1
Master Thesis
MSc. Finance
The Value of Leveraged Buyouts: Evidence From the Second U.S.
Leveraged Buyout Wave
Author: Rick Meijer
s1790013
Supervisor: Prof. Dr. W. Bessler
Abstract
Date: 14/01/2016
JEL classification
G32, G34
Keywords
LEVERAGED
BUYOUTS,
VALUE
CREATION,
RESTRUCTURING,
PUBLIC-TO-PRIVATE
TRANSACTIONS, PRIVATE EQUITY.
2
Table of contents
1. Introduction ... 3
2. Literature review ... 5
2.1 LBO value creation ... 5
2.2 Sources of LBO value creation ... 6
2.2.1 Increased operational performance ... 7
2.2.2 Tax benefit of debt ... 10
3. Data ... 12
3.1 Data gathering ... 12
3.2 Descriptive statistics ... 13
4. The value creation of LBOs ... 18
5. Sources of value ... 22
5.1 Increased operational performance ... 22
5.2 Tax benefits of debt ... 26
5.3 Changes in industry valuation multiples... 28
6. Conclusion ... 31
7. Acknowledgements ... 33
8. Reference list ... 34
3
1. Introduction
During the 1980s, Leveraged Buyouts (LBOs) became increasingly popular. Kaplan & Strömberg (2008: p. 2) define a LBO transaction as: “In a leveraged buyout, a company is acquired [..] using a relatively small portion of equity and a relatively large portion of outside debt financing.”. LBOs generally come in three variants, Management Buyouts (MBOs) in which the existing management team acquires the assets and operations of the company. Management Buy-ins (MBIs) are similar to MBOs only in MBIs management teams from outside the company buy the assets and operations and become the new management team of the company. Lastly, Institutional Buyouts (IBOs) are buyouts in which the company is acquired by an institutional investor, mostly Private Equity (PE) companies and venture capitalists1.
The wave of LBOs in the 1980s is widely researched by academics and practitioners and many studies find evidence that these transactions add value (e.g. Andrade & Kaplan, 1998; Kaplan, 1994) even though e.g. more than 30% of companies involved in MBOs went bankrupt in the economic downturn in the beginning of the 1990s (Andrade & Kaplan, 1998). A second LBO wave started after the dotcom bubble in the early 2000s. Guo, Hotchkiss & Song (2011) acknowledge, however, that there is little or no research conducted on the value creation and source of this value for the second LBO wave.
There are several theories that try to explain from where the value creation of LBO transactions originates. These can broadly be categorized into two groups, value creation due to increased operational performance, and value creation due to the tax benefits of debt. For LBOs in the 1980s, both groups of theories are found to be value adding [see e.g. Kaplan (1989a); Cotter & Peck (2001)]. However, recent studies suggest the factors that drive firms to go private have changed since the 1980s. During the 1980s, the LBO was an important corporate restructuring tool to transform low-growth sluggish public companies into efficient private companies (Mehran & Peristiani, 2010). In the more recent LBO wave, however, financial visibility seems to be a much larger factor driving companies to go private (Bharath & Ditmar, 2010; Mehran & Peristiani, 2010). Financial visibility is the capacity of the company to generate sufficient investor recognition, such as analyst coverage, institutional ownership and trading volume. During the 1990 to 2007 period, companies were more likely to exit the public markets when analyst coverage is low, institutional ownership is low and trading volume is low (Mehran & Peristiani, 2010; Bharath & Dittmar, 2010).
1
4 The recently increased LBO activity and changing forces that drive companies to exit public markets raise new questions regarding whether LBO transactions still create value and if so, what drives this value creation.
There are two main contributions of this study. First, few studies examine the most recent LBO wave. Guo, Hotchkiss & Song (2011) try to fill this gap by studying all public-to-private LBOs in the U.S. from 1990 to 2005. However, the second LBO wave hit its peak in 2006 and 2007 with U.S. deal numbers and values doubled in 2006 and fivefold in 2007 compared to 2005 (see chapter 3). The Guo, Hotchkiss & Song (2011) study therefor is incomplete. Next to this, the other paper that covers part of the second LBO wave by Acharya, Gottschalg, Hahn & Kehoe (2013) studies LBO transactions in Western European countries. LBOs in these countries may be structurally different from U.S. LBOs as ownership structures and the market for corporate control of these countries are often very different from the U.S. e.g. more closely-held companies with larger shareholders (Faccio & Lang, 2002). For these reasons more research is needed to adequately describe the most recent LBO wave in the U.S. I intend to fill this gap by studying all U.S. public-to-private LBOs during the second LBO wave from 2003 up to and including 2013. Next to studying the value creation of LBO transactions during this period, this thesis will also study the underlying sources of value creation, providing a useful comparison to the studies of 1980s LBOs.
The main research question of this thesis is:
“Do leveraged buyout transactions create value?”
I find that during the second LBO wave, U.S. LBO transactions still created significant value for all outcome groups of LBOs, apart from firms that end in bankruptcy. I find an average market- and risk-adjusted return on pre- (post-)buyout capital of 68.2% (30.2%). Looking at the source of this value, I find that both increased operational performance and the tax benefits of debt contribute an equal amount of value. Comparing these results with 1980s LBOs, I show that returns on capital are significantly smaller during the second wave compared to the first wave of LBOs. This difference is mainly attributable to declined improvements of operating performance. Although operating improvements are still responsible for 17.3% return on pre-buyout capital, this is as little as one fifth of what previous studies find during 1980s LBOs. The returns on capital due to the tax benefits of debt remain with 18.6% at similar levels compared to 1980s LBOs. Furthermore I show that increased industry multiples do not form a source of value during the most recent wave of LBOs.
5 empirical evidence of value creation. Chapter 3 discusses the dataset of this thesis and shows the summary statistics. Chapter 4 reports on LBO value creation, chapter 5 describes the main sources of value and chapter 6 concludes.
2. Literature review
This chapter provides the theoretic background of this thesis. The first part discusses the empirical evidence on value creation during LBO transactions, the second part explains the most important theories on the source of the created value.
2.1 LBO value creation
Kaplan (1989a) studies the value creation of 76 large U.S. MBOs during the 1980 to 1986 period. He finds a mean market-adjusted return of 96% from 2 months prior to the buyout date to the post-buyout exit. Andrade & Kaplan (1998) specifically look at the value creation of 31 U.S. LBO transactions completed between 1980 and 1989 in which the focal company ended in financial (not economical) distress. The authors find that the costs of financial distress after a LBO transaction are between 10% and 20% of firm value. More importantly, they find that the combined value creation from the LBO and financial distress is slightly positive. This leads to their conclusion that LBO transactions that do not become distressed create significant value.
In a special case study of Federated Department Stores, a LBO that went into Chapter 11 (reorganization) bankruptcy post-buyout, Kaplan (1989b) reports that the value of the assets of Federated increased with 3.4 billion dollars, and thus that the LBO generated value. Including the costs of financial distress and bankruptcy, Kaplan (1994) estimates that even though Federated entered Chapter 11 bankruptcy, the LBO still created 3.1 billion dollars in value. This special case offers the insight that LBO transactions can be value increasing but still be unable to meet the debt obligations. Furthermore it shows that costs of financial distress and bankruptcy are modest.
Trying to compare more recent LBOs with the LBO wave of the 1980s, Guo, Hotchkiss & Song (2011) examine U.S. LBOs from 1990 through 2005. They find that in the 1990-2005 period LBOs are more conservatively priced and are less leveraged compared to the LBOs in the 1980s. The LBOs generate a median market- and risk-adjusted return to pre- (post-) buyout capital invested of 72.5% (40.9%). Similarly, Acharya Gottschalg, Hahn & Kehoe (2013) study the value creation of 395 LBO transactions in Western Europe performed by 37 large, mature PE houses. Over this sample, they find a gross, mean IRR of 56.1% per transaction.
