This paper studies the main research question on whether private equity backed LBOs affect the innovation output of target companies. While many previous papers have analyzed private equity buyouts and their impact on operating performance, asymmetric information, or employment, only a few papers have analyzed the impact of private equity on target’s innovation. Based on a quasi-experiment consisting of closed leveraged buyouts and cancelled leveraged buyouts announced over the period from 1997 to 2019, this paper shows a negative impact of private equity backed LBOs. After the deal, targets of successfully completed LBOs exhibit a lower number of patent applications than the targets of cancelled LBOs. One possible explanation for the decrease in the number of patents can be the reduced agency problems as suggested by Jensen (1989). According to Bernstein et al. (2015), agency theory may explain the decrease of patent applications as a reduction of the overinvestment problem, where the management becomes more disciplined and focuses on more qualitative innovation. However, this argument is refuted by an additional test where it is shown that the quality of patenting activity as measured by the number of citations scaled by patent applications decreases following the LBO. These results also contradict Lerner (2011), who finds that the patent quality of patents increases after a leveraged buyout. Another test shows that there is no effect on the economic value of patents measure. Overall, the results provide evidence against the three hypotheses of the paper and in favor of the private equity short-termism theory, which suggests that this negative impact on the level and quality of innovation output may be caused due to the high pressure put on the management to increase returns.
Limitations
The main focus of this study is to measure the changes in the target’s innovation output levels arising due to the PE-backed LBO transactions, and for this purpose, data regarding patenting
38 activity measures are used. Even though information about patent applications, grants, citations and information about their corresponding assignee can be freely accessed from the USPTO (U.S Patent and Trademark Office), there is no readily available firm identifier that can be used to match the companies with other financial databases. One possible way to overcome this issue is by using fuzzy matching algorithms to match the companies with financial information based on name similarities. However, this method is not completely reliable as, in many cases, it may lead to incorrect matches which require to be manually checked. Due to time constraints, the previous method is not feasible, and consequently, this paper used an updated patent dataset from Kogan et al. (2017) to conduct the analysis. This dataset contains patent information matched with the corresponding CRSP identifiers (permno) for patents issued from 1926 to 2019. In order to match the LBO targets with their permnos, the PE-backed LBO transactions sample is restricted to include only public-to-private deals, and this, in turn, may present selection bias to the analysis as not all the types of private equity transactions are considered. However, this issue does not significantly affect the results as the public-to-private transactions make up for the largest fraction of private equity buyouts (Kaplan and Stromberg, 2009).
Following a private equity backed leveraged buyout, publicly listed targets are generally de-listed from the stock market and turned private, except for the so-called PIPEs (Private Investments in Public Equity), who still continue to remain public. When companies incorporated in U.S stock markets go private, they do not have an obligation to file their financial information with the SEC (U.S. Securities and Exchange Commission) anymore, unless these companies have public debt outstanding. Hence, this makes it difficult to collect the necessary data for the performance analysis of the private equity LBO targets following the transaction.
Consequently, due to the previously mentioned data limitation, the number of target companies that also satisfies the condition of having filed at least one patent application on the event period drops to 220 inclusive of the treatment and control group. Nevertheless, the panel data analysis augmented with fixed effects allows to reduce possible selection biases.
Recommendations for further research
While this paper only provides evidence on the target’s innovation levels and quality changes following the buyout, there is no empirical test that analyzes how this innovation is generated
39 and what its driving factors are. Hence, further research may be conducted in order to analyze these determinants.
For instance, a possible avenue for research would be to consider how innovation activity of leveraged buyout targets is related to human capital and, more specifically, the general partner’s background, skills, and knowledge. Since the private equity industry competition is increasing, the private equity firms may need to differentiate themselves by attracting highly-skilled general partners to improve their performance. Consistent with this argument, Acharya et al. (2012) provide evidence that there exists a relation between the general partner’s background and the private equity fund’s superior performance. Similarly, Zarutskie (2010) show that different human capital skill help explain the differences in the performance of venture capital funds. Following these findings on the relationship between human capital and fund performance, studying more in-depth how different skills of general partners affect target’s innovation would shed light on how private equity firms can create value in target companies. At the same time, this research would help understand the main cause behind the negative effect on innovation found in this paper and consequently suggest certain skills that would mitigate this effect.
