GGDC RESEARCH MEMORANDUM 146

Transcription

1 GGDC RESEARCH MEMORANDUM 146 Productivity spillovers of organization capital Wen Chen and Robert Inklaar February 2014 university of groningen groningen growth and development centre

2 Productivity spillovers of organization capital Wen Chen and Robert Inklaar * Groningen Growth and Development Centre University of Groningen, the Netherlands February 2014 Abstract Investments in organization capital may impact productivity of not just the investing firm but could also spillover to other firms like investments in research and development. In this paper, we test whether there are (positive) know-how spillovers and (negative) business-stealing productivity spillovers for a panel of 1266 U.S. manufacturing firms over the period We only find evidence of negative spillovers on product-market rivals. This implies that firms invest more in organization capital than would be socially optimal and that existing estimates of the importance of organization capital for productivity growth are overstated. Keywords: Organizational capital; Intangible assets; Spillovers. JEL Classification Numbers: D24, L22, O33. * Corresponding author: Faculty of Economics and Business, University of Groningen, PO Box 800, 9700 AV Groningen, the Netherlands; ; r.c.inklaar@rug.nl.

3 1. Introduction The role of knowledge-based assets for growth in advanced economies has drawn much recent interest from researchers and policy makers alike. 1 But while researchers are rapidly incorporating such assets into a standard sources-of-growth framework, 2 much is yet unknown about the productive impact of such assets. Since knowledge-based assets are typically intangible, they tend to be non-rival and non-excludable. This makes determining their (economy-wide) productive impact more difficult as there may be spillovers. 3 In the case of research and development (R&D) spending, this has long been known (e.g. Griliches, 1979, 1992) and the evidence for both positive knowledge and negative market-stealing spillovers is strong, see Bloom, Schankerman and van Reenen (2013, BSV henceforth). But R&D investment may not be the only type of knowledge-based investment to spill over. In this paper we analyze the possible spillovers from organization capital. Organization capital can be thought of as the information a firm has about its assets and how these can be used in production (Prescott and Visscher, 1980) or more specifically as the value of brand names and other knowledge embedded in firm-specific resources (Corrado et al., 2005). 4 There is a growing number of studies that show how organization capital is important for a firm s own productivity and overall performance, 5 and its intangibility makes it a potential source of spillovers. The contribution of this paper is that we are the first to look for spillovers from firm investment in organization capital. The literature on productivity spillovers from foreign direct investment (FDI) 6 is partly related since domestic firms could learn from the foreign multinational s superior organization. However, any learning taking place could also be on aspects of the multinational s productivity that are unrelated to organization capital. Goodridge, Haskel and Wallis (2012b) do specifically analyze the 1 See e.g. Corrado and Hulten (2010) and OECD (2013). 2 See e.g. Corrado, Hulten and Sichel (2009) and Corrado, Haskel, Jona-Lasinio and Massimiliano (2012). 3 See Nakamura (2010) for a general discussion on the difficulty of accounting for the productive impact of intangible assets. 4 See also Atkeson and Kehoe (2005). Conceptualizing organization capital as embedded in the organization distinguishes it from measures of human capital, see e.g. Jovanovic (1979) and Becker (1993). 5 See e.g. Eisfeldt and Papanikolaou (2013), Hulten and Hao (2008) and Lev and Radhakrishnan (2005). 6 See e.g. Liu (2008) and Keller and Yeaple (2009). 1

4 productivity spillovers from organization capital, yet their use of industry-level data leads to a limited sample size and a less clear identification of the pool of organization capital that firms could learn from or be otherwise affected by. Following the approach of BSV, we test for know-how spillovers that would enhance productivity in related firms, and market-stealing spillovers that would hurt productivity in related firms, defining related firms in the same way as BSV. Since organization capital aims to capture the way in which production in a firm is organized, we expect that firms are more likely to learn and benefit from the investments of firms that are close in technology space. Firms are likely to suffer, though, from investments in organization capital made by close competitors, i.e. firms that are close in product market space. By locating firms in these two spaces, we can distinguish between the two types of spillovers. We test these hypotheses using financial data for a sample of 1266 US manufacturing firms over the period and measures of proximity in technology space and in product market space. We measure investment in organization capital as selling, general and administrative (SGA) expenses, an approach followed by many in the firmlevel analysis of organization capital. 7 Past investments are cumulated into a stock of organization capital and added to a production function with (tangible) capital and labor. The proximity in technology space is determined using patent data an approach pioneered by Jaffe (1986) in the context of R&D spillovers. We assume that firms with more patents in similar technology fields have greater potential to learn from each other s organization capital. One example of such a spillover is Toyota s just-in-time system that quickly spread to other car manufacturers. 8 An example of cross-industry diffusion is the build-to-order (BTO) distribution system that originated with Dell Computers, but that has since been copied by firms in other industries, such as BMW. 9 Though patents may not perfectly reflect the scope for such copying, they may be useful 7 E.g. Eisfeldt and Papanikolaou (2013), Hulten and Hao (2008) and Lev and Radhakrishnan (2005). 8 See Liker and Morgan (2006). 9 See Gunasekaran and Ngai (2005). 2

