- Email: [email protected]
Diversification discount over the long run: New perspectives
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
JID: FRL
Finance Research Letters 000 (2015) 1–6
Contents lists available at ScienceDirect
Finance Research Letters journal homepage: www.elsevier.com/locate/frl
Diversification discount over the long run: New perspectives Mieszko Mazur a,∗, Shage Zhang b,1 a b
IESEG School of Management, 3 Rue de la Digue, 59000 Lille, France School of Business, Trinity University, One Trinity Place, San Antonio, TX 78209, United States
a r t i c l e
i n f o
Article history: Received 26 July 2015 Accepted 19 August 2015 Available online xxx JEL classification: G30 G31 G32 G34 Keywords: Diversification discount Internal capital market Capital allocation efficiency
a b s t r a c t We study the long-term trend of excess value and internal capital allocation of diversified firms from 1976 to 2013. The late 1970s and 1980s are characterized by large average diversification discount but narrow dispersion of excess value. Excess value of diversified firms becomes less negative on average after 1990, but its dispersion grows larger. In contrast, capital allocation efficiency of diversified firms converges significantly over time. Three quarters of diversified firms do not suffer from severe capital misallocation after the early 2000s. The effect of capital allocation efficiency on excess value varies over time and becomes larger in recent years. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Much of what we know about the value of diversification and the working of internal capital market is taken from research performed before the early 1990s, with only a handful of studies expanding the sample period to the early 2000s (Laeven and Levine, 2007; Datta et al., 2009; Schmid and Walter, 2009; Hoechle et al., 2012). Lang and Stulz (1994), Berger and Ofek (1995) and Servaes (1996) first document that Tobin’s Q of diversified firms is persistently lower than the sum of their segment values imputed from stand-alone firms in the same industry. These earlier results are verified in Rajan et al. (2000), Lamont and Polk (2002), Laeven and Levine (2007) and Schmid and Walter (2009), which find that diversification ∗
1
Corresponding author. Tel.: +33 3 20 54 58 92; fax: +33 3 20 57 48 55. E-mail addresses: [email protected], [email protected] (M. Mazur), [email protected] (S. Zhang). Tel.: +1 210 999 7083.
http://dx.doi.org/10.1016/j.frl.2015.08.008 1544-6123/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
2
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
causes value destruction and provide evidence against the argument that diversification discount is due to measurement errors (Whited, 2001) or firms’ endogenous choice of diversifying (Campa and Kedia, 2002). One prominent interpretation of such diversification discount is that, although the formation of an internal capital market provides a valuable real option in allocating capital across divisions in the presence of costly external financing (Matsusaka and Nanda, 2002), such internal capital market plays limited roles and fails to direct corporate resources to the best use (Berger and Ofek, 1995; Lamont, 1997; Shin and Stulz, 1998). The inefficient functioning of the internal capital market is likely due to internal politics and top management’s limited power over divisional managers (Rajan et al., 2000), or poor corporate governance and weak executive incentives (Scharfstein and Stein, 2000; Datta et al., 2009; Ozbas and Scharfstein, 2010; Saunter and Villalonga, 2010). In a recent study, Glaser and Riepe (2014) find that the risk-adjusted asset allocation in banks is also related to segments’ business activities. For the first time in the literature, this study provides an analysis on diversification discount and internal capital market efficiency over a long period from 1976 to 2013. Examining long-term trend sheds light on how the value creation (or destruction) of diversification evolves with the deepening of financial market, the decrease of market frictions, and the improvement in technology and productivity. Focusing beyond the late 1990s is particularly interesting, because financial markets have witnessed several significant events since then, including the dotcom bubble and subsequent burst, the booming of securitization and the recent financial crisis and economic recession. These events alter the availability and cost of external financing, which in turn influence the value-added from having an internal capital market. Moreover, thanks to recent regulatory changes, corporate governance has greatly improved to mitigate agency problems, and hence the efficiency of internal capital market is expected to increase. Prior studies find that diversification was perceived poorly by capital markets leading up to the previous refocusing wave in the 1980s. This study also contributes in providing evidence on how market perceives diversification during the two recent merger waves featuring mega cross-border deals. Several new patterns emerge. First, the diversification discount exists persistently but reduces over time. It disappears and even turns into a premium during the dotcom bubble burst and the financial crisis. Second, the dispersion of excess value among diversified