6 return on the S&P500 during the same period. This underperformance was mainly fueled by small and relatively young PE-funds, a subset of the largest and most mature PE-funds recorded a positive net IRR of 150% of the S&P500 return over the same period. The results of Ljungqvist, Richardson & Wolfenzon (2007) help explain these findings, they find that young PE-funds usually take higher risks than mature PE-funds, hoping that successful, risky LBOs positively change the market’s perception of their talent. Higher perceived talent then enables managers to raise more capital in subsequent investment funds. Acharya et al. (2013) further attenuate the results of Kaplan & Schoar (2005) by stating that the underperformance of the PE-funds does not necessarily has to mean that the underlying LBO transactions do not generate value, PE-funds might capture that value themselves via their fee-structure.
As most empirical studies provide existence of positive value creation in LBO transactions, their results lead to the following hypotheses:
HYPOTHESIS 1: LBO transactions create value.
2.2 Sources of LBO value creation
Existing literature provides multiple theories on the source of LBO value creation. These theories can be categorized in two categories: value from increased operational performance and value from the tax benefits of debt (Guo, Hotchkiss & Song, 2011). The increased operational performance originates from lowered agency costs and efficiency improvements due to industry knowledge. This is graphically depicted in figure 1.
Figure 1: Sources of LBO value
LBO value
Value from increased
operational
performance
Value from lowered
agency costs
Value from
operational
efficiencies
7 2.2.1 Increased operational performance
Several theories imply that the value created in LBO transactions comes from increased operational performance. According to Renneboog & Simons (2005) the most important ones are (i) the incentive realignment hypothesis, (ii) the free cash flow hypothesis, and (iii) the control hypothesis which all add value by lowering agency costs2. Apart from lowering agency costs, PE-companies have a strong industry focus, and use their industry knowledge to make operating improvements (Acharya & Kehoe, 2008) and use internal- or external consulting groups to increase their efficiencies (Kaplan & Stromberg, 2008).
2.2.1.1 Value from lowered agency costs
The problems that arise with the separation of ownership and control of public companies are already described in 1776. “The directors of such [public] companies, however, being the managers rather of other people’s money than of their own, it cannot well be expected, that they should watch over it with the same anxious vigilance with which the partners in a private copartnery frequently watch over their own” (Smith, 1776: p. 311). By going private, companies can theoretically lower the agency costs that originate from the separation of ownership and control. Jensen (2007) phrases it as “PE funds enable the capture of value destroyed by agency problems in public firms – especially failures in governance”.
The incentive realignment theory: Managers of public companies have incentives to grow the company beyond its optimal size and retain resources at the expense of the shareholders (Jensen, 1986; Murphy, 1985). This so called “empire building” behavior lowers firm value while it increases the managers’ private benefits (Renneboog & Simons, 2005). By aligning the interests of the managers to those of the owners of the company, maximizing company value, managers will, in theory, not partake in empire building but will pursue value maximization (Jensen & Meckling, 1976). In LBO transactions, the incentives of management and the owners are realigned by increasing equity ownership of management. Jensen & Murphy (1990) find that the median equity stake of a CEO in a LBO organization is 6.4% whereas the median equity stake of CEOs of public companies is 0.2%. Kaplan (1989a) reports a median 4.41% increase in equity ownership for the top two officers in LBO organizations and a 9.96% equity ownership increase for all of management.
Free cash flow theory: The control function of debt is another way of inhibiting managers from empire building, especially in organizations with large free cash flows (Jensen, 1986; Jensen
2
8 1989)3. Limiting the cash flows available to managers reduces the power of management and increases the monitoring of the capital markets when the company has to obtain new capital for investments (Easterbrook, 1984; Rozeff, 1982). Dividends and share repurchases are ways via which managers can transfer free cash flow to the shareholders, however, these measures are not permanent, as managers can announce divided cuts and stop repurchase programs. A more permanent way of binding managers to pay out their free cash flow in principal and interest payments is to substitute debt for equity and paying out the proceeds to the shareholders (Jensen, 1986). The threat of failing to make principal and interest payments helps motivating management to make organizations more efficient (Jensen, 1986). Guo, Hotchkiss & Song (2011) find that lowered agency costs due to the higher debt is important in generating value, 25% of the value created in LBO transactions is derived from lowered agency costs. Cotter & Peck (2001) find that tighter debt terms does lead to better LBO performance at LBO transactions in which no buyout specialists (mainly specialized PE-companies) are present.
The control theory: When ownership of a firm is dispersed, a free-rider problem arises. All shareholders benefit from an investment in monitoring by one shareholder and therefore (small) individual equity owners will underinvest in monitoring (Grossman & Hart, 1980). As ownership is more concentrated in a LBO, investors have stronger incentives to invest in monitoring (Admati, Pleiderer & Zechner, 1994). When buyout specialists control the majority of post-buyout equity, these specialists actively monitor management by having greater board representation on smaller boards (Cotter & Peck, 2001) and more frequent board meetings (Gertner and Kaplan, 1996; Acharya and Kehoe, 2008; Cornelli, 2008). Subsequently, when these buyout specialists are present, the focal companies are less likely to default post-buyout (Cotter & Peck, 2001).
Unfortunately, the dataset in this thesis does not allow me to distinguish between increased operational performance due to lowered agency costs and increased operational performance due to operational efficiencies. For this reason, I will test these theories at the operational performance level.
2.2.1.2 Empirical evidence of increased operational performance
In his study, Kaplan (1989a) provides evidence that in the three years after a LBO buyout, companies experience increases in operating income, decreases in capital expenditures and increases in net cash flow to investors. Furthermore the paper suggests that the increases in operating performance are mainly due to improved management incentives rather than layoffs. These results are confirmed by
3
9 Smith (1990), who finds that EBIT per employee and per dollar of assets increases significantly post-buyout in a sample of 58 public-to-private MBOs between 1977 and 1986. This increase is sustained in subsequent years. Adjustments in the management of working capital contribute to the increase in operational returns, while the increase is not the result of layoffs or smaller capital expenditures. Additional studies by Opler (1992), Smart & Waldfogel (1994) and Zahra (1995) all find increases in operational performance post-buyout in U.S. LBOs in the 1980s. Looking at reverse LBOs, companies that exit a LBO via an IPO, Muscarella & Vetsuypens (1990) and Singh (1990) find that companies sustain the improved operational performance after going public. Holthauser & Larcker (1996) add to these results that companies that went public after a LBO outperform their industry peers that remained public. Bruton, Keels & Scifres (2002) also find industry outperformance after reverse LBOs, but show that this outperformance starts to gradually erode after 3 years of going public.
Lichtenberg & Siegel (1990) analyze plant-level Total Factor Productivity increase after LBO transactions on a sample of over 12,000 manufacturing plants over the 1981 to 1986 period. They find a strong positive effect on productivity in the first three years after the LBO transaction. Total Factor Productivity increased from 2% above the industry mean pre-buyout to 8.3% above the industry mean post-buyout for the deals done in the 1983 to 1986 period. For the 1981 to 1982 period there are no significant changes. This study is extended by Harris, Siegel & Wright (2005), using a data sample of 36,000 U.K. manufacturing MBOs. Harris et al. find that the manufacturing MBOs originally were less productive than their peers pre-buyout, but showed significant improvement in productivity post-buyout both in the short-term (70.5%) as in the long run (90.3%). In other European LBOs similar results are found, Bergström, Grubb & Johnsson (2007) report significant improvements in EBITDA-margins and ROIC post-buyout in 69 Swedish LBOs, while Boucly, Sraer & Thesmar (2008) provide evidence of operating improvements in French LBOs.
