Non-Traded REITs. Olgun Fuat Sahin Minnesota State University Moorhead

Transcription

1 Non-Traded REITs Olgun Fuat Sahn Mnnesota State Unversty Moorhead In ths paper, I examne dfferences between non-traded and traded REITs by usng avalable fnancal data. Fndngs ndcate that there are many dfferences along a number of dmensons between nontraded and traded REITs. For example, non-traded REITs hold fewer propertes, employ lower leverage, generate lower return on nvested captal (ROIC) and funds from operatons (FFO) and pay a greater fracton of ther ncome n dvdends. There does not appear to be sgnfcantly dfferent dvdend yelds. These fndngs have publc polcy mplcatons especally consderng how ther dvdends may be funded and how non-traded REIT shares are purchased. INTRODUCTION A REIT s a publcly held or prvately owned company that nvests n and manages a portfolo of ncome producng commercal propertes or mortgages. A REIT s created based on the Internal Revenue Code (Sectons ) to become a pass-through entty that dstrbutes sgnfcant porton of ts earnngs and captal gans to ts shareholders. REITs do not pay corporate taxes as long as qualfcaton standards are met; however, dstrbuted earnngs and captal gans are taxed at the shareholder level. Non-traded REITs are organzed just lke publcly traded REITs n terms of ther legal and governance structures. They are both subject to the same regulatory oversght by the SEC and the Sarbanes Oxley Act of However, there are dfferences between the two types of REITs n terms of ther formaton, management and sze. Typcally a non-traded REIT s formed by a sponsor frm that packages propertes or captal from dfferent sources (ndvdual or nsttutonal nvestors) nto a REIT structure. Accordng to Farrell (2006) non-traded REITs may have a fnte lfe, about ten years, and cease to exst wth an Intal Publc Offerng (IPO), a merger or lqudaton of propertes. Non-traded REITs may be managed by a thrd party (REIT havng no employees) and/or ts shares are sold to nvestors by a broker-dealer that may be related to the sponsorng frm. A sponsor frm may receve fees for servces (day-to-day, asset management or advsory) t provdes to a non-traded REIT ether drectly or through ts subsdares. By comparson, most publcly traded REITs have ther own management teams and boards. In terms of sze, t s nterestng to note that whle aggregate total assets of traded REITs are much larger than that of non-traded REITs, the growth rate of aggregate total assets of non-traded REITs has been hgher n the recent past. Accordng to SNL Fnancal, aggregate total assets of non-traded REIT were $6.07 bllon compared to $ n aggregate total assets managed by publc REITs n Aggregate total assets of non-traded and traded REITs grew to $65.24 bllon and $ bllon, respectvely n These ncreases represent a 30 percent and 4 percent aggregate total assets growth rated of non-traded and traded REITs, respectvely. Non-traded REITs contnue to attract more captal. Accordng to Joanna Randell of REIT Café, non-traded REITs receved approxmately $8 bllon n Journal of Accountng and Fnance vol. 12(1) 2012

2 and expected to get more than $10 bllon n Fgure 1 reports the aggregate total assets of nontraded and traded REITs over tme. FIGURE 1 AGGREGATE TOTAL ASSETS OF NON-TRADED AND TRADED REITS Non-traded Traded Aggregate Total Assets (n bllons of USD) Source: SNL Fnancal The aggregate total assets growth rate of non-traded REIT accelerated after 2003 whle that of traded REITs have contracted snce The number of REITs n each group s shown n Fgure 2 and complements the trends observed wth ther total assets. Year FIGURE 2. NUMBER OF NON-TRADED AND TRADED REITS Number of REITs Non-traded Traded Year Source: SNL Fnancal In ths paper, I examne the dfferences between non-traded REITs and traded REITs along a number of fnancal dmensons based on avalable fnancal data. In addton, I also attempt to quantfy the effects of no lstng on REIT valuaton. The paper s organzed as follows. The next secton provdes Journal of Accountng and Fnance vol. 12(1)

