PRIVATE DEBT INVESTMENTS IN ASIA: VOLATILITY, CREDIT RISK, AND RETURNS Douglas Cumming, Grant Fleming and Frank Liu
Paper presented to the Second Annual Credit Risk Conference in China: Recent Advances in Credit Research Shanghai Advanced Institute of Finance, SJTU Shanghai, China May 16, 2015
Douglas Cumming is Professor and Ontario Research Chair, York University - Schulich School of Business, 4700 Keele Street, Toronto, Ontario M3J 1P3, Canada ( http://ssrn.com/author=75390;
[email protected]) Grant Fleming is Partner, Continuity Capital Partners, Level 8, 12 Moore Street, Canberra, ACT 2601, Australia (http://ssrn.com/author=1454188;
[email protected]) Frank Liu is Assistant Professor, Accounting and Finance, University of Western Australia Business School, 35 Stirling Highway, Crawley, WA 6009, Australia (http://ssrn.com/author=1948476;
[email protected])
PRIVATE DEBT INVESTMENTS IN ASIA: VOLATILITY, CREDIT RISK, AND RETURNS
ABSTRACT
We examine the performance of investments made by private credit fund managers into 321 private companies in 13 Asian countries from 2001 to 2014. We show that the returns to private debt investments are relatively uniform across size, country and industry despite diversity in legal and economic system, size and age of credit markets. We compare the returns to two investments strategies commonly adopted by credit fund managers – buy-and-hold and secondary trading strategies. We find that strategies which involve buying/selling private debt on the secondary market deliver higher returns than a strategy of buying-and-holding a primary issuance. Finally, we conduct time series analysis of the variation in the performance of private debt investments. We build a private credit return index from the underlying loan data and calculate excess returns to private debt investments. Excess returns are positive and stationary over time. Excess returns are positively related to volatility (as measured by ΔVIX), but are not influenced by credit risk (TED spread) or market liquidity.
JEL Codes: C53, D82, G23, G24 Keywords: Private debt; performance; trading strategies; excess returns; credit risk; liquidity; volatility
2
1.
Introduction
Private debt is the predominant source of debt financing for companies around the world. A number of studies have examined the borrower’s decision on the source of debt (public, bank or nonbank)(Denis and Mihov 2003; Lin, Ma, Malatesta and Xuan 2013), the characteristics of private debt issuers (Krishnaswami, Spindt and Subramaniam 1999; Cantillo and Wright 2000; Denis and Mihov 2003; Ackert, Huang and Ramirez 2007), private debt loan contracts (Strahan 1999; Rajan and Winton 1995; Carey, Post and Sharpe 1998; Dennis, Nandy and Sharpe 2000; Bradley and Roberts 2004; Ackert, Huang and Ramirez 2007) and the risk and return of private loans (Carey 1998; Cumming and Fleming 2013). Most of this research focuses on the United States and few studies examine the performance of private debt investments to the lender/investor.
This paper extends the empirical literature on loans to private companies in several ways. First, we examine the performance of mature private loan investments as measured by the internal rate of return and return on investment (return multiple) to non-bank lenders. Our hand-collected dataset comprises private debt investments made by specialist credit investment funds in 321 private companies in 13 Asian countries from 2001 to 2014. Seventy-five percent of the loans in the dataset are located in companies in Mainland China, Australia, Indonesia and Hong Kong providing diversity by legal and economic system, size and age of credit markets. The median sized investment was US$20 million (average US$28.5 million) delivering an investor a median internal rate of return of 20% (average 32%) with a return multiple of 1.23 (average 1.33). We find that there is some evidence of the return to private debt investments varying by size, but no statistically significant differences in returns by country or industry.
Second, we compare the returns to private loans for two investment strategies commonly adopted by specialist credit investment funds – buy-and-hold and secondary trading strategies – according to whether the loan is senior secured or subordinated in the capital structure of the borrower. Private debt investors have flexibility as to when they invest (at issuance or acquiring the 3
loan post-issuance) and into which level of seniority. Both strategies require investors to source private loans from borrowers and analyse credit quality under conditions of asymmetric information. Furthermore, secondary markets for private loans are typically over-the-counter markets, imposing search, due diligence and contracting costs on buyers and sellers. Our multivariate analysis shows there are statistical differences between buy-and-hold and secondary trading strategies in terms of rates of return and return multiple. Dynamic trading strategies which involve buying private debt on the secondary market deliver higher returns than primary issuance buy-and-hold strategies.
Third, we conduct time series analysis of the variation in the performance of private loan investments in Asia. We build a private credit return index from the underlying loan data using discretisation techniques and lattice models pioneered by Moody’s KMV in estimation of private company credit risk. We calculate excess returns to private debt investment as the difference between the private credit return series and a comprehensive public markets return series (J.P. Morgan Asia Credit Index). We find that excess returns are positive over time and that the excess return series is stationary as measure by standard unit root tests. Finally, we document a positive relation between the excess return to private debt investments and volatility (as measured by VIX), but find that returns are not influenced by credit risk (TED spread) or market liquidity.
The remainder of the paper is organised as follows. In Section 2 we briefly review existing literature on why firms issue private debt, the common features of debt investments and the performance of debt investments. We also outline the key questions examined in this paper. In Section 3 we describe the dataset and provide summary statistics. In Section 4 we present univariate and multivariate results on the performance of private debt investments by investment strategy. We then turn out attention to time series analysis of performance. Section 5 describes out methodology in constructing a private credit index and multivariate analysis of variations in the return and excess return indices. Section 6 presents our conclusions.
4
2.
Existing Literature and Research Questions
In this section we briefly review existing literature on why firms raise private debt and the common features of private debt contracts. We then discuss the performance of private debt investments and literature on time series variation in returns. Finally, we outline the research questions investigated in the paper.
Private Loans – Firm Characteristics and Contract Terms
Private companies have several options to raise debt financing from capital markets. Public debt markets provide a mechanism to issue bonds to public bondholders, companies could secure a bank loan or place debt privately to non-bank financial intermediaries (e.g. investment banks, hedge funds, pension funds). Denis and Mihov (2003) find that companies with higher credit quality borrow from public markets before banks and non-bank lenders. Indeed, there is a positive association between the use of public debt and company characteristics such as size, leverage, age and the amount of debt issued (see Krishnaswami, Spindt and Subramaniam 1999; Cantillo and Wright 2000; Denis and Mihov 2003; Ackert, Huang and Ramirez 2007). The degree of asymmetric information between a company and its lenders also influences the choice of type of debt. Firms with higher levels of idiosyncratic information are more likely to issue debt privately while those with lower information asymmetry issue public debt (Diamond 1991; Denis and Mihov 2003). Dennis and Milleneaux (2000) find a positive relation between information opaqueness and private issuances. It may be the case however that higher quality firms choose private over public debt to avoid disclosing information. Yosha (1995) shows that small and mid-sized private companies might be willing to incur higher financing costs to keep sensitive information away from competitors. These firms opt for bilateral financing (typically privately) over multilateral financing.
5
The issuing of private debt to banks and non-bank lenders typically involves greater levels of monitoring than in the case of “arms-length” investors (Diamond 1984). Banks and non-bank lenders are more likely to have access to information on the borrower, detect managerial expropriation and early warning signs of changes in ability to service and repay the loan. Lin, Ma, Malatesta and Xuan (2013) show that firms with large shareholders (and excess control rights) seek to avoid such monitoring by preferring public debt financing over bank debt. Banks and non-bank lenders are able to structure a loan to a company which incorporates price and non-price terms (collateral, covenants, information rights, control rights) in order to mitigate higher credit risk (Strahan 1999; Dennis, Nandy and Sharpe 2000; Ackert, Huang and Ramirez 2007). If the borrower is a long-term bank customer or repeat issuer on the private market, lenders are able to capture idiosyncratic fir information not available in financial statements – management expertise, their ability to respond to changes in market conditions or competitor threats, the nature of customer and supplier relationships (Dennis and Mullineaux 2000).