Looking at more recent LBOs, evidence suggests that increased operational efficiency might not add as much value in recent LBOs as it did during the LBOs in the late 1980s. Desbrières & Schatt (2002), Vinten (2007) and Weir, Jones & Wright (2008) even show declines in profitability post-buyout in French, Danish and British LBOs. These results are partly in line with the results of Guo, Hotchkiss & Song (2011) who find a positive effect of increased operating performance which is “substantially smaller than those documented for deals in the 1980s” (Guo, Hotchkiss & Song, 2011: 480). Using a different methodology, Acharya, Gottschalg, Hahn & Kehoe (2013) find that 34% of the value increase comes from abnormal performance post-buyout.
This leads to the following hypothesis:
10 2.2.2 Tax benefit of debt
Next to increased operational performance, other theories state that the created value in LBO transactions is derived from the tax benefit of debt.
Tax benefit theory: LBOs are financed by a large portion of outside debt. For most companies involved in LBOs this means a substantial increase in their debt-to-equity ratio. The tax deductibility of interest payments presents a tax shield, which increases company value (Jensen et al., 2006). Kaplan (1989c) estimates that the tax advantages of this higher leverage is between 21% and 143% of the premium paid to pre-buyout investors depending on assumptions on the tax rate and the rate of debt repayment. Guo, Hotchkiss & Song (2011) find evidence that realized tax benefits from increasing leverage is important in generating value, 25% of the value created in the LBO transactions originated from tax benefits from increased debt. Acharya, Gottschalg, Hahn & Kehoe (2013) show that 50% of the created value in LBO transactions is derived from the tax benefits of debt.
This leads to the following hypothesis:
HYPOTHESIS 3: The tax benefit of increased debt creates value in LBO transactions.
11
Table 1
Relevant Empirical Literature
This table provides an overview of the most relevant empirical literature on the value creation of LBOs, and the sources of value.
Author Sample Sample Period N Main Conclusion
Kaplan (1989a) U.S.
Pubic-to-private MBOs 1980-1986 76 The author finds a mean market-adjusted return of 96% from 2 months prior to the buyout date to the post-buyout sale.
Kaplan (1989b) U.S. LBO 1988 1 The LBO of Federated generated a total value of 3.4 Billion USD. Costs of financial distress and bankruptcy are not included.
Smith (1990) U.S. MBOs 1977-1986 58 EBIT per employee and per dollar of assets increases significantly post-buyout, mainly due to adjustments in working capital.
Lichtenberg &
Siegel (1990) U.S. LBOs 1981-1986 20,493
Productivity increased from 2% above the industry mean pre-buyout to 8.3% above the industry mean post-buyout for the deals done in the 1983 to 1986 period. For the 1981 to 1982 period there are no significant changes.
Kaplan (1994) U.S. LBO 1989-1992 1 Even though the company entered Chapter 11, the LBO still created 3.1 Billion USD in value. Costs of financial distress are negligible.
Andrade & Kaplan (1998)
U.S. LBOs which ended in financial distress
1980-1989 31 Costs of financial distress are 10-20% of firm value. Combined value from LBO and financial distress is slightly positive, so LBOs which do not encounter financial distress are significantly value adding.
Cotter & Peck
(2001) U.S. LBOs 1984-1989 64
Active monitoring by PE-firm leads to better LBO performance and lower chances of bankruptcy. For LBOs without a PE-firm, strict debt-terms also increases LBO performance.
Desbrières &
Schatt (2002) French MBOs 1988-1994 161
French companies involved with MBOs initially performed better than their peers pre-buyout, but performance deteriorated post-buyout.
Kaplan & Schoar
(2005) U.S. PE-funds 1985-2005 764
PE-funds only have a IRR that is 80% of the S&P500 return over the same period. Mainly due to relatively new PE-funds. Large, mature PE-funds generated an IRR of 150% of the S&P return over the same period.
Harris et al. (2005)
U.K. manufacturing
MBOs 1994-1998 35,752
Production plants involved in MBOs were less productive pre-buyout and showed a significant increase in productivity post-buyout in the short- and long-term of 70.5% and 90.3% respectively.
Vinten (2007) Danish LBOs 1991-2004 73 The main finding is that PE buyout fund ownership has a significant negative effect on firm performance relatively to similar firms.
Weir et al. (2008) U.K.
public-to-private LBOs 1998-2004 122
Performance deteriorates post-buyout compared to the pre-buyout situation. However, no clear evidence that LBOs perform worse than firms that remain public.
Bergström et al.
(2008) Swedish LBOs 1993-2006 69 Significant ROIC and EBITDA-margin improvements post-buyout, compared to industry peers. Guo et al. (2011) U.S.
public-to-private LBOs 1990-2005 192
Median market- and risk-adjusted returns to pre- (post-) buyout capital invested are 72.5% (40.9%). Operating efficiencies, increases in industry valuation multiples and realized tax benefits from increasing leverage are all equally important in generating value.
Acharya et al. (2013)
Western Europe
LBOs by PE-funds 1991-2007 364
12
3. Data
This chapter discusses the data that is used in this thesis. The first part describes the data gathering process. In the second part I present the descriptive statistics for the most important deal characteristics.
3.1 Data gathering
I use the MergerMarket database to identify all U.S. public-to-private leveraged buyouts with deal values over $100 million that are completed between January 1st 2003 and December 31st 2013. This results in a total of 322 buyouts. I obtain detailed deal structure information on these buyouts from the SEC’s EDGAR database. I eliminate the cases in which the target firm is purchased in a Chapter 11 restructuring, together with all cases for which the deal structure is not available in the SEC filings and deals with atypical deal structures4 which leaves a sample of 140 buyouts. Since this thesis studies the post-buyout performance of LBOs, I further focus on companies that publish their annual reports while being private. A final subsample of 59 companies provides their financial statements either because the company exits the LBO via an IPO or because the company has public debt. Table 2 presents a breakdown from the total sample of 322 buyouts to the final subset of 59 LBOs. I collect data for these companies on revenues, earnings before taxes, depreciation and amortization, impairments, interest expense, capital expenditures, long-term debt, total assets, taxes and tax loss carryforwards from the company’s balance sheet and profit and loss accounts on a yearly basis, while being private. I calculate EBIT adding back the interest payments to the earnings before income tax and subsequently, I calculate EBITDA by adding back depreciation and amortization to EBIT. Furthermore, to normalize EBITDA, I add back all incidental impairments. Lastly, I calculate net cash flow by subtracting capital expenditures from EBITDA. All post-buyout financial statements are gathered from annual reports (10-Ks), IPO prospectuses (424Bs) and other filings in the SEC’s EDGAR database.
Table 2 Dataset breakdown
This table presents a breakdown from the original 322 buyouts to the final subset of 59 buyouts that provide financial statements post-buyout.
U.S. LBOs between 2003 and 2013 with deal values over $100mln. 322 - LBOs bought in a Chapter 11 bankruptcy procedure. (5) - LBOs without detailed deal structure in SEC database. (149) - LBOs with a-typical deal structure. (23) +
Dataset of LBOs. 140
- LBOs that do not provide financial statements post-buyout. (81) + Subset of LBOs that provide financial statements post-buyout. 59
4
13 I download pre-buyout data from the COMPUSTAT database and annual reports (10-Ks) in the SEC’s EDGAR Database for each of the 140 companies in the total sample. This comprises the company’s 60 month stock return prior to the buyout announcement, total assets, depreciation and amortization, impairments, interest expense and capital expenditures, EBITDA of the 2 years prior to the buyout, the company’s market capitalization and its interest payments. The pre-buyout and post-buyout levels of debt are collected from the post-buyout’s SEC filing. I use the book value of debt for the post-buyout levels of debt, since the new debt is normally issued at a price closely to face value (Guo, Hotchkiss & Song, 2011). For LBOs that are still private but have post-buyout information available, I collect industry median EV/EBITDA multiples from the Infinancials and COMPUSTAT database.