Appendix
Table 1. This table contains the definitions of all the variables used in the empirical analysis, including the Innovations Measures and Control Variables.
Variable Name Description
Panel A: Innovation Measures
Patent Count This variable is calculated as the natural logarithm of the total number of successful patent applications filed in a given year.
Citations Count This variable is calculated as the natural logarithm of the total number of citations received in a given year.
Citations Weighted This variable is calculated as the natural logarithm of the citations scaled variable, which is in turn calculated as the number of total citations received by patents produced in a year divided by the number of total patents produced in that same year.
40 Xi Nominal This variable is pre-constructed by Kogan et al. (2017) as “the
product of the estimate of the stock return due to the value of the patent times the market capitalization M of the firm that is issued patent j on the day prior to the announcement of the patent issuance.”
Xi Real The Xi Nominal variable deflated to the 1982 inflation using the CPI index (Kogan et al., 2017).
Panel B: Control Variables
Age The firm age at the year of the transaction announcement, calculated from the founding year. The founding year data are obtained from the Field-Ritter Dataset.
Age Squared The variable Age squared.
R&D/Sales R&D Expense scaled by Total Sales.
R&D/AT R&D Expense scaled by the Book Value of Total Assets.
Size The natural logarithm of Total Book Value of Assets.
Cashflow The total cash flow of firm i scaled by the Book Value of Total Assets.
Operating Cash Total Cash and Short-Term Investments divided by the Book Value of Total Assets.
Return on Assets (ROA) EBITDA divided by the Book Value of Total Assets, where EBITDA stands for the Earnings Before Interest Taxes Depreciation and Amortization.
Leverage Ratio Total Debt divided by the Book Value of Total Assets.
Retained Earnings/AT Retained Earnings divided by the Book Value of Total Assets.
Herfindahl Index This variable is calculated as the sum of the squared market share of each firm in the sample.
Herfindahl Index Square Calculated as the Herfindahl Index variable squared.
Market-to-Book Ratio The sum of Market Value of Equity and Book Value of Total Debt divided by The Book Value of Total Assets.
Completed Completed is a dummy variable that equals one if the firm has been a target of a successful PE LBO deal and zero if the deal was withdrawn.
Post Post is a dummy variable that equals one on the deal announcement date and onwards and equals zero in the years prior to the announcement.
41 References
Acharya, V.V., Gottschalg, O.F., Hahn, M. and Kehoe, C., 2013. Corporate governance and value creation: Evidence from private equity. The Review of Financial Studies, 26(2), pp.368-402.
Aghion, P., Angeletos, G.M., Banerjee, A. and Manova, K., 2005. Volatility and growth: Credit constraints and productivity-enhancing investment (No. w11349). National Bureau of Economic Research.
Axelson, U., Strömberg, P. and Weisbach, M.S., 2009. Why are buyouts levered? The financial structure of private equity funds. The Journal of Finance, 64(4), pp.1549-1582.
Bertoni, F., 2017. Innovation in Private Equity Leveraged Buyouts. Available at SSRN 3054410.
Bertoni, F., Croce, A. and Guerini, M., 2015. Venture capital and the investment curve of young high-tech companies. Journal of Corporate Finance, 35, pp.159-176.
Bertrand, M., Duflo, E. and Mullainathan, S., 2004. How much should we trust differences-in-differences estimates?. The Quarterly journal of economics, 119(1), pp.249-275.
Boucly, Q., Sraer, D. and Thesmar, D., 2011. Growth lbos. Journal of Financial Economics, 102(2), pp.432-453.
Brown, J.R., Martinsson, G. and Petersen, B.C., 2012. Do financing constraints matter for R&D?. European Economic Review, 56(8), pp.1512-1529.
Chemmanur, T.J., Loutskina, E. and Tian, X., 2014. Corporate venture capital, value creation, and innovation. The Review of Financial Studies, 27(8), pp.2434-2473.
