The question of why productivity growth slowed from the early- to mid-2000s continues to interest policymakers, but researchers have paid less attention to the role of human capital accumulation. This column suggests that one-sixth of the productivity slowdown in OECD countries can be accounted for by slowing human capital accumulation, and that the use of digital devices in school is a major factor in children’s worsening test scores. Education policy reforms to reverse the trend will play out over long time horizons, but structural reforms to enhance labour market reallocation and adaptability could boost productivity in the shorter term.
While the question of why productivity growth slowed from the early- to mid-2000s continues to interest policymakers, researchers have paid less attention to the role of human capital accumulation (Biondi et al. 2024, Friesenbichler et al. 2024, Goldin et al. 2024, Ho et al. 2024). In new OECD research (Andrews et al. 2024), we show that around one-sixth of the productivity slowdown can be accounted for by weaker human capital accumulation and explore the scope for policy to lean against these headwinds.
The slowdown in human capital accumulation
To estimate the aggregate human capital stock, we take the cohort-weighted average of historical student test scores (i.e. the quality of education) and the corresponding mean years of schooling (i.e. the quantity of education) for the working age population. Across the OECD, the human capital stock expanded at an average rate of 0.11% between 1987 and 2005, but then slowed to 0.05% between 2005 and 2016 and remained flat after 2017. This trend is primarily due to the quality component of human capital, driven by a noticeable fall in student (PISA) test scores over the past 20 years (Figure 1B).
Figure 1 Human capital growth and its components in the OECD, 1987-2022
Note: Group70 includes countries starting in 1970 (AUS, CHL, DEU, FIN, FRA, GBR, HUN, ISR, ITA, JPN, NLD, NZL, SWE and USA). Group95 includes all OECD countries except BEL, COL, CRI, EST, POL, TUR. The data series are based on PISA data extended with two vintages of the World Bank Global Data Set on Education Quality (Altinok et al. 2018).
Source: Andrews et al. (2024).
How important was the human capital deceleration for the productivity slowdown?
To establish the contribution of human capital slowdown to productivity slowdown, the elasticity of human capital and productivity, as estimated in Égert et al. (2024), is applied to the average annual growth rate of the human capital stock. Total factor productivity (TFP) growth declined from 1.7% between 1987 and 2005 to 0.5% from 2005 to 2022. The actual slowdown in human capital accumulation implies that its contribution to annual TFP growth decreased from 0.24 to 0.06 percentage points. Consequently, the deceleration in human capital accumulation accounts for almost one-sixth of the aggregate TFP slowdown.
Why did student performance decline?
We use cross-country student-level regressions to investigate the forces behind the almost 15-point decline in PISA test scores observed on average across the OECD between 2009 and 2022, driving the human capital slowdown. Unsurprisingly, pandemic factors played a key role, accounting for almost one-half of the decline in PISA scores between 2018-2022 (Figure 2).
The longer-term decline in PISA scores can be potentially understood in terms of a common technological factor: the rapid diffusion of smartphone technology and related social media platforms from the late 2000s. Contemporary accounts argue that this trend dramatically transformed the lives of children and teenagers around the world, and in some cases for the worse (Haidt 2024). We find that classroom distractions associated with smartphone technology could plausibly account for a large share of the pre-pandemic decline and one quarter of the post-2018 decline in PISA scores (Figure 2).
Figure 2 PISA decline and estimated contribution of policy drivers
Average across the OECD, 2009-2022
Note: Implied changes in PISA scores are obtained by multiplying average changes in policies, digital device use and Covid-19 effects with the estimated coefficients derived from student-level regressions linking PISA scores to policies.
Source: Andrews et al. (2024).
What can policy do?
First, we find that education policy reforms – for example, early childhood education, teacher quality, homework assistance and responsible internet use in schools – could potentially yield a gain of 10 PISA points in the average OECD country.
Second, structural reforms to that support reallocation and adaptability can help allocate the existing stock of human capital more productively. In the Nordics and the US, where policy is more supportive of labour market matching efficiency, the positive link between human capital and TFP is stronger than in Eastern and Southern Europe and emerging market OECD economies, where structural policy rigidities loom larger.
The future of human capital and productivity
The harsh reality is that a generation of young workers who have performed more poorly on standardised testing since 2003 have entered the workforce and are now placing downward pressure on the human capital stock. And if future generations of students continue to perform poorly in education, simulations show that long-run TFP could be almost 3% lower as a result.
While an ambitious education policy reform package would materially boost human capital accumulation and thus TFP, these benefits would take a long time to materialise and only offset part of the projected productivity decline in the absence of reform.
But combining education policy reforms with structural reforms to enhance labour market reallocation and adaptability could turn these headwinds into tailwinds and boost long-run TFP by about 1.5%. Overall, policymakers should focus not only on growing the stock and the quality of workforce talent, but also on allocating existing talent more efficiently.
Source : VOXeu
