* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download PRIVATE EQUITY FOR THE COMMON MAN
Corporate venture capital wikipedia , lookup
Mark-to-market accounting wikipedia , lookup
Short (finance) wikipedia , lookup
Stock trader wikipedia , lookup
Capital gains tax in Australia wikipedia , lookup
Fund governance wikipedia , lookup
Investor-state dispute settlement wikipedia , lookup
Rate of return wikipedia , lookup
International investment agreement wikipedia , lookup
Internal rate of return wikipedia , lookup
Mutual fund wikipedia , lookup
Investment banking wikipedia , lookup
Environmental, social and corporate governance wikipedia , lookup
History of private equity and venture capital wikipedia , lookup
History of investment banking in the United States wikipedia , lookup
Socially responsible investing wikipedia , lookup
Special-purpose acquisition company wikipedia , lookup
Private equity wikipedia , lookup
Leveraged buyout wikipedia , lookup
Private equity secondary market wikipedia , lookup
Private equity in the 2000s wikipedia , lookup
Private equity in the 1980s wikipedia , lookup
Private money investing wikipedia , lookup
PRIVATE EQUITY FOR THE COMMON MAN Timothy J. Riddiough Department of Real Estate & Urban Land Economics School of Business University of Wisconsin – Madison 5262D Grainger Hall Madison, WI 53706 [email protected] Jonathan A. Wiley Department of Real Estate J. Mack Robinson College of Business Georgia State University 35 Broad St., Suite 1408 Atlanta, GA 30303 [email protected] June 2014 1 PRIVATE EQUITY FOR THE COMMON MAN Abstract We describe and analyze supply and demand characteristics of a private equity investment vehicle called the Unlisted REIT. Equity shares are offered through registered investment advisors to retail investors whose net worth and income level requirements are low relative to those seen with other private equity and hedge fund investment alternatives. Fee structures, including front-end loads, are shown to exceed (often far exceed) fee structures on mutual funds, private equity funds, venture capital funds, and hedge funds Unlisted REITs offer high dividend yields and post share prices that do not vary over time, and as such are promoted as high-yielding fixed-income substitutes even though they remain a residual claim on time varying cash flows produced by commercial real estate assets. We show that investment flows into the Unlisted REIT vehicle vary directly with fee percentages paid to the investment advisor, and inversely with interest rates and dividend yields on close and more liquid investment substitutes. Realized holding period returns are calculated and compared to direct liquid share substitutes that offer lower dividend yields and that experience standard share price volatility. Return differentials are shown to exceed 9 percent per year on average, in favor of the liquid share substitute, illustrating the deleterious effects of high fees and a conflicted management structure on realized investment outcomes. Alternative explanations are considered for the anomalous investment behavior (why would informed investors accept such low returns), where we conclude that macroeconomic conditions and monetary policy are at least in part responsible for the proliferation of the Ponzi-like Unlisted REIT investment vehicle. I. Introduction In this study we undertake a detailed analysis of the capital raising process for Unlisted Real Estate Investment Trusts (REITs). These firms have rather interesting and controversial business models, raising capital primarily from retail investors with only moderate levels of wealth and who are often retired. The Unlisted REIT promises high dividend yields (6-7 percent typically on initial investment, gross of fees), issuing common equity shares that are generally 2 illiquid and have little apparent volatility. As such, the investment offers the appearance of highyield bond-like returns made available to yield-hungry investors in a low interest rate environment. We refer to this modified cash flow delivery scheme as an “equity-for-debt” transformation process. At the same time, front-end fee loads for Unlisted REIT investment approach 14 percent, with ongoing operating fees further eroding cash available for investment and capital gains available for distribution upon the realization of a (highly uncertain) liquidity event. The relative success of Unlisted REITs is somewhat puzzling, however, given the high cost of investment and the close liquid alternative of Listed REITs.1 More specifically, the combined Listed and Unlisted equity REIT industry has now surpassed the $1 trillion mark for assets under management, where proceeds from new equity offerings by Unlisted REITs have outpaced those by Listed REITs every year over the 2006 to 2013 time frame. While the Listed REIT raises new equity through an initial public offering (IPO) – an event that occurs on a single date – the Unlisted REIT offers share subscriptions through independent broker-dealer networks in a continuous “blind pool” offering process. This equity raise channel is remarkably effective, with the average Unlisted REIT raising $150 million per quarter by Q4 2013. Fees paid to the investment advisor, which do vary but are generally in the range of 6 to 7 percent, essentially make Unlisted REITs the highest commission product available to offer their clients. The Unlisted REIT sector provides a unique setting to examine issues associated with private equity investment available to low-to-moderately wealthy retail investors—what we call “private equity for the common man.” While a large degree of heterogeneity exists among the assets and income sources for most private equity funds, REITs largely invest in income- 1 For both Listed and Unlisted REITs, we are referring specifically to public firms which register their securities and submit periodic filings to the Securities and Exchange Commission (SEC), as opposed to private REITs. 3 producing commercial real estate. The lengthy economic life and functioning secondary market for the underlying assets generates consistent patterns among fund cycles. The minimum investment in Unlisted REITs is $2,500. “High net worth” investors are qualified by either minimum net worth of $250,000 or minimum annual gross income of at least $70,000 plus a minimum net worth of $70,000. These investor suitability standards constitute strikingly low minimum thresholds when compared to restrictive criteria applied by most private equity and hedge funds. There are two main contributions of this study toward the understanding of private equity for retail investors. First, we collect detailed management compensation fee data along with key macroeconomic measures, and provide estimations for the determinants of common equity subscription flow at the fund level. Second, we introduce a novel methodology for calculating investor returns to date (as of Q4 2013) based on an Imputed share price if the security were liquid and valued as a going-concern. The Imputed share price is based on a ratio of total capitalization to book value of assets for comparable Listed REITs selected in a propensity-score matching. Holding period returns calculated as an IRR for the Unlisted REITs are then compared to returns over consistent horizons for Listed REITs. We show that investment flows into the Unlisted REIT vehicle vary directly with fee percentages paid to the investment advisor, and inversely with interest rates and dividend yields on close and more liquid investment substitutes. Realized holding period returns are calculated, where return differentials are shown to exceed 9 percent per year on average in favor of the Listed REIT. In comparison to the Unlisted REIT, the Listed REIT provides superior liquidity in its share price, greater transparency, and a lower risk asset investment structure, but with greater observable (but not actual) share price volatility. In short the Listed REIT dominates the Unlisted 4 REIT alternative along just about every dimension of preferred investment characteristic, where inferior Unlisted REIT investment performance differences illustrate the deleterious effects of high fees and a conflicted management structure on realized investment return outcomes. We wonder why informed investors would accept such low returns for a superficial equity-for-debt cash flow transformation, which could in fact be replicated at home by the investor. We consider an “ignorance is bliss”/”tree in the forest”/”head in the sand” rationale to explain this behavior in a world with high volatility-risk averse investors. But with greater than a 9 percent return difference it is hard to believe that a superficial reduction in share price volatility could be worth that much. The more likely explanation is effective marketing and promotion by the Listed REITs and their investment advisors, who, with the prospect of harvesting outrageously high fees, likely mislead or otherwise fail to properly disclose significant investment risks and fee structure. It is not a stretch to conclude that Unlisted REITs are a Ponzi structure that will eventually collapse under the weight of its high fee and conflicted management structure. Importantly, this investment vehicle has thrived in a low interest rate economy in which many volatility-risk averse investors exist that are desperate for yield. We conclude that the Unlisted REIT investment vehicle—private equity for the common man—is a Ponzi-like investment scheme that exists within the confines of the regulated financial investment environment. This is in contrast to the usual characterization of Ponzi schemes as idiosyncratic creations of narcissistic anti-social amoral individuals working alone or almost alone outside the reach of the regulatory community. The remainder of this study is organized as follows. Section II provides background information for the Unlisted REIT business model. Section III covers data, methods and results 5 for the equity issuance analysis. In Section IV, the propensity-score matching and Imputed share price methodology are explained, along with a comparison between firm characteristics in the Listed and Unlisted REIT sample. The IRR calculations are presented and discussed. Section V summarizes the behavioral and policy implications that can be drawn from this study. 