Davis, S.J., Haltiwanger, J., Handley, K., Jarmin, R., Lerner, J. and Miranda, J., 2014. Private equity, jobs, and productivity. American Economic Review, 104(12), pp.3956-90.
Dittmar, A., Li, D. and Nain, A., 2012. It pays to follow the leader: Acquiring targets picked by private equity. Journal of Financial and Quantitative Analysis, pp.901-931.
Engel, D. and Stiebale, J., 2014. Private equity, investment and financial constraints: firm-level evidence for France and the United Kingdom. Small Business Economics, 43(1), pp.197-212.
42 Field, L.C. and Karpoff, J.M., 2002. Takeover defenses of IPO firms. The Journal of Finance, 57(5), pp.1857-1889.
Griliches, Z., 1990. Patent Statistics as Economic Indicators: A Survey Part II. NATIONAL BUREAU OF ECONOMIC RESEARCH (NBER).
Hall, B.H., Jaffe, A. and Trajtenberg, M., 2005. Market value and patent citations. RAND Journal of economics, pp.16-38.
Harris, R., Siegel, D.S. and Wright, M., 2005. Assessing the impact of management buyouts on economic efficiency: Plant-level evidence from the United Kingdom. Review of Economics and Statistics, 87(1), pp.148-153.
Harris, R.S., Jenkinson, T. and Kaplan, S.N., 2014. Private equity performance: What do we know?. The Journal of Finance, 69(5), pp.1851-1882.
Jaffe, A.B. and Trajtenberg, M., 2002. Patents, citations, and innovations: A window on the knowledge economy. MIT press.
Jensen, M.C. and Murphy, K.J., 1990. Performance pay and top-management incentives. Journal of political economy, 98(2), pp.225-264.
Jensen, M.C., 1989. The Eclipse of the Public Corporation, Harvard Business Review September-October.
Kaplan, S., 1989. The effects of management buyouts on operating performance and value.
Journal of financial economics, 24(2), pp.217-254.
Kaplan, S.N. and Schoar, A., 2005. Private equity performance: Returns, persistence, and capital flows. The journal of finance, 60(4), pp.1791-1823.
Kogan, L., Papanikolaou, D., Seru, A. and Stoffman, N., 2017. Technological innovation, resource allocation, and growth. The Quarterly Journal of Economics, 132(2), pp.665-712.
Liu, Ping (2017), “Horse picker or right jockey? An examination of private equity value creation through the lens of withdrawn leveraged buyouts”
43 Long, W.F. and Ravenscraft, D.J., 1993. LBOs, debt and R&D intensity. Strategic management journal, 14(S1), pp.119-135.
Loughran, T. and Ritter, J., 2004. Why has IPO underpricing changed over time?. Financial management, pp.5-37.
McKinsey & Company, 2021. McKinsey’s Private Markets Annual Review. Retrieved from:
www.mckinsey.com [June 30, 2021]
Phalippou, L. and Gottschalg, O., 2009. The performance of private equity funds. The Review of Financial Studies, 22(4), pp.1747-1776.
Popov, A.A. and Roosenboom, P., 2009. Does private equity investment spur innovation?
Evidence from Europe.
Rappaport, A., 1990. The staying power of the public corporation. Harvard business review, 68(1), pp.96-104
Seru, A., 2014. Firm boundaries matter: Evidence from conglomerates and R&D activity.
Journal of Financial Economics, 111(2), pp.381-405.
Stock, J.H. and Watson, M.W., 2020. Introduction to Econometrics 4th ed.
Ughetto, E., 2010. Assessing the contribution to innovation of private equity investors: A study on European buyouts. Research Policy, 39(1), pp.126-140
Viviani, Diego, Marco Giorgino, and Roberto Steri. “Private equity-backed IPOs and long-run market performance analysis of Italian firms.” The Journal of Private Equity 11, no. 3 (2008):
50-60.
Zarutskie, R., 2010. The role of top management team human capital in venture capital markets:
Evidence from first-time funds. Journal of Business Venturing, 25(1), pp.155-172.
44
Appendix A
Table A1
LBO Announcements over the Sample Period
This table presents the full sample of LBO transactions included in the final sample and their distribution over time. In the first column are presented the announcement years covered in the analysis, and in the second column the total number of announced LBO deals for each of the announcement years. The third and fourth columns show the distribution of the Closed and Cancelled deals, respectively, for each year of the analyzed sample period.