5 in identifying the technological position of the firm in a broad sense. 10 We measure proximity in the product market space using information on the industries that each firm operates in and assume that greater overlap between firms makes them fiercer competitors. Increased investment in organization capital by competitors is likely to hurt firm performance: competitors may have to devote resources to copying successful business models such as the BTO system. Investment in organization capital also includes spending on marketing and sales, and while some of this spending may expand the market, another part is aimed at capturing market share from competitors. 11 Market share gains by competitors could leave firm productivity unaffected, but with factor adjustment costs or increasing returns to scale, market stealing may leave firms with underutilized inputs or operating at an inefficiently small scale. The marketstealing effect could also capture heterogeneity in the productive returns to organization capital across firms, as investments by competitors reduce the productive impact from the firm s own organization capital. Whatever the precise underlying reason, a significant negative market-stealing effect would mean that in a standard sources-of-growth analysis with assumed productive returns based on the user cost of capital the contribution from organization capital to productivity would be overstated. Our findings are, first, that organization capital contributes substantially to the firm s own productivity; second, that investment in organization capital by firms that are close in technology space has no effect on firm productivity; and third, that investment by firms that are close in product market space has a significant negative effect on firm productivity. These results are robust across industries and to alternative distance measures and assumptions regarding the capitalization of organization capital. Following the approach of BSV, we show that the private return to organization capital of 34 percent is much higher than the social return of 21 percent. The remainder of this paper will outline the data and methods (Section 2), present the results (Section 3) and conclude (Section 4). 10 In addition, some business methods can be and have been patented since the 1990s, see Hall (2009). 11 See Landes and Rosenfield (1994) on the long-lived nature of (some) advertising spending and, more broadly, Bagwell (2007) on the economics of advertising. 3

6 2. Data and methodology In this section we discuss the econometric approach in analyzing organization capital spillovers, followed by a detailed description of the data and the methods used to construct the measures of organization capital and the spillover pools. 2.1 Econometric specification Akin to BSV, the main model starts from a Cobb-Douglas production function for firm i at time t, extended to include organization capital: (1) Y it = A it L α it K β γ it G it where Y is a measure of output (i.e. real sales), A is Hicks-neutral technology, K is the physical capital as measured by the net stock of property, plant, and equipment, L denotes labor input measured by the number of employees and G is the stock of organization capital. To determine the importance of spillovers, equation (1) is further augmented to include the dual spillover pools of organization capital: (2) Y it = A it L α it K β it G γ φ it KH 1 φ it MKT 2 it where KHit captures the spillovers of management know-how in technology space and MKTit denotes the market-stealing effect of organization capital. After logdifferentiation, equation (2) is transformed into the following estimating equation: (3) lny it = γlng it 1 + φ 1 lnkh it 1 + φ 2 lnmkt it 1 + ωx it + η i + τ t + ε it where physical capital K and labor input L are combined into X. The technology term A is modeled as the residual, which is decomposed into a correlated firm fixed effect (η i ), a full set of time dummies (τ t ), and an idiosyncratic component (ε it ) that is allowed to be heteroskedastic and serially correlated. The main coefficients of interests are φ 1 and φ 2. Coefficient φ 1 tests for the presence of a spillover of management know-how and the latter examines whether there is a market-stealing effect of organization capital. Given our earlier discussion, we would expect to find φ 1 > 0 and φ 2 < 0. The estimation of equation (3) can be affected by structural identification problems related to measurement error and simultaneity bias. Measurement error arises because firm sales are deflated by an industry price index obtained from the Bureau of Economic 4

