is often presented as the ultimate driver of competitiveness, but it is not a direct measure of efficiency. To a large extent, TFP is an accounting residual: the portion of GDP growth that is not explained by labor, capital, and other factors included in the production function. For this reason, it depends on which variables are included, how physical capital, employment, human capital, or intangibles are measured, and how the production function is specified. As a result, TFP growth may occur without growth in productivity per employed person.

Because of this residual nature, an isolated improvement in TFP does not guarantee an increase in GDP per employed person if other factors deteriorate. An increase in organizational efficiency, as measured by TFP, is of little use if the stock of productive capital per worker declines. Likewise, intense capital accumulation does not constitute sustainable structural progress if declining TFP reflects deep inefficiencies in resource allocation. What matters is that the full set of determinants of productivity persistently raises income per worker.

Once these measures have been defined, we are better equipped to analyze the international evidence. From 1995 to 2025, a period that includes the more recently acceded EU countries, patterns of convergence are highly heterogeneous. The countries where productivity has grown the most—whether per hour, per employed person, or in terms of TFP—have been many of the new member states. Latvia, Lithuania, Estonia, and Poland rank among the economies with the largest cumulative gains, in some cases exceeding those of the United States. This performance reflects an intense process of real convergence, supported by investment, integration into the European market, and technological assimilation. Even so, these countries started from lower initial productivity levels, so their higher growth rates do not mean they have caught up with the most advanced economies. By contrast, the countries of Southern Europe, including Spain, Italy, and Greece, show growth below the EU27 average and a persistent structural gap.

If we restrict the analysis to the 2019–2025 period, from the year before the pandemic to the present, this heterogeneity reappears. Some Eastern European economies and Poland continue to post relatively robust growth in GDP per hour and per employed person. Spain, by contrast, shows more moderate progress, consistent with its difficulties in translating employment growth into sustained productivity gains.

To interpret these comparisons correctly, it is essential to take into account the debate over purchasing power parities, or PPPs. Current PPPs indicate how much income in a country can buy at each year’s relative prices and are useful for comparing living standards at a given point in time. Constant PPPs, with a fixed base year such as 2021, better isolate changes in relative prices and allow analysis of the evolution of real productive capacity.

This distinction is not minor. When productivity per employed person is assessed using current prices and current PPPs, the European Union appears to have maintained a relatively stable position vis-à-vis the United States since 1995. However, part of that apparent resilience is explained by Europe’s relative cheapening compared with the U.S. economy. Under this metric, Spain would have lost about 9 percentage points, falling from 83% to 74% of the U.S. level, while the EU27 would have gained 1 percentage point. When GDP per hour worked is measured, Spain would have lost only 4 points, while the EU27 would have gained 5 points.

In short, international productivity comparisons are useful, but only if there is a precise distinction between what is being measured. Productivity per hour provides information on the efficiency with which working time is used; productivity per employed person, on the ability to generate income per worker; and TFP, on the part of growth not explained by observed productive factors. The priority should not be to improve any one of these metrics in isolation, but to ensure that GDP per employed person rises on a sustained basis through factor accumulation and improvements in their efficiency, which would allow growth in GDP per hour worked to be even higher. For Spain, the comparative evidence suggests that this simultaneous improvement has not occurred and that the challenge remains to increase productive investment, improve the allocation of resources, raise human capital, and strengthen competition. Without these improvements, working less will not be a consequence of progress, but an additional constraint on achieving it.