3 characterstcs of non-traded REITs followed by a secton on data and methodology. Secton four presents results and secton fve provdes conclusons. CHARACTERISTICS OF NON-TRADED REITS Exstng lterature on non-traded REITs s very lmted. Stll ndustry experts and followers appear to agree on some characterstcs of non-traded REITs. I summarze those characterstcs from an nvestment standpont wthout categorzng them:. An mportant characterstc of non-traded REITs s that ther share prce does not fluctuate. Accordng to Farrell (2006) sponsor of a non-traded REIT establshes ts share prce. Mattox (2006) suggests that the share prce of a non-traded REIT would be reset f ndependent apprasals ndcate that underlyng property valuatons have changed.. Hogan (2010) clams that non-traded REITs ntroduce greater dversfcaton to a portfolo because they have low correlaton wth stocks and they provde a consstent dvdend yeld. The low correlaton n ths context reflects relatvely constant values of non-traded REITs smlar to drect real estate nvestments.. Non-traded REITs may have hgher dvdend yelds than that of publc REITs. Accordng to Kmbrell and Chtty (2010) the average dvdend yeld on non-traded REITs was 6.53% n Farrell (2006) and Morrssey (2009) report dvdend yeld ranges from 6% to 7% and from 6% to 9%, respectvely. v. Accordng to an ndustry executve, Ncholas S. Schorsch, nvestors purchasng shares of nontraded REITs can avod payng sgnfcant premums to net asset values when traded REITs can only be purchased at a premum n the stock market. 1 Even though ths s reasonable, t does not explan why a non-traded REIT sponsor would not ntate an IPO process to capture the premum. v. Non-traded REIT shares are not lqud, lmtng nvestors ablty to ext an nvestment f crcumstances change. Accordng to Farrell (2006) nvestor redemptons are very restrctve, typcally t s lmted to 3% of the fund per year and can be elmnated at any tme by the sponsor. Morrssey (2009) suggests that there are penaltes assocated wth early redemptons that can be 10% of an nvestor s ntal nvestment. v. Commssons assocate wth ntal purchase of non-traded REIT shares are relatvely hgh. Accordng to Farrell (2006), they are between 8% and 9%, but Morrssey (2009) suggests that ntal nvestment commssons pad to brokers and advsors can reach to 15%. v. A non-traded REIT s desgned to have a set termnaton date. Accordng to Farrell (2006), t s 10 to 12 years and nvestor s return depends on the ext prce. Investors would have to commt to a long tme perod before makng any non-traded REIT purchases. Typcally, shareholders ext ther nvestment va an IPO, a merger or lqudaton. The actual total return of a non-traded REIT nvestor s determned by, not only the ntal purchase prce and dvdends receved over tme, but also the ext prce. If the market condtons at the tme of termnaton are not favorable, then nvestors can experence sgnfcant losses. v. A sponsor of non-traded REITs may cut dvdends and or reduce the share value f property nvestments do not perform as expected. For example, Cole Credt Property Trust II, Inc. (Cole REIT II) reduced the value of ts commons stock to $8.05 from $10.00 n June 22, Cole REIT II made ths change n response to a new reportng requrement by the Fnancal Industry Regulatory Authorty (FINRA) that reported per share value estmate cannot be older than 18 months. x. In the extreme case, dvdends payments may come from new captal rased through share sales or borrowng. Currently there s no mandate as to source of dvdends payment. Some ndustry advocates support a mandate that dvdend payment be funded through operatng cash flows. x. There s also the ssue of conflct of nterest between non-traded REIT nvestors and a sponsor/manager. A sponsor/manager may earn property acquston, fnance and management 50 Journal of Accountng and Fnance vol. 12(1) 2012