7 Analysis (BEA). When prices vary across firms within industries, part of the variation in sales is due variation in prices rather than quantities (Foster, Haltiwanger and Syverson, 2008). Simultaneity bias causes concern because there might be unobserved productivity shocks that are known to the firms when they choose their input levels (Griliches and Mairesse, 1998). To deal with the measurement error problem, we include the industry output index and price index as part of the control variables X following BSV. 12 The error term is assumed to include a firm fixed effect (η i ) because if the deviation between firm and industry prices is largely time-invariant, this should go a long way towards dealing with the problem of firm-specific prices (see BSV). Moreover, to the extent that unobserved, firm-specific productivity is also timeinvariant, the simultaneity problem should also be controlled for. Since spillovers might take time, the organization capital and spillover terms in equation (3) are included with a one-period lag, again following BSV. 2.2 Data sources For this paper, firm-level data is obtained by matching two major sources: patents data from Bureau van Dijk s Orbis database and company accounts data from Datastream. This paper focuses on manufacturing firms as these are the most intensive investors in intangible assets (Goodridge, Haskel and Wallis, 2012a). Manufacturing firms are also the most active in taking out patents, which is important for locating firms in technology space and thus for identifying spillovers of management know-how. For this reason, we also restrict our sample to manufacturing firms with at least one patent. This leads to data on the patenting activity of 1722 U.S. manufacturing firms, obtained from Orbis. These patent data are matched to company account data from Datastream using firm international securities identification number (ISIN) codes as the unique firm identifier. From Datastream we collect information on the number of employees (WC07011), total sales (WC01001), the stock of physical capital (net property, plant, and equipment, WC02501), and investment in organization capital (selling, general and administrative expenses, WC01101), all for the period Of the 1722 firms from Orbis, 212 were not covered in Datastream and a further See also Klette and Griliches (1996) and De Loecker (2011). 5

8 firms had missing values for one or more of the company account data items. Dropping these firms results in an unbalanced panel of 1266 U.S. manufacturing firms with over 18,000 usable observations. Table 1 provides some basic descriptive statistics on the key variables. Table 1, Descriptive statistics Median Mean SD Observations Sales SGA expenses Physical capital Employment External OC stock (tech. space) External OC stock (market space) Technology fields Product markets Notes: SGA: selling, general and administrative; OC: organization capital SD: standard deviation. Sales are deflated by the industry price index and SGA expenses are deflated by the implicit GDP price deflator; all price indices are from the Bureau of Economic Analysis. Sales, SGA expenses, physical capital and the external OC stock are all in millions of dollars; employees, technology fields and product markets are in numbers. Computation of the external OC stocks, technological fields and number of markets is explained in Sections Figure 1, Distribution of firms across industries Figure 1 shows the distribution of firms across 19 broader (2-digit) manufacturing 6

9 industries. The sample of firms is fairly concentrated in the more high-tech sectors of the economy, such as computers & electronics and chemicals & pharmaceuticals, with the top-five industries accounting for around 80 percent of the firms. However, as shown in the sensitivity analysis below, this concentration does not bias the results. 2.3 Measuring organization capital Investment in organization capital has been measured in a number of ways in the literature. These include using business surveys (Black and Lynch, 2005), part of the wage bill of managers (Squicciarini and Mouel, 2012), as the residual from a production function (Lev and Radhakrishnan, 2005) and using sales, general and administrative expenses (Tronconi and Marzetti, 2011; Eisfeldt and Papanikolaou, 2013). Given the availability of the required data, we opt to use SGA expenses for measuring investment in organization capital. Lev and Radhakrishnan (2005) present detailed arguments and examples of how resources allocated to this expense item can yield improvements in employee incentives, distribution systems, marketing technologies, and a wide range of other organizational structures. Further evidence is from Eisfeldt and Papanikolaou (2013) who find that their measure of organization capital based on SGA expenses correlates highly with the managerial quality scores constructed by Bloom and Van Reenen (2007). This evidence suggests that using SGA expenses to measure organization capital is informative about the quality of management practices across firms. SGA expenses includes R&D expenditure, so to focus specifically on organization capital we subtract of R&D expenditure to get our measure of investment in organization capital. 13 To convert this flow measure into organization capital stock, we apply the perpetual inventory method: (4) G i,t = (1 δ)g i,t 1 + SGA i,t p t where p t is the implicit GDP price deflator retrieved from the Bureau of Economic Analysis. To implement the law of motion in equation (4), an initial stock and a rate of 13 Tronconi and Marzetti (2011) measure investment as 20 percent of this amount to reflect that not all SGA expenses add to organization capital. This is irrelevant from an econometric point of view. 7