firms grows wider over time. Its top quartile is twice as high in the late 2000s as it was in the late 1970s, while its bottom quartile becomes more negative. Third, the internal capital allocation efficiency slightly improves on average, but its distribution converges greatly over time. Capital misallocation is less severe in about three quarters of diversified firms after early 2000. Last, the effect of allocation efficiency on improving excess value is much larger and significant during the recent two merger waves, compared to the earlier period of refocusing merger wave. 2. Data The data are drawn from Compustat Segment and Industrial Annual data files for 1976–2013. To examine whether diversification creates or destroys corporate value, we calculate the excess value (EXV) of a firm’s Tobin’s Q over its imputed Q using the algorithm from Berger and Ofek (1995) based on median asset multiples.2 We use relative value added by allocation (RVA) from Rajan et al. (2000) to capture the efficiency of capital allocation. Firms in the financial industry are not included in the sample because Glaser and Riepe (2014) point out that banks allocate “risk-bearing” capacity instead of investment budgets. Hence, general allocation measures do not work well for financial companies. We also exclude heavily regulated utility firms from the sample. Table 1 reports the summary statistics of key variables. During the sample period, the Statement of Financial Accounting Standards 131 (SFAS 131) issued in 1997 leads to changes in the reporting of segment data starting in 1998. Hoechle et al. (2012) pointed out that such reporting changes lead to an increase in the number of reported segments. It is not clear how these changes affect the comparability of excess value and allocation efficiency measures before and after SFAS 131. Both the trend analysis and regression analysis do not reveal systematic changes in the two variables and their relationship after the implementation of the new rule. Nevertheless, we try to 2 As a robustness check, we follow Berger and Ofek (1995) and repeat the analysis using sales and EBIT multiples. The results remain qualitatively similar in each case. We also find similar results when using the excess value measure constructed following the Rajan et al. (2000) approach.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
JID: FRL
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
3
Table 1 Summary statistics. This table reports mean, median, standard deviation, first, and third quartile of the variables used in the empirical analysis. The sample consists of all diversified U.S. firms in the Compustat Historical Segments database between 1976 and 2013. Financial data come from Compustat North America. Detailed descriptions of all variables and data sources are provided in Appendix A. Variable
Mean
St.dev.
Q1
Median
Q3
Obs.
Excess value Allocation efficiency × 100 Firm size Capital expenditures Cash holdings Segment diversity
−0.0791 −0.0184 6.2460 0.0642 0.0681 0.2859
0.4775 1.2036 1.8748 0.0662 0.0951 0.1953
−0.3771 −0.0711 4.7436 0.0247 0.0101 0.1278
−0.0872 0.0000 6.1695 0.0454 0.0303 0.2484
0.2006 0.0486 7.5617 0.0795 0.0861 0.4145
20,527 20,527 20,527 20,527 19,339 9,853
Fig. 1. Excess value. The figure plots excess value for the sample of diversified U.S. firms from 1976 through 2013. We calculate excess value using the Berger and Ofek (1995) procedure. The data come from Compustat Historical Segments and Compustat North America. Detailed descriptions of all variables and data sources are provided in Appendix A.
avoid making direct inferences from observed changes immediately around 1998. Instead, we focus on the general trend in the pre- and post-SFAS 131 periods separately. Moreover, in regression analysis we include year dummy variables to pick up the changes in reporting as Hoechle et al. (2012) do. 3. Empirical results 3.1. The excess value Fig. 1 shows the trend of the excess value of diversified firms. Consistent with the large extant literature documenting diversification discount, we find the average excess value is negative most of the time from 1976 to 2013. The diversification discount is noticeably larger through the late-1970s to the mid-1980s preceding the wave of corporate refocus and it becomes lower in early 1990s. In the post-SFAS 131 era, the diversification discount becomes larger on average during the boom of securitization from 2003 to 2006 when there is a high credit supply and better access to financing. In contrast, the diversification discount turns to a premium on average during the dotcom bubble burst of 2001–2002 and becomes zero during the financial crisis of 2007–2008. This suggests that diversified firms do not underperform and even outperform stand-alone firms on average during these two periods when borrowing and external financing are much more difficult. Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
4
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
Fig. 2. Allocation efficiency. The figure plots allocation efficiency computed for our sample of diversified U.S. firms between 1976 and 2013 using the approach of Rajan et al. (2000). The data come from Compustat Historical Segments and Compustat North America. Detailed descriptions of all variables and data sources are provided in Appendix A.