6 II. Background and the Basic Business Model A primary contribution of this paper is our analysis of dual investment vehicles for a particular type of asset: commercial real estate. The vehicles are the listed or public equity REIT (L-REIT) and the unlisted or private REIT (UL-REIT), where the UL-REIT vehicle can be characterized as private equity. Both vehicles must follow a specified set of rules laid out by regulation in order to avoid taxation at the entity level, including minimum requirements on assets invested in real estate, income sources, investor holdings and dividend payouts. The L-REIT has been studied extensively, so we will not focus on the details in this section.2 The UL-REIT has been studied much less extensively, so we will provide additional background and a description of the basic business model. Other than the exchange-listed versus non-listed nature of equity share liquidity, the most important differences between investment vehicles are that L-REITs are internally-managed going concerns at the time of investment with standard fee structures, management costs and corporate governance that are typical of a public company, whereas the UL-REIT is an externally-managed blind-pool private offering with an offering period can remain open for several years. As we will discuss in some detail shortly, fees associated with investment in the UL-REIT are substantially higher than those associated with the L-REITS. External management and the high fee structure, along with weak corporate governance, produce firms that are rife with conflicts of interest between outside equity investors and managers. At the time of organization, both L-REITs and UL-REITs will submit an S-11 filing with the SEC. During the period from January 1994 to May 2014, there were 981 such S-11 filings 2 For surveys of the literature on L-REITs see Corgel, McIntosh and Ott (1995) and Chan, Erickson, and Wang (2002). 7 submitted to the SEC. Not all these filings mention an intent to qualify as a REIT, however. Among the subset are 699 S-11 filings that state an intention to qualify as a REIT. Of this subset, 429 described the intent to have shares listed on an organized stock exchange immediately following the IPO, while 270 of the REIT filings explicitly state that they do not intend to apply for exchange listing and they will utilize a continuous offering process. The former describes the L-REIT sample, while the latter identifies the UL-REIT sample. Both groups are subject SEC, NASD and NASAA reporting requirements, as well as compliance with all tax-related REIT restrictions as noted earlier. Rather than issuing shares through an IPO on a single-day exchange event, new equity shares for UL-REITs are sold through traditional broker-dealer/investment advisor channels using a continuous offering process. As Corgel and Gibson (2008) describe: “During the early 1990s, real estate fund sponsors in the US began raising equity for commercial real estate investment using broker-dealer channels. Replicating dividend networks established by mutual fund groups, sponsors offer real estate investment programs to the population of mostly smallunit investors who rely on the services of financial planners” (p.132-133). A nominal share price (typically set at $10 or $11) is held constant during the offering process. Corgel and Gibson (2008) present a model to demonstrate that the fixed share price and continuous offering provide an opportunity for follow-on investors to observe performance before investing. Under a more traditional IPO, strong demand for shares following the IPO would benefit early investors by the bidding up of share prices, creating capital gains. In the ULREIT space these benefits, at least in the short and medium term, are unrealized due to the lack of share liquidity and price discovery, as well as the steep up-front loads paid to the brokerdealers and sponsors. 8 A “blind pool” offering is the industry standard, meaning that there are no assets under management at the time of the initial offering. Lacking assets, the fund is externally managed by the Advisor at inception. It is generally disclosed that significant conflicts of interest exist due to the Advisor’s role as a fiduciary for other real estate funds. More specifically, the unobservable non-traded market share price, blind pool offering and external management intensify agency problems for UL-REITs relative to their exchange-listed counterparts. Anti-takeover provisions are commonplace, and include staggered board of director elections, expansive boards, poison pills and high fees to the external advisor should management be internalized. Exacerbating agency issues is the general void of participation by institutional investors in UL-REIT offerings, which limits monitoring, analyst coverage and shareholder influence. Corgel and Gibson (2008) outline the marketing approach to retail investors: “The chain that connects REITs to investors… begins with the establishment of an in-house broker-dealer affiliate that develops contractual relationships with national and regional managing dealers (e.g., A.G. Edwards, UBS Warburg, and Paine Weber). Local financial planners acting as soliciting dealers with or without affiliations to these national and regional broker-dealers then sell investment programs of unlisted REITs to the public” (p.134). UL-REITs organize with finite-life and mention in their offering documents target liquidation dates (typically within 6-8 years). Targeted liquidation dates are non-binding and loosely worded. Illiquidity risk is an important issue in this environment, and UL-REITs almost uniformly attempt to address investor concerns through inclusion of a share redemption plan. The most common UL-REIT offering price is at $10 per share, and the most common share redemption plan allows investors to return their shares to the firm for $9.50 per share. However, the fine print of these share redemption plans reveals that they may not open the program until 9 such time that there is adequate subscription to the dividend reinvestment plan (which allows investors to elect for stock dividends at $9.50 per share in lieu of cash dividends). In addition, once the redemption program has opened, it may be suspended or cancelled at any point subject to managerial discretion. The finite-life structure implies that the continuous offering process will ultimately end at some point in time. Wiley (2013) provides evidence that roughly one-third of UL-REITs had never opened their share redemption program to investors by the end of 2011. Of those that did open, half of the open redemption programs ultimately became suspended, cancelled, recast at a lower price or constrained due to excessive redemption request volume. Wiley (2013) further notes that these adverse changes are often closely timed with the end of the equity offering. When the UL-REIT closes the gate on the share redemption program, investors are effectively locked-in and must wait until management can accomplish a comprehensive liquidity event.3 Wiley (2013) shows that the most common form of liquidity event is acquisition by another UL-REIT (typically from the same Sponsor). Other options include exchange-listing, individual asset sales in the private real estate market and bankruptcy. Marketing of equity interests occurs as follows. Brokers/investment advisors/financial planners counsel potential retail investors that commercial real estate generates stable cash flows over time which generally move with inflation. This implies stable and gradually-increasing asset values that offer something of an inflation hedge. As a result of the bond-like structure of fundamental cash flows, share prices are suggested to experience little volatility in the short run. At the same time, the investment offers a high dividend yield, typically 6 to 7 percent based on 3 There are two exceptions to the investor lock-in: (1) online auctions for UL-REIT shares, and (2) tender offers that appear (not coincidentally) immediately after the share redemption program is suspended. In both cases, the share prices are deeply discounted (relative to the initial investment price) due to the thinness of buyers who participate in the online auctions, and the reality that those tender offers are made by vulture capital funds seeking high investor returns. 10 the “constant” posted share price (gross of up-front fees).4 Given little or no observed volatility in asset values and share prices combined with high and stable dividend payouts, investment takes on the character of a high-yielding medium-duration bond. Retail investors with only moderate wealth and income requirements are the primary if not exclusive investor clientele, where it is thought that retirees on fixed incomes are the largest sub-investor clientele. Promotional materials and verbal promises of dividend payouts of 6 to 7 percent with stable and increasing share prices that provide inflation protection is clearly an attractive product in a low interest rate environment. These returns noted above are quoted prior to inclusion of transaction costs and fees. Front-end loads on these investments are remarkably high (typically 14 percent), with an initial booked share price that is net of the up-front fees (say $8.60 per share based on a posted share price of $10 per share). It is not clear whether the transaction costs and the net booked investment amount are verbally disclosed by the marketing agent, where the posted $10 per share price creates the illusion of the “true” or fundamental value of investment. The $10 per share price thus creates a mental anchor for the investor, who mentally sets $10 per share as a lower bound on value that is expected to be realized at the time of fund liquidation (analogous to the full repayment of the principal investment amount on a bond). From the time of initial investment to liquidation, investors are further comforted by the stated redemption policy in which they are told they can simply sell their shares back (typically at $9.50 per share) if there is any unanticipated demand for liquidity by the investor. As noted earlier, however, liquidity realized through the share redemption programs is largely illusory. 4 See Seguin (2012). 