WPs I5q POLICY RESEARCH WORKING PAPER 2659 Firm Entry and Exit, Labor Firms entering and eating a market contribute almost as Demand, and Trade Reform much to employment changes as firms continuing Evidence from Chile and Colombia in a market. As much effort shoufd be made to understanding sensitivity to Pablo Fajnzylber wage changes in entering William F. Maloney and exiting firms as to Eduardo Ribeiro understanding wage elasicities in continuing firms. The World Bank Latin America and the Caribbean Region Poverty Sector Unit August 2001 l POLICY RESEARCH WORKING PAPER 2659 Summary findings There are increasing fears that trade reform-and Estimates of labor demand elasticities of entering and globalization generally-will increase the uncertainty the exiting firms were surprisingly similar in Chile and average (especially less skilled) worker faces. If product Colombia and somewhat higher than elasticities for firms markets become more competitive and the access to that survived. foreign inputs is increased, will demand for workers Estimates of the effect of trade liberalization offer only among existing firms become more elastic? Will labor ambiguous lessons on trade reform's probable impact on markets become more volatile because bad shocks to these elasticities. The data suggest that in Chile greater output will translate into greater impacts on wages and exchange rate protection does reduce the wage- employment? employment elasticitiy of entering and exiting plants, but So far the literature on this question has focused the opposite effect is found for continuing plants. And almost entirely on labor demand within continuing firms. the results are reversed in Colombia's case. But much of the movement in the job market arises from Moreover, in Colombia higher import penetration the entry and exit of firms. lowers the elasticity of labor demand and in Chile higher Fajnzylber, Maloney, and Ribeiro show that firms tariffs increase it. entering and exiting a market contribute almost as much These findings, combined with very ambiguous results to employment changes as firms continuing in a market. from probit regressions on the determinants of plant In several samples, firms entering and exiting affected the exit, suggest that circumspection is warranted in asserting net change in-positions more than the expansion of that trade liberalization will increase the wage elasticity continuing plants did, although contributions varied of labor demand. greatly across the business cycle and period of adjustment. This paper-a product of the Poverty Sector Unit, Latin America and the Caribbean Region-is part of a larger effort in the region to study the impact of liberalization on labor market risk. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contactAnne Pillay, room 18-104, telephone 202-458-8046, fax 202-522-2119, email address apillay@worldbank.org. Policy Research WorkingPapers are also posted on theWeb at http:/ /econ.worldbank.org. The authors may be contacted at pablo@cedeplov.ufmg.br or wmaloney@worldbank.org. August 2001. (32 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center Firm Entry and Exit, Labor Demand, and Trade Reform Evidence from Chile and Colombia' Pablo Fajnzylber Universidad Federal de Minas Gerais Belo Horizonte, Brazil William F. Maloney World Bank Washington, DC Eduardo Ribeiro Universidad Federal de Rio Grande do Sul, Porto Alegre, Brazil Our appreciation to Marcelo Tokman, the Ministerio de Hacienda, and the Instituto Nacional de Estadistica, Santiago Chile for help updating the data series. They of course, bear no responsibility for erroneous analysis or conclusions. Our thanks for helpful disucussions to Luis Serven. I. Introduction There are increasing fears that trade reform -- and globalization more generally -- will increase the uncertainty faced by the average worker, particularly those with fewer skills. Rodrik (1997) suggests that the increased competitiveness of product markets and the greater access to foreign inputs may lead to a more elastic demand for workers among existing firms. This leads to greater volatility in the labor market since bad shocks to output translate to larger impacts on wages and employment than formerly was the case. To date, the literature has focused virtually entirely on movements in labor demand within continuing firms with both theory and empirical specifications derived from neo-classical profit or cost functions. However, as has now been documented extensively, a large fraction of movements in the stock of jobs arise from the entry and exit of firms. Dunne, Roberts and Samuelson (1989) and Davis and Haltiwanger (1990), for instance, show that up to 25% overall job changes in the US were due to firm births and deaths. Roberts (1996) finds that in Chile (1979-86), Colombia (1977-91), Morocco (1984-89) and the US (1973-86) entry and exit contributed more to the net change in positions than did the expansion of continuing plants although the contribution varied greatly across business cycle and period of adjustment. What this suggests is that an understanding of the overall labor demand elasticities that workers face requires not only understanding