4 x. fees. If these transactons are not at arm s length, then nvestors of non-traded REITs would suffer possble losses. There s lmted transparency wth non-traded REIs makng t dffcult to assess and montor ther nvestments and operatons. Even though they are subject to the same regulaton as publcly traded RIETs, there are not many analyst specalzed n non-traded REITs. DATA AND METHODOLOGY The data set used n ths study s provded by SNL Fnancal and covers a tme perod from 1990 to There are 460 REITs n the sample; 86 non-traded and 374 traded. The dstrbuton of non-traded and traded REITs across property types are shown n Table 1. Non-traded REITs appear to concentrate n dversfed, offce, hotel and retal (other than regonal malls and shoppng centers) type property holdngs. TABLE 1 PROPERTY TYPES OF NON-TRADED AND TRADED REITS Ths Table reports the dstrbuton of non-traded and traded REITs across property types. The data s provded by SNL Fnancal and covers a tme perod from 1990 to There are 460 REITs n the sample; 86 non-traded and 374 traded. Property Focus Nontraded Percent of nontraded REITs Traded Percent of traded REITs Dversfed % % Health Care % % Hotel % % Industral % % Manufactured Homes % % Mult-Famly % % Offce % % Regonal Mall % % Retal: Other % % Self-Storage % % Shoppng Center % % Specalty % % Total Table 2 reports descrptve statstcs of varables based on fnancal statements. I use these varables to compare non-traded and traded REITs. Most of these varables do not meet the propertes of a standard normal dstrbuton; therefore, n my comparsons I use t-tests as well as Wlcoxon sgned rank tests on medans. Journal of Accountng and Fnance vol. 12(1)

5 TABLE 2 DESCRIPTIVE STATISTICS Ths Table reports the dstrbuton of non-traded and traded REITs across property types. The data s provded by SNL Fnancal and covers a tme perod from 1990 to There are 460 REITs n the sample; 86 nontraded and 374 traded. Varable Mean Medan Standard Devaton Mnmum Maxmum N Total Equty-to-Total Assets (%) ,944 Common Equty-to-Total Assets (%) ,944 Propertes-to-Assets (%) ,949 Total Debt-to-Total Equty (x) ,864 Return on Average Assets (%) ,592 Return on Average Equty (%) ,526 Return on Invested Captal (%) ,124 Number of Propertes (Total) , ,506 Portfolo Occupancy (%) ,531 Gross Leasable Area (sq. ft.) 31,273,975 16,510,726 52,169,448 56, ,779, Expense Remburse-to- Rental Rev (%) ,268 Total Debt-to-Gross Propertes (%) ,747 Interest Income-to-Total Revenue (%) ,420 Loans-to-Assets (%) ,583 Interest Expense-to-Total Expenses (%) ,822 Property Expense-to-Total Expenses (%) ,834 Funds from Operatons-to- Total Revenue (%) ,993 Dvdend Payout Rato (%) ,008 Dvdend Yeld (%) ,956 Common Dvdends Padto-Common Equty (%) , Journal of Accountng and Fnance vol. 12(1) 2012

6 I also estmate values of non-traded REITs based on comparable multples from traded REITs and estmate the possble loss of value due to no lstng of shares. I use only the 2009 data and estmate the followng regresson for ths analyss: ( MB) = α + ( Status) + ( Infnte) + ( Age) + ( Number of Propertes) + ( ROIC) + ( TDTE) + ε Where, MB s the natural logarthm of market value of common stock per share-to-book value of common stock per share. Market value per share of traded REITs s based on the market prce at the end of I estmate market values of non-traded REITs usng a medan Funds From Operatons (FFO) multple of the same property type REITs that are publcly traded. I estmate market values of non-traded REITs only f they have postve FFO n I also consder alternatve defntons based on IPO prce as well. Status and Infnte are dummy varables that have values of 1 f a REIT s non-traded and set up to exst forever, respectvely. Age and Number of Propertes are natural logarthms of age of a REIT snce ts ntal publc offerng and number of propertes held as of 2009, respectvely. ROIC and TDTE refer to Return on Invested Captal and Total Debt-to-Total Equty n 2009, respectvely. I expect Status to affect MB negatvely because non-traded REITs are not lsted and therefore nvestors face a sgnfcant loss of lqudty. In addton, the structure of nvestng n non-traded REITs rase concern about agency problems and cost of nvestng. I expect all these negatve factors outwegh any beneft nvestors would get due to non-lstng of stocks. In addton, I expect ROIC and TDTE to have postve effect on MB because greater return on nvested captal and hgher leverage typcally lead to hgher MB ratos. RESULTS I report unvarate comparsons of non-traded and traded REITs on Table 3. The Table shows the results of both t-tests on means and Wlcoxon snged rank tests on medans. I observe that non-traded REITs use less leverage as measured by Total Equty-to-Total Assets and Common Equty-to-Total Assets, but ther nterest expense-to-total expense rato s hgher than that of traded REITs. All these dfferences are statstcally sgnfcant. These fndngs are consstent wth the noton that non-traded REITs may be exposed to a greater cost of borrowng. Such a fndng would suggest that non-traded REITs would be less lkely to leverage ther assets to seek growth. Non-traded REITs are not as proftable as traded REITs when comparng Return on Average Assets, Return on Average Equty and Return on Invested Captal. Except for Return on Average Assets, they are all statstcally sgnfcant. I also observe that traded REITs typcally nvest n three tmes as many propertes and hold property portfolos three tmes as large as non-traded REITs. These dfferences are statstcally sgnfcant. In addton, Funds from Operatons-to-Total Revenue s lower whle Dvdend Payout Rato s hgher for non-traded REITs than that of traded REITs. Both of these dfferences are also statstcally sgnfcant. Comparson of dvdend yelds between non-traded and traded REITs are affected by how the dvdend yeld for non-traded REITs s defned. If I use SNL Fnancal reported dvdend yelds, nontraded REITs have lower dvdend yelds than traded REITs. However, ths s affected by many 0% dvdend yeld observatons for non-traded REITs reported by SNL Fnancal. Ths s true even though cash flow statements ndcate that there were dvdend payments to common stockholders. Ths reduces the usefulness of SNL Fnancal reported dvdend yeld to compare non-traded and traded REITs. Therefore, I also consder an alternatve defnton of dvdend yeld (Common Dvdends Pad-to- Common Equty) that follows the same procedure for all REITs. Common Dvdends Pad-to-Common Equty s computed as common stock dvdends pad by a REIT dvded by ts book value of common equty; both varables are based on SNL Fnancal data. The result of ths comparson fnds no statstcal dfference between the two groups n terms of ther dvdend yeld. Although not reported n Table 3, I also compare dvdend yeld of non-traded REITs computed as mentoned before and dvdend yeld of Journal of Accountng and Fnance vol. 12(1)