11 In reality early dividends paid out to incumbent shareholders are usually funded through new capital raised – analogous to a Ponzi scheme. Those early dividends are at elevated levels and prove highly effective at attracting additional new equity from investors (see Wiley, 2014), but these elevated dividends also widen the gap on the firm’s ability to return the initial equity investment by the point of liquidation. Fund behavior changes abruptly following the close of the offering period. Dividend cuts become more common and dividend volatility increases drastically (Wiley, 2014). Share redemption programs are significantly more likely to become suspended or cancelled (Wiley, 2013). Investors are then often left at the mercy of management to accomplish a meaningful liquidity event. III. Net Equity Proceeds III.a. Data and Mechanics This study focuses on the capital raise process for UL-REITs. As noted previously, equity offerings are predominantly “blind pool” offerings, meaning that the fund has no assets under management at conception. This is different from most other comparable investment vehicles, which generally attract capital as seasoned enterprises (e.g., mutual funds, hedge funds, SEOs by L-REITs) or disclose a list of existing assets in place at the time of the initial offering (e.g., many types of private equity, IPO by L-REIT). Venture capital and some private equity funds are, in contrast, blind pool investments (Gompers and Lerner, 1999). As a blind pool, and thus in lieu of observing any existing performance, the only public information at the beginning of the offering can be obtained from the S-11 filing – the securities registration form for real estate companies submitted to the SEC. Of interest, the S-11 form includes content for management compensation structures and associated fees, among other 12 information. To define the sample, we use the SNL Financial database under Real Estate, REIT & Property Company, selecting all Non-Listed companies in United States and Canada. In April 2014, at the time of data collection, the query yields 173 UL-REITs, of which 57 are deleted because there is no financial information available for these firms in SNL. Panel A of Table 1 provides summary statistics for the fee structures of UL-REITs. Taking modes, the most common front-end loads include a 7 percent commission to the investment adviser who sells their client UL-REIT shares, 3 percent to the dealer-manager to cover administrative and marketing costs associated with distributing the shares, 2 percent acquisition fees for all properties acquired (paid up-front),5 0.5 percent reimbursement for acquisition expenses, and 1.5 percent reimbursement for organization and offering expenses. Thus, these numbers imply front-end loads of 14 percent. This in turn means that, for every $1 of shareholder equity contributed, the typical UL-REIT has about $0.86 available to invest. By comparison, of funds that have front-end loads, the average front-end load in the mutual fund industry is 3 percent (Gil-Bazo and Ruiz-Verdú, 2009). But many mutual funds have no frontend load whatsoever. Taking averages across all mutual funds, the dealer-manager fee is 9 basis points and the total front-end load is 52 basis points (Adams, Mansi and Nishikawa, 2012). In private equity, the typical front-end load is 2 percent, after a 1 to 2 percent transaction fee is levied for acquisitions (Metrick and Yasuda, 2010). Hedge funds typically do not have front-end loads and use promotes and hurdles instead to incentivize management. Once the Unlisted REIT brings in sufficient funds to begin to acquire assets, it moves into the operational stage (while still marketing shares during its offering period). During this stage, assets will have been acquired with management collecting fees on at least an annual basis. The 5 For exchange-listed REITs in Singapore, acquisition fees are almost always 1% (Ooi, 2009). 13 most common operational fees include 1 percent asset management fee, 4.5 percent property management fee, and 1 percent financing coordination fee. Comparing to mutual fund expense ratios, the asset management fee line averages 0.24 percent (Adams, Mansi and Nishikawa, 2012), although real estate mutual funds generally levy higher management fees, averaging 0.8 percent (Philpot and Peterson, 2006). Average asset management fees for hedge funds are 1.52 percent (Cassar and Gerakos, 2010). In private equity, asset management fees are typically 2 percent (Metrick and Yasuda, 2010). Other Unlisted REIT operational fees provide for construction and development fees (typically 5 percent of costs) or oversight fees for third-party property management (typically 1 percent of gross property revenues). These fees are somewhat less consistent in the sample, however. By comparison, private equity funds typically assess 2 percent monitoring fees for externally managed investments – where applicable (Metrick and Yasuda, 2010). In a few cases, UL-REIT management participates in all distributions made to shareholders, with fees that range from 0.7 to 10 percent. In other cases, management earns incentive compensation ranging between 0.5 and 35 percent when shareholder returns have exceeded a specified return hurdle. At the liquidation stage there are incentives for management to exceed return hurdles once a comprehensive liquidity event is achieved. The liquidity event can take the form of an exchange listing (or an IPO if new shares are also issued), a merger with an existing firm, or asset liquidation in the private market (including bankruptcy, when administered by the court). The most common structure for liquidation incentives is subordinate participation after shareholders realize a stated return hurdle. For example, the mode return hurdle is an 8 percent shareholder return per year, with managers receiving 15 percent of all cash available for distribution in excess of the amount required to meet the return hurdle. These incentives compare 14 to an average 19.34 percent participation with zero return hurdle for hedge funds (Cassar and Gerakos, 2010), or to the typical 20 percent participation rate after an 8 percent hurdle in private equity (Metrick and Yasuda, 2010). In the few cases where subordinate participation is not used on UL-REITs, the Sponsor holds preferred shares that are convertible to common equity based on a sliding scale which is linked to shareholder performance. Management is also entitled to a disposition fee, which is typically 3 percent of the contract sales price for every asset sold. Hartzell, Kim, Kimbrell and Sprow (2013) provide IRR calculations for Unlisted REITs that achieved full-cycle events. Merging their calculations with our data on initial hurdle rates, we find that less than one-half of UL-REITs are able to achieve liquidation values that exceed the stated return hurdle. Thus, performance incentives appear less relevant than they would be for hedge fund and private equity (as shown in Phalippou and Gottschalg, 2009). Instead, it would appear the dominant incentive with UL-REITs is to maximize equity subscription to produce high dealer and acquisition fees on the front-end. Those front-end fees are then replaced by operating fees, including high asset management and property management fees – thereby generating annual income streams for management, which becomes a major disincentive for liquidation. Panel B of Table 1 describes offering outcomes. The average UL-REIT raises over $76 million per quarter during a continuous offering process that runs nearly four years (11.7 quarters). There are a number of Repeat Sponsors of UL-REIT funds which produce multiple offerings over time. More specifically, Repeat Sponsors typically generate one highly similar offering following another with the closure its immediate predecessor. These Repeat Sponsors are dominant in equity fundraising, averaging over $91 million per quarter, as compared to $20 15 million per quarter for One-off Sponsors (defined as completing an equity offering only once in the SNL sample). Figure 1 illustrates the historic development of the Unlisted REIT equity formation process. The gray portion of each bar represents net equity proceeds from Repeat Sponsors, while the black portion represents the same for One-off Sponsors. In cumulative amounts, Oneoff Sponsors have raised $3.4 billion of the total $69 billion raised by the sector thus far – less than 5 percent of the total equity capital raised. Also shown in Figure 1 are the Fed Funds rate and the NAREIT dividend yield (a weighted-average dividend yield of Listed REITs). Equity proceeds for the UL-REIT sector appear much higher in recent years, corresponding with historic lows in both the Fed Funds rate and NAREIT dividend yield. Figure 2 shows equity fundraising for the Unlisted REIT sector in context of equity raised by Listed REITs, including through IPOs and SEOs. UL-REITs are represented by the black line and appear to have exceeded the aggregate capital raised in L-REIT IPOs for nearly every quarter since 2005. However, SEOs by L-REITs have become much more frequent in recent years and represent the largest channel for equity fundraising by L-REITs. We note that a handful of L-REITs began as UL-REITs and later succeeded in executing an exchange-listing.6 Any IPO or SEO net proceeds that occur for these firms are counted for L-REITs upon the date of their exchange-listing. Figure 3 illustrates that with success comes competition. While the UL-REIT sector as a whole has experienced mounting success at equity fundraising, net proceeds for the average fund appear heavily influenced by the number of competing firms that have open capital offerings during the same quarter. Prior to 2007, there were seldom more than 10 UL-REITs competing 6 Examples include American Realty Capital Trust, CatchMark Timber Trust, Cole Real Estate Investments, Columbia Property Trust, DCT Industrial Trust, Healthcare Trust of America, Piedmont Office Realty Trust, Retail Properties of America, and Whitestone REIT. 16 for new equity in a given quarter. Since 2008, however, the number of UL-REITs with open equity offerings in a given quarter has fluctuated in the range of 15 to 35. Post-2008, the average net equity proceeds per fund appear inversely related to the number of competing offerings. The red line in Figure 3 depicts net flows into commercial mortgages from the Fed Flow of Funds Accounts. Commercial mortgage flows are correlated with commercial property price indices, but not highly correlated with L-REIT equity flows. In the REIT industry, underlying asset valuations in respective property markets can become somewhat disconnected from valuations in the stock market (e.g., see the series on premiums/discounts to NAV by GreenStreet Advisors). Thus, commercial mortgage flows provide a measure of the attractiveness of yield-based investments that are backed by commercial real estate, and are not necessarily closely dependent of stock market valuations of property. Panel C of Table 1 presents a selection of macroeconomic measures that will be evaluated for their impact on UL-REIT equity proceeds. The number of competing UL-REITs with open equity offerings is determined by examining the Net equity proceeds measure for each firm to distinguish periods of equity access from closed offerings. The Fed Funds rate, 3-Year Treasury, 10-Year Treasury and credit spreads are each extracted from the Selected Interest Rates (H.15) series from the Federal Reserve. Core inflation is from the Bureau of Labor Statistics (BLS), and includes the Consumer Price Index (CPI) for all items, less food and energy. Non-core inflation is the difference between CPI and Core inflation. Commercial mortgage flows and L-REIT equity flows are from the Fed’s Flow of Funds (Z.1) series. NAREIT total returns to L-REITs and NAREIT dividend yields are from the FTSE NAREIT US Real Estate Index series. 