To calculate the market- and risk-adjusted discount rates the risk free rate is proxied by using the 1-month T-bill rate, downloaded from Ibbotson, and the market return is proxied using the total return index of the S&P500. This is downloaded from the Thompson-Reuters DATASTREAM database.
3.2 Descriptive statistics
Comparing the types of companies in the full sample of this study with the buyouts of the 1980s, this sample relatively holds more services companies (50.7%) and fewer manufacturing companies (21.4%). In total, the 140 transactions are backed by 131 unique PE-companies as equity providers, with half of the deals being backed by multiple companies (known as club deals). Several PE-companies are involved in multiple LBOs, with the Texas Pacific Group (TPG) leading with providing equity in 16 LBOs5.
Table 2 presents the median deal pricing statistics for both the total sample, the subsample that provides post-buyout data and the subsample that does not provide post-buyout data. The second LBO wave is divided in two smaller waves, with deal volumes peaking in 2007 with 46 deals. During the global financial crisis deal volumes went down to only 1 deal in 2009. In 2010 the second wave started with its peak in 2011 with 20 LBOs. For the total sample of buyouts, the median deal price, in this thesis further referred to as post-buyout capital, is $1686.8 million. This represents a median 24.7% premium over the market value of the company two months before the LBO announcement, which in this thesis is further referred to as the pre-buyout capital). Comparing this premium with the median deal premium during the 1980s of 43% (Kaplan & Stein, 1993), it shows that more recent LBOs are more conservatively priced than LBOs during the 1990s. The fact that smaller premiums are paid on average suggests that buyout companies expect smaller returns on capital on LBO transactions and are thus less willing to pay large premiums. Comparing the
5
14 subsample that provides post-buyout data with the subsample that does not provide post-buyout data yields no statistical significance differences, apart from median deal price. The median deal price for companies that do not provide post-buyout data is $1,226.8 million while the median deal price for companies that do provide post-buyout data is $2,826.5 million. This in itself is no surprise, as larger companies tend to have public debt more often than smaller companies and thus have to provide post-buyout data6. Not only are the deals in my sample more conservatively priced, they are also funded with significantly less debt comparing to LBOs in the 1980s. Table 4 presents the summary statistics for the debt usage and financial structure.
Table 3
Descriptive Statistics for Deal Pricing
This table presents the descriptive statistics on pricing for all LBOs in my sample. It shows median post-buyout capital, EBITDA to post-buyout capital % and premium paid for each share compared to the pre-buyout stock price. These statistics are presented for both the entire sample as well as the subsample of LBOs which provide financial information while being private. The Mann-Whitney U (Wilcoxon rank-sum) test is used to compare the median of both the sample with post-buyout data and without post-buyout data. ***, ** and * indicate statistical significance differences in medians at respectively the 1%, 5% and 10% significance levels.
Year N Post-buyout Capital ($mil) EBITDA to Capital % Premium % Full sample 2003 2 578.8 17.68% 36.5% 2004 5 1,250.0 9.52% 15.3% 2005 10 3,083.5 10.81% 29.0% 2006 18 1,749.5 9.50% 20.5% 2007 46 2,803.9 7.44% 27.7% 2008 9 2,008.0 9.49% 21.1% 2009 1 539.0 0.00% 60.1% 2010 13 878.6 10.67% 17.8% 2011 20 1,634.0 8.74% 24.0% 2012 12 959.9 9.09% 23.3% 2013 4 10,499.3 13.90% 33.6% Total 2003-2013 140 1,686.8 8.95% 24.7%
Subsample with post-buyout data available
2003 1 1,048.0 20.61% 14.1% 2004 2 1,825.5 11.02% 21.8% 2005 6 3,637.5 9.78% 29.3% 2006 7 3,400.0 9.26% 21.0% 2007 22 4,985.2 7.98% 29.5% 2008 5 1,571.2 8.39% 24.4% 2009 0 2010 6 1,269.9 12.08% 39.1% 2011 7 2,913.9 9.30% 23.7% 2012 2 1,573.1 4.84% 28.8% 2013 1 27,130.7 6.71% 25.7% Total (2003-2013) 59 2,826.5 9.30% 24.8%
(1) Total (2003-2013) with post-buyout data
59 2,826.5 9.30% 24.8%
(2) Total (2003-2013) without post-buyout data 81 1,226.8 8.33% 24.2% Difference in medians (1) - (2) (+)*** (+) (+) 6
16
Table 4
Debt Levels and Debt Coverage Ratios
This table presents median statistics on debt levels and debt coverage ratios. All EBITDA ratios are based on pre-buyout EBITDA (LBO year -1). Descriptive statistics are presented for both the full sample of LBOs as well as the subsample that provides post-buyout financial information. No significant differences between the sample with and without post-buyout data are found. The number of observations for each statistic is depicted between parenthesis.
Year Pre-Buyout Debt
to Capital Post-Buyout Debt to Capital Change in Debt to Capital Post-Buyout Equity to Capital Pre-Buyout Debt to EBITDA Post-Buyout debt to EBITDA EBITDA to Pre-Buyout Interest EBITDA to Post-Buyout Interest Full sample 2003 0.00% (2) 62.20% (2) 62.20% (2) 37.80% (2) 0.00 (2) 4.08 (2) 15.23 (1) 3.69 (1) 2004 15.63% (5) 67.26% (5) 54.35% (5) 32.74% (5) 1.76 (5) 6.22 (5) 3.31 (5) 2.04 (2) 2005 15.37% (10) 71.34% (10) 52.84% (10) 28.66% (10) 1.16 (10) 6.28 (10) 9.41 (9) 1.73 (6) 2006 17.62% (18) 69.40% (18) 50.00% (18) 30.60% (18) 1.48 (18) 6.95 (18) 10.68 (14) 1.65 (6) 2007 15.81% (46) 69.35% (46) 47.05% (46) 30.65% (46) 1.62 (45) 9.25 (46) 6.65 (41) 1.33 (19) 2008 33.07% (9) 56.52% (9) 41.63% (9) 43.48% (9) 2.20 (9) 7.86 (9) 13.15 (7) 1.36 (5) 2009 38.74% (1) 50.03% (1) 11.29% (1) 49.97% (1) (0) (0) (0) (0) 2010 27.60% (13) 62.25% (13) 36.79% (13) 37.75% (13) 1.95 (13) 6.13 (13) 5.84 (9) 1.72 (5) 2011 0.32% (20) 63.21% (20) 56.67% (20) 36.79% (20) 0.14 (20) 7.76 (20) 8.61 (18) 1.56 (7) 2012 11.90% (12) 60.25% (12) 46.55% (12) 39.75% (12) 1.64 (11) 6.67 (12) 2.62 (11) 0.86 (2) 2013 8.66% (4) 57.05% (4) 40.12% (4) 42.95% (4) 0.24 (4) 4.96 (4) 15.76 (3) 2.79 (1) Total (2003-2013) 13.64% (140) 67.05% (140) 48.32% (140) 32.95% (140) 1.27 (136) 7.73 (139) 7.18 (119) 1.56 (54)
Subsample with post-buyout data available
17 Table 5
Post-buyout deal outcomes
This table reports LBO exits as of October 2015 for the full sample of 140 buyouts, as well as for the subsample that provides financial information. The first number is the number of observations in the full sample, followed by the number of observations having post-buyout information in parenthesis.