Variation in excess value is large, with the top quartile around 20% and the bottom quartile around 37%. About 34–56% of the diversified firms are traded at a premium each year during the sample period. As shown in Fig. 1, the dispersion of excess value grows wider since the 1990s with the top quartile steadily improving and the bottom quartile going down. By the late 2000s, the top performing diversified firms are traded more than 30% higher than stand-alone firms while the bottom performing diversified firms are traded at more than a 37% discount. 3.2. The efficiency of internal capital market The trend in capital allocation efficiency shows a clearly different picture. The distribution of RVA becomes significantly tighter over time. Fig. 2 shows that RVA has a larger dispersion from the late-1970s to the mid-1980s and it shrinks significantly during the 1990s. After early 2000, the bottom quartile of RVA moves very close to zero, indicating that roughly three quarters of the diversified firms do not suffer from much value loss due to inefficient allocation. One potential explanation of such improvement is the series regulatory changes that improve corporate governance and reduce agency problems, which can have greater impact on bottom quartile firms. 3.3. Effect of capital allocation on excess value The divergence in excess value and the convergence in capital allocation efficiency among diversified firms, especially in more recent years, is an interesting phenomenon. To explore further, we test the relationship between allocation efficiency and excess value using multivariate regressions as shown in Table 2. We control for variables that likely affect excess value, including firm size measured by the logarithm of total book assets, firm capital expenditures over total book assets, level of diversity, and firm cash holding over total book assets. Firm and year fixed effects are also included. Using the full sample from 1976 to 2013, we find that RVA is positive and highly significant, suggesting that more efficient capital allocation is associated with higher excess value, largely consistent with what previous literature documented (Rajan et al., 2000; Ahn and Denis, 2004). In further analysis over the three merger waves shown in Table 3, we find the marginal effect of RVA on EXV varies over time. RVA is positive, albeit insignificant, during the refocusing merge wave of 1984–1989 when most conglomerate Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49] M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
5
Table 2 Multivariate regressions of excess value on allocation efficiency. This table reports estimates obtained from OLS regressions with firm and year fixed effects performed on a sample of all diversified U.S. firms between 1976 and 2013. Standard errors are robust to heteroscedasticity and clustered at the firm level, p-values are reported in parentheses. ∗ , ∗∗ , and ∗∗∗ denote significance at 10%, 5%, and 1% level, respectively. Detailed descriptions of all variables and data sources are provided in Appendix A. Variable
(1)
Allocation efficiency
(2) ∗∗∗
(3)
1.4555 (0.000)
1.5558 (0.000) −0.0611∗∗∗ (0.000) 0.8653∗∗∗ (0.000) 0.7797∗∗∗ (0.000)
1.5765∗∗∗ (0.005) −0.0819∗∗∗ (0.000) 0.7994∗∗∗ (0.000) 0.5863∗∗∗ (0.000) −0.1922∗∗∗ (0.000)
Yes Yes Yes 0.59 0.18 20,527
Yes Yes Yes 0.60 0.21 19,339
Yes Yes Yes 0.63 0.21 9,237
Firm size Capital expenditures Cash holdings
∗∗∗
Segment diversity Year fixed effects Firm fixed effects Clustered standard errors R-squared Adjusted R Obs.
Table 3 Relationship between excess value and allocation efficiency over different subperiods. This table reports the estimated coefficients, along with the p-values from regressing excess value on allocation efficiency and various control measures (firm size, capital expenditures, and cash holdings). The coefficients of control variables are not shown here. We use OLS regressions over different sample subperiods controlling for firm and year fixed effects. Standard errors are robust to heteroscedasticity and clustered at the firm level. ∗ , ∗∗ , and ∗∗∗ denote significance at 10%, 5%, and 1% level, respectively. Detailed descriptions of all variables and data sources are provided in Appendix A. Time period 1976–2013 1984–1989 1993–2000 2002–2008 1997–2000 2001–2002 2007–2008
Coefficient 1.5558 0.1014 0.9781 3.0251 0.6138 2.6124 2.0187
p-value ∗∗∗
0.000 0.955 0.051∗ 0.000∗∗∗ 0.272 0.030∗∗ 0.069∗
R-squared
Adjusted R
Obs.