17 III.b. Estimation To model the Unlisted REIT equity raise process, net equity proceeds per firm-quarter are estimated for all periods when the respective firm has an open offering – i.e., one that it currently accepting new equity subscriptions. The estimation for net equity proceeds includes a set of RHS variables for firm characteristics and a set of variables for macroeconomic conditions during the offering period. Firm characteristics are drawn from the S-11 filing based on information observable at the time of initial offering and before the firm has any assets under management or performance record. ln(Net equity proceeds) = β0 + β1∙Broker commission + β2∙Dealer-manager fee + β3∙Acquisition fee + β4∙Asset management fee + β5∙Subordinate participation + β6∙One-off Sponsor + β7∙CEO age + β8∙Number of UL-REITs + β9∙Fed Funds rate + β10∙3-Year Treasury spread + β11∙10-Year Treasury spread + β12∙Baa-Aaa spread + β13∙Core inflation + β14∙Non-core inflation + β15∙Commercial mortgage flows + β16∙L-REIT equity flows + β17∙NAREIT total returns + β18∙NAREIT dividend yield + + Σβk∙Offering maturityk + ɛ. (1) Variables for Unlisted REIT characteristics include the Broker commission rate, Dealer-manager fees, Acquisition fees, Asset management fees, Subordinate participation, One-off Sponsor indicator, and CEO age at the time of offering. Aggregate-level measures include the Number of UL-REITs with active equity offerings, the Fed Fuds rate, the 3-Year Treasury spread (net of the Fed Funds rate), the 10-Year Treasury spread (net of the 3-Year Treasury rate), the Baa-Aaa credit spread, Core and Non-core inflation, Commercial mortgage flows, L-REIT equity flows, NAREIT total returns and dividend yields. Fixed effects for the number of quarters since offering inception are also included (as indicators variables over a maximum of 27 quarters). 18 Results from the estimation of Equation (1) are presented in Table 2. Broker commission rates and Dealer-manager fees have positive and significant impacts on equity subscription. Increasing the broker commission by 1 percent (e.g., from 6 to 7 percent) increases the quarterly offering proceeds by an estimated 57.3%.7 Increasing dealer-manager fees, which are largely spent on marketing and promotion for the offering, by 1 percent increases the quarterly equity raise by 55.3 percent.8 When average quarterly proceeds per offering are in the ballpark of $150 million at year-end 2013, the potential consequences from these fee adjustments are obviously quite significant. Mutual funds with front-end loads have been found to underperform no-load mutual funds (Carhart, 1997). In Barber, Odean and Zheng (2005), mutual fund investors are found to be sensitive to high front-end expenses, including brokerage commissions and acquisition fees, resulting in significant less funds flow. For UL-REITs, the opposite is true. Investor subscriptions are highest at funds with the highest front-end loads. As Golec (2003) points out, if equity flows fail to respond negatively to high front-end fees, it is likely due to the limited ability of investors to distinguish marketing efforts from performance. In any case, we are able to directly link front-end marketing fees to investor subscriptions. We are not aware of any other work that establishes this type of strong positive relationship between front-end load and funds flow. This in turn presents something of a new puzzle that is difficult to reconcile with existing rational or behavior-based investor theory. Investors are observed to display some sensitivity to ongoing Asset management fees. These fees are disclosed in the offering documents, and are likely highlighted by the investment advisor to investors prior to investment. Increasing asset management fees by 25 basis points 7 8 exp{45.31/100} – 1 = 57.3% exp{44/100} – 1 = 55.3% 19 (0.25 percent) reduces quarterly equity proceeds by 16.5 percent.9 One-off Sponsors indicate relatively unknown managers, and as such lack reputation in the UL-REIT space. Their success in raising new equity is reduced by 75 percent as a result, as compared to Repeat Sponsors.10 Competition from other UL-REITs reduces an individual firm’s equity proceeds in a given quarter. Adding one additional UL-REIT reduces net equity proceeds for incumbent funds by an estimated 2.8 percent.11 This number approximates an equal market share when there are 35 funds competing (1/35 = 2.86 percent). Success in equity offerings for Unlisted REITs is strongly enhanced by low interest rates. A one percent reduction in the Fed Funds rate increases net equity proceeds for the average ULREIT by 32 percent per quarter.12 A related result obtains for dividend yields on Listed REITs. The coefficient for NAREIT dividend yield is negative and significant, where a one percent reduction in the NAREIT dividend yield increases equity proceeds by 18 percent.13 We do not find any relationship between funds flow and realized inflation, but this may be the result of persistently low rates of inflation realized during the sample period. In summary, our main findings are that the rate of funds flow into the Unlisted REIT investment vehicle is sensitive to broker-dealer/investment manager load fees (positive relation), the level of interest rates as measured by the Fed Funds rate (negative relation), and commercial property fundamentals as proxied by debt flows into commercial property markets (positive relation). Recall the UL-REIT business model, which is presented to potential investors as a high-yield fixed-income alternative with underlying commercial real estate assets that promise to 9 exp{-72.14/400} – 1 = -16.6% exp{-1.38} – 1 = -74.9% 11 exp{-0.03} – 1 = -2.8% 12 exp{-38.57/100} – 1 = -32.0% 13 exp{-19.92/100} – 1 = -18.1% 10 20 generate stable cash flows that move with inflation, a 7 percent dividend yield, posted share prices that remain stable over time, a promise of a liquidity event in the next five years or so, and the stated option to liquidate early if necessary through a share redemption program. This investment as characterized is clearly attractive to certain retail investors in a low interest rate environment in which most fixed-income investment alternatives generate significantly lower promised returns over similar investment horizons. The extremely high fee structure for Unlisted REITs, together with the positive relation between investor advisor fees and the rate of offering funds flow, raises some concerns that retail investors may be overpaying for this high dividend yielding “fixed-income substitute” (that is, they apparently cannot distinguish between marketing effort and fund performance). Yet the Listed REIT exists as a direct equity investment substitute, with a much lower investment fee structure. The UL-REIT in comparison provides the illusion of reduced or even non-existent stock price volatility (at least until a liquidation event occurs) together with higher dividend yields (about 7 percent for UL-REITs versus 4 percent on average for L-REITs) that are also more stable over time. But at what price does this increased stability come (even though stability in the investment value is illusory)? And are the promises made to investors realized, in that they earn returns on average and after fees that compensate them for the risk of investment? This is the topic of the next section of the paper. IV. Unlisted REIT Holding Period Returns In this section we analyze the comparative investment performance of Listed and Unlisted REITs. In doing so we quantify the implied benefit to UL-REIT investors of “financial claim transformation”, in which a standard equity claim is superficially turned into a high 21 yielding debt claim. After quantifying this return differential, in the subsequent section we then ask whether the compensation appears reasonable or not. To the extent that the differential seems unreasonably high, we consider alternative explanations for the finding. To set the stage for analysis, consider first the standard equity claim as exemplified by the exchange-traded REIT share. The listed firm targets dividend payouts as a percentage of its share price and experiences stock price volatility that goes with being listed on an exchange. Shareholders can sell their stakes at any time into a (typically) liquid secondary market. The (pre-tax) holding period return realized for a representative L-REIT share depends on the price paid for the share, dividends received during the holding period, and the price at which the share is sold at the end of the holding period. In contrast, the Unlisted REIT posts a price for its shares that does not vary over time. This posted price is analogous to the par value of a bond. Up-front fees are subtracted from the posted price paid, where the remaining amount is available for investment. Dividend payout percentages are fixed as a percentage of the non-varying posted share price that is gross of fees, at least during the offering period (Wiley, 2014). After the offering period closes, dividends are often cut in recognition of the unsustainably high rate. The dividend payments promised at the time of the offering are analogous to interest payments that reference