Outcome: IPO Trade Sale 2nd LBO Chapter 11 Still Private Total
LBO year 2003 0 (0) 1 (0) 1 (1) 0 (0) 0 (0) 2 (1) 2004 0 (0) 2 (2) 0 (0) 0 (0) 3 (1) 5 (2) 2005 3 (3) 1 (1) 1 (1) 3 (0) 2 (1) 10 (6) 2006 6 (6) 3 (1) 0 (0) 0 (0) 9 (1) 18 (7) 2007 13 (13) 11 (1) 0 (0) 3 (2) 19 (6) 46 (22) 2008 2 (2) 2 (0) 1 (0) 0 (0) 4 (3) 9 (5) 2009 0 (0) 0 (0) 0 (0) 0 (0) 1 (0) 1 (0) 2010 0 (0) 4 (2) 0 (0) 0 (0) 9 (4) 13 (6) 2011 1 (1) 2 (1) 0 (0) 0 (0) 17 (5) 20 (7) 2012 0 (0) 1 (0) 0 (0) 0 (0) 11 (2) 12 (2) 2013 0 (0) 0 (0) 0 (0) 0 (0) 4 (1) 4 (1) Total (2003-2013) 24 (24) 27 (8) 3 (2) 6 (2) 80 (23) 140 (59) Percent of deals, N/140 (N/59) 17.1% (40.7%) 19.3% (13.6%) 2.1% (3.4%) 4.3% (3.4%) 57.1% (39.0%) Median months to outcome 55.5 56.7 52 58.8 56
18
4. The value creation of LBOs
In order to calculate the value creation of LBOs, I follow the methodology of Andrade & Kaplan (1998) used to calculate the excess returns to LBO investors of the LBOs during the first LBO wave in the 1980s. In this analysis time is measured as depicted in figure 2.
Figure 2:
This figure depicts the timeline of a LBO transaction.
T=1
T=2
T=3
T=4
Two months before LBO announcement LBO Completed Exit LBO announcement
The total capital value of the company at time T is calculated by taking the sum of the values of the equity, long-term debt, short-term debt and capitalized leases. As can be found in formula 1.
𝑇𝐶𝐴𝑃𝑇 = 𝑀𝑎𝑟𝑘𝑒𝑡 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑇
+ 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐿𝑜𝑛𝑔 − 𝑇𝑒𝑟𝑚 𝑎𝑛𝑑 𝑆ℎ𝑜𝑟𝑡 − 𝑇𝑒𝑟𝑚 𝐷𝑒𝑏𝑡𝑇 + 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑒𝑑 𝐿𝑒𝑎𝑠𝑒𝑠𝑇
+ 𝐿𝑖𝑞𝑢𝑖𝑑𝑎𝑡𝑖𝑜𝑛 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑃𝑟𝑒𝑓𝑒𝑟𝑟𝑒𝑑 𝑆𝑡𝑜𝑐𝑘𝑇
(1)
The nominal total return to pre- and post-buyout capital (T=1 & T=3) is calculated as the total capital value at T=4 plus interim payments to capital . This figure is divided by the total capital value at T=1 to find the return percentage. This formula can be found in formula 2.
𝑁𝑅𝐸𝑇𝑇 = ∑ 𝐼𝑛𝑡𝑒𝑟𝑖𝑚 𝑃𝑎𝑦𝑚𝑒𝑛𝑡𝑠 𝑡𝑜 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 + 𝑇𝐶𝐴𝑃4
𝑇𝐶𝐴𝑃𝑇 − 1 (2)
In order to calculate 𝑇𝐶𝐴𝑃𝑇=4 the Terminal Value (TV) at the exit of the LBO. For firms in which the TV of the firm is not directly observable7 I follow the procedure of Guo, Hotchkiss & Song (2011) using the average of the EV/EBITDA multiple and the EV/Sales multiple to estimate TV. The EBITDA multiple is calculated as the median ratio of market value of debt plus equity to EBITDA for a sample of comparable companies. For the Sales multiple the same procedure is performed but then with net sales instead of EBITDA. As these returns are hypothetical returns, all return analyses are performed with and without this sample of estimated returns.
Apart from the nominal returns on the LBOs also the market- and risk-adjusted returns on capital are calculated in line with Gilson, Hotchkiss & Ruback (2000). The interim payments and TV
7
19 are discounted to the pre- or post-buyout date (respectively T=1 or T=3) using formula 3. In the formula 3, 𝑟𝑓 is the risk-free rate, proxied by the 1-month T-bill return. 𝛽𝑢 is the asset beta. I unlever the company’s equity beta into an asset beta which I use to discount cash flows, as capital exists of both equity and debt financing. Finally, 𝑟𝑚 is the return on the S&P500 from the time of the cash flow to the pre- or post-buyout date. As exact cash flow dates are unknown, all cash flows to capital are assumed to be made mid-year. An exception to this is for cash flows made in the same year of the LBO, in LBOs that happen in the second half of the year. These cash flows are assumed to be made at the end of the year. In formula 4, 𝐷𝐸 is the pre-buyout debt to equity ratio, T is the company’s marginal tax rate and 𝛽𝑒 is the equity beta based on the 60 months of stock returns prior to the buyout announcement.
𝑟 = 𝑟
𝑓∗ (1 − 𝛽
𝑢) + 𝛽
𝑢∗ 𝑟
𝑚(3)
𝛽
𝑢=
1−(1−𝑇)∗(𝛽𝑒 𝐷 𝐸)(4)
Table 6 reports the median and average nominal return on pre- and post-buyout capital, as well as the median and average market- and risk-adjusted return on pre- and post-buyout capital. It shows that the average and median nominal returns are positive and statistically significant for all outcome groups apart from the company that exited via a Chapter 11 bankruptcy procedure. The average market- and risk-adjusted returns are also positive for all outcome groups apart from the Chapter 11 outcome group, while the median returns to post-buyout capital (the LBO price) is slightly negative for the still private outcome group. Furthermore, it is notable that LBOs that exit via a secondary LBO show considerably higher returns than the other outcome groups, however as this is such a small subsample of LBOs (N=2), there is a possibility that this difference erodes over a larger sample.
20
Table 6
Nominal Return and Market- and Risk-Adjusted Return to Pre- and Post-Buyout Capital
This table presents the average and median nominal and market- and risk-adjusted returns to both the pre-buyout capital (enterprise value at T=1) as well as the post-buyout capital (enterprise value at T=3). The returns are grouped per exit category, numbered 1 to 5. Also the average and mean returns for the total sample (1-5) and the subsample with an exit (1-4) are presented. Next to this the table also presents the Internal Rate of Return (IRR) for both the total sample (1-5) and the subsample with an exit (1-4). The statistical significance levels of the returns are calculated via a two-tailed t-test for the mean returns and a Wilcoxon Signed Rank test for the median returns. ***, ** and * indicate significant results at the 0.01, 0.05 and 0.10 level, respectively.
Nominal Return Market- and Risk-Adjusted Return
Outcome Capital N Mean Median
# of Positive Returns Mean Median # of Positive Returns 1. IPO Pre 25 90.4% *** 75.8% *** 24 68.0% ** 35.9% *** 21 Post 25 42.1% *** 38.7% *** 21 28.6% ** 9.5% * 16 2. Trade sale Pre 8 99.3% ** 95.0% ** 7 48.5% * 32.9% ** 6
Post 8 45.1% * 46.7% * 7 15.3% 19.2% 5
3. 2nd LBO Pre 2 286.6% ** 286.6% 2 146.4% ** 146.4% 2
Post 2 128.2% 128.2% 2 107.6% 107.6% 2
4. Chapter 11 Pre 1 -62.8% -62.8% 0 -68.5% -68.5% 0
Post 1 -64.9% -64.9% 0 -70.5% -70.5% 0
5. Still private Pre 22 160.7% *** 65.4% *** 16 74.8% ** 2.5% 11 Post 22 102.4% ** 36.5% ** 15 35.0% -15.5% 9 Total (1-5) Pre 58 122.4% *** 77.5% *** 49 68.2% *** 31.6% *** 40 Post 58 66.5% *** 39.8% *** 44 30.2% ** 6.8% 32 Total (1-4) Pre 35 103.7% *** 80.8% *** 33 64.2% *** 37.0% *** 29 Post 35 47.7% *** 43.3% *** 30 27.3% ** 10.7% ** 23 IRR (1-5) Pre 58 19.1% *** 12.8% *** 48 Post 58 10.1% *** 7.0% *** 44 IRR (1-4) Pre 35 22.6% *** 14.0% *** 33 Post 35 11.9% *** 7.2% *** 30
When omitting companies that have not exited yet, the group with the largest difference between average and median returns, the results do not differ significantly from the results I find when including this group. The only important difference is that the median market- and risk-adjusted return is statistically different from zero.