0.59 0.75 0.78 0.70 0.83 0.89 0.86
0.18 0.13 0.35 0.30 0.32 0.46 0.36
19,339 1,227 4,188 6,180 2,858 1,830 1,673
firms reduced their level of diversity, and hence the effect of capital allocation is less significant (Rajan et al., 2000).3 The effect of RVA on EXV becomes highly significant and much larger in magnitude during the 1993–2000 and 2002–2008 periods. The former represents an era of mega cross-border M&A deals and the latter features large deals with active participation of shareholder activists and private equity funds. Capital allocation efficiency is much more important to firm value in these gigantic conglomerate firms arising from large M&A deals. Meanwhile, value-reducing allocation is also less frequent during this period as shown in Fig. 2.4 A detailed analysis in the post-2000 period suggests that the marginal effect of 3 Rajan et al. (2000) find that RVA is positively and significantly associated with EXV in 1980–1993. Using data in the same period, we are able to replicate their results. 4 To make sure this result is not distorted by the change of reporting rules in 1998, we run the same test for two years before and after (including) 1998. The marginal effect of RVA on excess value is comparable in these two regressions.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
6
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
RVA is substantially larger over the 2001–2002 period of the dotcom bubble burst and the 2007–2008 period of financial crisis. The estimated coefficient of RVA is 2.61 and 2.01 during these two periods, respectively. In comparison, the coefficient of RVA is much smaller and insignificant during the dotcom bubble from 1997 to 2000, only around 0.61. This pattern is strongly consistent with the theoretical argument in Matsusaka and Nanda (2002) that the functioning of the internal capital market is more important and value-enhancing for diversified firms when external financial market condition is worse. 4. Conclusions We conduct a trend analysis on the excess value and capital allocation efficiency of diversified firms over the long period of 1976–2013. Diversification discount exists persistently but its magnitude is smaller in recent years. It even turns into a premium during the dotcom bubble burst and becomes zero during the financial crisis, suggesting that diversified firms are less affected by shocks in external financial markets. Perhaps this can be attributed to the benefit of having an internal capital market. We find capital allocation efficiency has a much stronger effect on excess value during recent periods. Correspondingly, the trend analysis suggests that severe capital misallocation is less frequent after the early 2000s. Appendix A. Variable definitions and sources of data Variable
Definition
Data source
Excess value
Value of the diversified firm computed following the procedure described in more detail in Berger and Ofek (1995) Value added from internal capital allocation between business segments computed as in Rajan et al. (2000) Natural logarithm of the book value of total assets Expenditures on property, plant, and equipment scaled by the book value of total assets Sum of cash and marketable securities scaled by the book value of total assets Standard deviation of segment asset-weighted Tobin’s Q scaled by equally weighted average Tobin’s Q for all business segments in the firm
Compustat
Allocation efficiency Firm size Capital expenditures Cash holdings Segment diversity
Compustat Compustat Compustat Compustat Compustat
References Ahn, S., Denis, D.J., 2004. Internal capital markets and investment policy: evidence from corporate spinoffs. J. Finan. Econ. 71 (3), 489–516. Berger, P.G., Ofek, E., 1995. Diversification’s effect on firm value. J. Finan. Econ. 37 (1), 39–65. Campa, J.M., Kedia, S., 2002. Explaining the diversification discount. J. Finance 57 (4), 1731–1762. Datta, S., D’Mello, R., Iskandar-Datta, M., 2009. Executive compensation and internal capital market efficiency. J. Finan. Intermediation 18 (2), 242–258. Glaser, M., Riepe, J., 2014. Internal capital market studies in empirical banking: biases due to usage of assets instead of risk capital? Finan. Res. Lett. 11 (1), 47–53. Hoechle, D., Schmid, M., Walter, I., Yermack, D., 2012. How much of the diversification discount can be explained by poor corporate governance? J. Finan. Econ. 103 (1), 41–60. Laeven, L., Levine, R., 2007. Is there a diversification discount in financial conglomerates? J. Finan. Econ. 85 (2), 331–367. Lamont, O., 1997. Cash flow and investment: evidence from internal capital markets. J. Finance 52 (1), 83–109. Lamont, O.A., Polk, C., 2002. Does diversification destroy value? Evidence from the industry shocks. J. Finan. Econ. 63 (1), 51–77. Lang, L.H., Stulz, R.M., 1994. Tobin’s q, corporate diversification, and firm performance. J. Polit. Economy 102 (6), 1248–1280. Matsusaka, J.G., Nanda, V., 2002. Internal capital markets and corporate refocusing. J. Finan. Intermediation 11 (2), 176–211. Ozbas, O., Scharfstein, D.S., 2010. Evidence on the dark side of internal capital markets. Rev. Finan. Stud. 23 (2), 581–599. Rajan, R., Servaes, H., Zingales, L., 2000. The cost of diversity: the diversification discount and inefficient investment. J. Finance 55 (1), 35–80. Sautner, Z., & Villalonga, B. (2010). Corporate governance and internal capital markets. Harvard Business School Finance Working Paper Scharfstein, D.S., Stein, J.C., 2000. The dark side of internal capital markets: divisional rent-seeking and inefficient investment. J. Finance 55 (6), 2537–2564. Schmid, M.M., Walter, I., 2009. Do financial conglomerates create or destroy economic value? J. Finan. Intermediation 18 (2), 193– 216. Servaes, H., 1996. The value of diversification during the conglomerate merger wave. J. Finance 51 (4), 1201–1225. Shin, H.H., Stulz, R.M., 1998. Are internal capital markets efficient? Q. J. Econ. 113 (2), 531–552. Whited, T.M., 2001. Is it inefficient investment that causes the diversification discount? J. Finance 56 (5), 1667–1691.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