10 depreciation must be chosen. Assuming a steady-state relationship from the Solow growth model, the initial stock can be calculated according to: (5) G 0 = SGA 1 g+δ where g denotes the steady-state growth rate of organization capital and δ is the rate at which organization capital become obsolete. According to the aggregate estimates of the INTAN-Invest database compiled by Corrado et al. (2012), organization capital grows at an average rate of 6 percent per year, so we use this value for g in equation (5). Organization capital can depreciate over time for a variety of reasons. The existing management practices become obsolete if improvements come along. Moreover, organization capital can also erode through work attrition and the adoption of new products and/or production processes (Hulten and Hao, 2008). In the existing empirical works, the rate of depreciation various from a slow rate at 10 percent (Tronconi and Marzetti, 2011) to a much larger rate at 40 percent (Corrado, Hulten and Sichel, 2009). Given the fact that organization capital has two contrasting components: a long-lasting learning-by-doing element which depreciates like R&D; and a short-lived organizational forgetting dynamic which depreciates like advertising, a rate that lies in the mid-range is chosen as the benchmark rate; that is, δ = The alternative rates of 10 and 40 percent will be considered in the robustness analysis. 2.4 Technological proximity We assume that firms are more likely to learn about from firms organization capital if those firms are closer in technology space and the technology space is defined using patent data. This assumes that firms developing similar technologies, have organized their organizations similarly. As discussed earlier, the diffusion of just-in-time production system and build-to-order supply chain management are two good cases in point. We use the patents data provided in Orbis, which originates from the European Patent Office s PATSTAT database. This database covers over 80 percent of the world s patents to date and these patents are classified by four-digit international patent classification (IPC) code. This means that even if the firm had been awarded patents from patenting offices in different countries, their patents can be compared. Our sample of

11 manufacturing firms obtained around half a million patents spanning 612 technology fields, as defined by the first three digits of the IPC code. 9 All patents of a firm are included, partly for the practical reason that it is not possible to select patents for a specific time frame, but also because it is helpful to define the average technological position of each firm over time, rather than focusing only on activity for a specific period. Define the vector T i = (T i1, T i2, T i3 T i612 ) where T iτ indicates the number of patents of firm i in technology class τ. Firm s knowledge diffusion distance is then calculated as the uncentered correlation between all firm i, j paring as in Jaffe (1986): (6) Proximity KH i,j = T i T j (T i T i ) 1 2(T j T j ) 1 2 The larger the number the more effective the knowledge of (management) know-how can be diffused between firms i and j (or vice versa). As indicated in Table 1, the media firm is active in 29 technological fields, providing ample opportunity for learning from other firms in any of these fields. Analogous to BSV, the spillover pool of management know-how available to firm i at time t is calculated as: (7) KH i,t = Proximity KH j,j i i,j G j,t 2.5 Product market proximity We also locate firms in product market space, using information on the industries in which firms are active. Datastream provides up to eight industry codes for each firm at the four-digit standard industrial classification (SIC) level, which means that a firm can be active in up to eight different markets. As shown in Table 1, firms on average report sales activities in 3 different markets out of a total of 569 different four-digit SIC industries. 11 Define the vector S i = (S i1, S i2, S i3 S i569 ) where S ik indicates whether or not firm i sells goods in market k. In contrast to BSV, we have no information on the 9 The level of disaggregation of a 3-digit IPC code generates a workable and comparable amount of technology classes to that of BSV. A further breakdown of the classification codes to the fourth digit is not pursued Henderson, Jaffe and Trajtenberg (2005) argue that a finer disaggregation of patent classes is not necessarily superior because the classification is subject to a greater degree of measurement error. 11 Only 61 firms (5 percent) are active in eight different markets, while more firms (29 percent) are active in just two markets. 9