Comparing these results to the returns of 1980s LBOs, it shows that the returns are considerably smaller than the results documented by Kaplan (1989a). The 1980s public-to-private LBOs had an average (median) market- and risk-adjusted return on pre-buyout capital of 96% (77%), and an average (median) market- and risk-adjusted return on post-buyout capital of 41.9% (28%). Unfortunately Kaplan (1989a) does not report returns per outcome group, so therefore no comparison can be made.
22
5.
Sources of value
The results in table 6 indicate that considerable value is created in LBO transactions, which raises the question what the source of this value is. First I report the proportion of return on capital due to increased operational performance and the tax benefit of debt individually. Furthermore, I look whether the change in industry and market valuation multiples present a form of value creation.
5.1 Increased operational performance
To test the hypothesis that companies become more efficient after LBO transactions, I use the procedure by Kaplan (1989a). This research focusses on changes in two cash flow measures, Earnings Before Interest, Taxes, Depreciation and Amortization (EBITDA) and Net Cash Flow (NCF), which is calculated by subtracting capital expenditures from EBITDA. EBITDA measures the cash that is generated by the company’s operations before depreciation and amortization and all theories on increased operational performance imply an increase in EBITDA after the LBO transaction. NCF measures the cash generated by operations minus the capital expenditures to generate this cash. The theories on increased operational performance that focus on lowering agency costs state that managers of public firms invest in value destroying projects for their own private benefits. These theories imply that after a LBO, managers do not invest in these projects, which causes capital expenditures to decrease and subsequently NCF to increase compared to the pre-buyout levels.
For these two cash flow measures, I report change in absolute value as well as the change in a profitability ratio and a return on asset ratio, to control for acquisitions, divestitures and differences in growth. The profitability ratio is calculated by dividing EBITDA and NCF by that year’s sales. To calculate the return on assets ratio, EBITDA and NCF are divided by the average assets of the company that year. I adjust pre-buyout assets in line with Kaplan (1989a) and Guo, Hotchkiss & Song (2011) by adjusting them with the size with which the assets increased at the time of the buyout due to accounting principles. This size is gathered from the SEC proxy statement of the buyout. When this size is not presented, I estimate it by taking the difference between the purchase price of equity at time of the buyout and the pre-buyout book value of equity.
23 I also calculate the industry adjusted change in operating performance, to control for market- and industry wide effects. This methodology follows Kaplan (1989a) and thus provides a useful comparison to the change in operating performance of 1980s LBOs. To do this, I subtract the industry median change of each measure of operational efficiency over the relevant period from the company’s change. The industry median change is calculated by taking the median change of all companies in the Compustat database with the same 4-digit Standard Industry Code (SIC) with total assets of at least $20 million. When there are less than three companies that satisfy these criteria, the search is widened by matching industry on the first 3 digits of the SIC.
Lie (2001) argues that the method of Kaplan (1989a) produces biased test statistics. He shows that making a small sample of control firms with similar prior levels and changes of performance together with similar market-to-book ratios yields the most reliable test statistics for detecting abnormal operating performance. Therefore, this paper also presents the industry, performance, and market-to-book adjusted change. I create a benchmark by identifying companies that have (i) the same 2 digit SIC as the sample company; (ii) a return on assets in the last full year before the buyout within ± 20% of the sample company; (iii) a change in return on assets in the last full year before the buyout within ± 20% of the sample company; and (iv) a market-to-book ratio in the last full year before the buyout within ± 20% of the sample company. To form the final benchmark companies, I select the three companies with the lowest sum of absolute differences on return on assets in year t-1, change in return on assets from year t-2 to t-1 and market-to-book ratio in year t-1. In formula this is defined as formula 5.
𝑇𝑜𝑡𝑎𝑙 𝑠𝑢𝑚
𝑓𝑖𝑟𝑚 𝑖=
|𝑅𝑂𝐴
𝑆𝑎𝑚𝑝𝑙𝑒 𝑓𝑖𝑟𝑚,𝑡−1− 𝑅𝑂𝐴
𝐹𝑖𝑟𝑚 𝑖,𝑡−1| + |∆𝑅𝑂𝐴
𝑆𝑎𝑚𝑝𝑙𝑒 𝑓𝑖𝑟𝑚,𝑡−2 𝑡𝑜 𝑡−1−
∆𝑅𝑂𝐴
𝐹𝑖𝑟𝑚 𝑖,𝑡−2 𝑡𝑜 𝑡−1| + |𝑀/𝐵
𝑆𝑎𝑚𝑝𝑙𝑒 𝑓𝑖𝑟𝑚,𝑡−1− 𝑀/𝐵
𝐹𝑖𝑟𝑚 𝑖,𝑡−1|
(5)
When less than 3 companies are found that match all criteria, I repeat the same procedure with 1 digit SIC.
24 EBITDA and the hypothetical EBITDA to arrive at earnings before taxes (EBT). Hypothetical taxes are then calculated for both the actual EBT and hypothetical EBT, based on the federal marginal tax rate. The sum of differences in after-tax cash flows for the full period of the LBO are divided by either pre- or post-buyout capital to calculate the proportion of return on capital that is generated by the changes in operating performance. If changes in operating performance are sustained after the exit of the LBO, these changes are reflected in the terminal value of the LBO company. To estimate the part of terminal value that is attributable to changes in operating performance, the differences in actual and hypothetical EBITDA of the last year before the exit of the LBO are multiplied by the company’s industry EV/EBITDA multiplier. This value is divided by pre- and post-buyout capital to calculate the portion of returns that are generated by the change in terminal value due to changes in operating performance. Lastly, hypothetical returns on capital are presented if the company had maintained its pre-buyout level of operating performance. These are calculated by subtracting the proportion of return due to the annual benefit of increased operating performance and the proportion of return due to change in terminal value
Table 8 in the appendix presents the median changes in operating performance between the last two full years before the LBO (-2 to -1) and between the last full year before the LBO to the years after the LBO. Because not all companies exit the LBO after the same number of years, the change between the last year pre-buyout to the last year before the exit presents the most objective statistic. The table presents the unadjusted change in EBITDA and Net Cash Flow, as well as in two profitability ratios, EBITDA/Sales and Net Cash Flow/Sales, and two return on assets ratios, EBITDA/total assets and Net Cash Flow/total assets. Furthermore the table presents the industry adjusted change and the industry- performance- and market-to-book ratio-adjusted change, which is the best comparison according to Lie (2001). The unadjusted change in EBITDA and Net Cash Flow between the last 2 full years before the LBO is positive and statistically significant. The unadjusted change in EBITDA and Net Cash Flow from the last year pre-buyout to the first full year after the LBO, the second year after the LBO and the final year before the exit of the LBO are all positive as well, but statistically insignificant. Furthermore, there is a substantial difference in change in EBITDA and Net Cash Flow from the last full year before the LBO to the last year for the total sample and the subsample that has exited the LBO.