Private debt is typically more costly for a company to issue and private debt contracts will usually include more restrictive terms and conditions. Bradley and Roberts (2004) argue that the agency theory of covenants provides a rationale as to why debt contracts contain covenants (see also Rajan and Winton 1995). Covenants allow lenders to mitigate potential conflicts between themselves and managers who act on behalf of shareholders. They find that covenants are more likely in smaller, higher growth firms with less leverage and fewer tangible assets. Notably, Bradley and Roberts show empirically that the inclusion of covenants and the pricing of a loan are determined simultaneously. Ackert, Huang and Ramirez (2007) show that private loan terms are driven by the degree of asymmetric information between borrower and lender, contracting costs and credit risk. Loans to small private firms also tend to have shorter maturities than public firms as lenders limit their exposure to the most risky firms (Berger and Udell 1999; Denis, Nandy and Sharpe 2000; Hubbard, Kuttner and Palia 2002). Finally, loans to private firms tend to have higher levels of collateral (Berger and Udell 1999) although whether this is associated with higher credit risk (Rajan and Winton 1995)
6
or lower credit risk (higher quality firms signalling to lenders by willing to post collateral)(Bester 1985; Besanko and Thakor 1987) is an open empirical question.
The Performance of Private Loans
The performance of private loans and returns to private lenders has received less attention in the finance literature. Lenders to private firms receive financial return from the loan from the cash coupon (typically a fixed rate, paid regularly), payment-in-kind (interest accrued and paid at maturity), upfront fees associated with providing the loan and early repayment penalties (penalties stipulated in loan agreements should the firm repay the loan prior to maturity). Banks and non-bank lenders will take all features of the loan into account when evaluating a new loan and when calculating a fair value of the loan during its holding period (Tschirhart, O’Brien, Moise and Yang 2007). Carey (1998) has shown that a portfolio of private loans has lower default and higher recovery rates than a riskequivalent portfolio of public bonds, and that the difference increases with credit risk. That is, there is good evidence that the highly structured nature of private loans (collateral, covenants etc), close monitoring and scrutiny by private lenders has value which lowers the ex-ante riskiness of the borrower.
Another strand of finance literature examines the relation between bond yields and legal institutions (in particular, creditor rights). Qian and Strahan (2007) and Bae and Goyal (2009) show that bank loan yields are negatively related to the quality of a country’s legal institutions. This body of work draws on the “law matters” finance literature which has established a positive relation between the strength of a country’s legal system, credit rights, structure of covenants, and and the size of corporate bond markets (La Porta, Lopez-de-Silanes, Shleifer and Vishny, 1998; Djankov, McLiesh and Schleifer, 2007; Djankov, Hart, McLiesh and Schleifer, 2008; Qi, Roth and Wald 2008; 2010; 2011). Cumming and Fleming (2013) extend the law and finance literature by examining the returns to private debt investments in 25 countries. They show that there is no relation between returns to
7
private debt investments and a country’s legal system, suggesting that borrowers and lenders negotiate terms and conditions in loan agreements which mitigate specific country/jurisdictional risk.
Macroeconomic and credit market factors have been identified as important determinants of the variation in credit spreads and public bond performance over time. Greenwood and Hanson (2013) show that the quantity of credit (market liquidity) is negatively related to credit quality and lower excess returns to public bondholders. The average quality of issuances on public bond markets deteriorates during the credit boom, resulting in significant underperformance of public corporate bonds against Treasury bonds of similar maturity. Similarly, Collin-Dufresne, Goldstein and Spencer Martin (2001) find that monthly credit spread changes are largely driven by local demand/supply shocks rather than by idiosyncratic default risk. Tang and Yan (2010) document a positive association between credit spreads and volatility in the growth (or change) in gross domestic product (GDP). They also show that credit spreads widen when investors are more risk averse (as measured by investor sentiment). Cumming and Fleming (2013) provide, to our knowledge, the only analysis of private debt returns and macroeconomic and credit market factors. Using panel data over ten years they find no cross-sectional relationship between in private debt returns and GDP per capita of the borrower’s location, or between private debt returns and credit market risk as measured by the TED spread (levels or changes).
Research Questions
We draw upon the existing literature to formulate three sets of research questions.
1. Size, Geography and Industry
Credit quality and firm risk tend to be negatively associated with the size of the firm. The literature on private loans indicates that size is a proxy for credit quality, information opaqueness and associated information asymmetries (Krishnaswami, Spindt and Subramaniam 1999; Cantillo and Wright 2000; 8
Denis and Mihov 2003; Ackert, Huang and Ramirez 2007). We postulate a negative relation between the size of a private debt investment (a proxy for firm size) and returns (as measured by internal rate of return and return multiple). In terms of country of private debt issuer, we have no prior as to whether returns vary by the location of the issuer. Cumming and Fleming find that there is no relationship between private debt returns and the legal jurisdiction of the company issuing the debt. By contrast, Qian and Strahan (2007) and Bae and Goyal (2009) show that legal system can influence credit spreads. Finally, we expect that investment returns vary by industry given differences in levels of tangible assets, revenue and earnings volatility.
2. Investment strategies: Buy-and-hold versus dynamic trading strategies
Our data allows an examination of the returns to buy-and-hold versus “dynamic trading strategies”. Does a trading strategy of buying and selling loans in the secondary market add or detract value from a primary loan only (buy-and-hold) strategy? Our a priori view is that credit fund managers are rational and that compensation structures encourage value enhancing behaviour. On this basis, we expect the investment returns to trading on the secondary market to be at least as high as those available from primary placements (buy-and-hold). Second, we measure the extent to which private debt investment returns vary by a particular ownership type – a leveraged buyout (LBO) debt issuer. To our knowledge our study is the first to examine the private debt returns associated with LBO and non-LBO backed firms. We have no prior view as to whether debt issued by LBO-backed firms performs differently from non-LBO backed private debt. LBO backed firms have large shareholders (typically an LBO firm will own in excess of 90% of the equity of a private company) which are motivated to maximise equity value and use debt to disciple managers by limiting free cashflow (see Kaplan and Stromberg 1999; Axelson, Jenkinson, Strömberg and Weisbach 2007). Also, LBO firms may be incentivised to ensure that private firm issuers do not renege on debt contracts in order to build reputation on debt markets. In cases when LBO firms default on debt, Cressy and Farag (2012) find that LBO backed firms have higher recover rates than non-LBO backed firms during periods of high credit availability. On the other hand, LBO firms also tend to use debt to bolster returns during 9
periods of high credit availability. Greenwood and Hanson (2013) show that the quantity of credit (market liquidity) is negatively related to credit quality and lower excess returns to public bondholders. Thus, it is possible that LBO backed firms have higher leverage levels and higher default risk than non-LBO backed firms during credit booms.
3. Excess returns
Our third set of research questions relate to the time series features of private debt returns. We develop a private credit return index from underlying private debt information to chart whether and how private debt returns vary over time. We first examine whether there is excess returns to private debt investing over and above publicly traded debt. We draw inspiration from the alternative assets literature which has found excess returns (alpha) in private equity (Kaplan and Schoar 2005; Fan, Fleming and Warren 2013) and private real estate (Kaiser 2005; Alcock, Baum, Colley and Steiner 2013) and postulate positive excess returns to private debt. Second, are excess returns stationary over time? Finally, we examine the time series variation in our excess returns series. Following CollinDufresne, Goldstein and Spencer Martin (2001), Greenwood and Hanson (2013) and Tang and Yan (2009) we expect excess returns to be positively related to credit risk (as measured by the TED spread), positively related to volatility (VIX) and negatively related to market liquidity.