Announcement Year
Total Announced Deals
Closed Deals
Cancelled Deals
1997 14 13 1
1998 11 8 3
1999 25 22 3
2000 26 21 5
2001 8 3 5
2002 16 9 7
2003 19 13 6
2004 9 7 2
2005 24 12 12
2006 28 21 7
2007 38 24 14
2008 15 9 6
2009 10 5 5
2010 15 12 3
2011 18 14 4
2012 16 8 8
2013 18 15 3
2014 12 11 1
2015 9 8 1
2016 14 11 3
2017 12 10 2
2018 12 10 2
2019 6 5 1
Total 375 271 104
45 Figure 1. LBO Announcements Distribution over the Sample Period.
This graph shows the distribution of the private equity backed LBO transactions included in the full sample covering the period starting in 1997 until 2019. The solid line shows the total number of deal announcements, and the other two dotted lines represent the cancelled and closed transactions.
46 Table A2. Summary Statistics
This table provides the summary for all the variables considered in the analysis. All the variables except for the innovation measures are estimated with data collected over the five years prior to the announcement of the transactions and are winsorized at a 0.5% level in order to control for possible outliers. Patent Count, Citations Count, Citations Weighted, Xi Real and Xi Nominal are estimated only for firms with at least one patent application over five years before or five years after the deal announcement year and are estimated with data covering the period before the transaction announcement. Panel A presents the summary statistics of the variables mentioned above for the completed transaction targets, and Panel B presents the summary statistics for the cancelled transaction targets.
Closed Transactions
Mean Median SD p 25 p 75 N
Paten Count 0.847 0.693 1.000 0 1.386 2498
Citations Count 1.995 1.242 2.208 0 3.807 2498
Citations Weighted 1.409 0.693 1.549 0 2.799 2498
Xi Nominal 8.292 2.420 25.509 0.763 7.565 1396
Xi Real 4.531 1.396 14.567 0.534 3.979 1396
Age 20.75 16 16.280 10 26 2319
Age Squared 695.6 256 1198.6 100 676 2319
R&D/AT 0.069 0.043 0.072 0.017 0.099 3211
Size 4.719 4.770 1.798 3.518 6.065 4707
Operating Cash 0.165 0.086 0.180 0.026 0.261 4704
Leverage Ratio 0.220 0.185 0.203 0.027 0.349 4691
Retained Earnings/At -0.043 0.175 0.754 -0.095 0.393 4673
Return on Assets (ROA) 0.109 0.126 0.126 0.062 0.189 4685
Herfindahl Index 0.016 0.012 0.015 0.005 0.019 4802
Herfindahl Index Square 0.000 0.000 0.001 0.000 0.0004 4802
Market-to-Book Ratio 1.796 1.381 1.154 1.019 2.090 4350
Cancelled Transactions
Mean Median SD p 25 p 75 N
Patent Count 0.762 0 1.045 0 1.099 1198
Citations Count 1.784 0 2.265 0 3.738 1198
Citations Weighted 1.237 0 1.569 0 2.639 1198
Xi Nominal 6.676 2.217 14.37 0.950 6.662 566
Xi Real 3.563 1.413 6.988 0.564 3.615 566
Age 21.34 15 19.98 8 24 1023
Age Squared 854.0 225 1559 64 576 1023
R&D/AT 0.075 0.037 0.085 0.010 0.116 1165
Size 4.923 4.909 1.878 3.614 6.126 1886
Operating Cash 0.178 0.087 0.199 0.027 0.266 1886
Leverage Ratio 0.228 0.192 0.208 0.028 0.353 1880
Retained Earnings/At -0.166 0.115 0.820 -0.241 0.294 1878
Return on Assets (ROA) 0.078 0.105 0.140 0.030 0.167 1862
47
Herfindahl Index 0.016 0.012 0.014 0.008 0.020 1913
Herfindahl Index Square 0.000 0.000 0.001 0.000 0.000 1913
Market-to-Book Ratio 1.682 1.275 1.091 0.966 1.940 1761
Table A3
The Effect of LBO Transactions on the Economic Value of Innovation
This table presents the resulting coefficients of the difference in difference regression analyzing the impact of the private equity backed LBO deals on the economic value of innovation measure of the target company (Xi Real) over the period 1997-2019. The sample used to run this regression covers observations over the event window of 5 years before and after the transaction's announcement. Full Sample includes the full sample of companies with at least one patent application over the event window, including deals that failed for endogenous reasons.
