Multifactor productivity over the cycle

A number of studies indicate that multifactor productivity growth (MFP) behaves cyclically, i.e., it increases in upturns and declines in downturns. This has sometimes been interpreted as a paradox, as MFP has traditionally been perceived as exogenous technological change, which should typically not behave cyclically.

Key findings

The empirical evidence confirms the cyclical pattern of MFP. In fact, MFP follows GDP growth very closely, not only in terms of the direction but also in terms of the size of the change. While the contribution of labour fluctuated relatively strongly for most G7 countries, up to 2007, adjustments in labour input typically lagged. The contribution of capital input changed little over the cycle, possibly reflecting adjustment costs. Capital input reflects the accumulation of past investment of all firms in the economy. Hence, although investment is typically relatively volatile, capital stock and capital services estimates are less so. However, the contribution of capital input to GDP growth declined significantly after the crisis, possibly reflecting the sluggish recovery of investment.


Four factors help explain this cyclical movement and each of them is related to the definition of MFP as the part of GDP growth that cannot be explained by changes in labour and capital inputs (Chapter 8. ). First, cycles in productivity growth may relate to imperfect competition and the potential to capitalise on increasing returns to scale during upturns. Second, labour input typically adjusts with a lag in downturns, as firms seek to retain workers even if not needed for current production so as to keep the human capital (labour hoarding). Third, adjustment costs prevent an immediate up- or downsizing of production and capital, resulting in lower utilisation of existing capital stock in downturns. Fourth, the reallocation of resources to production of goods and services with higher or lower marginal productivities may be pro or counter cyclical.


The appropriate measure of capital input for productivity analysis and within the growth accounting framework is capital services (Chapter 8. ). While these take into account the productivity of the different capital assets, no account is taken of the extent to which the existing capital stock is actually used, i.e. the rate of capital utilisation, which may affect comparability over time and space.

Theoretically, measuring labour input by the total actual hours worked of persons employed should capture the rate of labour utilisation and hence account for the cyclical effects of labour input. Continuous labour force surveys provide a basis for measuring this. However, in practice, total hours worked are often measured based on hours typically worked or actual hours worked during a reference week, which are then extrapolated over the year using additional data sources. These may not capture sufficiently variations in actual hours worked over the cycle (Chapter 8. ).

Official data for Germany after unification are available only from 1991 onwards. Estimates for Germany as a whole back to 1970 have been derived by applying the relevant growth rates for West Germany to 1991 data.


OECD Productivity Statistics (database),

OECD (2009), Measuring Capital – OECD Manual,

OECD (2001), Measuring Productivity – OECD Manual,

Wölfl, A. and D. Hajkova (2007), “Measuring multifactor productivity growth”, OECD Science, Technology and Industry Working Papers, No. 2007/05,

Figure 7.9. Contributions to GDP growth over time in G7 countries
Total economy, percentage point contributions at annual rate
Figure 7.9. Contributions to GDP growth over time in G7 countries


End of the section – Back to iLibrary publication page