10
3.
Data and Summary Statistics
The dataset comprises private debt investments made by thirteen specialist credit investment funds in 321 private companies in 13 Asian countries from 2001 to 2014. The data were hand-collected from confidential, private placement memorandum issued by Asia-based credit fund managers which raised capital from “sophisticated” (or wholesale) institutional investors. The data represent the total investment track record for each credit fund manager, typically audited by a reputable accounting firm. The median fund manager had been investing in Asian credit markets for 13 years (average 11.9 years), had invested US$1.7 billion (average US$2.2 billion) and had 10 investment professionals (average 32 investment professionals). The institutional investor which provide the dataset only invested in a subset of the credit fund managers (2 out of 13 managers), reducing any selection bias in the collection of private placement memorandum.
Each private placement memorandum provides prospective investors with the historical track of the credit fund manager at the individual private debt investment level. The data typically includes the following information: •
Issuance and realisation data of the private debt investment;
•
Location (country) of company issuing the private debt;
•
A company description and industry in which the issuing company operates;
•
The type of debt instrument – senior secured loan or subordinated loan;
•
Private debt investment metrics for the credit fund manager – the amount of capital invested in the debt instrument; the realised component of the investment and total return; and
•
Private debt investment returns: an internal rate of return for the investment (based in audited cashflows), and the return on investment (or return multiple)(defined as the total amount of capital returned – principal, coupon and additional payments (e.g. upfront arrangement fees; early prepayment fees) divided by the initial investment outlay).
11
Our process for data collection and verification involved double-checking the entry of all data, crosschecking investment returns with each credit manager and against audited financial statement (where possible), and re-calculating internal rates of return.
The summary statistics for dataset are provided in Table 1 below.
TABLE 1 ABOUT HERE
The median sized investment was US$20 million (average US$28.5 million) delivering an investor a median internal rate of return of 20% (average 32%) with a return multiple of 1.23 (average 1.33). Private debt investment returns range from an internal rate of return of 1,310% to -100%, and a return multiple of 3.97 times investment to 0.00 times investment (that is, a total loss of the loan). The relatively higher internal rates of return at the right-hand side of the distribution are likely explained by the fact that the dataset includes investments which are secondary trades of private debt. Secondary trading strategies involve a credit fund managers acquiring private debt investments over-the-counter at discount to par at times when liquidity is at a premium or a specific holder of the debt needs to sell the debt instrument. Short holding periods and low acquisition prices can result in high internal rates of return as compared with buy-and-hold investment strategies (see Duffie, Gârleanu and Pedersen 2007 for a discussion of how such “price jumps” can occur in over-the-counter markets). Similar cross-sectional variation in returns has been observed in hedge fund studies on dynamic trading strategies across various styles (see, for example, Fung and Hsieh 1997; Sadka 2010). Given such a large range, we winsorize the dataset to account for outliers/influential points later in our analysis.
Table 2 reports the country (location) of each private debt issuer and the industry in which the issuer operates (as defined by the Global Industry Classification Code; GICS). Over eighty percent of the loans in the dataset are located in companies in five countries – Mainland China (38.0%), Australia (14.6%), Indonesia (14.0%), Hong Kong companies with operations in Mainland China 12
(8.7%) and India (6.2%). The data provides diversity by legal and economic system, size and age of credit markets.
TABLE 2 ABOUT HERE
Sixty-eight percent of the loans in the dataset are located in three industries – financials (which includes real estate)(38.0%), industrials (15.9%) and consumer discretionary (14.0%).
Our first set of research questions relate to private debt investment returns and the size, geography and industry of the private debt issuer. We hypothesized a negative relation between size of investment (a proxy for issuer size) and investment returns, due to smaller firms being lower credit quality and having higher degrees of information opaqueness. We found no differences in returns by size as they relate to the internal rate of return on the investment, but a weak significant negative association between size and return multiple. 1 We conclude that there is unsufficient statistical evidence to suggest that private debt investments vary by size. The law and finance literature shows that bank loan yields are negatively related to the quality of a country’s legal institutions (see Qian and Strahan 2007; Bae and Goyal 2009). By contrast, Cumming and Fleming (2013) find no relation between private loans and location of private debt issuers. We might also expect to find differences in private debt returns across industry, as levels of tangible assets, revenue and earnings volatility varies by industry. We investigated the variation in returns by geography and industry at the univariate level through tests for equality of country/industry means and medians using analysis of variable (ANOVA)(means) and Chi-Squared/Kruskal-Wallis (medians) tests. We found no statistically significant differences in means or medians across the dataset for either country or industry.
1
We tested for differences in investment returns by size using ordinary least squares regressions on size (investment cost) and log of size. Both full sample and winsorized samples produced similar results (although with extremely low model adjusted R2 (approximately 1-2%) and F-statistics (significant at the 10% level only).
13
4.
Trading Strategies – Buy-and-Hold versus Secondary Trading
Private credit managers have several ways in which they can invest in private company debt. The credit manager can participate in the primary debt issuance (solely as a bilateral loan or as part of a private syndicate) and hold the investment to maturity (or early repayment). In this scenario the private credit manager is party to negotiation of price and non-price terms of the loan agreement (collateral, covenants, information rights, control rights) in order to mitigate credit risk (Strahan 1999; Dennis, Nandy and Sharpe 2000; Ackert, Huang and Ramirez 2007). An alternative investment strategy is for the private credit manager to acquire the private debt instrument on the secondary market. Such a “dynamic trading strategy” involves the credit fund manager acquiring the private loan over-the-counter, usually brokered by an investment bank (see Duffie, Gârleanu and Pedersen 2007; Duffie 2010).
We have stratified the data on the basis of whether investment returns were generated by the credit fund manager by a primary issuance (buy-and-hold) or secondary (trading) strategy, and whether the debt instrument was senior secured or subordinated.
TABLE 3 ABOUT HERE
Each quadrant shows the average internal rate of return and average return multiple for the various combinations. Primary issuance investments in senior secured loans generated an average return of 31.2% and an average return multiple of 1.26 times, as compared with average secondary returns of 46.3% and 1.75 times. We also note that returns to subordinated loans appear to be lower than those for senior loans despite being lower in the capital structure of the firm (and by definition, having higher credit risk).
14
We next estimate OLS regressions on a winsorized dataset to examine whether there are statistical differences to returns in investment strategies. As noted above, we observe large variation in the returns to private debt investments, resulting in several influential points which bias estimates in ordinary least squared regressions. We adopt a 95% winsorized approach for our regressions, excluding the upper and lower 2.5% of data points. Table 4 reports results for various estimates of the generalised regression model:
Returns = f(Secondary, Subordinated, LBO)
where the base return is a buy-and-hold investment at primary issuance and indicator variables equal one for the type of investment, zero otherwise.
TABLE 4 ABOUT HERE
The regression results indicate that secondary trading generates additional returns over above returns to a buy-and-hold strategy. The secondary coefficient is positive and statistically significant at the 1% level in all model estimations using internal rate of return and return multiples. We also find that return multiples for subordinated debt are higher than senior secured debt, but not for internal rates of return. We find that there is no difference between LBO and non-LBO private debt issuances.
15
5.