12 share of sales in each market, but that information was not crucial to their results. 13 Despite of this deficiency, firms could still be positioned in the market space (with somewhat less variation admittedly). Analogous to the know-how proximity measure, the market proximity measure for any two firms i and j is calculated as: (8) Proximity MKT i,j = S i S j (S i S i ) 1 2(S j S j ) 1 2 The spillover pool of product market for firm i in year t is then constructed as: (9) MKT i,t = Proximity MKT j,j i i,j G j,t For the separate identification of know-how and product-market spillovers we rely on differences in the two proximity measures. The correlation between the proximity metrics in technology and product-market space is 0.196, indicating substantial variation between the two proximity measures. 3. Results In this section, we discuss the main empirical findings, with first results of production function estimates without spillovers, followed by the evidence on the presence of spillovers, including the robustness of that evidence. Finally, we discuss what the spillover results imply in terms of the private and social return to investment in organization capital. 3.1 Production function estimates without spillovers Table 2 shows production function estimates, where we consider estimation with firm and year fixed effects (FE) and instrumental variables (IV). In the IV estimates, we follow Tronconi and Marzetti (2011) and use two lagged values of the inputs as instruments. The first two columns of Table 2 show production function results with only capital and labor as inputs. Both are highly significant and the sum of their coefficients is very close to unity. This implies that when taking only traditional factor inputs into account, returns to scale appear approximately constant. In the next two 13 For a further comparison, we constructed a market spillover variable for R&D stock like the one used by BSV but based on our information on the number of active markets. For 237 firms, this market spillover variable can be compared to the corresponding BSV variable. The correlation coefficient is high at 0.68, giving confidence that our market spillover measure is comparable to theirs. 10

13 columns, the stock of organization capital is added to the production function and it enters with a highly significant coefficient. This finding is in line with the earlier firmlevel analyses of organization capital and provides further support for considering intangible assets as factors in production alongside tangible capital (Corrado, Hulten and Sichel, 2005, 2009; Van Ark et al., 2009). Table 2, Firm production function estimates, without and with organization capital Physical capital and labor Augmented by organization capital FE IV FE IV Physical capital (K) 0.160*** 0.134*** 0.102*** 0.087*** (0.023) (0.031) (0.024) (0.032) Labor (L) 0.782*** 0.835*** 0.528*** 0.635*** (0.034) (0.036) (0.041) (0.048) Organization capital (G) 0.561*** 0.380*** (0.044) (0.033) Sum of coefficients Observations Adjusted R Note: FE: fixed effects; IV: instrumental variables. Dependent variable in all estimations is real sales and all specifications include firm and year fixed effects and the industry output index. Robust standard errors, clustered by firm, are shown in parentheses. In the IV columns, lagged values of the explanatory variables are used as instruments. *** p<0.01, ** p<0.05, * p<0.1. The output elasticity of organization capital is substantial in size and the sum of coefficients now point to increasing returns to scale. The IV coefficient of 0.38 is also very comparable to the coefficients found by Tronconi and Marzetti (2011): the average of their coefficients for R&D and for non-r&d firms is With these results, a necessary condition for there to be any scope for spillovers has been satisfied: organization capital contributes systematically to firm productivity. 3.2 Production function estimates with spillovers Table 3 presents the main spillover results. The FE and IV estimates are highly comparable with no significant effect on productivity from the diffusion of 14 In some of their specifications, namely when estimating a translog rather than a Cobb-Douglas function and when estimating in first differences rather than levels, they find somewhat higher coefficients. 11

14 organizational know-how from firms that are close in technology space but negative and significant effects of organizational capital of close competitors in the product market. 15 Taken at face value, this implies that investment in organization capital is too high, since there are no positive know-how spillovers to offset the negative marketstealing effects. Note that this is under the assumption that the technology space is the relevant space for diffusion of organizational capital know-how; we discuss this issue further in the conclusions. We first proceed by discussing the robustness of the results in Table 3 and the implications of these findings for social and private returns to organization capital. Table 3, Organization capital spillover estimates FE IV Physical capital (K) 0.114*** 0.101*** (0.023) (0.031) Labor (L) 0.595*** 0.616*** (0.039) (0.052) Organization capital (G) 0.210*** 0.153*** (0.037) (0.044) Know-how diffusion (KH) (0.258) (0.294) Marketing-stealing (MKT) *** *** (0.050) (0.063) Observations Adjusted R Note: FE: fixed effects; IV: instrumental variables. Dependent variable in all estimations is real sales and all specifications include firm and year fixed effects, the current and lagged value of industry sales and the industry price index. Robust standard errors, clustered by firm, are shown in parentheses. In the IV columns, two lagged values of the explanatory variables are used as instruments. *** p<0.01, ** p<0.05, * p< Sensitivity analysis We first consider the sensitivity of the results to the exclusion of individual industries. Figure 1 showed that the firms in our sample are strongly concentrated in a few, mostly high-tech, industries. In Table 4, we therefore consider the sensitivity of the results in Table 3 when excluding one at a time the five most heavily-represented industries, 15 Regressions using Newey-West standard errors as in BSV show very similar results. 12