Finance Research Letters 000 (2015) 1–6
Contents lists available at ScienceDirect
Finance Research Letters journal homepage: www.elsevier.com/locate/frl
Diversification discount over the long run: New perspectives Mieszko Mazur a,∗, Shage Zhang b,1 a b
IESEG School of Management, 3 Rue de la Digue, 59000 Lille, France School of Business, Trinity University, One Trinity Place, San Antonio, TX 78209, United States
a r t i c l e
i n f o
Article history: Received 26 July 2015 Accepted 19 August 2015 Available online xxx JEL classification: G30 G31 G32 G34 Keywords: Diversification discount Internal capital market Capital allocation efficiency
a b s t r a c t We study the long-term trend of excess value and internal capital allocation of diversified firms from 1976 to 2013. The late 1970s and 1980s are characterized by large average diversification discount but narrow dispersion of excess value. Excess value of diversified firms becomes less negative on average after 1990, but its dispersion grows larger. In contrast, capital allocation efficiency of diversified firms converges significantly over time. Three quarters of diversified firms do not suffer from severe capital misallocation after the early 2000s. The effect of capital allocation efficiency on excess value varies over time and becomes larger in recent years. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Much of what we know about the value of diversification and the working of internal capital market is taken from research performed before the early 1990s, with only a handful of studies expanding the sample period to the early 2000s (Laeven and Levine, 2007; Datta et al., 2009; Schmid and Walter, 2009; Hoechle et al., 2012). Lang and Stulz (1994), Berger and Ofek (1995) and Servaes (1996) first document that Tobin’s Q of diversified firms is persistently lower than the sum of their segment values imputed from stand-alone firms in the same industry. These earlier results are verified in Rajan et al. (2000), Lamont and Polk (2002), Laeven and Levine (2007) and Schmid and Walter (2009), which find that diversification ∗
1
Corresponding author. Tel.: +33 3 20 54 58 92; fax: +33 3 20 57 48 55. E-mail addresses: [email protected], [email protected] (M. Mazur), [email protected] (S. Zhang). Tel.: +1 210 999 7083.
http://dx.doi.org/10.1016/j.frl.2015.08.008 1544-6123/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
2
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
causes value destruction and provide evidence against the argument that diversification discount is due to measurement errors (Whited, 2001) or firms’ endogenous choice of diversifying (Campa and Kedia, 2002). One prominent interpretation of such diversification discount is that, although the formation of an internal capital market provides a valuable real option in allocating capital across divisions in the presence of costly external financing (Matsusaka and Nanda, 2002), such internal capital market plays limited roles and fails to direct corporate resources to the best use (Berger and Ofek, 1995; Lamont, 1997; Shin and Stulz, 1998). The inefficient functioning of the internal capital market is likely due to internal politics and top management’s limited power over divisional managers (Rajan et al., 2000), or poor corporate governance and weak executive incentives (Scharfstein and Stein, 2000; Datta et al., 2009; Ozbas and Scharfstein, 2010; Saunter and Villalonga, 2010). In a recent study, Glaser and Riepe (2014) find that the risk-adjusted asset allocation in banks is also related to segments’ business activities. For the first time in the literature, this study provides an analysis on diversification discount and internal capital market efficiency over a long period from 1976 to 2013. Examining long-term trend sheds light on how the value creation (or destruction) of diversification evolves with the deepening of financial market, the decrease of market frictions, and the improvement in technology and productivity. Focusing beyond the late 1990s is particularly interesting, because financial markets have witnessed several significant events since then, including the dotcom bubble and subsequent burst, the booming of securitization and the recent financial crisis and economic recession. These events alter the availability and cost of external financing, which in turn influence the value-added from having an internal capital market. Moreover, thanks to recent regulatory changes, corporate governance has greatly improved to mitigate agency problems, and hence the efficiency of internal capital market is expected to increase. Prior studies find that diversification was perceived poorly by capital markets leading up to the previous refocusing wave in the 1980s. This study also contributes in providing evidence on how market perceives diversification during the two recent merger waves featuring mega cross-border deals. Several new patterns emerge. First, the diversification