Development and Reform Research Team University of Bologna

Development and Reform Research Team University of Bologna Lectures on Job Creation and Job Destruction With applications to transition economies Prof. Hartmut Lehmann Dipartimento di Scienze Economiche Università di Bologna

Topics discussed • Measures of job and worker flows • A look at job and worker flows in Poland, Russia (pre- and post-transition), and Ukraine. • Stylized facts about job creation and destruction in transition countries. • Is there creative destruction in transition countries? Looking at the example of Russia. Development and Reform Research Team – University of Bologna

Why do we want to look at gross job flows and not only on changes in employment stocks – in a western and in a transition context? DE = entry (jc) + exit (jd) + expansion (jc)+ contraction (jd) Development and Reform Research Team – University of Bologna

Measures of job gross flows • Net employment growth rate at firm/establishment level. Following Davis and Haltiwanger (1999) we define this growth rate as git = nit – ni, t-1 / xit where nit stands for employment of firm i at time t and xit = (nit + ni, t-1) / 2 is the average employment size of the establishment/firm. This net employment growth rate, being symmetric and lying in the interval [2, 2], can take account of entry, expansion, contraction and exit of firms. Development and Reform Research Team – University of Bologna

Net Growth rates of employment: Russia 1997 Development and Reform Research Team – University of Bologna

Numerical example of net employment growth rate • git Let nit = 100 and ni, t-1 = 0, then = (100 -0)/(100 + 0)/2=2 entry Let nit = 0 and ni, t-1 = 100, then = (0 – 100)/(0 +100)/2= -2 exit Let nit = 100 and ni, t-1 = 50, then = (100 -50)/(100 + 50)/2=0. 666 expansion Let nit = 50 and ni, t-1 = 100, then = (50 -100)/(50 + 100)/2= -0. 666 contraction So, in an economy we have: E = entry (jc) + exit (jd) + expansion (jc)+ contraction (jd) Development and Reform Research Team – University of Bologna

Job creation rate • Let Xt be total average employment of the economy or of the sector under investigation, i. e. let Xt = i I xit, where I stands for the set of all firms in the economy or in the sector. The job creation rate is then defined as: pos i I+ git(xit/Xt) = i I+ (nit – ni, t-1) / Xt, where I+ is the subset of expanding/entering firms. The job creation rate is thus defined as the weighted sum of all positive net growth rates in the economy or in the sector. Alternatively we can think of this rate as the increase in employment in expanding firms expressed as a proportion of total employment. Development and Reform Research Team – University of Bologna

Job destruction rate • The job destruction rate is defined analogously neg i I- |git|(xit/Xt) = i I- |nit – ni, t-1| / Xt, here we now sum over the subset of contracting/exiting firms, I-. The job destruction rate, normally expressed in absolute value, can also be interpreted as the absolute value of the decrease in employment in contracting firms as a proportion of total employment. Development and Reform Research Team – University of Bologna

Other job flow rates • The gross job reallocation rate is gross = pos + neg, • The net change of employment is represented by net = pos – neg. If net employment changes are very large and mainly driven by contraction and exit, as will be the case particularly during the early phase of transition, then gross might not capture the reallocation of jobs very well. For example, if, like in U. S. , pos= 10% and neg=10%, then gross=20 % reflects true job reallocation; if on the other hand, pos=0% and neg = 20%, then gross=20%, but there has not been any job reallocation at all (jobs have only disappeared in this scenario). Development and Reform Research Team – University of Bologna

Other job flow rates, cont. The alternative measure of job reallocation, the so-called excess job reallocation rate, excess = gross - |net|, is therefore often used to catch job reallocation in excess of the amount necessary to accommodate a net aggregate employment change. We can also think of excess as an index of firm heterogeneity with respect to job creation and destruction in an economy or a given sector. For example when pos=10% and neg=10%, then |net| = 0% and excess = 20% - 0% = 20% true job reallocation of 20%; if, on the other hand pos=0% and neg=20%, then |net| = 20% and excess = 20% - 20% = 0% no job reallocation at all. Development and Reform Research Team – University of Bologna

Job flows in Poland before and after transition Development and Reform Research Team – University of Bologna

Job and worker flows by ownership Development and Reform Research Team – University of Bologna

Job flows by size – Poland 1991 Development and Reform Research Team – University of Bologna

Results of regression determining factors that have impact on net growth rate of employment – Konings/Lehmann/Schaffer Development and Reform Research Team – University of Bologna

Job Flows in Russian Manufacturing Firms, Brown and Earle (2002) Development and Reform Research Team – University of Bologna

Decomposition of ecxess job reallocation rate by 2 -digit sector in Ukraine – Konings, Kupets and Lehmann (2003) Development and Reform Research Team – University of Bologna

Job flows in transition: stylized facts o The patterns of job creation and job destruction vary over the transition period. In early transition: job destruction dominates job creation, In later stage of transition: economy roughly destroys as many jobs as it creates. o The empirical evidence for CEE countries : whether countries embark on a course of rapid or slow reform, eventually they show job reallocation rates similar to those in mature capitalist economies, with roughly equal creation and destruction rates. Development and Reform Research Team – University of Bologna

Job flows in transition: stylized facts, cont. • The worker flows mirror the pattern of job flows. • In most CEE countries, the increment in worker flows is dominated by the separation rate in the early stages of transition, while the hiring rate outpaces the separation rate in the latter stages. • A large part of separations and hirings are driven by job destruction and job creation, suggesting that a major factor underlying worker mobility is that workers are moving because the allocation of jobs across businesses is changing as opposed to workers reallocating themselves for a given allocation of jobs across businesses. Development and Reform Research Team – University of Bologna

Job flows in transition: stylized facts, cont. o Small and new private firms contribute disproportionately to job creation while stateowned firms are responsible for most of the job destruction. o The heterogeneity of job creation and destruction within narrowly defined sectors is tremendous. The vast majority of job reallocation at any point in the transition is not across sectors but within sectors. So, while some industries contract and others expand their employment shares, most of job reallocation takes place within industries. Even so, the pace of between industry reallocation is higher in the transition economies than in mature, developed economies. Development and Reform Research Team – University of Bologna

Is job destruction “creative” in transition economies? • • Joseph Schumpeter’s concept of creative destruction Brown and Earle (2002) test whether job reallocation is enhancing productivity as predicted by theory. data: ‘traditional’ manufacturing sector in Russia spanning the years 1985– 99 findings: a negligible association of job flows with productivity before 1992, which turns strongly positive after the beginning of reforms. implication: even as the Russian manufacturing sector goes through a difficult period of downsizing, the job reallocation process has taken such a form as to make job destruction more creative. Development and Reform Research Team – University of Bologna

Productivity Decomposition using decomposition suggested by Olley and Pakes (1996) Development and Reform Research Team – University of Bologna