25 percentage points respectively. Only the -1.54 percentage point change in EBITDA/Sales is statistically significant. The industry adjusted change in profitability ratios indicate that the LBO companies outperformed the industry in the years before the LBO (-2 to -1) with 0.20 and 0.86 percentage point on respectively the EBITDA/Sales and Net Cash Flow/Sales ratios. From the last year before the LBO to the last year before the exit the change in EBITDA/Sales is slightly less negative, -0.47 percentage point, while the Net Cash Flow/Sales ratio is more negative, -0.45 percentage point. The industry- performance- and market-to-book-adjusted change in these ratios is considerably better than both the unadjusted change and the industry-adjusted change. Over the total sample of LBOs, the industry- performance- and market-to-book-adjusted change in EBITDA/Sales and Net Cash Flow/Sales is 1.59 and 2.24 percentage points respectively. These results are not statistically significant. Apart from the industry adjusted change in Net Cash Flow, all ratios are substantially better for the subsample that has exited the LBO than for the full sample.
The unadjusted return on assets ratios also show a decline from pre-buyout to the last full year before the exit. The EBITDA/Total Assets ratio declined with -1.07 percentage point, while the Net Cash Flow/Total Assets ratio declined with -0.42 percentage point. The industry adjusted change for these two ratios is positive, 0.41 and 0.50 percentage points respectively. The industry- performance- and market-to-book-adjusted change in EBITDA/Total Assets for the last year before the LBO to the last year before the exit is -0.31 percentage points, while the change in the Net Cash Flow ratio is positive, 0.37. However, most of these results are statistically insignificant. Furthermore, the subsample that has exited the LBO performs again considerably better than the full sample of LBOs.
If we compare these results with the results Kaplan (1989a) reports on 1980s LBOs, it becomes clear that increased operational performance was a far larger contributor to value during the 1980s. Kaplan (1989a) reports industry-adjusted increases in net cash flow, net cash flow/total assets and net cash flow/total sales of 43.1, 85.4 and 72.5 percentage points respectively from the year before the buyout to two years after the buyout (years -1 to +2). Industry-adjusted change in EBITDA, EBITDA/total assets and EBITDA/total sales was 0.7, 36.1 and 23.3 percentage points over the same period.
26 Flow/Total Assets ratio. However, most outcome groups have small samples, so not much value can be derived from these results.
Although most changes in operating performance are statistically insignificant, they are economically significant. Table 9 in the appendix presents the median proportion of return on pre- and post-buyout capital that is due to changes in operating performance. The proportion of return due to annual operational efficiency benefit is the proportion of return on capital that is due to changes in operational performance while being private. The proportion of return on capital that is due to the terminal value operational performance benefit is the increase in terminal value due to increased operational performance in the last year pre-buyout. For the total sample of buyouts, the median proportion of return on pre- (post-)buyout capital due to the annual operational performance benefit is 0.62% (0.4%). The proportion of return on pre- (post-)buyout capital due to the terminal value performance benefit is 16.29% (12.91%). The median total operational performance benefit to pre- (post-)buyout capital is 17.31% (13.31%).
Thus, over the total sample, 17.31% (13.31%) return on pre- (post-)buyout capital is attributable to increased operational performance, which confirms my second hypothesis that increased operational performance are a source of value in LBO transactions.
5.2 Tax benefits of debt
To calculate the effects of the value of the tax benefits of debt on LBO return, the difference between the actual tax payments and hypothetical tax payments if the company would have sustained its pre-buyout debt levels are calculated. This is in line with Guo, Hotchkiss & Song (2011). To find the hypothetical taxes, hypothetical interest payments are calculated based on the company’s actual earnings before interest and taxes (EBIT) and tax loss carryforwards for each year following the buyout. For each year with a positive EBIT, it is assumed that the company pays hypothetical interest payments such that the interest coverage ratio stays equal to the pre-buyout interest coverage ratio. For years with a negative EBIT, it is assumed that the company’s interest payments are equal to the interest payments in the last year pre-buyout. These assumptions are in line with Graham (2000, 2001). These hypothetical interest payments are deducted from the company’s actual EBIT to obtain a hypothetical earnings before taxes (EBT). To calculate the hypothetical taxes, the hypothetical EBT is multiplied with the company’s marginal federal corporate income tax rate8. In order to arrive at the difference in taxes, the hypothetical tax payments are
8
27 subtracted from the actual tax payments. The net present value of the difference in taxes is calculated by discounting them to the pre- or post-buyout date using the median interest rate on debt of 10.35%. To estimate the portion of return on capital that is generated by the tax shield on debt, sum of present values of the differences in taxes is divided by pre- or post-buyout capital.
If the company would sustain its debt-to-equity ratio after the exit of the LBO, part of the terminal value of the company would reflect the tax benefits of debt (Guo, Hotchkiss & Song, 2011). Guo, Hotchkiss & Song (2011) advise to calculate this benefit separately from the tax benefit while being private, as the assumption that a company maintains its debt-to-equity ratio post-LBO is not so likely for all LBO outcomes. For instance, for a secondary LBO it is reasonable to assume that debt-to-equity ratios will not change drastically after the exit of the LBO. However, companies that enter a Chapter 11 bankruptcy code are likely to delever significantly. To estimate the effect of tax benefits on terminal value, it is assumes that the difference between the hypothetical taxes and actual taxes in the last year of the buyout are sustained into perpetuity, discounted by the median interest rate of 10.35%. This perpetuity is then discounted to the pre- or post-buyout date. A total tax benefit is calculated by adding the annual tax benefit while being private with the terminal value tax benefit. Furthermore, I present hypothetical returns on capital if the company had maintained its pre-buyout level of debt by subtracting the proportions of return on capital due to the annual tax benefit and the terminal value tax benefit from actual return on capital.
Table 10 in the appendix reports the proportion of return on capital due to the tax benefits on debt. The median tax benefit while the company is private is 6.8% (5.2%) return on pre- (post-)buyout capital for the total sample of LBOs. The median terminal value tax benefit is 10.5% (8.6%) return on pre- (post-)buyout capital for the total sample of LBOs, however, as explained in the methodology section, this is subjective to assumptions on the leverage percentage and debt repayments after the exit. The median total tax benefit is 18.6% (14.1%) return on pre- (post-)buyout capital. The secondary LBO outcome group has a substantially higher median tax benefit while the company is private compared to the full sample, 14.7%. Unsurprisingly, this subsample also has the largest increase in leverage from pre-buyout to post-buyout. The hypothetical terminal value tax benefit is the largest for the Chapter 11 outcome group, however, during a Chapter 11 procedure companies delever significantly and will thus not be able to fully capture this benefit.
The results show that the median total tax benefit for the full sample is attributable for 18.6% (14.1%) of return on pre- (post-)buyout capital and I can therefor confirm my hypothesis that the tax benefits of debt provide value in LBO transactions.
28 5.3 Changes in industry valuation multiples
If the market- and industry valuation multiples increase during the period that the company is private, company value increases without changing underlying cash flows. To test to which extend changes in these multiples influence the return on capital, I calculate EV/EBITDA multiples for each of the sample companies on the LBO announcement date (T=1), LBO completion date (T=3), and terminal value date (T=4). For the LBO announcement date multiple and the LBO completion date multiple, I divide the pre- or post-buyout capital by the last reported EBITDA before that date. Next, I divide the exit value of the company by the last reported EBITDA before the exit to calculate the terminal value date multiple. Over the same time period, also the change in industry multiple from the pre- and post-buyout date to the terminal value date for each sample company is calculated. To calculate the industry multiple, the median EV/EBITDA multiple for all companies in the COMPUSTAT database with the same 4-digit SIC, excluding the sample company is calculated. If there are less than five companies with the same 4-digit SIC, companies with the same 3-digit SIC are selected. Furthermore, I calculate the difference in market multiple over the period of the LBO by calculating the median EV/EBITDA multiple for all companies in the S&P500.