discount exists persistently but reduces over time. It disappears and even turns into a premium during the dotcom bubble burst and the financial crisis. Second, the dispersion of excess value among diversified firms grows wider over time. Its top quartile is twice as high in the late 2000s as it was in the late 1970s, while its bottom quartile becomes more negative. Third, the internal capital allocation efficiency slightly improves on average, but its distribution converges greatly over time. Capital misallocation is less severe in about three quarters of diversified firms after early 2000. Last, the effect of allocation efficiency on improving excess value is much larger and significant during the recent two merger waves, compared to the earlier period of refocusing merger wave. 2. Data The data are drawn from Compustat Segment and Industrial Annual data files for 1976–2013. To examine whether diversification creates or destroys corporate value, we calculate the excess value (EXV) of a firm’s Tobin’s Q over its imputed Q using the algorithm from Berger and Ofek (1995) based on median asset multiples.2 We use relative value added by allocation (RVA) from Rajan et al. (2000) to capture the efficiency of capital allocation. Firms in the financial industry are not included in the sample because Glaser and Riepe (2014) point out that banks allocate “risk-bearing” capacity instead of investment budgets. Hence, general allocation measures do not work well for financial companies. We also exclude heavily regulated utility firms from the sample. Table 1 reports the summary statistics of key variables. During the sample period, the Statement of Financial Accounting Standards 131 (SFAS 131) issued in 1997 leads to changes in the reporting of segment data starting in 1998. Hoechle et al. (2012) pointed out that such reporting changes lead to an increase in the number of reported segments. It is not clear how these changes affect the comparability of excess value and allocation efficiency measures before and after SFAS 131. Both the trend analysis and regression analysis do not reveal systematic changes in the two variables and their relationship after the implementation of the new rule. Nevertheless, we try to 2 As a robustness check, we follow Berger and Ofek (1995) and repeat the analysis using sales and EBIT multiples. The results remain qualitatively similar in each case. We also find similar results when using the excess value measure constructed following the Rajan et al. (2000) approach.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
JID: FRL
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
3
Table 1 Summary statistics. This table reports mean, median, standard deviation, first, and third quartile of the variables used in the empirical analysis. The sample consists of all diversified U.S. firms in the Compustat Historical Segments database between 1976 and 2013. Financial data come from Compustat North America. Detailed descriptions of all variables and data sources are provided in Appendix A. Variable
Mean
St.dev.
Q1
Median
Q3
Obs.
Excess value Allocation efficiency × 100 Firm size Capital expenditures Cash holdings Segment diversity
−0.0791 −0.0184 6.2460 0.0642 0.0681 0.2859
0.4775 1.2036 1.8748 0.0662 0.0951 0.1953
−0.3771 −0.0711 4.7436 0.0247 0.0101 0.1278
−0.0872 0.0000 6.1695 0.0454 0.0303 0.2484
0.2006 0.0486 7.5617 0.0795 0.0861 0.4145
20,527 20,527 20,527 20,527 19,339 9,853
Fig. 1. Excess value. The figure plots excess value for the sample of diversified U.S. firms from 1976 through 2013. We calculate excess value using the Berger and Ofek (1995) procedure. The data come from Compustat Historical Segments and Compustat North America. Detailed descriptions of all variables and data sources are provided in Appendix A.
avoid making direct inferences from observed changes immediately around 1998. Instead, we focus on the general trend in the pre- and post-SFAS 131 periods separately. Moreover, in regression analysis we include year dummy variables to pick up the changes in reporting as Hoechle et al. (2012) do. 3. Empirical results 3.1. The excess value Fig. 1 shows the trend of the excess value of diversified firms. Consistent with the large extant literature documenting diversification discount, we find the average excess value is negative most of the time from 1976 to 2013. The diversification discount is noticeably larger through the late-1970s to the mid-1980s preceding the wave of corporate refocus and it becomes lower in early 1990s. In the post-SFAS 131 era, the diversification discount becomes larger on average during the boom of securitization from 2003 to 2006 when there is a high credit supply and better access to financing. In contrast, the diversification discount turns to a premium on average during the dotcom bubble burst of 2001–2002 and becomes zero during the financial crisis of 2007–2008. This suggests that diversified firms do not underperform and even outperform stand-alone firms on average during these two periods when borrowing and external financing are much more difficult. Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
4
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
Fig. 2. Allocation efficiency. The figure plots allocation efficiency computed for our sample of diversified U.S. firms between 1976 and 2013 using the approach of Rajan et al. (2000). The data come from Compustat Historical Segments and Compustat North America. Detailed descriptions of all variables and data sources are provided in Appendix A.