15 who together account for 80 percent of the sample. 16 The table shows that the main findings are unchanged: there is no positive know-how spillover and significantly negative spillovers from product-market rivals. We also get very similar results for the IV specification. Table 4, Organization capital spillovers sensitivity to exclusion of individual industries Excluding industry: Chemicals Computers Electrical Machinery Miscellaneous Physical capital (K) 0.123*** 0.132*** 0.109*** 0.115*** 0.113*** (0.024) (0.033) (0.024) (0.025) (0.024) Labor (L) 0.561*** 0.559*** 0.592*** 0.594*** 0.623*** (0.040) (0.056) (0.041) (0.041) (0.039) Organization capital (G) 0.201*** 0.234*** 0.218*** 0.216*** 0.180*** (0.034) (0.057) (0.037) (0.039) (0.038) Know-how diffusion (KH) (0.273) (0.300) (0.271) (0.320) (0.247) Marketing-stealing (MKT) ** *** *** *** *** (0.041) (0.051) (0.045) (0.048) (0.046) Observations Adjusted R Note: Each column runs the FE regression from Table 3, excluding the industry named in the column title; see notes to Table 3 for details on the estimation. *** p<0.01, ** p<0.05, * p<0.1. To rule out the possibility that the results are specific to the definition of proximity in the technology and product market space, we consider two alternatives: (1) IPC code at 2-digit with SIC code at 3-digit (denoted Proximity (2-3) ) and (2) IPC code at 1-digit with SIC code at 2-digit (denoted Proximity (1-2) ). As can be in Table 5, the negative effect of product-market spillovers remains strongly significant. Defining productmarket proximity at the 2-digit level even leads to much larger negative effects. This large difference in magnitude is mainly due to a much smaller variance of the market spillover pool variable MKT i,t, since firms are by definition much closer competitors. Since the rate at which organization capital becomes obsolete is far from settled, we repeat the analysis using two other depreciation rates proposed in the literature: a slower rate at 10 percent and a faster one at 40 percent. As shown in the third and fourth column of Table 5, the market stealing effect is robust to these alternative rates of depreciation. The final column shows results are also unchanged when imposing 16 Results for the other 14 industries are very similar. 13

16 constant returns to scale on the production function by normalizing all variables by employment. In contrast to the robust evidence on the presence of a negative marketstealing effect of organization capital, no evidence is found in support of a positive spillover of organizational know-how across technology space. We obtain very similar results for the IV specification. Table 5, Organization capital spillovers sensitivity to measurement of proximity, depreciation of organization capital and returns to scale Proximity (2-3) Proximity (1-2) =10% =40% Per worker Physical capital (K) 0.114*** 0.113*** 0.128*** 0.102*** 0.073*** (0.023) (0.023) (0.023) (0.023) (0.023) Labor (L) 0.597*** 0.599*** 0.631*** 0.566*** (0.039) (0.039) (0.039) (0.039) Organization capital (G) 0.206*** 0.205*** 0.154*** 0.256*** 0.215*** (0.037) (0.036) (0.039) (0.036) (0.034) Know-how diffusion (KH) (0.328) (0.664) (0.272) (0.252) (0.049) Marketing-stealing (MKT) ** *** *** *** *** (0.055) (0.104) (0.047) (0.044) (0.049) Observations Adjusted R Note: Each column runs the FE regression from Table 3; see notes to Table 3 for details on the estimation. Proximity (2-3) defines technology proximity at the 2-digit IPC and product market proximity at the 3-digit SIC level; Proximity (1-2) is defined analogously. The subsequent two columns vary the assumption regarding the depreciation of organization capital and the final column imposes constant returns to scale by normalizing all variables by employment. *** p<0.01, ** p<0.05, * p< Private and social returns to organization capital Our results imply that the private returns to investments in organization capital are lower than the social returns. To see this, consider the output elasticity γ = ρ (G/Y), where ρ is the marginal productivity of organization capital G. If one assumes a constant marginal product γ and a constant discount rate r along with an infinite planning horizon, then ρ can be given the economic interpretation of a marginal gross internal rate of return. 15 Following BSV, the marginal social return (MSR) to organization capital of firm i is defined as the increase in aggregate output generated by a marginal increase in firm i s organization capital stock: (10) MSR = (Y/G)(γ + φ 1 ) 15 For detailed derivation and discussion, see Hall, Mairesse and Mohnen (2010). 14