The Returns to Private Credit Investing Over Time
There is little empirical evidence on the returns to private debt investing over time. In this section we describe the methodology used to construct an Asia private credit return index (APCR), examine the characteristics of the index and provide multivariate analysis of variations in the return and excess return indices as they relate to changes in macroeconomic and credit market factors.
Private Credit Return Index – Methodology
Private loans can be valued several different ways, each imposing a valuation model which attempts to estimate credit risk at a point in time and a net present value of expected cashflows under no default and default scenarios (Turnbull 2003; Dwyer, Kocagil and Stein 2004; Tschirhart, O’Brien, Moise and Yang 2007; Agrawal, Korablez and Dwyer 2008). The challenge to building a return index with our dataset is the lack of loan revaluations information between the start and maturity dates. That is, only the amount of the investment on the start date and the final realized and unrealized return on investment are recorded. In other words, the actual performance trajectory of each investment is unknown. In order to conduct the time-series analysis on the relationship among Asia private credit returns and market volatility, we employ discretisation techniques and apply a lattice model to construct the credit return index.
First, we discretise the time interval between the maturity or valuation date and the start date into T days. At each time period t, there are a finite number of credit states, N, where the investment can be. In a classic lattice model analysis, when modelling a T-day loan investment as having N credit states, there are NT possible paths for this investment. These credit states include default, nondefault and prepaid. It is a general practice to consider prepayment options when evaluating a loan (for example, see Agrawal, Korablez and Dwyer 2008). However, we do not directly observe
16
sufficient information to infer whether a loan was prepaid in our data set. 2 This reduces the possible paths in the lattice analysis down to 2T.
For each credit state in each time period, being default or non-default, there is a risk-neutral probability of moving from this state to the next. This probability could be approximated by using the expected default frequency (EDF), which is a firm specific and forward-looking measure of actual default probability (Kealhofer 2003). A common practice is to use the Moody’s KMV model to estimate EDF, which requires inputs of the value of equity and other items from the borrower’s balance sheet (Dwyer, Kocagil and Stein 2004; Agrawal, Korablez and Dwyer 2008). Without access to such variables, we are not able to estimate the firm specific EDF and instead we use the cumulative default rates among speculative-grade ratings in Asia-Pacific region (as provided by S&P 2013) to approximate the individual default probability.
The second step is to determine the value of the investment for each credit state. Let us use Si,t 𝑁𝑁 to represent the value of an investment i at time t in [0, Ti], where Ti is the maturity day; and 𝑆𝑆𝑖𝑖,𝑡𝑡 and
𝐷𝐷 to represent the value at time t if it is non-default and default from previous time t-1, respectively. 𝑆𝑆𝑖𝑖,𝑡𝑡
In this setting, only Si,0 and Si,T are known. We start at the maturity date or the last report date. In the
credit state where the investment has not gone into default from the previous period Ti-1, the value of the investment at Ti is: 𝑁𝑁 𝑆𝑆𝑖𝑖,𝑇𝑇 = 𝑆𝑆𝑖𝑖,𝑇𝑇
In the alternate credit state where the investment has been defaulted from the previous period, the value of the investment at Ti is: 𝐷𝐷 𝑆𝑆𝑖𝑖,𝑇𝑇 = (1 − 𝐿𝐿𝐿𝐿𝐿𝐿)𝑆𝑆𝑖𝑖,0
It is important to note that here we assume a fixed proportion of the original investment can be recovered from the original investment in the event of default. An alternative assumption could be 2
One may argue that a loan could be assumed to be prepaid if the maturity of the loan was shorter than a certain length of time; however, there is no empirical finding to support any of arbitrary time periods.
17
to assume a recovery of the investment value from the previous period i.e. Si,T-1 in this case. However, 𝑁𝑁 we do not directly observe the true value Si,T-1. If we used 𝑆𝑆𝑖𝑖,𝑇𝑇−1 to proxy for the true value, it would
𝑁𝑁 is determined by itself. We then step back one day to Ti – 1. For have induced a loop such that 𝑆𝑆𝑖𝑖,𝑇𝑇−1
each credit state, we compute the expected value of the next period’s cash flows under the risk-neutral measure. In the credit state that the investment has not gone default from the previous period Ti – 2, the value of the investment at Ti-1 is:
𝑁𝑁 𝐷𝐷 𝑁𝑁 = 𝑝𝑝𝑖𝑖,𝑡𝑡 𝑆𝑆𝑖𝑖,𝑇𝑇 + �1 − 𝑝𝑝𝑖𝑖,𝑡𝑡 �𝑆𝑆𝑖𝑖,𝑇𝑇 = 𝑝𝑝𝑖𝑖,𝑡𝑡 𝑆𝑆𝑖𝑖,0 (1 − 𝐿𝐿𝐿𝐿𝐿𝐿) + �1 − 𝑝𝑝𝑖𝑖,𝑡𝑡 �𝑆𝑆𝑖𝑖,𝑇𝑇 𝑆𝑆𝑖𝑖,𝑇𝑇−1
where pi,t is the probability of default for investment i, LGD is the loss given default rate and Si,0 is the value of investment at the start, which is known in this setting. Given the lack of firm specific information we set LGD as 20%. Our approach is consistent with Kealhofer (2003) and Gupton and Stein (2005) who argue that LGD values should be set with reference to historical averages to avoid endogeneity issues in estimating the probability of default. As Kealhofer (2003, 84), states, a “…problem arises if LGD has cross-sectional variation that correlates with default probability; this characteristic makes it difficult to separately identify the effect of default probability... this choice of specification makes the default probability to do all the work in fitting the bond prices.” In the alternate credit state where the investment has been defaulted from the previous period, the value of the investment at Ti-1 is:
𝐷𝐷 𝑆𝑆𝑖𝑖,𝑇𝑇−1 = (1 − 𝐿𝐿𝐿𝐿𝐿𝐿)𝑆𝑆𝑖𝑖,0
That is, the investment would terminate at Ti-1 and no further movement in valuation will be observed.
We continue to work backward and track the value of the investment at each credit state
until time 1, assuming that the loan investment would stop following a default. It follows some basic algebra to show that at any time t in [1, …, T-1],
18
𝑁𝑁 = (1 − 𝑝𝑝𝑖𝑖,𝑡𝑡 )𝑇𝑇−𝑡𝑡 𝑆𝑆𝑖𝑖,𝑇𝑇 + 𝑝𝑝𝑖𝑖,𝑡𝑡 𝑆𝑆𝑖𝑖,0 (1 − 𝐿𝐿𝐿𝐿𝐿𝐿)�(1 − 𝑝𝑝𝑖𝑖,𝑡𝑡 )𝑇𝑇−𝑡𝑡−1 + (1 − 𝑝𝑝𝑖𝑖,𝑡𝑡 )𝑇𝑇−𝑡𝑡−2 + ⋯ + 1� 𝑆𝑆𝑖𝑖,𝑡𝑡 𝐷𝐷 = (1 − 𝐿𝐿𝐿𝐿𝐿𝐿)𝑆𝑆𝑖𝑖,0 𝑆𝑆𝑖𝑖,𝑡𝑡
𝑁𝑁 where in the final parenthesis of 𝑆𝑆𝑖𝑖,𝑡𝑡 , there are a total of (T-t) terms.
In the third step, we incorporate coupon payments during the life of an investment. Most private credit investments provide an investor with the combination of cash and non-cash interest (payment-in-kind), with the proportion negotiated as part of the terms of the loan agreement between the borrower and the private credit lender at the start of the loan period. Due to the lack of detailed information on the actual amount of coupon payment and the frequency of payments for each investment, we choose 3 coupon frequencies – quarterly, semi-annually and annually – with 6 coupon rates. Given that the median IRR in our dataset is around 1.20, a 20% yearly coupon yield is used in the optimal case below. This is provided in Table 5 below and applies to all investments in the sample.