Variation in excess value is large, with the top quartile around 20% and the bottom quartile around 37%. About 34–56% of the diversified firms are traded at a premium each year during the sample period. As shown in Fig. 1, the dispersion of excess value grows wider since the 1990s with the top quartile steadily improving and the bottom quartile going down. By the late 2000s, the top performing diversified firms are traded more than 30% higher than stand-alone firms while the bottom performing diversified firms are traded at more than a 37% discount. 3.2. The efficiency of internal capital market The trend in capital allocation efficiency shows a clearly different picture. The distribution of RVA becomes significantly tighter over time. Fig. 2 shows that RVA has a larger dispersion from the late-1970s to the mid-1980s and it shrinks significantly during the 1990s. After early 2000, the bottom quartile of RVA moves very close to zero, indicating that roughly three quarters of the diversified firms do not suffer from much value loss due to inefficient allocation. One potential explanation of such improvement is the series regulatory changes that improve corporate governance and reduce agency problems, which can have greater impact on bottom quartile firms. 3.3. Effect of capital allocation on excess value The divergence in excess value and the convergence in capital allocation efficiency among diversified firms, especially in more recent years, is an interesting phenomenon. To explore further, we test the relationship between allocation efficiency and excess value using multivariate regressions as shown in Table 2. We control for variables that likely affect excess value, including firm size measured by the logarithm of total book assets, firm capital expenditures over total book assets, level of diversity, and firm cash holding over total book assets. Firm and year fixed effects are also included. Using the full sample from 1976 to 2013, we find that RVA is positive and highly significant, suggesting that more efficient capital allocation is associated with higher excess value, largely consistent with what previous literature documented (Rajan et al., 2000; Ahn and Denis, 2004). In further analysis over the three merger waves shown in Table 3, we find the marginal effect of RVA on EXV varies over time. RVA is positive, albeit insignificant, during the refocusing merge wave of 1984–1989 when most conglomerate Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49] M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
5
Table 2 Multivariate regressions of excess value on allocation efficiency. This table reports estimates obtained from OLS regressions with firm and year fixed effects performed on a sample of all diversified U.S. firms between 1976 and 2013. Standard errors are robust to heteroscedasticity and clustered at the firm level, p-values are reported in parentheses. ∗ , ∗∗ , and ∗∗∗ denote significance at 10%, 5%, and 1% level, respectively. Detailed descriptions of all variables and data sources are provided in Appendix A. Variable
(1)
Allocation efficiency
(2) ∗∗∗
(3)
1.4555 (0.000)
1.5558 (0.000) −0.0611∗∗∗ (0.000) 0.8653∗∗∗ (0.000) 0.7797∗∗∗ (0.000)
1.5765∗∗∗ (0.005) −0.0819∗∗∗ (0.000) 0.7994∗∗∗ (0.000) 0.5863∗∗∗ (0.000) −0.1922∗∗∗ (0.000)
Yes Yes Yes 0.59 0.18 20,527
Yes Yes Yes 0.60 0.21 19,339
Yes Yes Yes 0.63 0.21 9,237
Firm size Capital expenditures Cash holdings
∗∗∗
Segment diversity Year fixed effects Firm fixed effects Clustered standard errors R-squared Adjusted R Obs.
Table 3 Relationship between excess value and allocation efficiency over different subperiods. This table reports the estimated coefficients, along with the p-values from regressing excess value on allocation efficiency and various control measures (firm size, capital expenditures, and cash holdings). The coefficients of control variables are not shown here. We use OLS regressions over different sample subperiods controlling for firm and year fixed effects. Standard errors are robust to heteroscedasticity and clustered at the firm level. ∗ , ∗∗ , and ∗∗∗ denote significance at 10%, 5%, and 1% level, respectively. Detailed descriptions of all variables and data sources are provided in Appendix A. Time period 1976–2013 1984–1989 1993–2000 2002–2008 1997–2000 2001–2002 2007–2008
Coefficient 1.5558 0.1014 0.9781 3.0251 0.6138 2.6124 2.0187
p-value ∗∗∗
0.000 0.955 0.051∗ 0.000∗∗∗ 0.272 0.030∗∗ 0.069∗
R-squared
Adjusted R
Obs.