the bond’s par value, and are high relative to interest payments on standard debt claims. The high dividend yield seems to acknowledge that there are equity risks contained in the UL-REIT share claim that exceed those of a standard debt claim. At the same time the high payouts dissipate capital that could have been retained within the firm to put to productive use. Unlike the L-REIT, there is little true liquidity in the UL-REIT share claim prior to a liquidation event. And like a bond, a “maturity date” is stated in offering documents (typically five to eight years). But unlike a bond the maturity date is 22 a non-binding liquidity event that is left completely to the discretion of management. Investors are led to believe at the time of investment that they will receive a liquidation payoff that equals or exceeds the par value or posted share price of UL-REIT investment. In sum, this equity-to-debt claim transformation process takes cash flows and prices from a given set of assets and reconfigures them by increasing dividend payouts and replacing markto-market share pricing with a constant posted price that—net of fees—in most cases and for most of the time overstates true share value. The posted price thus creates a mental anchor for the UL-REIT investor, much like the par value of a bond, say at $10 per share, where a liquidity event presumably results in the realization of a $10 price per share or more.14 A holding period return for the UL-REIT can be calculated as with the L-REIT, where the terminal value of the UL-REIT share must be estimated if a liquidation event has not occurred prior to the end of the specified holding period.15 14 It is relevant to note that the apparent transformation of an equity claim into a debt claim does not involve increasing the priority of the claim in the event of financial distress. The shareholder remains the residual claimant in the UL-REIT investment structure. 15 The list of UL-REITs [quarter of first dividend] are American Realty Capital - Retail Centers of America, Inc. [2012Q2], American Realty Capital Daily Net Asset Value Trust, Inc. [2012Q1], American Realty Capital Global Trust, Inc. [2012Q4], American Realty Capital Healthcare Trust II, Inc. [2013Q2], American Realty Capital Healthcare Trust, Inc. [2011Q2], American Realty Capital New York Recovery REIT, Inc. [2010Q3], American Realty Capital Trust V, Inc. [2013Q2]. Apple Hospitality REIT, Inc. [2008Q3], Apple REIT Ten, Inc. [2011Q1], Behringer Harvard Multifamily REIT I, Inc. [2009Q3], Behringer Harvard Opportunity REIT I, Inc. [2006Q4], Behringer Harvard Opportunity REIT II, Inc. [2008Q4], Bluerock Residential Growth REIT, Inc. [2009Q4], Carey Watermark Investors Incorporated [2011Q2], Carter Validus Mission Critical REIT, Inc. [2011Q3], CNL Growth Properties, Inc. [2011Q4], CNL Healthcare Properties, Inc. [2012Q1], CNL Lifestyle Properties, Inc. [2005Q1], Cole Corporate Income Trust, Inc. [2011Q2], Cole Credit Property Trust IV, Inc. [2012Q2], Cole Real Estate Income Strategy (Daily NAV), Inc. [2011Q4], Corporate Property Associates 17 – Global Inc [2008Q2], Dividend Capital Diversified Property Fund Inc. [2006Q1], Global Income Trust, Inc. [2011Q2], Griffin Capital Essential Asset REIT, Inc. [2010Q1], Griffin-American Healthcare REIT II, Inc. [2010Q1], Hartman Short Term Income Properties XX, Inc. [2011Q1], Hines Global REIT, Inc. [2010Q2], Hines Real Estate Investment Trust, Inc. [2004Q4], Industrial Income Trust Inc. [2010Q3], Inland American Real Estate Trust, Inc. [2006Q1], Inland Diversified Real Estate Trust, Inc. [2009Q4], Inland Real Estate Income Trust, Inc. [2012Q4], Jones Lang LaSalle Income Property Trust, Inc. [2008Q1], KBS Legacy Partners Apartment REIT, Inc. [2011Q1], KBS Real Estate Investment Trust II, Inc. [2008Q3], KBS Real Estate Investment Trust III, Inc. [2011Q3], KBS Real Estate Investment Trust, Inc. [2008Q1], KBS Strategic Opportunity REIT, Inc. [2010Q4], Landmark Apartment Trust of America, Inc. [2007Q4], Lightstone Value Plus Real Estate Investment Trust [2010Q2], Lightstone Value Plus Real Estate Investment Trust [2006Q1], Moody National REIT I, Inc. [2010Q2], MVP REIT, Inc. [2012Q4], Phillips Edison – ARC Shopping Center REIT Inc. [2011Q1], Resource Real Estate Opportunity REIT, Inc. [2010Q3], 23 With this background we will now compare investment performance of our sample of Unlisted REITs that have not experienced a liquidity event as of Q4 2013 with the alternative investment of holding an index of Listed REIT stocks over an identical holding period. To do this comparison we calculate holding period returns as follows. Assume there are T+1 cash flows to analyze, Cash flow 0 is the initial investment on a per share basis (an outflow) that occurs one quarter prior to the first dividend payment made by the UL-REIT.16 For the UL-REIT, the initial share offering price (typically $10) is obtained from the S-11 filing. In contrast, at time 0 an investment is made into a basket of L-REIT stocks that serve as a comparable substitute investment for the UL-REIT, as proxied by the NAREIT index. Cash flows 1 through T-1 are dividend payments per share (inflows) actually made to investors. The dividend history is collected from the SNL database for each UL-REIT. Dividends are automatically reinvested into the index in the case of investment into the NAREIT index. Cash flow T is the last dividend payment plus the share liquidation or sales price. Since liquidation has not yet occurred by Q4 2013, we approximate the share price of the UL-REIT by applying the Market-to-Book ratio as of Q4 2013 from a comparable L-REIT. With respect to the basket of liquid L-REIT stocks, the NAREIT index is simply sold at its end-of-quarter market value at that time. Liquidation values of non-liquidated Unlisted REITs must be estimated because only asset book values are available in the data (the shares are non-traded and there is no other independent valuation of the firms asset values). To establish an appropriate match with the ULREIT to estimate its liquidation value, we first identify 246 Listed REITs in the SNL database with available information in Q4 2013. A simple linear regression of market value of total assets Sentio Healthcare Properties, Inc. [2009Q1], Signature Office REIT Inc. [2010Q4], Steadfast Income REIT, Inc. [2010Q3], Strategic Realty Trust, Inc. [2010Q1], Strategic Storage Trust, Inc. [2008Q3], Summit Healthcare REIT, Inc. [2006Q4], and TIER REIT, Inc. [2004Q1]. 16 This conservative assumption increases the holding period return to the UL-REIT investor, as early-stage investors will not realize a dividend payment for more than one quarter relative to the time of investment. 24 against book value of total assets for L-REITs yields an intercept of zero, slope of 1.32, and adjusted R2 of 84.7 percent. In these data the Market-to-Book ratios for L-REITs tend to fluctuate within a relatively narrow band—50 percent of observations are in the range of 1.0 and 1.4—over the sample period. As a basis of comparison, among the UL-REITs that have subsequently exchange-listed, Market-to-Book ratios range from 0.933 to 1.286 at the earliest point of available capitalization data from SNL.17 To establish a best match for the Unlisted REIT in question, we apply a propensity score matching approach, implemented as follows. First, a probit model is estimated for the combined sample of L-REITs and UL-REITs (the UL-REIT is indicated with a 1), where matching is based four firm characteristics as of Q4 2013: firm age, total assets, leverage ratio, type of assets held by property type. In particular, the model to be estimated is: Pr{UL-REIT = 1} = β0 + β1∙Firm age + β2∙Total assets + β3∙Leverage ratio + Σβj∙%Property typej + ɛ. (2) The property type percentage establishes the primary uses of commercial property held by a particular REIT, where the calculation method comes from Geltner and Kluger (1998) and Riddiough, Moriarty and Yeatman (2005). The probit estimation generates a propensity score for each REIT in the sample. Based on its propensity score, each UL-REIT is then one-to-one matched with the nearest-neighbor propensity score of a L-REIT. Then the corresponding Market-to-Book ratio of the L-REIT is collected, where Market value equals the market value of 17 Those Market-to-Book ratios are American Realty Capital Trust = 1.218 (Q2 2012), CatchMark Timber Trust = 1.117 (Q4 2013), Cole Real Estate Investments = 1.153 (Q2 2013), Columbia Property Trust = 1.03 (Q4 2013), DCT Industrial Trust = 1.183 (Q4 2006), Healthcare Trust of America = 1.286 (Q2 2012), Piedmont Office Realty Trust = 1.091 (Q1 2010), Retail Properties of America = 0.933 (Q2 2012), and Whitestone REIT = 0.965 (Q3 2010). 25 all equity shares (inclusive of UPREIT ownership units) as well as the book value of all other reported financial claims. Book value is calculated net of depreciation and amortization. Table 3 compares summary statistics of the matching characteristics for the sample of Unlisted REITs, the full sample of Listed REITs prior to matching (Pre-Match), and for the matched sample of Listed REITs (Post-Match). UL-REITs have very similar property type configurations to L-REITs overall, even though individual firms in both categories often elect to focus investment on a single property type. UL-REITs tend to be younger than the full sample of L-REITs (average age: 4.1 years versus 10.8 years), where the age gap in large part can be attributed to the finite life structure of an UL-REIT that is under pressure to accomplish a liquidity event (which in turn causes the firm to exit the sample as a liquidated firm). No similar pressure exists for L-REITs. UL-REITs are considerably smaller (at just under $1.3 billion in assets under management on average) than the typical L-REIT in the full sample (at over $4.3 billion on average). There are not large differences between UL-REITs and L-REITs based on average leverage. Finally, and not surprisingly, the Post-Match sample of L-REITs possess matching characteristics that are on average much closer to comparable UL-REIT averages than the full sample of L-REITs. Table 4 presents estimation results from the probit model of Unlisted REITs versus Listed REITs. The estimation confirms discrepancies noted from the summary statistics. ULREITs are significantly younger, smaller and use somewhat less leverage. Following the outcome of the propensity-score matching, the