On exports stability: the role of product and geographical diversification

On exports stability: the role of product and geographical diversification Marco Grazzi 1 and Daniele Moschella 2 1 Department of Economics - University of Bologna, Bologna, Italy. 2 LEM - Scuola Superiore Sant Anna, Pisa, Italy. March 24, 2014 Abstract Using firm-level data on all exporting Italian firms, this paper addresses how product and geographical diversification are related to the stability of exports. We find that more diversified firms in terms of product exhibit a higher volatility of exports. On the contrary, firms that are more diversified in terms of destination countries are more able to stabilize their exports. One interpretation of this result is that firms might insure themselves against demand shocks by sellig their products to different national markets. On the other hand, economies of scope may give firms the incentive to diversify across similar products, which are more likely to be affected by correlated shocks. Keywords: exports stability, international trade, product diversification, geographical diversification JEL classification: D22, F10, F14, L25 1

1 Introduction Exporting firms differ widely along several dimensions: total and unit export value, number of countries served, number of products exported (Mayer et al.; 2011; Bernard et al.; 2012). In this paper, we analyze another characteristic of exporting firms, usually neglected in the trade literature: exports stability. We relate exports stability to two of its possible determinants: product diversification and geographical diversification. Selling in different international markets might help to stabilize exports and, more in general, sales, as shown in the classical Hirsch and Lev (1971) and, more recently, in Kim et al. (1993), who explained the Bowman (1980) s pradox showing that firms can achieve a high return low-risk profile through global market diversification. Less clear-cut is the evidence about the positive role of product diversification. In fact, Braakmann and Wagner (2011) find that in a sample of German entreprises more diversified firms in terms of products exhibit a higher variability of sales and employment. In this paper, we focus on both dimensions of diversification. We find that product diversification has a negative but negligible impact on exports stability, while selling in more countries help to reduce the volatilty. One possible explanation of this result is that firms diversify in products that are related each other, because they are more likely to exploit economies of scope and technological similarities. On the other hand, diversifying across different countries might help to insure firms against demand shocks. 2 Data The analysis is based upon two firm-level datasets collected by the Italian statistical office (ISTAT), namely Statistiche del Commercio Estero (COE) and Archivio Statistico Imprese Attive (ASIA). The COE dataset consists of all cross-border transactions performed by Italian firms and it covers the period 2000-2007. COE includes the annual value and quantity of export transactions by the firm for product-country destination pairs. A product is defined as a six digit category in the Harmonized System (HS6). Using the unique identification code of the firm, we link the firm-level export data to ISTATs archive of active firms, ASIA. In ASIA, firms are classified according to their main activity, as identified by ISTATs standard codes for sectoral classification of business (5-digit ATECO). This information allows us to distinguish between four broad categories of firms: manufacturers, wholesalers, retailers, and a residual group including the remaining sectors. ASIA also contains information on firms operations including the number of employees and total turnover. The combined dataset used for the analysis is not a sample but rather includes all active firms. In table 1, we report the distribution of Italian exporting firms, their export value, and their employment, according both to the number of products and the number of 2