how own wage elasticities in continuing plants change, but also how the own wage elasticities of exit and entrance may change. As Hamermesh (1993) notes, this is relatively unexplored territory empirically with only a handful of estimates of dubious comparability available. llis survey of the literature reveals only Carleton (1979) as generating elasticity of firm birth rate with the respect to the wage rate ranging from 1 to 1.5. His own calculations for the US generate elasticities of exit between .86 and .37. Only Berger and 1 Garen (1990) estimate the elasticity of entry/exit related employment change, .94. All of these estimates are at the high end of those standard within firm reallocation elasticities and suggest that, given the potentially large contribution to employment changes, this may be the more important elasticity to understand and quantify.2 This paper attempts to make four contributions. First, using establishment level data from Chile and Colombia, it documents patterns of job creation, destruction, net job creation and turnover and establishes the share of employment changes due to within-firm vs entry and exit effects. Second, it sketches some possible channels through which liberalization might alter these elasticities and turnover more generally. Third, it offers among the first estimates of own wage elasticities of employment due to entry and exit and the first that are comparable among countries. Fourth, it attempts to link any observed changes to measures of openness and liberalization. II. Conceptual Issues Identifying any links between firm turnover and labor demand and trade liberalization, requires entering the complex theoretical literature on firm dynamics which, unfortunately, enjoys nowhere near the consensus of that of the neo-classical model of the firm. Only Hamermesh (1988), using a fairly traditional model of the firm, has focused specifically on the issue of how wage changes may induce exit, and nowhere does the literature make the linkage to openness. Jovanovic (1982), Pakes and Ericson (1990) and Hopenhayn (1992) have offered the most sophisticated, although distinct, models of firm entry and exit but none preoccupies itself 2 A related literature examines what variables, including wages may affect firm location but, as Hamermesh notes, this is not strictly comparable since the elasticity of firm location is almost certainly larger than that of the decision to open the firm in the first place. 2 especially with entry/exit response to factor price changes. What generally does emerge is that reductions in the fixed cost of entry and exit will make firms more sensitive to changes in costs. The lessons from the by now vast literature on irreversible investment under uncertainty also offer some insights. (See Dixit and Pindyck, 1994). The inability of investors to predict the future path of returns to investments with certainty, and the fact that much of an investment, whether in plant and equipment, marketing, information collection etc. is not recoupable if the firm goes bankrupt, makes firms react far more sluggishly to changes in expected profitability than they would in the absence of any uncertainty. The analytics are sufficiently complicated as to make closed form solutions for entry and exit elasticities difficult. But both simulations and empirical evidence suggest that the elasticity of the response of investment to price or cost changes varies greatly dependent on the degree of uncertainty faced.3 Further, there need not be a symmetric impact on entry and exit. Caballero (1991) has introduced an important caveat to these findings noting that implicit in the literature is the assumption of imperfect competition. As market structure becomes more competitive, he finds that uncertainty begins to increase investment (see also the models of Abel 1983). Consistent with this, Guiso and Parigi (1999) find the negative effect of uncertainty is stronger in Italian industries where firms have more market power. Since one presumed impact of trade liberalization is to increase overall competition, the potential sign switch makes theoretical inference even more difficult. Below, therefore, are several possible impacts of reform measures and potential impacts on own wage net job growth elasticities for entering and exiting firms with only tentative assertions about what the effects mighlt be. 3 1. Reduction of barriers to imported capital may lower its price and thereby reduce the irreversible component of costs. More generally, freer trade may improve the secondary market for machinery and have the same effect. 2. Increased labor intensity of production with trade liberalization, as predicted by standard Heckscher-Ohlin-Samuelson models of international trade, may lead to lower irreversible fixed costs in capital and perhaps lower investment in human capital that would be lost should the firm go out of business. 3. Lower barriers to foreign investment might lead to lower fixed costs of entry for foreign firms. Taken as a global phenomenon, this also leads to a possible "substitution" effect where firms will shift, for instance maquila production, to other countries quickly in response to small changes in the wage. This may be compounded if higher labor intensity lowers irreversible capital costs. 