sample of L-REITs have similar firm age, size and leverage ratios to the UL-REITs. The average Market-to-Book ratio of the matched Listed REIT is 1.1—somewhat lower than those of the full sample average. This outcome is not unexpected, since book values are 26 often closer to market values for young firms (before non-cash expense items like depreciation have more pronounced effects), where the matched sample of L-REITs is much younger than the unmatched sample. The Market-to-Book ratios collected for the matched L-REITs are then applied to the UL-REIT sample to calculate the Imputed share price in the following way: Imputed share price = (Book value of total assets)(M/B ratio†) – total liabilities – preferred equity – OP units) # of common equity shares outstanding (3) All values are taken as of year-end 2013, where again M/B ratio† is the market-to-book ratio of the nearest-neighbor L-REIT drawn from a one-to-one propensity matching procedure. Results from the Imputed share price calculations are summarized in Panel A of Table 5. The average Imputed share price for UL-REITs at year-end 2013 is $8.97, down from the average offering price of $10.93 per share by 18 percent. The largest negative price difference in the data is a decline of 83.2 percent from the offering price. This UL-REIT had very limited success in raising new equity, failed to realize its investment objectives, and has not paid a dividend to shareholders in several years. In some cases share prices appear to have increased substantially, particularly for firms that were created in late-2008 and 2009, having added assets whose book values (entered at historic cost) were low when commercial real estate prices were temporarily severely depressed. With Imputed share prices providing liquidation values as of Q4 2013, holding period returns (IRRs) can be calculated for the sample of Unlisted REITs together with the matched 27 Listed REITs.18 The UL-REIT IRR calculations are summarized in Panel A of Table 5. The average annual holding period return for UL-REITs is -0.7 percent and the median is -1.3 percent. In comparison, returns to holding a basket of L-REITs through the NAREIT index over identical time horizons average 8.4 percent. The difference between the average UL-REIT and the average NAREIT benchmark return provides a point-estimate of -9.1 percent. These numbers are not adjusted for differences in liquidity or other investment risks. Figure 4 plots the Unlisted -REIT IRRs against returns from the Listed REIT NAREIT index benchmark. UL-REITs that under or outperform the L-REIT benchmark over the respective investment time horizon are separated by the 45º diagonal. Of the 53 UL-REITs for which IRR calculations are available, nine are above the line (indicating outperformance of the NAREIT index on a holding period return basis). The remaining 44 UL-REIT IRRs fall below the NAREIT series, indicating negative return differentials. Panel B of Table 5 provides results for a simple regression of UL-REIT IRRs against the NAREIT benchmark returns. The alpha from this simple regression model is estimated to be -7.4 percent for UL-REITs. To provide a sense of the cash flow magnitudes realized from an investment into a representative Unlisted REIT, consider the following example. The average representative ULREIT returns just under zero percent, meaning that the initial investment amount is returned over the holding period in dividends and liquidation value with roughly no return on investment. Given a holding period of five years (about the average in our data), a $10/share posted share price, and a dividend yield of 7 percent, an IRR of zero follows from the following stream of 18 To assess the robustness of our liquidation value estimation methodology, we estimate the Probit model for Unlisted REITs that subsequently exchange listed, using their final non-listed quarter financial information. With this information we obtain propensity scores, match with nearest-neighbor Listed REIT for Market-to-Book ratio and obtain the Imputed share price. Those Imputed share prices are all within 3 percentage points of the IPO or tender offer price for the UL-REITs that achieved full-cycle liquidation events, using values reported in Hartzell, Kim, Kimbrell and Sprow (2012). 28 cash flows, with cash flows stated on an annual basis: CF0 = −10; CF1-CF4 = .70; CF5 = 6.5+.70. That is, with metrics that are typical of the data we analyze, we find that the representative ULREIT returns only about 65 percent of initial invested capital at the 5-year holding period mark. Analysis is over a time frame in which commercial property prices have on average been flat or have increased. Where did the UL-REIT’s invested shareholder capital go? Assumed no increase in values of real investment into commercial real estate assets and enough cash flow from investment to fund the dividend payments, it went to pay up-front fees of about 14 percent with another 21 percent going to pay other operating fees. To the extent the operating cash flows were insufficient to fund dividend payments, the 21 percent attributed to operating fees is in part a “fee” to fund dividend payments. V. Assessment, Analysis and Some Preliminary Conclusions The holding period returns calculated above are not risk-adjusted for liquidity or other potentially relevant differences. Adjusting for liquidity would increase the return differential to well more than 9.1% on average. This adjusted difference would increase further if we accounted for fees associated with a UL-REIT liquidation event, which as noted can be substantial. In contrast, the only fees we ignore with the Listed REIT investment alternative are standard broker fees associated with purchase and sale of index shares—costs which are so low as to be almost trivial. We also believe we have properly controlled for investment risk, as we have matched firms based on size, asset type (in the percent assets held variables) and leverage. Thus we don’t believe that the large holding period return differentials can be explained by Unlisted REIT share 29 ownership being a fundamentally lower risk proposition than investment in the Listed REIT share counterpart. Finally, also note that Unlisted REIT investment is into a blind pool, with invulnerable feedriven external managers that are particularly susceptible to conflicts. The Listed REIT investment alternative is, in contrast, an existing pool of internally managed assets with good-toexcellent transparency, generally high-quality corporate governance and well-incentivized managers—all of which spells significantly higher relative risk of investment with the UL-REIT vehicle. Note also that a representative investor could accomplish the same equity-for-debt transformation at home through investment in the liquid Listed REIT vehicle, as follows. After the initial investment in the L-REIT vehicle, whatever shortfall (or overage, which could occur later) in the realized periodic dividend payouts relative to the UL-REIT (to be used to achieve a desired level of consumption or to fund alternative investments) could be funded from the partial sale of L-REIT stock. If over the holding period consistent shortfalls are realized in dividend payments, capital gains to investment in the L-REIT vehicle would be reduced. But on average exactly the same total investment return would be realized with this homemade investment strategy, with the same differential in relative realized returns, at an identical level of risk, where in the process the representative investor achieves the desired level of period cash payout to fund consumption or alternative investments in a substantially better managed and more liquid investment vehicle. The only thing missing from the above scenario is that by holding the Listed REIT shares the investor experiences mark-to-market share price volatility. In contrast, there is no apparent share price volatility in the posted UL-REIT share price, a completely superficial construct 30 which the UL-REIT investor appears to value quite highly. It would, however, seem implausible that informed UL-REIT investors value this superficial equity-for-debt claim transformation at more than 9 percent per year on average. Our findings thus suggest something of an anomaly or puzzle, which could variously be described as the “ignorance is bliss”/”tree in the forest”/”head in the sand” fallacy of excess share price volatility reduction. Investors would appear particularly vulnerable to this effect in the period leading up to the financial crisis and the post-crisis period, where understandably many investors became rather averse of market volatility. This identified investment behavior also illustrates apparent time inconsistency in investor behavior, in which UL-REIT investors surrender investment liquidity and back-end load their investment risk in hopes of a successful liquidity event. The bargained gain appears to be the superficial elimination of share price volatility created by the mental anchoring of liquidation value to the share’s posted price. The sheer magnitude of fees associated with Unlisted REIT investment together with the related massive underperformance of these firms relative to the their nearly identical exchangetraded cousins lead us to conclude that most UL-REIT investors are less than fully informed. Instead, evidence points to the previously discussed issue of an inability to distinguish between marketing efforts from investment performance. The direct and economically meaningful relationship between brokerage fee percentages (apparently the highest of any investment alternative available through registered investment advisors) and UL-REIT investment funds flow cannot be ignored, suggesting to us very close monitoring of this private equity investment vehicle by the regulatory community as well as perhaps changing the net worth and annual income thresholds required for investment. Indeed, the Unlisted REIT investment vehicle can be characterized as private equity investment for the common man, where it appears to us that the 31 common man is typically less than fully informed about fee structure and may not be able to comfortably bear the high risk of loss. It furthermore does not seem a stretch to argue that Unlisted REITs are a Ponzi scheme that will collapse in the not-too-distant future under the weight of its fee structure. Importantly, the emergence of this Ponzi structure