7 traded REITs based on ther market prces. I observe that non-traded REITs have hgher dvdend yelds. However, ths comparson s based on varables that are not computed n a smlar manner. TABLE 3 UNIVARIATE ANALYSIS Ths Table reports results of unvarate analyss based on REIT status. The data s provded by SNL Fnancal and covers a tme perod from 1990 to There are 460 REITs n the sample; 86 non-traded and 374 traded. I report p-values for dfferences n means and Wlcoxon sgned rank test n medans. Standard Varable Status Mean Medan Total Equty-to- Total Assets (%) Common Equtyto-Total Assets (%) Propertes-to- Assets (%) Total Debt-to-Total Equty (x) Return on Average Assets (%) Return on Average Equty (%) Return on Invested Captal (%) Funds from Op.-to- Total Revenue (%) Devaton Mnmum Maxmum N Non-traded Traded ,537 p-value Non-traded Traded ,537 p-value Non-traded Traded ,539 Non-traded Traded ,480 p-value Non-traded Traded ,253 p-value Non-traded Traded ,208 Non-traded Traded ,823 Non-traded Traded , Journal of Accountng and Fnance vol. 12(1) 2012

8 TABLE 3 (CONTINUED) Varable Status Mean Medan Standard Devaton Mnmum Maxmum N Non-traded Number of Traded ,203 Propertes (Total) Non-traded Portfolo Traded ,366 Occupancy (%) p-value Gross Leasable Area (sq. ft.) Expense Remburse-to- Rental Rev (%) Total Debt-to- Gross Propertes (%) Interest Income-to- Total Revenue (%) Loans-to-Assets (%) Interest Expenseto-Total Expenses (%) Property Expenseto-Total Expenses (%) Funds from Operatons-to-Total Revenue (%) Dvdend Payout Rato (%) Dvdend Yeld (%) Common Dv. Pad-to-Common Equty (%) Non-traded 11,885,954 5,010,478 13,452,595 56,336 45,957, Traded 35,395,101 18,008,838 56,270,117 95, ,779, Non-traded Traded ,020 p-value Non-traded Traded ,391 p-value Non-traded Traded ,047 Non-traded Traded ,210 p-value Non-traded Traded ,429 p-value Non-traded Traded ,439 Non-traded Traded ,791 Non-traded Traded ,783 p-value Non-traded Traded ,544 Non-traded , Traded , , ,590 p-value Journal of Accountng and Fnance vol. 12(1)