TABLE 5 ABOUT HERE
Let us use ci to represent the coupon payment rate for investment i and Ii,t as an indicator function that equals to 1 if t is a coupon paying day. The value of the investment at any time t in [1, …, T-1] is a sum of the investment value in the non-default credit state and an accrued amount of coupons received up until t, 𝑆𝑆� 𝚤𝚤,𝑡𝑡 =
𝑁𝑁 𝑆𝑆𝑖𝑖,𝑡𝑡
𝑡𝑡
+ � 𝑐𝑐𝑖𝑖 𝑆𝑆𝑖𝑖,0 𝐼𝐼𝑖𝑖,𝑗𝑗 𝑗𝑗=1
After the third step, we are able to recover one particular trajectory of investment i:
19
� � �𝑆𝑆𝑖𝑖,0 , 𝑆𝑆� 𝚤𝚤,1 , 𝑆𝑆𝚤𝚤,2 … , 𝑆𝑆𝚤𝚤,𝑇𝑇−1 , 𝑆𝑆𝑖𝑖,𝑇𝑇 � The last step is to construct the Asia private credit return index, which is capitalizationweighted and requires a minimum of two active investments. More specifically, the individual’s weight, wi,t, is determined by the size of the total investment at its inception Si,0. If there are N investments underlying the index on time t, the weight for investment i is,
𝑤𝑤𝑖𝑖,𝑡𝑡 =
𝑆𝑆𝑖𝑖,0 𝑁𝑁 ∑𝑗𝑗=1 𝑆𝑆𝑗𝑗,0
For any t in the sample period from 2006 to 2013, the return index can be calculated as
𝑁𝑁
𝑆𝑆� 𝚤𝚤,𝑡𝑡 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑡𝑡 = � 𝑤𝑤𝑖𝑖,𝑡𝑡 ( − 1) � 𝑆𝑆𝚤𝚤,𝑡𝑡−1 𝑖𝑖=1
In this model, the loan investment values are quite stable such that a daily change can be close to zero. Such observations are not uncommon in private debt valuations. Agrawal, Korablez and Dwyer (2008) find that monthly changes in loans quotes are equal to zero around 47% of the time in their sample of LPC loans quotes from 2002 to 2006.
Private Credit Returns, Public Credit Returns and Stationary
Our private credit return index provides a monthly return series for Asian private credit investments between 2005 and 2014. In order to investigate whether private debt returns differ to public debt returns we calculate an excess return series as the difference between our APCR index and the J.P. Morgan Asia Credit Index (JACI). The JACI is a broad public credit markets index comprising 705 U.S. dollar denominated bonds issued by 312 sovereign, quasi-sovereign and corporates in 15 Asian countries, excluding Japan and Australia/New Zealand. The index is market capitalisation-weighted 20
(market capitalisation of US$445 billion at 31 December 2013) and is 76% investment grade debt and 24% non-investment grade debt. We take the moving average monthly return for each of our six estimations of the Asian private credit index and subtract the monthly JACI. The summary statistics for the excess returns private credit index are shown in Table 6.
TABLE 6 ABOUT HERE
Table 6 shows that monthly average excess returns have a mean between 1.3% and 1.6% per month, with a median between 1.2% and 1.5% per month. However, we can also note periods of private credit underperformance, with minimum monthly returns ranging between -4.1% and -4.8% per month. Skewness and kurtosis statistics indicate that the distribution of excess monthly returns contains a higher proportion of positive excess returns. We test for the stability of the return series using an Augmented Dickey-Fuller test. All test statistics indicate that we cannot reject the null hypothesis that the series is stationary. Figure 1 shows the time series variation in the excess return series.
FIGURE 1 ABOUT HERE
In summary, our excess return series shows that Asian private credit returns deliver out-performance over public market credit returns, that the excess returns are on average between 1.2% and 1.5% per month, and that positive excess returns are stationary over time. These findings are consistent with excess returns documented for private equity (Kaplan and Schoar 2005; Fan, Fleming and Warren 2013) and private real estate (Kaiser 2005; Alcock, Baum, Colley and Steiner 2013). We turn next to examine in more detail the variations in the excess return series.
21
Time Series Variation in Asian Private Credit Returns
Our third set of research questions relate to the time series features of private debt returns. We regress the excess private credit index against three variables measuring credit risk, volatility and liquidity. The generalised form of our regression is as follows:
EXCESS = a + b(TED) + c(ΔVIX) + d(Liquidity) + e
Global credit risk is defined as the TED spread, the daily percentage spread between 3-Month LIBOR rate (based on U.S. dollars) and the 3-Month Treasury bill rate, as calculated by the Federal Reserve Bank of St. Louis. Financial market volatility is measured as the change in the volatility index (ΔVIX) as calculated by the Chicago Board Options Exchange. We adopt an Asia-specific measure of market liquidity using the quarterly year-on-year percentage change in cross-border and domestic credit, using data from the Bank of International Settlements. An increase in the liquidity measure indicates that there is greater amount of credit available in the Asian region as compared with the previous year, due to domestic and/or cross-border capital inflows. We hypothesise that excess returns are positively related to credit risk and volatility, and negatively related to market liquidity. Ceteris paribus, an increase in global credit risk indicates higher levels of investor risk aversion which require higher excess returns as compensation. Similarly, times of higher volatility in finance markets will be associated with higher excess returns (Tang and Yan 2009; Greenwood and Hanson 2013). Finally, we hypothesise that increases in market liquidity in the Asian region will result in excess supply of credit for private firms, and lower excess returns (Collin-Dufresne, Goldstein and Spencer Martin 2001).
The correlation probabilities for the excess return series and explanatory variables used in our regressions are shown in Table 7.
22
TABLE 7 ABOUT HERE
The correlation probabilities show that there are statistically significant correlations between each of the explanatory variables and the excess private credit index. The TED spread and ΔVIX are significantly positively correlated with each of the excess return series, in most cases at the 1% or 5% significance level. The TED spread correlations range from 0.172 (Model e, significant at 10%) to 0.263 (Model b, significant at 1%). The ΔVIX is highly correlated (all at 1% significance) with all excess return series, ranging between 0.315 (Model f) and 0.472 (Model d). The liquidity index is significantly negatively correlated with the excess return index, although usually only at the 10% level. We also note a positive significant correlation between the TED spread and ΔVIX (correlation of 0.315, significant at 1%).
The results of regression analysis are reported in Table 8.
TABLE 8 ABOUT HERE
The consistent result across all regression models is a significant positive association between the excess return index and ΔVIX. We find no association between the TED spread or liquidity and excess returns.
23
6.
Concluding Remarks
Private debt is the predominant source of debt financing for companies around the world. A number of studies have examined the borrower’s decision on the source of debt, the characteristics of private debt issuers, private debt loan contracts and the risk and return of private loans. Most of this research focuses on the United States and few studies examine the performance of private debt investments to the lender/investor. Our paper provides the first analysis of the cross-sectional and time series returns to private debt investments in Asian companies, using a sample of credit fund manager investments across the region. Our data provides insights into private debt investments in diverse economic systems and finance markets including Mainland China, Australia and South East Asia. We show that the returns to private debt investments are relatively uniform across size, country and industry despite country diversity. We find no evidence that “laws matter” for private debt returns; rather if laws do matter we suggest that borrowers and lenders negotiate terms and conditions in loan agreements which mitigate specific country/jurisdictional risk.