Sub Sample instead excludes deals that are cancelled due to endogenous reasons from the control group, and only deals that failed due to exogenous reasons are kept. Completed is a dummy variable that equals one if the LBO transaction was closed and zero if the deal was cancelled.
Post is a dummy variable that equals zero before the announcement day and equals one on the deal announcement day and onwards. All regressions are augmented by control variables, except for the regressions of Column (1) and Column (2). In Column (1)-(4), year and industry fixed effects are included, while in Column (5) and Column (6), year and firm fixed effects are used.
In Column (5) and Column (6), the variable Completed is omitted from the regressions in order to prevent multicollinearity with firm fixed effects. In parenthesis are shown the standard errors clustered at the firm level and ***, **, * denote significance level at 1%, 5% and 10%, respectively.
(1) (2) (3) (4) (5) (6)
VARIABLES Full
Sample
Sub Sample
Full Sample
Sub Sample
Full Sample
Sub Sample
Completed x Post 1.457 1.581 -0.272 -1.380 -0.079 -1.804
(0.90) (0.76) (-0.16) (-0.71) (-0.06) (-1.32)
Completed 1.404* 0.542 1.205* 1.113
(1.74) (0.47) (1.77) (1.17)
Post 0.530 0.155 0.113 0.412 1.293* 2.491**
(0.75) (0.14) (0.18) (0.40) (1.80) (2.39)
Observations 743 577 438 350 421 337
Adjusted R-squared 0.598 0.610 0.445 0.487 0.577 0.602
Year Fixed Effects YES YES YES YES YES YES
Industry Fixed Effects YES YES YES YES NO NO
Firm Fixed Effects NO NO NO NO YES YES
Standard Control Variables
NO NO YES YES YES YES
48 Table A4
The Effect of LBO Transactions on Corporate Innovation Measures
This table presents the resulting coefficients of the difference in difference regression, analyzing the impact of the private equity backed LBO deals on the three different innovation measures for transactions announced over the period 1997-2015. This robustness check will ensure that for all the transactions announced, at least four years have elapsed. The sample used to run this regression covers observations over the event window of 5 years before and after the transaction's announcement. The control group in these regressions includes only deals cancelled due to exogenous reasons. Completed is a dummy variable that equals one if the LBO transaction was closed and zero if the deal was cancelled. Post is a dummy variable that equals zero before the announcement day and equals one on the deal announcement day and onwards. All regressions are augmented by control variables. In Column (1), Column(3) and Column (5), year and industry fixed effects are included, while in Column (2), Column (4) and Column (6), year and firm fixed effects are used. In Column (2), Column (4) and Column (6), the variable Completed is omitted to prevent multicollinearity with the firm fixed effects. In parenthesis are shown the standard errors clustered at the firm level and ***, **, * denote significance level at 1%, 5% and 10%, respectively.
