Do Productivity Laggards Ever Catch Up With Leaders?†

• Published date: 01 April 2022
• Author: Evguenia Bessonova,Anna Tsvetkova
• Date: 01 April 2022
• DOI: http://doi.org/10.1111/roiw.12539

© 2021 International Association for Research in Income and Wealth

S71

DO PRODUCTIVITY LAGGARDS EVER CATCH UP WITH LEADERS?†

by EvguEnia bEssonovaand anna TsvETkova*

The Bank of Russia

National Research University Higher School of Economics

The paper focuses on trends in the convergence of labor and multifactor productivity in Russia. Using

firm- level data for the period 2011– 2016, we show that firms with low- productivity grow faster than

those with high- productivity. This result is, however, mostly driven by new entrants. The catch- up

momentum fades after the first few years of a firm’s life, so it is not capable of closing the gap between

the most and least productive firms in the Russian economy. We show that the gap widened over the

period 2011– 2016, suggesting major divergence in productivity levels of Russian firms. We also use

stochastic frontier analysis to verify the divergence within narrowly defined industries. Our estimates

confirm divergence in most industries.

JEL Codes: D24, E22, O47

Keywords: productivity gap, β- convergence, σ- convergence, stochastic frontier analysis, Russia

1. inTroducTion

A large number of studies examining productivity dynamics in various coun-

tries provide evidence of a significant slowdown in productivity growth after the

2008 crisis. In recent years, advanced economies have experienced slower growth

of both labor and multifactor productivity (MFP) (Syverson, 2017, for the US;

Goodridge et al., 2018, for the UK; Ollivaud et al., 2016, for OECD countries;

Bergeaud et al., 2016, for advanced economies). These trends are observed on both

aggregate and micro- level data. In fact, Cette et al. (2018), using macro- and micro-

economic data for France, found downward structural breaks in productivity levels

even several years before the crisis.

Annual changes in Russia’s productivity growth rates are similar to the world-

wide trends. Official statistics indicate that since 2009 labor productivity growth

rates at an aggregate level have been significantly lower than in the earlier years

of this century, which saw rapid growth (Timmer and Voskoboynikov, 2014) (see

Figure 1). Negative rates of growth of aggregate productivity in Russia in 2015 and

2016 give particular cause for concern.

†Note: The views expressed in this paper are solely those of the authors and do not necessarily

reflect the official position of the Bank of Russia. The Bank of Russia assumes no responsibility for the

contents of the paper.

*Correspondence to: Anna Tsvetkova, Research and Forecasting Department, Bank of Russia,

Neglinnaya, 12, Moscow 107031, Russia (tsvetkovaan@cbr.ru).

Review of Income and Wealth

Series 68, Number S1, April 2022

DOI: 10.1111/roiw.12539

bs_bs_banner

Review of Income and Wealth, Series 68, Number S1, April 2022

S72

© 2021 International Association for Research in Income and Wealth

Figure 1 also shows our estimates of labor productivity growth in recent years

based on micro- level data.1 Our estimations and aggregate labor productivity

growth have similar trends. However, both simple and weighted averages of firms’

productivity growth rates2 are lower than the aggregated indicator. This suggests

high heterogeneity of Russian firms in terms of productivity growth and size. The

gap between aggregated productivity growth and sample estimations reflects the

fact that there is huge number of small firms with relatively low productivity

growth, while the small group of large and more efficient firms make a sizeable

contribution to the aggregated indicator.

In recent literature, there is no generally accepted explanation for the causes

of slowdown in productivity growth. The availability of firm- level data makes it

possible to study this question at the micro level and to analyze the evolution of

productivity distribution as well as changes in aggregate indicators.