Private credit fund managers commonly execute two investment strategies in Asian debt markets. The first involves investment at primary issuance in a private company’s debt and holding that debt to maturity. The second involves a more active dynamic strategy where credit fund managers buy and sell debt over-the-counter. We find that strategies which involve buying/selling private debt on the secondary market deliver higher returns than a strategy of buying-and-holding a primary issuance. The regression results indicate that secondary trading returns are positive and statistically significant at the 1% level in all model estimations using internal rate of return and return multiples. We find that there is no difference between LBO and non-LBO private debt issuances. Our results suggests that credit fund manager trading skills are important in assessing excess returns, over and above the skills involved in the evaluation of private debt opportunities at issuance no matter whether debt is senior secured or subordinated or in LBO-backed or non-LBO backed private companies. Further research is required on how private credit manager trade on the secondary market through the
24
credit cycle and on whether the success or regularity of secondary trading strategies varies due to macroeconomic and credit market factors.
Our private credit return index is the first index to show excess portfolio returns to Asian private credit investments. We have used discretisation techniques and lattice models pioneered by Moody’s KMV to estimate private company credit risk and backwards induce credit returns during the holding period of the investment. We find that excess returns are on average between 1.2% and 1.5% per month, and that positive excess returns are stationary over time. Excess returns are positively related to volatility (as measured by ΔVIX), but are not influenced by credit risk (TED spread) or market liquidity.
25
References
Ackert, L., Huang, R. and Ramírez, G. 2007. “Information opacity, credit risk, and the design of loan contracts for private firms.” Financial Markets, Institutions & Instruments 16(5), 221–242. Agrawal, D., Korablez, I. and Dwyer, D. 2008. Valuation of Corporate Loans: A Credit Migration Approach. Moody’s KMV, January 25. Alcock, J., Baum, A. Colley, N. and Steiner, E. 2013. “The role of financial leverage in the performance of private equity real estate funds.” Journal of Portfolio Management 39(5), 99-110. Axelson, U., Jenkinson, T., Strömberg, P. and Weisbach, M. 2007. “Leverage and pricing in buyouts: An empirical analysis.” Working Paper, SIFR. Bae, K. and Goyal, V. 2009. “Creditor rights, enforcement, and bank loans.” Journal of Finance 64, 823-860. Berger, A. and Udell, G. 1990. “Collateral, loan quality, and bank risk.” Journal of Monetary Economics 25(1), 21–42. Besanko, D. and Thakor, A. 1987 “Collateral and rationing: Sorting equilibria in monopolistic and competitive credit markets.” International Economic Review 28(3), 671-689. Bester, H. 1985. “Screening vs. rationing in credit markets with imperfect information.” American Economic Review 75(4), 850-855. Bradley, M. and Roberts, M. 2004. “The structure and pricing of corporate debt covenants.” 6th Annual Texas Finance Festival. Available at SSRN: http://ssrn.com/abstract=585882. Cantillo, M. and Wright, J. 2000. “How do firms choose their lenders? An empirical investigation.” Review of Financial Studies 13(1), 155-189. Carey, M. 1998. “Credit risk in private debt portfolios.” Journal of Finance, 53(4), 1363-1387. Carey, M., Post, M. and Sharpe, S. 1998. “Does corporate lending by banks and finance companies differ? Evidence on specialization in private debt contracting.” Journal of Finance 53(3), 845–878. Collin-Dufresne, P., Goldstein, R. and Martin, J. 2001. “The determinants of credit spread changes.” Journal of Finance 56, 2177–207. Cressy, R. and Farag, H. 2012. Do private equity-backed buyouts respond better to financial distress than PLCs? European Journal of Finance 18(3-4), 239-259. Cumming, D. and Fleming, G. 2013. “Debt securities in private firms: Types, institutions and performance in 25 countries” Journal of Fixed Income 23(1), 102-123. Denis, D. and Mihov,V. 2003 “The choice among bank debt, non-bank private debt, and public debt: Evidence from new corporate borrowings.” Journal of Financial Economics 70, 3–28.
26
Dennis, S. and Mullineaux, D.J. 2000. “Syndicated loans.” Journal of Financial Intermediation 9(4), 404–426. Dennis, S., Nandy, D. and Sharpe, I. 2000. “The determinants of contract terms in bank revolving credit agreements.” Journal of Financial and Quantitative Analysis 35(1), 87–100. Diamond, D. 1984. “Financial intermediation and delegated monitoring.” Review of Economic Studies 51, 393–414. Diamond, D. 1991. “Monitoring and reputation: The choice between bank loans and directly placed debt.” Journal of Political Economy 99, 689–721. Djankov, S., McLiesh, C., and Schleifer, A. 2007, Private credit in 129 countries. Journal of Financial Economics 84, 299–329. Djankov, S., Hart, O., McLiesh, C., and Schleifer, A. 2008, Debt enforcement around the world. Journal of Political Economy 116, 1105–1149. Duffie, D. 2010. “Presidential address: Asset price dynamics with slow-moving capital.” Journal of Finance 65(4), 1237-1267. Duffie D., Gârleanu, N. Pedersen, L.H. 2007. “Valuation in over-the-counter markets.” Review of Financial Studies 20(6), 1865-1900. Dwyer, D., Kocagil, A. and Stein, R. 2004. Moody’s KMV RiskCalc v3.1 Model, Moody’s KMV, April 5. Fan, F.J., Fleming, G. and Warren, G. 2013. “The alpha, beta and consistency of private equity reported returns.” Journal of Private Equity 16(4), 21-30. Fung, W. and Hsieh, D. 1997. “Empirical characteristics of dynamic trading strategies: The case of hedge funds.” Review of Financial Studies 10(2), 275-302. Greenwood, R. and Hanson, S. 2013. “Issuer quality and corporate bond returns.” Review of Financial Studies 26(6), 1483-1525. Gupton, G. and Stein, R. 2005. LossCalc V2: Dynamic Prediction of LGD, Moody’s KMV, January. Hubbard, G., Kuttner, K. and Palia, D. 2002. “Are there bank effects in borrowers’ cost of funds? Evidence from a matched sample of borrowers and banks.” Journal of Business 75(4), 559– 581. Kaiser, R. 2005. “Analyzing real estate portfolio returns.” Journal of Portfolio Management 31(5), 134-142. Kaplan, S. and Schoar, A. 2005. “Private equity performance: Returns, persistence, and capital flows.” Journal of Finance 60, 1791-1823. Kaplan, S. and Stromberg, P. 1999. ‘Leveraged buyouts and private equity.’ Journal of Economic Perspectives 23(1), 121–146. Kealhofer, S. 2003. “Quantifying credit risk II: Debt valuation.” Financial Analysts Journal 59(3), 78-92. 27
Krishnaswami, S., Spindt, P. and Subramaniam, V. 1999 “Information asymmetry, monitoring, and the placement structure of corporate debt.” Journal of Financial Economics 51(3), 407–434. Lin, C., Ma, Y., Malatesta, P. and Xuan, Y. 2013 “Corporate ownership structure and the choice between bank debt and public debt.” Journal of Financial Economics 109(2), 517–534. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. 1998. “Law and finance.” Journal of Political Economy 106, 1113–1155. Mansi, S., Maxwell, W., & Zhang, A. 2011, “Bankruptcy prediction models and the cost of debt, Journal of Fixed Income 21(4), 25-42. Qi, Y., Roth, L., & Wald, J. 2008. “State laws and debt covenants.” Journal of Law and Economics 51(2), 179-207. Qi, Y., Roth, L., & Wald, J. 2010. “Political rights and the cost of debt.” Journal of Financial Economics 95, 202-226. Qi, Y., Roth, L., & Wald, J. 2011. “How legal environments affect the use of covenants.” Journal of International Business Studies 42, 235-262. Qian, J. & Strahan, P. 2007. “How laws and institutions shape financial contracts: The case of banks loans.” Journal of Finance 62, 2803-2834. Rajan, R. and Winton, A.1995. “Covenants and collateral as incentives to monitor.” Journal of Finance 50, 1113–1146. Sadka, R. 2010. “Liquidity risk and the cross-section of hedge-fund returns.” Journal of Financial Economics 98(1), 54–71. Strahan, P. 1999. “Borrower risk and the price and non-price terms of bank loans.” Working Paper, Banking Studies Function, Federal Reserve Bank of New York. Tang, D. and Yan, H. 2010. “Market conditions, default risk and credit spreads.” Journal of Banking and Finance 34(4), 743–753. Tschirhart, J., O'Brien, J. Moise, M. and Yang, E. 2007. “Bank commercial loan fair value practices.” FEDS Working Paper No. 2007-29. Available at SSRN: http://ssrn.com/abstract=1017604. Turnbull, S. 2003. “Pricing loans using default probabilities.” Economic Notes 32(2), 197-217. Yosha, O. 1995 “Information disclosure costs and the choice of financing source.” Journal of Financial Intermediation 4(1), 3–20.