must be considered in the context of the broader economic environment and monetary policies of the Federal Reserve. That is, strong funds inflow to these Ponzi-like investment vehicles is strongly inversely related to Fed Funds interest rates and dividend yields on the Listed REIT alternative. Many investors, particularly retirees and others living on a fixed income, are not only volatility-risk averse but are desperate for yield. Desperation unfortunately produces strong neurotic behavioral tendencies, making investors particularly susceptible to promotional schemes that in the longer run produce unfortunate but (in hindsight) rather predictable outcomes. It follows therefore that these Ponzi-like investment vehicles are a direct consequence of the current economic environment. We consider this connection to be an important contribution of this paper, in that Ponzi schemes are typically identified as idiosyncratic edifices of narcissistic anti-social amoral individuals working alone or almost alone outside the reach of regulators. We instead show that the emergence of Ponzi schemes can be much more systematic—a product of the current economic environment—existing within a prescribed institutional framework as the intended consequence of industry specialists that know how to package and promote deeply flawed investment structures that in the process enrich themselves and agents along the supply chain. Finally, we note that apparent lack of concern by financial regulators about the emergence of these vehicles may be because they are deemed as “not systemically important.” We are not so 32 convinced. What we now well know is that the financial system is highly interconnected, where systemically important financial entities regularly interact with, or are only thinly separated from, less systemically important but nevertheless high risk entities. And further note that high leverage and/or deployment of repo debt funding are not lone red flags—vehicle management, investment and fee structure are also relevant considerations, particularly when the assets held by firms are real estate. 33 References Adams, J.C., Mansi, S.A., & Nishikawa, T. (2012). Are mutual fund fees excessive? Journal of Banking & Finance, 36(8), 2245–2259. Barber, B.M., Odean, T., & Zheng, L. (2005). Out of sight, out of mind: the effects of expenses on mutual fund flows. Journal of Business, 78(6), 2095–2119. Carhart, M.M. (1997). On persistence in mutual fund performance. Journal of Finance, 52(1), 57–82. Cassar, G., & Gerakos, J. (2010). Determinants of hedge fund internal controls and fees. Accounting Review, 85(6), 1887–1919. Chan, S.H., Erickson, J., & Wang, K. (2002). Real Estate Investment Trusts: Structure, Performance, and Investment Opportunities. Oxford University Press. Corgel, J.B., & Gibson, S. (2008). Real estate private equity: The case of US unlisted REITs. Journal of Property Investment & Finance, 26(2), 132–150. Corgel, J.B., McIntosh, W., & Ott, S.H. (1995). Real estate investment trusts: A review of the financial economics literature. Journal Real Estate Literature, 3(1), 13–43. Geltner, D., & Kluger, B. (1998). REIT-based pure-play portfolios: The case of property types. Real Estate Economics, 26(4), 581–612. Gil-Bazo, J., & Ruiz-Verdú, P. (2009). The relation between price and performance in the mutual fund industry. Journal of Finance, 65(5), 2153–2183. Golec, J. (2003). Regulation and the rise in asset-based mutual fund management fees. Journal of Financial Research, 26(1), 19–30. Gompers, P., & Lerner, J. (1999). An analysis of compensation in the U.S. venture capital partnership. Journal of Financial Economics, 51(1), 3–44. Hartzell, J.C., Kim, J., Kimbrell, V., & Sprow, J. (2012). Returns in the nontraded REIT industry: Evidence from full-cycle events. Working paper, University of Texas. Accessed at: http://blogs.mccombs.utexas.edu/realestate/files/2012/10/NT-REIT-White-Paper-forDistribution-zkcnpx.pdf on June 1, 2014. Metrick, A., & Yasuda, A. (2010). The economics of private equity funds. Review of Financial Studies, 23(6), 2303–2341. Ooi, J.T.L. (2009). The compensation structure of REIT managers: impact on stock valuation and performance. Journal of Property Research, 26(4), 309–328. 34 Phalippou, L., & Gottschalg, O. (2009). The performance of private equity funds. Review of Financial Studies, 22(4), 1747–1776. Philpot, J., & Peterson, C.A. (2006). Manager characteristics and real estate mutual fund returns, risk and fees. Managerial Finance, 32(12), 988–996. Riddiough, T.J., Moriarty, M., & Yeatman, P.J. (2005). Privately versus publicly held asset investment performance. Real Estate Economics, 33(1), 121–146. Seguin, P. (2012). Relative value of public non-listed REITs: An evaluation of equity return sustainability, liquidity, and organizational issues. Working paper. Presented at American Real Estate Society, Annual Meeting in St. Pete Beach, Florida on April 20, 2012. Wiley, J.A. (2013). Illiquidity risk in non-listed funds: Evidence from REIT fund exits and redemption suspensions. Journal of Real Estate Finance and Economics, forthcoming. DOI: 10.1007/s11146-013-9422-3. Wiley, J.A. (2014). Dividend manipulation at non-listed REITs. Working paper. Presented at 2014 AREUEA International Conference in Reading, United Kingdom on July 10, 2014. 35 One-off Sponsors Fed Funds rate NAREIT dividend yield $4,000,000 16.0% $3,500,000 14.0% $3,000,000 12.0% $2,500,000 10.0% $2,000,000 8.0% $1,500,000 6.0% $1,000,000 4.0% $500,000 2.0% $0 0.0% Yield Repeat Sponsors 1993Q1 1993Q4 1994Q3 1995Q2 1996Q1 1996Q4 1997Q3 1998Q2 1999Q1 1999Q4 2000Q3 2001Q2 2002Q1 2002Q4 2003Q3 2004Q2 2005Q1 2005Q4 2006Q3 2007Q2 2008Q1 2008Q4 2009Q3 2010Q2 2011Q1 2011Q4 2012Q3 2013Q2 Net Equity Proceeds (in $Thousands) Figure 1. UL-REIT Sector, Net Equity Proceeds Quarter Notes: Figure 1 displays the industry-aggregated proceeds from new equity issues per quarter (in $Thousands) for the UL-REIT sector. Proceeds from new equity are hand-collected from 10-Q and 10-K filings in the SEC Edgar database. The line includes proceeds from new equity issues, net of broker commissions and dealer-manager fees, from the Consolidated Statement of Cash Flows, under Cash Flows from Financing. Sponsors who appear in the SNL sample with more than one UL-REIT equity offering are termed “Repeat Sponsors” – represented by the gray portion of the bars in Figure 1. Sponsors who appear to have only one successful UL-REIT equity offering in the SNL sample are termed “One-off Sponsors” – depicted by the black portion of the bars in Figure 1. On the right-axis are annualized yields for the Fed Funds rate (depicted by the red line) and the NAREIT dividend yield (depicted by the blue line). 36 Figure 2. UL-REITs vs. L-REITs, Net Equity Proceeds Listed: SEOs Listed: IPOs Unlisted REITs $14,000,000 Proceeds from New Equity (in $Thousands) $12,000,000 $10,000,000 $8,000,000 $6,000,000 $4,000,000 $2,000,000 $0 Quarter Notes: Figure 2 shows net equity proceeds per quarter for the UL-REIT sector (depicted by the black line), compared to IPO proceeds by L-REITs (depicted by the gray line) and SEO proceeds by L-REITs (depicted by the blue line). Data for L-REIT equity offerings are from the NAREIT database, and only common equity offerings are represented here. 37 Figure 3. UL-REITs, Average Net Equity Proceeds & Competition Equity-raising Unlisted REITs Commercial Mortgage Flows $400,000 40 $350,000 35 $300,000 30 $250,000 25 $200,000 20 $150,000 15 $100,000 10 $50,000 5 $0 0 -$50,000 -5 -$100,000 -10 Number of Firms Raising Equity Avg Net Equity Proceeds Notes: Figure 3 displays the average net equity proceeds per firm for each quarter, depicted by the shaded gray area (in $Thousands). The red line shows Commercial mortgage flows from the Fed’s Flow of Funds (Z.1) series (measured in $Millions). On the right-axis and in blue bars is the total number of UL-REITs with open equity offerings during each quarter. 38 Figure 4. UL-REIT IRRs vs. NAREIT Returns 30% under 3 years 3 to 6 years 20% over 6 years 10% 0% -30% NAREIT returns -20% -10% 0% 10% 20% 30% -10% -20% -30% Unlisted REIT returns Notes: Figure 4 plots the calculated UL-REIT IRRs on the y-axis against the NAREIT total return index for L-REITs on the x-axis. Returns are measured as of Q4 2013 and the NAREIT benchmark is adjusted to the quarter of inception for the respective UL-REIT. The black line is the 45º diagonal, distinguishing UL-REITs that outperform vs. underperform the benchmark. Of the 53 UL-REITs shown, 9 are above the line, while the remaining 44 are below the line. 39 Table 1. Summary Statistics, UL-REIT Fees A. Fees Front-end load Broker commission Dealer-manager fee Acquisition fee Acquisition expenses Offering expenses Operational fees Asset management fee Property management fee Financing coordination fee Liquidation fees Disposition fee Subordinate participation Return hurdle Mode Mean Std Dev Min Max 7.0% 3.0% 2.0% 0.5% 1.5% 6.6% 2.4% 2.1% 0.5% 2.1% 1.4% 0.9% 0.9% 0.1% 1.0% 0.0% 0.0% 0.0% 0.3% 0.5% 8.0% 3.5% 4.5% 1.0% 5.5% 1.0% 4.5% 1.0% 0.8% 4.1% 0.9% 0.3% 0.9% 0.3% 0.0% 2.5% 0.1% 1.5% 6.0% 2.0% 3.0% 15.0% 8.0% 2.5% 12.3% 6.3% 0.8% 6.4% 2.7% 0.0% 0.0% 0.0% 3.0% 35.0% 10.0% Mean $76,336 11.7 18.8% 53.9 Std Dev $126,039 6.4 39.3% 7.5 Min -$9,873 1 0 36 Max $1,420,205 26 1 73 Mean 21.7 3.1% 0.6% 1.0% 1.0% 2.2% 0.2% $17.3 $4.8 1.0% 6.2% Std Dev 20.8 2.2% 0.8% 0.8% 0.4% 0.6% 2.5% $29.7 $3.9 3.2% 1.5% Min 1 0.1% -1.0% -0.3% 0.6% 0.7% -12.8% -$51.5 $0.1 -12.0% 4.0% Max 70 6.5% 2.3% 2.3% 3.0% 3.5% 6.7% $84.2 $16.1 9.6% 10.2% B. Equity Offerings Net equity proceeds (per quarter) Offering duration One-off sponsor CEO age C. Macro Variables Number of UL-REITs Fed Funds rate 3-Year Treasury spread 10-Year Treasury spread Baa-Aaa spread Core inflation Non-core inflation Commercial mortgage flows L-REIT equity flows NAREIT total returns NAREIT dividend yield Notes: Table 1 presents summary statistics for the sample of UL-REIT equity offerings. Panel A includes a set of fees commonly embedded in the UL-REIT offering, which are hand-collected from management compensation tables in the S-11 filings for each of the 86 funds in the SNL sample. Broker commission is paid to investment adviser who sells UL-REIT shares to their clients, as a percentage of gross offering proceeds. Dealer-manager fees are paid to the captive dealer-manager for marketing and administrative expenses associated with the offering, as a percentage of gross offering proceeds. Acquisition fees and Disposition fees are analogous to real estate commissions paid to the real estate brokers (except paid to management here), as a percentage of the contract purchase price. Acquisition expenses cover reimbursements for costs involved in the pursuit of an asset purchase. Offering expenses are reimbursements for organizational and offering costs for third party due diligence fees, as a percentage of gross offering proceeds. Asset management fees are paid annually, or more frequently, and based on the estimated value of average assets under management during the period. Property management fees are paid annually, or more frequently, and based on gross property revenues. Financing coordination fees are paid for each loan or refinancing, as a percentage of the loan amount. Subordinate participation measures the percentage share that management receives when a liquidity event is achieved and after the Return hurdle has been satisfied. Return hurdle is the cumulative, pre-tax, non-compounded return that shareholders must receive before Subordinate participation is activated. 