4. Inflation reduction that is both requisite and result of trade reform may lead to less variance in predicted returns to investment. In particular, the reduction in level and variance of inflation may lead to a reduction in the variance of real wages which are frequently nominally sluggish. In a lower inflation environment, a given rise in real wages may be seen as more persistent than in high inflation environments and therefore have greater impact on entry and exit decisions. As examples, Hasset and Metcalf (1992) show that investment in household energy saving technologies in response to a 15% tax credit almost doubles from their base case in the absence of uncertainty. Serven and 4 5. Privatization and fiscal rejorm may lead LDC governments cease to be willing to finance inefficient public firms (preventing exit) and eliminate barriers to investment (preventing entrance) (Roberts 1996). It may also signal greater commitment to market friendly policies and hence reduce uncertainty. However, other effects could lead to the reverse results. 6. Greater exposure to global shocks may lead to a more uncertain business environment and less predictability about output prices. However, with Caballero in mind, the combined impact of higher variance and greater competition might lead to increased investment and sensitivity to cost movements. Serven (2000) finds that in the aggregate there is no increase in uncertainty after trade reform in Latin America. 7. Shifts in trade regime may lead to greater uncertainty about the permanence of the new set of relative prices faced, regardless of the direction of movement. Bernanke (1983) shows that such uncertainty leads to waiting and overall lower responsiveness to return changes. In the present case, it is unclear whether entrepreneurs interpreted the return to protectionism in both Colombia and Chile as more credible than the subsequent attempts to re-open the economies. 8. Exchange rate realignments designed to provide some offset to tariff reductions may not be credible. As Chilean entrepreneurs expressed, it is exchange rate stability over the Solimano (1993) find investment in LDCs to be unresponsive to changes in profitability in the face of uncertainty. 5 medium term rather than the short term that is important to investors.4 In both countries, very large movements of the real exchange rate would be followed by extreme movements in the opposite direction in under five years. In addition to influencing demand elasticities, these changes may also affect overall job reallocation. Hopenhayn, for example, shows that turnover increases with a fall in entry costs. Again, however, there is no systematic theoretical treatment of the relationship with liberalization. The empirical evidence is limited. Dunne, Roberts, and Samuelson (1989) found no strong trend in plant-level turnover in US manufacturing from 1963-1982, a period of substantial tariff reduction and technological progress. Davis, Haltiwanger and Schuh (1996) find no relationship between US job flows and either import penetration or export share although Klein, Schuh and Triest (2000) find that the responsiveness of job flows to the real exchange rate varies with the industries openness to international trade. For LDCs, Roberts and Tybout (1996) find high turnover in Chile, Colombia and Morocco relative to the US (Davis and Haltiwanger 1990), but no obvious relation with trade reform. Tybout (1996) finds very high exit rates following the Chilean liberalization. However, Roberts (1996) finds that average entry and exit actually rose with trade restrictions in Colombia 1983-1985 relative to the previous period of relative openness. Levinsohn (1999) confirms Tybout's results for Chile using the same data, and also establishes that turnover is somewhat higher among tradeables than non-tradeables. This suggests that, to the degree that trade liberalization expands the share of tradeables in total output, it may lead to more churning in the job market. 4See Maloney (1999). 6 IV. Definitions We follow Davis and Haltiwanger (1992) and others in defining the relevant variables. Let nit be the employment stock of firm i at period t. We define employment growth of firm i as g11=(nit - nit- )Ixit, where xit=(nirnit l)/2. It is clear that g,, e[-2, 2], with the lower bound indicating that a firm exited and the upper bound that a firm entered the industry (or sample). Job Creation (JC) is defined as the size-weighted sum of the firm employment, or job, growth rates summing over the firms (i=l,...,N) that experienced non-negative employment growth: iJ,= , gi, wit I(gi,>O0), where wt=xi1/(E,7i xid=xi/X, is the firm weight and I( is the indicator function. Conversely, Job Destruction (JD) can be defined as the size-weighted sum of firm employment growth, summing over the firms that experienced negative employment growth: JD, = , IgtlI Wit 1(gi1 O)I(git#2) + yNI g1 wit I(git> O)I(g't=2) = F + B, where E stands for firm expansion and B for firm birth, and JD, =N I Igill Wit I(gi<0)I(g,t-2) + Ig |&| w,, I(git