28
TABLE 1 Asia Private Debt Investment Returns Summary Statistics
Table 1 shows the summary statistics for 321 private debt investments made by thirteen specialist credit investment funds in private companies in 13 Asian countries from 2001 to 2014. Investment is the amount of money outlayed for the private debt investment. Realised is the return to the private debt investment comprising principal, coupon and additional payments (e.g. upfront arrangement fees; early prepayment fees). Unrealised is the credit manager assessed fair market value of the remaining loan. IRR is the internal rate of return to the private debt investment, calculated from the audited cashflows of the credit fund manager. ROI is the return on investment (or return multiple)(defined as the total amount of capital returned divided by the initial investment outlay). All figures are current US dollars.
Investment
Realised
Unrealised
Total (Realised and Unrealised)
IRR
ROI
Average
28,564,428
24,675,000
16,438,431
34,507,139
32%
1.33
Median
20,000,000
11,452,528
-
20,274,416
20%
1.23
Stdev
32,448,382
34,002,898
36,566,424
42,097,239
84%
0.52
Max
300,000,000
203,879,478
332,438,356
332,385,737
1310%
3.97
Min
200,000
(747,916)
(3,977,002)
-
-100%
0.00
319
299
N Total
8,197,990,737
4,811,624,908
2,646,587,451
29
9,731,013,118
TABLE 2 Asia Private Debt Investment Returns by Geography and Industry Summary Statistics Table 2 shows the summary statistics for 321 private debt investments made by thirteen specialist credit investment funds in private companies in 13 Asian countries from 2001 to 2014. Panel A shows the data by the country of the company issuing the private debt (country was defined by the credit fund manager). Panel B shows the data by industry classification, using Global Industry Classification Code (GICS). IRR is the internal rate of return to the private debt investment, calculated from the audited cashflows of the credit fund manager. ROI is the return on investment (or return multiple)(defined as the total amount of capital returned divided by the initial investment outlay).
Panel A Country Frequency Mainland China 122 Australia 47 Indonesia 45 Greater China (Hong Kong/Mainland China) 28 India 20 South East Asia 12 Korea 8 New Zealand 6 Thailand 6 Hong Kong 5 Global 4 Singapore 4 Asia Pacific 3 Miscellaneous 3 Philippines 3 Taiwan 3 Japan 1 Malaysia 1 Panel B GICS Code/Sector 10 Energy 24 15 Materials 28 20 Industrials 51 25 Consumer Discretionary 45 30 Consumer Staples 22 35 Health Care 4 40 Financials 122 45 Information Technology 6 50 Telecommunication Services 3 55 Utilities 13 Miscellaneous 3
30
IRR Percent 38.0% 14.6% 14.0% 8.7% 6.2% 3.7% 2.5% 1.9% 1.9% 1.6% 1.2% 1.2% 0.9% 0.9% 0.9% 0.9% 0.3% 0.3%
Mean 42.1% 22.2% 25.7% 20.9% 29.6% 35.9% 30.2% 12.7% 30.4% 20.9% 18.0% 25.1% 78.3% 19.3% 34.0% 21.7% 18.0% 29.0%
Median 19.5% 16.5% 16.5% 23.9% 21.0% 26.0% 30.0% 26.5% 24.7% 21.8% 17.5% 22.5% 22.0% 6.0% 32.0% 16.0% 18.0% 29.0%
Mean 1.33 1.26 1.39 1.22 1.31 1.38 1.31 1.43 1.25 1.73 1.30 1.53 1.30 1.54 2.23 1.35 NA 2.30
7.5% 8.7% 15.9% 14.0% 6.9% 1.2% 38.0% 1.9% 0.9% 4.0% 0.9%
21.7% 21.8% 36.8% 32.3% 31.4% 34.0% 33.9% 19.1% 41.6% 38.5% 19.3%
14.9% 20.0% 20.2% 19.1% 23.6% 30.5% 19.6% 17.9% 41.0% 28.0% 6.0%
1.19 1.30 1.22 1.38 1.26 1.91 1.34 1.29 1.21 1.58 1.54
ROI Median 1.23 1.23 1.24 1.13 1.34 1.16 1.21 1.43 1.25 1.35 1.30 1.45 1.30 1.12 2.20 1.34 NA 2.30
1.13 1.23 1.16 1.39 1.22 1.60 1.25 1.20 1.29 1.50 1.12
TABLE 3 Investment Returns to Buy-and-Hold and Secondary Trading Strategies Table 3 shows univariate returns by investment strategy and seniority of debt instrument. Primary indicates that the investment made by the credit fund manager was at the primary issuance of the loan, and that the investment remained in the portfolio until realisation. Secondary indicates that the investment made by the credit fund manager was acquired on the secondary market. Senior secured and subordinated refers to whether the loan was senior or subordinated in the capital structure. IRR is the internal rate of return to the private debt investment, calculated from the audited cashflows of the credit fund manager. ROI is the return on investment (or return multiple)(defined as the total amount of capital returned divided by the initial investment outlay).
Panel A
Primary (0) Secondary (1)
Senior (0) IRR (%) N
31.2% 218
46.3% 30
28.4% 58
27.4% 11
Subordinated (1) IRR (%) N Panel B
Primary (0) Secondary (1)
Senior (0) ROI N
1.26 206
1.75 30
1.26 50
1.70 11
Subordinated (1) ROI N
31
TABLE 4 Regressions Results for Buy-and-Hold and Secondary Trading Strategies Table 4 reports results for four estimations of the generalised regression model Returns = f(Secondary, Subordinated, LBO). The base return (0,0) is a buy-and-hold investment at primary issuance. Indicator variables equal one for the type of investment, zero otherwise. IRR is the internal rate of return to the private debt investment, calculated from the audited cashflows of the credit fund manager. ROI is the return on investment (or return multiple)(defined as the total amount of capital returned divided by the initial investment outlay).
IRR Model Intercept
(1)
ROI (2)
(3)
(4)
0.236*** 0.236*** 1.250*** 1.250*** (0.012) (0.012) (0.024) (0.024) Secondary 0.093*** 0.093*** 0.244*** 0.246*** (0.031) (0.031) (0.063) (0.063) Subordinated -0.030 -0.036 0.163*** 0.146** (0.025) (0.030) (0.056) (0.065) 0.047 LBO 0.014 (0.042) (0.092)
Adj R
2
2.52%
2.23%
8.15%
*p < 0.10, **p < 0.05, ***p < 0.01. Standard errors are reported in parentheses.