Table 1: Distribution of exporters by number of products and export destinations, 2000 Share of exporting firms Number of countries Number of products 1 2 3 4 5+ All 1 29.83 2.17 0.72 0.36 0.86 33.95 2 7.12 4.94 1.30 0.60 1.43 15.39 3 2.79 2.42 1.61 0.77 1.90 9.51 4 1.41 1.32 1.06 0.70 2.13 6.61 5+ 2.97 2.64 2.56 2.26 24.10 34.54 All 44.13 13.50 7.25 4.70 30.42 100 Share of export value Number of countries Number of products 1 2 3 4 5+ All 1 0.54 0.24 0.15 0.13 0.62 1.69 2 0.29 0.28 0.23 0.19 1.26 2.25 3 0.17 0.22 0.24 0.17 1.60 2.41 4 0.12 0.16 0.16 0.13 1.00 2.58 5+ 0.66 0.81 0.82 1.09 87.68 91.07 All 1.80 1.72 1.60 1.71 93.16 100 Share of employment Number of countries Number of products 1 2 3 4 5+ All 1 10.31 1.05 0.28 0.17 0.51 12.32 2 2.64 3.28 0.93 0.71 0.96 8.52 3 1.02 2.67 1.10 0.65 1.59 7.04 4 0.55 0.52 0.59 0.55 2.08 4.29 5 1.50 1.30 2.68 6.99 55.36 67.82 All 16.02 8.82 5.59 9.08 60.50 100 Note. Table displays the joint distribution of Italian firms that export (top panel), their export value (middle panel), and their employment (bottom panel), according to the number of products firms export (rows) and their number of export destinations (columns). Products are defined as six-digit Harmonized System categories. export destinations, in 2000. The top panel shows that around 2/3 of all exporting firms are multiproduct firms, and half of these export more than four products. As for countries, more than half of firms export to at least two countries, with 1/3 exporting to more than four countries. The middle and the bottom panel say that multiproduct and multicountries firms account disproportionally for the export value and the employment: in particular, firms that export at least five products produce more than 90% of all total exports, and employ more than 65% of all employees; and similar numbers hold for firms that export to at least five countries. 3

3 Diversification measures We start by analyzing the bivariate relationship between export variability and our two dimension of diversification, across products and across countries. Our proxy for export variability is the standard deviation of exports growth rate, calculated as the log difference of export value between two consecutive years. This is equivalent to consider as unsystematic changes all deviations from an exponential growth pattern (see Hirsch and Lev; 1971). As for the diversification measure, we choose three different variables. The first one is a modified version of the Herfindahl index, which is defined as (see Aw and Batra; 1998): 1 s 2 i where s i indicates the share of export value accounted for by a product or by a country. The second measure is the entropy measure used in Hirsch and Lev (1971), which is also known as Shannon index: s i lns i The third measure is just the share of export value accounted for by the most sold product or by the country where most exports are direct: this is a quite intuitive measure of diversification, used in Braakmann and Wagner (2011). Tables 2 and 3 report the OLS coefficients, respectively for product and country diversification, from three bivariate regressions, each one containing as a regressor a different measure of diversification. Both product and country diversification increase export stability: the coefficients are negative and significant in the case of Herfindahl and Shannon index, since their increase show an increasing diversification; on the conrary, the coefficient is positive and significant for the maximum share measure, which is decreasing in diversification. On the whole, both diversification patterns appear to be relevant to stabilize exports, even if country diversification is more strongly correlated to export variability, as denoted by the R 2. In the next section, we will see that the multivariate analysis will partially change this conclusion. In what follows, we are going to show only the results obtained using the Herfindahl measure. Using the other measures does not change our main results. 4 Multivariate analysis In this section, we refine our previous analysis by considering the joint role of product and country diversification and, at the same time, by controlling for the size and the age of the firm. Concerning the size, there is a robust evidence showing a clear negative dependence 4

Table 2: Product diversification and export variability (1) (2) (3) Herfindahl Shannon Max share Herfindahl -0.4604 (0.0094) Shannon -0.2643 (0.0038) Max share 0.4424 (0.0111) N 91662 91662 91662 R 2 0.072 0.095 0.064 Note. Coefficients are from OLS regressions. The dependent variable is the standard deviation of export growth. All firms with at least four year export values are used. Standard errors in parentheses. p < 0.10, p < 0.05, p < 0.01 Table 3: Country diversification and export variability (1) (2) (3) Herfindahl Shannon Max share Herfindahl -1.0816 (0.0080) Shannon -0.4937 (0.0030) Max share 1.2379 (0.0094) N 91662 91662 91662 R 2 0.205 0.261 0.198 Note. Coefficients are from OLS regressions. The dependent variable is the standard deviation of export growth. All firms with at least four year export values are used. Standard errors in parentheses. p < 0.10, p < 0.05, p < 0.01 5