A large body of literature shows heterogeneous productivity levels even in nar-

rowly defined industries (for example, Hsieh and Klenow, 2009). This suggests that

aggregate productivity growth depends not only on firms that operate at the pro-

ductivity frontier, but also on a mass of non- frontier firms. While the productivity

frontier is pushed forward by technological progress, the performance of laggards

is also determined by the gap to the frontier. As Akcigit and Ates (2019) point out,

knowledge diffusion between the frontier and the rest makes laggard firms more

1The first estimate is the unweighted mean of growth rates of firms’ labor productivity in the sam-

ple. The second estimate is weighted productivity growth calculated as the difference between weighted

average growth of value- added and the weighted growth rate of average employment.

2In line with Decker et al. (2017) and Foster et al. (2018) we find that weighted average productivity

growth is higher than the simple unweighted mean since the correlation between size and labor produc-

tivity growth is positive. Popova (2019) shows similar results for correlation between output growth

rates and firm size in Russia.

Figure 1. Labor Productivity Growth in Russia

-12 -8-4 0 4 8

%

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

year

Firm sample, simple average

Firm sample, weighted

Aggregated data

Review of Income and Wealth, Series 68, Number S1, April 2022

S73

© 2021 International Association for Research in Income and Wealth

receptive to technological progress. So knowledge diffusion reduces dispersion of

productivity levels, and makes industry more homogeneous.

Andrews et al. (2016) and Cette et al. (2018) show that the gap between fron-

tier and laggard firms increases, because the frontier firms increase their produc-

tivity, while laggards drag aggregate productivity down. Decker et al. (2017) also

point out that the dispersion of productivity levels within industries has increased

in recent years. Akcigit and Ates (2019) argue that the reason for increasing hetero-

geneity of productivity is the reduction of knowledge diffusion.

However, another large body of literature analyzes convergence from a dif-

ferent perspective. Several papers document that low- productivity firms grow

faster than high- productivity firms (see Chevalier et al., 2012; Brown et al., 2016;

Bournakis and Mallick, 2018; Gemmell et al., 2018). These papers show positive

correlation between the initial gap to the technological frontier and productivity

growth, which implies catch- up to the frontier.

This conclusion is consistent with economic intuition for a number of reasons.

Firstly, growth rates are affected by the low- base effect. Given the same absolute

change, the lower the initial productivity level, the higher the growth rates. Secondly,

observational errors also result in positive correlation between the initial produc-

tivity level and its rates of growth. If a firm presents annual financial reporting that

shows productivity lower than its true value, then this firm will tend to report bet-

ter performance in the next year (the productivity growth rates reported in the next

year will reflect productivity improvements for two years instead of one). Thirdly,

productivity laggards exit the market more often than efficient firms (Linarello

and Petrella, 2017), but the mean growth rate among laggards does not account for

market exits and for new low- productivity firms entering the market. It is therefore

skewed towards surviving laggards with better performance that those, which exit.

The two strands in the literature might seem contradictory at first sight. The

first body of papers claims divergence, while the other argues for convergence. The

strands have hardly ever been linked, but some authors have noted that catch- up

is consistent with persistent heterogeneity within narrowly defined industries

(Griffith et al., 2009; Berlingieri et al., 2020). Several papers document positive

correlation between initial gap to the frontier and productivity growth and simul-

taneous increase of the gap between leaders and laggards (Andrews et al., 2016;

Cette et al., 2018). As Young et al. (2008) point out, catching up is not always

accompanied by decrease of dispersion.

In the present paper we combine different approaches to convergence analysis

and bring together the two strands in the literature. We study both catch- up and

the dispersion. Using Russian micro- level data we analyze how these approaches

to convergence analysis relate to each other. In particular, we test what catch- up

to the frontier means in terms of dispersion. We show explicitly that the results

obtained using the two approaches, which appear contradictory at first sight, may

in fact coexist.

In line with the literature, we find that productivity growth rates in a sample

of Russian firms are positively correlated with the initial productivity gap to the

frontier. We show that this result is robust for different specifications and holds

in all sectors and all years. We also consider the literature on the labor market

(Haltiwanger et al., 2013) and firm life- cycle (Akcigit et al., 2021) and find that