32
7.90%
TABLE 5 Private Credit Return Index Coupon Frequency and Coupon Rate Assumptions
Table 5 shows assumptions adopted in the private credit index with regards to the frequency of coupon payments and coupon rates. There are 3 coupon frequencies – quarterly, semi-annually and annually – with 6 coupon rates. Half coupon means that 50% of the assumed return (a return of 20% per annum) is paid as cash coupon; full coupon means that 100% of the return is paid as a cash coupon. Letters a – f denote six different models with corresponding assumptions. For example, Model a assumes that the investments in the private credit index pay coupons to investors every 90 days at a rate of 2.5% per quarter (or 10% per annum, half the total return of the investment).
Coupon Frequency (days)
Half Coupon
Full Coupon
90
a. 2.50%
b. 5%
180
c. 5%
d. 10%
360
e. 10%
f. 20%
33
TABLE 6 Asia Private Credit Excess Return Series Summary Statistics Table 6 shows summary statistics of the private credit return index. There are six different Models a – f adopting various assumptions with regards to the frequency of coupon payments and coupon rates. Excess is measured as the difference between the various APCR models and the J.P. Morgan Asia Credit Index (JACI). Excess a, c and e shows the excess return to the private credit index over the JACI, assuming that the investments in the private credit index pay coupons to investors every 90, 180 and 360 days at a rate of 2.5%, 5% and 10% per quarter. Excess b, d, and f shows the excess return to the private credit index over the JACI, assuming that the investments in the private credit index pay coupons to investors every 90, 180 and 360 days at a rate of 5%, 10% and 20% per quarter. ADF t-stat is the Augmented Dickey-Fuller t-statistics for the null hypothesis that the series is stationary.
Summary Stats Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis ADF t-stats
Excess_a Excess_b Excess_c Excess_d Excess_e Excess_f 0.014 0.016 0.014 0.016 0.013 0.016 0.014 0.015 0.012 0.013 0.012 0.013 0.167 0.178 0.166 0.177 0.168 0.183 -0.046 -0.044 -0.048 -0.041 -0.046 -0.044 0.028 0.028 0.028 0.028 0.029 0.029 1.890 2.150 1.868 2.090 1.881 2.152 12.005 14.305 11.663 13.487 11.442 13.053 -7.089 -8.041 -7.033 -8.014 -7.267 -8.706
34
FIGURE 1 Asia Private Credit Excess Return Series, Moving Average Monthly Excess Returns 2005 - 2014
Figure 1 shows excess return time series graphs under six different assumptions (Models a – f) with regards to the frequency of coupon payments and coupon rates. Excess is measured as the difference between the various models and the J.P. Morgan Asia Credit Index (JACI). Excess a, c and e shows the excess return to the private credit index over the JACI, assuming that the investments in the private credit index pay coupons to investors every 90, 180 and 360 days at a rate of 2.5%, 5% and 10% per quarter. Excess b, d, and f shows the excess return to the private credit index over the JACI, assuming that the investments in the private credit index pay coupons to investors every 90, 180 and 360 days at a rate of 5%, 10% and 20% per quarter.
0.2
0.15
0.1
Excess_a Excess_b Excess_c
0.05
Excess_d Excess_e Excess_f
2005M12 2006M02 2006M04 2006M06 2006M08 2006M10 2006M12 2007M02 2007M04 2007M06 2007M08 2007M10 2007M12 2008M02 2008M04 2008M06 2008M08 2008M10 2008M12 2009M02 2009M04 2009M06 2009M08 2009M10 2009M12 2010M02 2010M04 2010M06 2010M08 2010M10 2010M12 2011M02 2011M04 2011M06 2011M08 2011M10 2011M12 2012M02 2012M04 2012M06 2012M08 2012M10 2012M12 2013M02 2013M04 2013M06 2013M08 2013M10 2013M12 2014M02
0
-0.05
-0.1
35
TABLE 7 Asia Private Credit Excess Return Series Correlation Probability Matrix
Table 7 shows the correlation probabilities between the excess return time series (a – f), TED, ΔVIX and Liquidity. TED is measured as the daily percentage spread between 3-Month LIBOR rate (based on U.S. dollars) and the 3-Month Treasury bill rate, as calculated by the Federal Reserve Bank of St. Louis. VIX is measured as the change in the volatility index (VIX) as calculated by the Chicago Board Options Exchange. Liquidity is measured as the quarterly year-on-year percentage change in cross-border and domestic credit, using data from the Bank of International Settlements.
Correlation Probability
Excess_a
Excess_a
1.000
Excess_b
0.984***
1.000
Excess_c
0.995***
0.977***
1.000
Excess_d
0.969***
0.982***
0.981***
1.000
Excess_e
0.980***
0.958***
0.987***
0.964***
1.000
Excess_f
0.927***
0.932***
0.942***
0.955***
0.974***
1.000
Ted
0.206**
0.263***
0.196*
0.244**
0.172*
0.197*
1.000
ΔVIX
0.459***
0.467***
0.463***
0.472***
0.457***
0.454***
0.315***
1.000
Liquidity
-0.223**
-0.199*
-0.212**
-0.175*
-0.190*
-0.124
-0.008
-0.165
Excess_b
Excess_c
Excess_d
*p < 0.10, **p < 0.05, ***p < 0.01.
36
Excess_e
Excess_f
Ted
ΔVIX
Liquidity
1.000
TABLE 8 Regressions Results of Asia Private Credit Excess Return Series, Credit Risk, Volatility and Liquidity
Table 8 shows various results for estimations of the generalised regression model EXCESS = a + b(TED) + c(VIX) + d(Liquidity) + e. There are six different return series (a – f) adopting various assumptions with regards to the frequency of coupon payments and coupon rates. Excess is measured as the difference between the various models and the J.P. Morgan Asia Credit Index (JACI). Excess a, c and e shows the excess return to the private credit index over the JACI, assuming that the investments in the private credit index pay coupons to investors every 90, 180 and 360 days at a rate of 2.5%, 5% and 10% per quarter. Excess b, d, and f shows the excess return to the private credit index over the JACI, assuming that the investments in the private credit index pay coupons to investors every 90, 180 and 360 days at a rate of 5%, 10% and 20% per quarter. TED is measured as the daily percentage spread between 3-Month LIBOR rate (based on U.S. dollars) and the 3Month Treasury bill rate, as calculated by the Federal Reserve Bank of St. Louis. VIX is measured as the change in the volatility index (VIX) as calculated by the Chicago Board Options Exchange. Liquidity is measured as the quarterly year-onyear percentage change in cross-border and domestic credit, using data from the Bank of International Settlements.
Excess_a
Excess_b
Excess_c
Excess_d
Excess_e
Excess_f
Intercept
0.008** (0.004)
0.010** (0.004)
0.008** (0.004)
0.010*** (0.004)
0.009** (0.004)
0.012*** (0.004)
Ted
0.004 (0.005)
0.007 (0.005)
0.003 (0.005)
0.005 (0.005)
0.002 (0.005)
0.003 (0.005)
ΔVIX
0.060*** (0.014)
0.060*** (0.014)
0.061*** (0.014)
0.063*** (0.015)
0.063*** (0.015)
0.067*** (0.016)
Liquidity
-0.023*
-0.021
(0.014)
-0.020 (0.014)
(0.014)
-0.016 (0.014)
-0.018 (0.014)
-0.008 (0.015)
21.3%
22.5%
21.2%
21.9%
19.8%
18.6%
Model
Adj R 2
*p < 0.10, **p < 0.05, ***p < 0.01. Standard errors are reported in parentheses.
37