of the firm growth variance on size, usually interpreted as a violation of Gibrat s Law (Stanley et al.; 1996; Bottazzi and Secchi; 2006): one implication of this is that bigger firms enjoy less volatility in their sales or employment. We take care of this effect by controlling for the (log) export value. As for age, standard models of industry dynamics imply that younger firms should exhibit a higher volatilty in their growth rate (see Jovanovic; 1982; Cabral and Mata; 2003); this is confirmed by some empirical evidence. Moreover, older firms might have learnt both to diversify more their export and to manage better demand and supply shocks. As a consequence, age is a crucial dimension to control for in analyzing the different roles of diversification on exports stability. Table 4 reports the OLS coefficient for a multivariate regression, in which the dependent variable is, as before, the standard deviation of export growth rates and the regressors are the herfindahl indexes of product and country diversification, and as controls the (log) of export value and the age of the firm. Table 4: Country and product diversification (1) (2) (3) (4) Herfindahl product 0.0097 0.1358 0.0125 0.1354 (0.0095) (0.0087) (0.0094) (0.0087) Herfindahl country -1.0852-0.4211-1.0410-0.4048 (0.0087) (0.0094) (0.0088) (0.0094) log export value -0.1573-0.1546 (0.0012) (0.0012) log age -0.1415-0.0887 (0.0043) (0.0039) N 91662 91662 91662 91662 R 2 0.205 0.336 0.215 0.340 Note. Coefficients are from OLS regressions. The dependent variable is the standard deviation of export growth. All firms with at least four year export values are used. Standard errors in parentheses. p < 0.10, p < 0.05, p < 0.01 Both the size and the age show the expected, negative sign: bigger and older firms exhibit more stable exports. The coefficient on the Herfindahl index for geographical diversification is negative and significant in all the four specifications, even if point estimates are more than half smaller with the size control. The more striking result concerns the coefficient on product diversification, which is either not significant, or positive and significant when controlling for the size of the firm. So, product diversification does not appear to be an effective way to stabilize exports; on the contrary, firms exporting more products exhibit an higher export variability. 6

5 Conclusions In this paper, we started to investigate the relationship between epxort volatility on one hand, and product / geographical diversification on the other hand. Our preliminary results show that diversifying across countries is strongly correlated to exports stability. One interpretation of this result is that firms might insure themselves against demand shocks by sellig their products to different national markets. On the other hand, product diversification does not appear to be positively correlated to exports stability. In fact, firms the diversify more in terms of products show an higher volatility, when controlling for their size. This effect might be due to the fact that firms, in the attempt to exploit economies of scope, diversify in similar products, which are likely to be affected by correlated shocks. 7

References Aw, B.-Y. and Batra, G. (1998). Firm size and the pattern of diversification, International Journal of Industrial Organization 16(3): 313 331. Bernard, A. B., Jensen, J. B., Redding, S. J. and Schott, P. K. (2012). The empirics of firm heterogeneity and international trade, Annual Review of Economics 4(1): 283 313. Bottazzi, G. and Secchi, A. (2006). Gibrat s law and diversification, Industrial and Corporate Change 15(5): 847 875. Bowman, E. H. (1980). A risk/return paradox for strategic management. Braakmann, N. and Wagner, J. (2011). Product diversification and stability of employment and sales: first evidence from German manufacturing firms, Applied Economics 43(27): 3977 3985. Cabral, L. and Mata, J. (2003). On the evolution of the firm size distribution: Facts and theory, American Economic Review 93(4): 1075 1090. Hirsch, S. and Lev, B. (1971). Sales Stabilization Through Export Diversification, The Review of Economics and Statistics 53(3): 270 77. Jovanovic, B. (1982). Selection and the evolution of industry, Econometrica 50(3): 649 70. Kim, W. C., Hwang, P. and Burgers, W. P. (1993). Multinationals diversification and the risk-return trade-off, Strategic Management Journal 14(4): 275 286. Mayer, T., Melitz, M. J. and Ottaviano, G. I. (2011). Market size, competition, and the product mix of exporters, NBER Working Papers 16959, National Bureau of Economic Research, Inc. Stanley, M., Amaral, L., Buldyrev, S., Havlin, S., Leschhorn, H., Maass, P., Salinger, M. and Stanley, H. (1996). Scaling behaviour in the growth of companies, Nature 379: 804 806. 8