(1) (2) (3) (4) (5) (6)
VARIABLES Patent
Count
Patent Count
Citations Weighted
Citations Weighted
Xi Real
Xi Real Completed x Post -0.276 -0.705*** -0.543* -0.835** -1.921 -0.918
(-1.08) (-2.74) (-1.77) (-2.45) (-1.02) (-0.69)
Completed 0.094 -0.033 0.517
(0.30) (-0.09) (0.62)
Post 0.280 0.820*** 0.296 0.794*** 0.533 2.230**
(1.37) (4.38) (1.21) (3.49) (0.58) (2.19)
Observations 517 517 517 517 296 285
Adjusted R-squared 0.553 0.738 0.413 0.514 0.503 0.601
Year Fixed Effects YES YES YES YES YES YES
Industry Fixed Effects
YES NO YES NO YES NO
Firm Fixed Effects NO YES NO YES NO YES
Standard Control Variables
YES YES YES YES YES YES
49 Table A5
The Effect of LBO Transactions on Corporate Innovation Measures
This table presents the resulting coefficients of the difference in difference regression, analyzing the impact of the private equity backed LBO deals on the three different innovation measures for transactions announced over the period 1997-2019. The sample used to run this regression covers observations over the event window of 3 years before and 5 years after the transaction's announcement. The control group in these regressions includes only deals cancelled due to exogenous reasons. Completed is a dummy variable that equals one if the LBO transaction was closed and zero if the deal was cancelled. Post is a dummy variable that equals zero before the announcement day and equals one on the deal announcement day and onwards. All regressions are augmented by control variables. In Column (1), Column(3) and Column (5), year and industry fixed effects are included, while in Column (2), Column (4) and Column (6), year and firm fixed effects are used. In Column (2), Column (4) and Column (6), the variable Completed is omitted to prevent multicollinearity with the firm fixed effects. In parenthesis are shown the standard errors clustered at the firm level and ***, **, * denote significance level at 1%, 5%
and 10%, respectively.
(1) (2) (3) (4) (5) (6)
VARIABLES Patent
Count
Patent Count
Citations Weighted
Citations Weighted
Xi Real
Xi Real Completed x Post -0.157 -0.627** -0.337 -0.791** -0.732 -0.972
(-0.54) (-2.43) (-1.02) (-2.32) (-0.41) (-0.38)
Completed 0.027 -0.150 0.841
(0.08) (-0.43) (0.85)
Post 0.270 0.696*** 0.130 0.628** -0.161 1.489
(1.26) (4.02) (0.51) (2.53) (-0.14) (1.25)
Observations 423 423 423 423 209 189
Adjusted R-squared 0.548 0.753 0.431 0.581 0.425 0.503
Year Fixed Effects YES YES YES YES YES YES
Industry Fixed Effects
YES NO YES NO YES NO
Firm Fixed Effects NO YES NO YES NO YES
Standard Control Variables
YES YES YES YES YES YES
50 Table A6
The Effect of LBO Transactions on Corporate Innovation Measures
This table presents the resulting coefficients of the difference in difference regression analyzing the impact of the private equity backed LBO deals on patent applications number of the target company for transactions announced over 1997-2007. This robustness check will test whether the results still hold in the first half of the time period analyzed. The sample used to run this regression covers observations over the event window of 5 years before and after the transaction's announcement. The control group in these regressions includes only deals cancelled due to exogenous reasons. Completed is a dummy variable that equals one if the LBO transaction was closed and zero if the deal was cancelled. Post is a dummy variable that equals zero before the announcement day and equals one on the deal announcement day and onwards. All regressions are augmented by control variables. In Column (1), Column(3) and Column (5), year and industry fixed effects are included, while in Column (2), Column (4) and Column (6), year and firm fixed effects are used. In Column (2), Column (4) and Column (6), the variable Completed is omitted to prevent multicollinearity with the firm fixed effects. In parenthesis are shown the standard errors clustered at firm level and ***, **, * denote significance level at 1%, 5% and 10%, respectively.
(1) (2) (3) (4) (5) (6)
VARIABLES Patent
Count
Patent Count
Citations Weighted
Citations Weighted
Xi Real
Xi Real Completed x Post -0.538** 0.048 -1.275*** -0.543 -6.178** -2.484
(-2.05) (0.19) (-3.07) (-1.23) (-2.19) (-1.13)
Completed -0.090 0.224 0.623
(-0.30) (0.47) (0.64)
Post 0.590** 0.106 0.949** 0.508 4.391 5.148**
(2.43) (0.63) (2.31) (1.62) (1.56) (2.33)
Observations 317 317 317 317 156 151
Adjusted R-squared 0.706 0.789 0.444 0.512 0.505 0.568
Year Fixed Effects YES YES YES YES YES YES
Industry Fixed Effects
YES NO YES NO YES NO
Firm Fixed Effects NO YES NO YES NO YES
Standard Control Variables
YES YES YES YES YES YES