17 where γ and φ 1 are the coefficients from estimating equation (3) as given in Table 3. The MSR can be interpreted as the marginal product of a firm s organization capital contributed: (1) directly from firm s own organization capital stock (γ) and (2) indirectly from the external stock of management know-how φ 1. The marginal private return (MPR) is defined as the increase in firm i s output generated by a marginal increase in its own stock of organization capital: (11) MPR = (Y/G)(γ φ 2 ) Own organization capital increase a firm s own sales, thus γ is also included in calculating the MPR. Also included is φ 2 since the firm s own organization capital also has a business stealing effect in the product market. This effect increases the private incentive to invest in organization capital by redistributing output between firms. Using the FE estimates from Table 3 and using the median value of median value of Y/G of 1.008, the MSR is 21 percent (1.008 ( )) and the MPR is 34 percent (1.008 ( )). The IV estimates imply an MSR of 15 percent and an MPR of 32 percent so for this specification, the private return is double the social return. Given the strong assumptions outlined above, though, these returns estimates should be seen as indicative (see also BSV). 4. Conclusions This paper is the first that attempts to link the studies finding productivity benefits of investments in organization capital to the aggregate sources-of-growth studies that impute a rate of return to organization capital. In the case of traditional tangible capital, aggregate productivity benefits are simply a summation of firm benefits, but when the asset is intangible, such as organization capital, there may be spillovers across firms that drive a wedge between the private and social returns of investment. Our analysis is based on a sample of 1266 US manufacturing firms. We locate each of these firms in technology space, to capture potential (positive) spillovers of organizational know-how across technologically similar firms, and in product market space to capture negative market-stealing spillovers from competitors. We find clear support for these negative spillovers and none for positive spillovers of know-how. 15

18 The absence of a significant positive know-how effect could mean that a) organizational know-how does not spillover between firms, perhaps because it is too tacit or firm-specific (Lev and Radhakrishnan, 2005); b) the negative spillovers between product-market rivals are actually a net effect of negative market-stealing and positive know-how effects; or c) firms do learn from the organization capital of other firms, but these other firms are neither close in technology nor in product-market space. While our results do not shed light on which of these three alternative explanations is most plausible, our findings do limit the scope for where positive productivity spillovers from organization capital could be found. The negative market-stealing spillovers imply that firms face a higher private return to investment in organization capital than the social return: the social return considers only the productivity improvement for the firm itself, while the investing firm also takes the negative effect on its competitors into account. Our estimates show that the negative effect on its competitors can be as large as the positive effect for the firm itself. As a result, firms can be expected to keep investing in organization capital beyond the point where social returns equal (social) marginal cost. This implies that existing studies that impute a productive return to organization capital considerably overestimate its contribution to aggregate growth. Acknowledgments The authors would like to thank Marcel Timmer for useful comments and suggestions. 16

19 References Atkeson, A. and P. Kehoe Modeling and Measuring Organization Capital. Journal of Political Economy, 113: Bagwell, K The Economic Analysis of Advertising in Handbook of Industrial Organization volume 3, ed. M. Armstrong and R. Porter, Amsterdam: Elsevier Becker, G Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education. Chicago: Chicago University Press. Black, S. and L. Lynch Measuring Organizational Capital in the New Economy. In Measuring Capital in New Economy, ed. J. Haltiwanger, C. Corrado, and D. Sichel. Chicago, IL: University of Chicago Press Bloom, N. and J. Van Reenen Measuring and Explaining Management Practices across Firms and Countries. Quarterly Journal of Economics, 122: Bloom, N., M. Schankerman and J. Van Reenen Identify Technology Spillovers and Product Market Rivalry. Econometrica, 81: Corrado, C.A., C. Hulten and D. Sichel Measuring Capital and Technology: An Expanded Framework. In Measuring Capital in New Economy, ed. John Haltiwanger, Carol Corrado, and Dan Sichel. Chicago, IL: University of Chicago Press pp Corrado, C.A., C. Hulten and D. Sichel Intangible Capital and U.S. Economic Wealth. Review of Income and Wealth, 55: Corrado, C.A., J. Haskel, C. Jona-Lasinio and I. Massimiliano Intangible Capital and Growth in Advanced Economies: Measurement Methods and Comparative Results. IZA Discussion Paper no Corrado, C.A. and C.R. Hulten (2010), How do you Measure a Technological Revolution? American Economic Review 100(2): De Loecker, J Product Differentiation, Multi-Product Firms and Estimating the Impact of Trade Liberalization on Productivity. Econometrica, 79: Eisfeldt, A. and D. Papanikolaou Organization Capital and Cross-section of Expected Returns. Journal of Finance, 68:

20 Foster, L., J. Haltiwanger and C. Syverson Reallocation, Firm Turnover, and Efficiency: Selection on Productivity or Profitability? American Economic Review, 98: Goodridge, P., J. Haskel and G. Wallis. 2012a. UK Innovation Index: Productivity Growth in UK Industries. CEPR Discussion Papers no Goodridge, P., J. Haskel and G. Wallis. 2012b. Spillovers from R&D and other intangible investment: evidence from UK industries Imperial College Business School Discussion paper 2012/9 Griliches, Z Issues in assessing the contribution of research and development to productivity growth. Bell Journal of Economics, 10: Griliches, Z The Search for R&D Spillovers. Scandinavian Journal of Economics, 94: Griliches, Z. and J. Mairesse Production Function: The Search for Identification. In Econometrics and Economic Theory in the Twentieth Century: The Ragnar Frisch Centennial Symposium, ed. S. Strom. Cambridge, UK: Cambridge University Press Gunasekaran, A. and E. Ngai Build-to-order Supply Chain Management: A Literature Review and Framework for Development. Journal of Operations Management, 23: Hall, B. H The Manufacturing Sector Master File: NBER Working Paper no Hall, B. H Business And Financial Method Patents, Innovation, And Policy Scottish Journal of Political Economy, 56(s1): Hall, B. H., J. Mairesse and P. Mohnen Measuring the Returns to R&D. In Handbook of the Economics of Innovation, ed. Bronwyn. H. Hall and Nathan Rosenberg. Vol. 2 Elsevier, Henderson, R., A. Jaffe and M. Trajtenberg Patent Citations and the Geography of Knowledge Spillovers: A Re-assessment - Comment. American Economic Review, 95: Hulten, C. and X. Hao What is a company really worth? Intangible capital and the market to book value puzzle. NBER Working Paper No Jaffe, A Technological Opportunity and Spillovers of R&D: Evidence from Firms Patents, Profits, and Market Value. American Economic Review, 76:

21 Jovanovic, B Job Matching and the Theory of Turnover. Journal of Political Economy, 87: Keller, W. and S. R. Yeaple Multinational enterprises, international trade, and productivity growth: firm level evidence from the United States Review of Economics and Statistics 91(4): Klette, T. and Z. Griliches The Inconsistency of Common Scale Estimators When Output Prices Are Unobserved and Endogenous. Journal of Applied Econometrics, 11: Landes, E, M. and A. M. Rosenfield (1994). The Durability of Advertising Revisited, Journal of Industrial Economics, 42(3): Lev, B. and S. Radhakrishnan The Valuation of Organization Capital. In Measuring Capital in the New Economy, ed. John Haltiwanger, Carol Corrado, and Dan Sichel. Chicago, IL: University of Chicago Press pp Liker, J. and J. Morgan The Toyota Way in Services: The Case of Lean Product Development. Academy of Management Perspectives, 20: Liu, Z Foreign direct investment and technology spillovers: theory and evidence Journal of Development Economics 85: Nakamura, L Intangible Assets and National Income Accounting. Review of Income and Wealth, 56: OECD (2013), Supporting Investment in Knowledge Capital, Growth and Innovation, OECD: Paris. Prescott, E. and M. Visscher Organization Capital. Journal of Political Economy, 88: Squicciarini, M. and M. Le Mouel Defining and Measuring Investment in Organizational Capital: Using US Microdata to Develop a Task-based Approach. OECD Working Paper no. 2012/5. Tronconi, C. and V. Marzetti Organization Capital and Firm Performance: Empirical Evidence for European Firms. Economics Letters, 112: Van Ark, B., J. Hao, C. Corrado and C. Hulten Measuring Intangible Capital and its Contribution to Economic Growth in Europe. European Investment Bank Working paper, 14:

22