40 Panel B of Table 1 presents characteristics of the offering. Net equity proceeds are hand-collected from 10-Q and 10-K filings in the SEC Edgar database. The line includes proceeds from new equity issues, net of broker commissions and dealer-manager fees, from the Consolidated Statement of Cash Flows, under Cash Flows from Financing. Offering duration counts the number of quarters from the date that the continuous offering is declared effective by the SEC until the quarter when no new equity is raised (for firms that have closed the equity offering). One-off sponsors appear in the SNL sample on only one occasion. CEO age is collected from the S-11 filing, at the moment that the offering initiates. Panel C of Table 1 summarizes quarterly measures for aggregate variables during the period 1993Q1 to 2013Q4. Number of UL-REITs provides a tally for firms in the SNL database that are actively raising new equity during the respective period. The Fed Funds, 3-Year Treasury, 10-Year Treasury, Aaa and Baa rates are from the Selected Interest Rates (H.15) series of the Board of Governors of the Federal Reserve System. The 3-Year Treasury spread subtracts the Fed Funds rate. The 10-Year Treasury spread subtracts the 3-Year Treasury rate. The Baa-Aaa spread is the credit spread on Moody’s seasoned corporate bonds. Core inflation is from the Bureau of Labor Statistics, Consumer Price Index (CPI) for all items, less food and energy. Non-core inflation is the difference between CPI and Core inflation. Commercial mortgage flows and L-REIT equity flows are from the Fed’s Flow of Funds (Z.1) series (measured in $Billions). NAREIT total returns and NAREIT dividend yield are from the FTSE NAREIT US Real Estate Index series. 41 Table 2. Estimation: Net Equity Proceeds, UL-REITs Variable Constant Broker commission Dealer-manager fee Acquisition fee Asset management fee Subordinate participation One-off Sponsor CEO age No. of UL-REITs Fed Funds rate 3-Year Treasury spread 10-Year Treasury spread BBB-AAA spread Core inflation Non-core inflation Commercial mortgage flows L-REIT equity flows NAREIT total returns NAREIT dividend yield Offering maturity fixed effects Adjusted R2 Observations Coefficient 8.02 *** 45.31 *** 44.00 *** -6.40 -72.14 *** -0.76 -1.38 *** 0.02 *** -0.03 *** -38.57 *** -25.37 6.46 6.25 5.70 1.07 0.02 *** -0.01 -1.85 -19.92 *** t-stat 6.1 5.0 6.3 -0.8 -3.5 -0.3 -9.7 3.1 -3.0 -2.6 -1.5 0.3 0.3 0.4 0.5 7.5 -0.3 -1.1 -2.6 Coefficient 5.55 *** 47.51 *** 40.54 *** -6.30 -74.75 *** -0.75 -1.37 *** 0.02 *** -0.01 5.33 55.69 31.55 7.19 0.30 0.02 0.00 -1.37 -22.54 *** * *** *** t-stat 6.1 5.3 5.9 -0.7 -3.7 -0.3 -9.6 3.1 -1.6 0.4 3.8 1.9 0.5 0.1 7.9 -0.3 -0.8 -3.0 Coefficient 7.28 *** 44.15 *** 46.53 *** -9.15 -67.80 *** -0.41 -1.40 *** 0.02 *** -0.02 ** -43.67 *** -35.37 ** 3.95 -19.40 -1.10 1.97 0.02 *** 0.00 -0.35 t-stat 5.7 4.9 6.7 -1.1 -3.3 -0.2 -9.8 3.0 -2.5 -3.0 -2.2 0.2 -1.2 -0.1 0.9 8.8 0.1 -0.2 Coefficient 4.32 *** 46.51 *** 42.93 *** -9.46 -70.16 *** -0.36 -1.38 *** 0.02 *** 0.00 -1.50 60.30 5.87 -0.41 1.21 0.02 0.00 0.44 *** *** t-stat 5.3 5.1 6.3 -1.1 -3.4 -0.1 -9.7 3.0 -0.1 -0.1 4.1 0.4 0.0 0.6 9.4 0.2 0.3 Included [27 quarters] Included [27 quarters] Included [27 quarters] Included [27 quarters] 36.6% 809 36.1% 809 36.1% 809 35.5% 809 Notes: Table 2 presents results from the least squares estimation for Net equity proceeds, logged. Observations are in UL-REIT firmquarters and only periods where a fund is actively raising new equity are considered. Offering maturity fixed effects describes a set of 27 indicator variables that control for offering seasoning, measured in quarters relative to inception (27 quarters is the lengthiest offering period in the sample). The first column lists the variable name. All variables are defined in the notes to Table 1. The second and third columns report the estimated coefficient and corresponding t-statistic. ***, ** and * denote statistically significant coefficients at the 1%, 5% and 10% levels respectively. 42 Table 3. Summary Statistics, L-REITs vs. UL-REIT Characteristics: Q4 2013 Variable %Multifamily %Industrial %Office %Hotel %Retail %Theater %Healthcare %Storage %Manufactured %Restaurant Firm age Total assets Leverage ratio M/B ratio UL-REITs Mean Std Dev 15.8% 34.0% 11.4% 23.3% 20.6% 32.0% 9.1% 26.6% 21.9% 36.0% 0.0% 0.0% 11.5% 30.2% 2.7% 13.4% 1.3% 9.9% 0.3% 1.2% 4.1 2.3 $1.277 $1.573 0.491 0.209 L-REITs: pre-match Mean Std Dev 14.9% 34.1% 12.2% 27.9% 17.1% 31.4% 12.9% 32.9% 20.8% 36.2% 0.8% 7.3% 11.3% 31.1% 1.5% 12.1% 1.6% 12.1% 0.7% 4.7% 10.8 7.8 $4.334 $6.279 0.567 0.168 1.243 0.263 L-REITs: post-match Mean Std Dev 27.9% 44.6% 8.8% 24.4% 18.6% 35.1% 6.7% 25.1% 15.2% 33.9% 0.0% 0.0% 15.1% 36.0% 1.7% 12.9% 0.3% 2.3% 0.2% 1.2% 3.8 4.6 $1.127 $1.022 0.525 0.172 1.107 0.201 Notes: Table 3 provides summary statistics for characteristics of UL-REITs and L-REITs, as of year-end 2013. %PropertyType are hand-collected from SNL data for each firm. Property type allocations are based on property square footage (where available), acquisition prices (when square footage is unavailable), and total number of properties in each category (when acquisition prices and square footage are both unavailable). Firm age is measured in years. Total assets is based on book values, measured in $Billions. Leverage ratio equals total liabilities divided by Total assets. M/B ratio is the market value of total assets divided by book value of total assets. The L-REIT sample is Windorized, deleting observations with M/B ratios beyond the 5 th and 95th percentiles. The first column lists the variable. Sample Mean and standard deviations (Std Dev) are presented sequentially for the sample of UL-REITs (second and third columns), L-REITs (fourth and fifth columns), and L-REITs following the propensity-score matching procedure (sixth and seventh columns). 43 Table 4. Probit for Propensity-score Matching Variable Constant %Multifamily %Industrial %Office %Hotel %Retail %Theater %Healthcare %Storage %Manufactured %Restaurant Firm age Total assets Leverage ratio psuedo-R2 Observations Coefficient 0.337 0.459 0.149 0.966 * 0.147 0.711 -79.816 0.386 1.103 1.067 0.971 -0.098 *** -0.154 ** -1.012 * (Wald χ2) (0.4) (0.8) (0.1) (3.1) (0.1) (1.9) (0.9) (0.5) (1.8) (0.9) (0.1) (21.5) (6.6) (3.0) 27.3% 264 Notes: Table 4 presents results from the probit estimation used to generate propensity scores for the matching procedure. The dependent variable is an indicator, taking on a value of one for UL-REITs and zero for L-REITs. The first column lists the variable name. All variables are defined in the notes to Table 3. The second and third columns report the estimated coefficient and corresponding tstatistic. ***, ** and * denote statistically significant coefficients at the 1%, 5% and 10% levels respectively. 44 Table 5. Return Comparison: UL-REITs vs. NAREIT index A. Prices & Returns Offering price Imputed share price Price difference Mean $10.93 $8.97 -18.0% Median $10.00 $7.95 -22.3% Std Dev $3.65 $4.55 32.3% Min $8.00 $1.68 -83.2% Max $25.00 $26.54 69.3% UL-REIT IRR NAREIT returns Return difference -0.7% 8.4% -9.1% -1.3% 7.9% -9.5% 9.9% 5.2% 9.1% -22.0% -6.5% -37.1% 24.1% 24.3% 12.9% B. Estimation Coefficient -0.074 *** 0.803 *** 16.1% 53 α β Adjusted R2 Observations (t-stat) (-3.1) (3.3) Notes: Panel A of Table 5 displays the result of the Imputed share price calculations. Offering price is from the S-11 filing for each UL-REIT. Imputed share price equals the book value of total assets multiplied by M/B ratio* minus preferred equity and OP units, then divided by the total number of common equity shares outstanding at yearend 2013. M/B ratio* is the market-to-book ratio of the nearest-neighbor L-REIT drawn from a one-to-one propensity matching procedure. Price difference is the percentage difference in the Imputed share price relative to the Offering price. ULREIT returns are produced from an IRR calculation on the series of investor cash flows that includes the Offering price, all quarterly distributions to-date and the Imputed share price for year-end 2013. NAREIT returns are calculated from the FTSE NAREIT US Real Estate Index series for “All REITs”, adjusted to the quarter of inception for each UL-REIT return observation. Return difference equals UL-REIT return minus NAREIT returns. Panel B provides results from a simple estimation for: UL-REIT returns = α + β∙NAREIT returns + ɛ. The first column lists the parameter. The second and third columns report the estimated coefficient and corresponding t-statistic. *** denotes statistically significant coefficients at the 1% level. 45 Appendix. Typical Fees Front-end load Broker commission Dealer-manager fee Acquisition/transaction fee Other Total front-end load Operational fees Asset management fee Property management fee Monitoring fee Other Total expense ratio Liquidation fees Disposition fee Subordinate participation Return hurdle UL-REITs Typical Mutual Funds Typical Hedge Funds Typical Private Equity Typical 7% 3% 2% 2% 14% --0.09%a ----0.46%a ----------- ----1-2%c --2%c 1% 4.5% --1% --- 0.24%a ------0.50%a 1.52%b --------- 2%c --2%c ----- 3% 15% 8% ------- --19.34%b 0%b --20%c 8%c a Adams, Mansi and Nishikawa (2012), Table 2. Cassar and Gerakos (2010), Table 1. c Metrick and Yasuda (2010), Table 2. b 46