0.59 0.75 0.78 0.70 0.83 0.89 0.86
0.18 0.13 0.35 0.30 0.32 0.46 0.36
19,339 1,227 4,188 6,180 2,858 1,830 1,673
firms reduced their level of diversity, and hence the effect of capital allocation is less significant (Rajan et al., 2000).3 The effect of RVA on EXV becomes highly significant and much larger in magnitude during the 1993–2000 and 2002–2008 periods. The former represents an era of mega cross-border M&A deals and the latter features large deals with active participation of shareholder activists and private equity funds. Capital allocation efficiency is much more important to firm value in these gigantic conglomerate firms arising from large M&A deals. Meanwhile, value-reducing allocation is also less frequent during this period as shown in Fig. 2.4 A detailed analysis in the post-2000 period suggests that the marginal effect of 3 Rajan et al. (2000) find that RVA is positively and significantly associated with EXV in 1980–1993. Using data in the same period, we are able to replicate their results. 4 To make sure this result is not distorted by the change of reporting rules in 1998, we run the same test for two years before and after (including) 1998. The marginal effect of RVA on excess value is comparable in these two regressions.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008
JID: FRL
ARTICLE IN PRESS [m1G;September 10, 2015;8:49]
6
M. Mazur, S. Zhang / Finance Research Letters 000 (2015) 1–6
RVA is substantially larger over the 2001–2002 period of the dotcom bubble burst and the 2007–2008 period of financial crisis. The estimated coefficient of RVA is 2.61 and 2.01 during these two periods, respectively. In comparison, the coefficient of RVA is much smaller and insignificant during the dotcom bubble from 1997 to 2000, only around 0.61. This pattern is strongly consistent with the theoretical argument in Matsusaka and Nanda (2002) that the functioning of the internal capital market is more important and value-enhancing for diversified firms when external financial market condition is worse. 4. Conclusions We conduct a trend analysis on the excess value and capital allocation efficiency of diversified firms over the long period of 1976–2013. Diversification discount exists persistently but its magnitude is smaller in recent years. It even turns into a premium during the dotcom bubble burst and becomes zero during the financial crisis, suggesting that diversified firms are less affected by shocks in external financial markets. Perhaps this can be attributed to the benefit of having an internal capital market. We find capital allocation efficiency has a much stronger effect on excess value during recent periods. Correspondingly, the trend analysis suggests that severe capital misallocation is less frequent after the early 2000s. Appendix A. Variable definitions and sources of data Variable
Definition
Data source
Excess value
Value of the diversified firm computed following the procedure described in more detail in Berger and Ofek (1995) Value added from internal capital allocation between business segments computed as in Rajan et al. (2000) Natural logarithm of the book value of total assets Expenditures on property, plant, and equipment scaled by the book value of total assets Sum of cash and marketable securities scaled by the book value of total assets Standard deviation of segment asset-weighted Tobin’s Q scaled by equally weighted average Tobin’s Q for all business segments in the firm
Compustat
Allocation efficiency Firm size Capital expenditures Cash holdings Segment diversity
Compustat Compustat Compustat Compustat Compustat
References Ahn, S., Denis, D.J., 2004. Internal capital markets and investment policy: evidence from corporate spinoffs. J. Finan. Econ. 71 (3), 489–516. Berger, P.G., Ofek, E., 1995. Diversification’s effect on firm value. J. Finan. Econ. 37 (1), 39–65. Campa, J.M., Kedia, S., 2002. Explaining the diversification discount. J. Finance 57 (4), 1731–1762. Datta, S., D’Mello, R., Iskandar-Datta, M., 2009. Executive compensation and internal capital market efficiency. J. Finan. Intermediation 18 (2), 242–258. Glaser, M., Riepe, J., 2014. Internal capital market studies in empirical banking: biases due to usage of assets instead of risk capital? Finan. Res. Lett. 11 (1), 47–53. Hoechle, D., Schmid, M., Walter, I., Yermack, D., 2012. How much of the diversification discount can be explained by poor corporate governance? J. Finan. Econ. 103 (1), 41–60. Laeven, L., Levine, R., 2007. Is there a diversification discount in financial conglomerates? J. Finan. Econ. 85 (2), 331–367. Lamont, O., 1997. Cash flow and investment: evidence from internal capital markets. J. Finance 52 (1), 83–109. Lamont, O.A., Polk, C., 2002. Does diversification destroy value? Evidence from the industry shocks. J. Finan. Econ. 63 (1), 51–77. Lang, L.H., Stulz, R.M., 1994. Tobin’s q, corporate diversification, and firm performance. J. Polit. Economy 102 (6), 1248–1280. Matsusaka, J.G., Nanda, V., 2002. Internal capital markets and corporate refocusing. J. Finan. Intermediation 11 (2), 176–211. Ozbas, O., Scharfstein, D.S., 2010. Evidence on the dark side of internal capital markets. Rev. Finan. Stud. 23 (2), 581–599. Rajan, R., Servaes, H., Zingales, L., 2000. The cost of diversity: the diversification discount and inefficient investment. J. Finance 55 (1), 35–80. Sautner, Z., & Villalonga, B. (2010). Corporate governance and internal capital markets. Harvard Business School Finance Working Paper Scharfstein, D.S., Stein, J.C., 2000. The dark side of internal capital markets: divisional rent-seeking and inefficient investment. J. Finance 55 (6), 2537–2564. Schmid, M.M., Walter, I., 2009. Do financial conglomerates create or destroy economic value? J. Finan. Intermediation 18 (2), 193– 216. Servaes, H., 1996. The value of diversification during the conglomerate merger wave. J. Finance 51 (4), 1201–1225. Shin, H.H., Stulz, R.M., 1998. Are internal capital markets efficient? Q. J. Econ. 113 (2), 531–552. Whited, T.M., 2001. Is it inefficient investment that causes the diversification discount? J. Finance 56 (5), 1667–1691.
Please cite this article as: M. Mazur, S. Zhang, Diversification discount over the long run: New perspectives, Finance Research Letters (2015), http://dx.doi.org/10.1016/j.frl.2015.08.008