Changes in market valuation multiples can be ignored, since these are already accounted for in the market- and risk adjusted returns. However, these returns do not account for higher industry valuations relative to the market. For this reason, the change in market multiple are deducted from the change in industry multiple, to arrive at the (industry – market) multiple. The change in this multiple shows how companies in this industry are valuated relative to the market. For each company in my sample, I multiply the change in (industry – market) multiple for both the pre-buyout date to the terminal value date and the post-buyout date to the terminal value date with the actual EBITDA at T=4. I then divide the cash flow generated by the change in (industry – market) multiple by either the pre- or post-buyout capital to calculate the proportion of return on capital that is attributable to changes in industry valuation multiples.
29 do not create value. Changes in the (industry – market) multiple explain -1.6% (0.1%) of return to pre- (post-)buyout capital. The IPO outcome group is the only outcome group for which changes in the (industry – market) multiple adds substantial value, 9.5% (5.7%) return on pre- (post-)buyout capital. For the companies that exit via a trade sale and the companies that are still private, changes in the (industry – market) multiple creates no value or slightly negative value. For the LBOs that exit via a secondary LBO or via Chapter 11 changes in these multiples create a considerable negative value.
30
Table 7
Summary of the Sources of Value Creation
This table presents a summary of the results of tables 8, 10 and 12 in the appendix. It shows the return to pre- and post-buyout capital for each of the outcome groups, and the proportion of that return that is attributable to the changes in operating performance (see table 8), changes in multiples (see table 12), and to the tax benefit of debt (see table 10).
Proportion of
Return Due to: Proportion of Return Due to: Proportion of Return Due to:
Outcome Capital Return to capital Change in Operating Performance Change in Industry Multiple Change in Market Multiple Change in (Industry - Market) Multiple Tax Benefit
While Private TV Tax Benefit Total Tax Benefit
1. IPO Pre 35.91% 27.31% -21.29% -29.61% 9.47% 6.39% 11.54% 19.79%
Post 9.54% 19.38% -17.36% -23.76% 5.70% 5.16% 8.99% 15.11%
2. Trade Sale Pre 32.91% -36.09% -8.85% -12.49% -0.44% 4.85% 4.74% 12.90%
Post 19.18% -29.49% -4.19% -8.49% 1.19% 3.97% 4.12% 9.86%
3. 2nd LBO Pre 146.42% 83.08% -34.36% -11.83% -22.53% 14.65% 24.43% 39.09%
Post 107.58% 33.33% -32.90% 3.22% -36.12% 6.58% 18.71% 25.28%
4. Chapter 11 Pre -68.55% -1.59% -42.68% -19.11% -23.57% 4.85% 36.36% 41.21%
Post -70.54% -1.50% -40.28% -18.04% -22.24% 4.58% 36.74% 41.32%
5. Still private Pre 2.49% -0.64% -11.74% 6.06% -4.13% 7.22% 10.06% 17.80%
Post -15.47% -0.59% -8.05% 5.07% -2.22% 5.55% 8.57% 14.65%
Total (1-5) Pre 31.58% 17.31% -16.84% -14.19% -1.55% 6.82% 10.46% 18.63%
Post 6.83% 13.31% -10.15% -9.13% 0.05% 5.16% 8.57% 14.09%
Total (1-4) Pre 37.04% 25.60% -19.25% -25.29% 1.40% 6.37% 10.47% 19.24%
31
6. Conclusion
This thesis studies the value creation of U.S. leveraged buyouts completed during the second LBO wave. Previous literature mainly focuses on U.S. LBOs during the first LBO wave in the 1980s and found these LBOs to be highly value creating. Using a sample of 140 U.S. public-to-private LBOs completed between 2003 up to and including 2013, I show that the LBOs during the most recent wave are both more conservatively priced and more conservatively leveraged than 1980s LBOS. The usage of less debt results in a considerably smaller proportion of LBOs that end in Chapter 11 bankruptcy.
The main research question of this thesis is “do leveraged buyout transactions create value?”. I find that during the second LBO wave, U.S. LBO transactions still create significant value for all outcome groups of LBOs, apart from firms that end in bankruptcy. I find an average market- and risk-adjusted return on pre- (post-)buyout capital of 68.2% (30.2%). Looking at the source of this value, I find that both increased operational performance and the tax benefits of debt create an equal amount of value. Comparing these results with 1980s LBOs, I show that returns on capital are significantly smaller during the second wave compared to the first wave of LBOs. This difference is mainly attributable to declined improvements of operating performance. Although operating improvements are still responsible for 17.3% return on pre-buyout capital, this is significantly less than what previous studies find during 1980s LBOs. The returns on capital due to the tax benefits of debt remain with 18.6% at similar levels compared to 1980s LBOs. Furthermore I show that increased industry multiples do not form a source of value during the most recent wave of LBOs.
post-32 buyout effects may not be statistically significant, while they are economically very significant. These factors are the most important to keep in mind while interpreting the results.
33
7. Acknowledgements
Finally, I have to acknowledge that it was not possible for me to write this thesis without the help of several people. First of all I would like to thank my supervisor, Professor Doctor Wolfgang Bessler, whose knowledge and enthusiasm fueled my motivation with every meeting we had. Secondly, I’m very grateful to Ewout Brouwers, my supervisor at KPMG, whose critical eye helped me a lot. Furthermore, a special thanks to my colleagues at KPMG, who made me feel that I was never on my own during the whole process of writing this thesis. Lastly I would like to thank my family and friends, especially my dear friend Tim Wijnsma, whose support was never-ending.
34
8. Reference list
Acharya, V., & Kehoe, C. (2008). Corporate Governance and Value Creation Evidence from Private Equity. Working paper, London Business School.
Acharya, V., Gottschalg, O., Hahn, M., & Kehoe, C. (2013). Corporate Governance and Value Creation: Evidence from Private Equity. Review of Financial Studies, 26(2), 369-402.
Admati, A., Pleiderer, P., & Zechner, J. (1994). Large shareholder activism, risk sharing, and financial market equilibrium. Journal of Political Economy, 102, 1097-1130.
Andrade, G., & Kaplan, S. (1998). How Costly is Financial (not Economic) Distress? Evidence from Highly Leveraged Transactions That Became Distressed. Journal of Finance, 53, 1443-1493. Bergström, C., Grubb, M., & Jonsson, S. (2007). The operating impact of buyouts in Sweden: A study
of value creation. Journal of Private Equity, Winter 2007.
Bharath, S., & Dittmar, A. (2010). Why do firms use private equity to opt out of public markets? Review of Financial Studies, 23, 1771-1818.
Boucly, Q., Sraer, D., & Thesmar, D. (2008). Do leveraged buyouts appropriate worker rents? Evidence from French data. working paper, HEC Paris.
Bruton, G., Keels, J., & Scifres, R. (2002). Corporate resturcturing and performance: an agency perspective on the complete buyout cycle. Journal of Business Research, 59, 709-724. Cornelli, F. (2008). Private equity and corporate governance: Do LBOs have more effective boards?
Working paper, London Business School.
Cotter, J., & Peck, S. (2001). The sturcture of debt and active equity investors: The case of the buyout specialist. Journal of Financial Economics, 59, 101-147.
Desbrières, P., & Schatt, A. (2002). The Impacts of LBOs on the Performance of Acquired Firms: The French Case. Journal of Business, Finance & Accounting, 29, 695-729.
Easterbrook, F. (1984). Two agency-cost explanations of dividends. American Economic Review, 74, 650-659.
Faccio, M., & Lang, L. (2002). The Ultimate Ownership of Western European. Journal of Financial Economics, 65, 365-395.
Gertner, R., & Kaplan, S. (1996). The value-maximizing board. Available at SSRN: http://ssrn.com/abstract=10975.
Gilson, S., Hotchkiss, E., & Ruback, R. (2000). Valuation of Bankrupt Firms. The Review of Financial Studies, 13(1), 43-74.
Graham, J. (2000). How big are the tax benefits of debt? Journal of Finance, 55, 1901-1941.