9 I use a regresson analyss to examne the effects of loss of value due to tradng status of REITs. I restrct the sample for ths analyss to the latest year of avalable data, The characterstcs of ths sub-sample are provded on Table 4. Non-traded REITs n ths sub-sample have lower MB and ROIC, hold fewer propertes, use lower leverage and are relatvely new when compared to traded counterparts. TABLE 4 DESCRIPTIVE STATISTICS OF VARIABLES IN MULTIVARIATE ANALYSIS I restrct the sample for the regresson analyss to the latest year of avalable data, Ths Table report characterstcs of the sub-sample. MB s the natural logarthm of market value of common stock per shareto-book value of common stock per share. Market value per share of traded REITs s based on the market prce at the end of I estmate market values of non-traded REITs usng a medan funds from operatons (FFO) multple of the same property type REITs that are publcly traded. I estmate market values of non-traded REITs only f they have postve FFO n Age and Number of Propertes are natural logarthms of age of a REIT snce ts ntal publc offerng and number of propertes held as of 2009, respectvely. ROIC and TDTE refer to Return on Invested Captal and Total Debt-to-Total Equty n 2009, respectvely. Varable Mean Medan Panel A. All observatons Standard Devaton Mnmum Maxmum N MB Age Number of propertes , ROIC TDTE Panel B. Non-traded REITs MB Age Number of propertes ROIC TDTE Panel C. Traded REITs MB Age Number of propertes , ROIC TDTE Journal of Accountng and Fnance vol. 12(1) 2012

10 TABLE 5 MULTIVARIATE ANALYSIS I also estmate values of non-traded REITs based on comparable multples from traded REITs and estmate the possble loss of value due to no lstng of shares. I use only the 2009 data and estmate the followng regresson for ths analyss: ( MB) = α + ( Status) + ( Infnte) + ( Age) + ( Number of Propertes) + ( ROIC) + ( TDTE) + ε Where, MB s the natural logarthm of market value of common stock per share-to-book value of common stock per share. Market value per share of traded REITs s based on the market prce at the end of I estmate market values of non-traded REITs usng a medan funds from operatons (FFO) multple of the same property type REITs that are publcly traded. I estmate market values of non-traded REITs only f they have postve FFO n I also consder alternatve defntons based on IPO prce as well. Status and Infnte are dummy varables that have values of 1 f a REIT s set up to exst forever and non-traded, respectvely. Age and Number of Propertes are natural logarthms of age of a REIT snce ts ntal publc offerng and number of propertes held as of 2009, respectvely. ROIC and TDTE refer to Return on Invested Captal and Total Debt-to-Total Equty n 2009, respectvely. Varable Model 1 Model 2 Model 3 Intercept *** *** *** Non-traded *** *** *** Age Number of propertes ROIC *** *** *** TDTE ** ** ** Adjusted R The results of the regresson analyss reported n Table 5. I consder alternatve specfcatons n sx dfferent forms. Each of these models accounts for about half of the varaton n MB. Generally, I observe that the results of regresson results support the expectatons. Infnte lfe, age and number of propertes do not affect MB n a statstcally sgnfcant manner. I observe that ROIC and TDTE affect MB postvely. These effects are also statstcally sgnfcant. The coeffcent for status dummy s negatve and statstcally sgnfcant n all specfcatons. I determne that MB rato would be about 68% lower for nontraded REITs than that of traded REITs based on the transformed coeffcent of status dummy f they were publcly traded. Journal of Accountng and Fnance vol. 12(1)

11 TABLE 5 (CONTINUED) Varable Model 4 Model 5 Model 6 Intercept *** *** *** Non-traded *** *** *** Infnte lfe Age Number of propertes ROIC *** *** *** TDTE ** Adjusted R CONCLUSIONS In ths paper, I examne dfferences between non-traded REITs and traded REITs based on avalable fnancal data. In addton, I attempt to quantfy the effects of no lstng on REIT valuaton. I fnd that many dmensons of non-traded and trades REITs are dfferent. For example non-traded REITs hold fewer propertes, employ lower leverage, generate lower ROIC and FFO and pay greater fracton of ncome n dvdends. I do not observe sgnfcantly dfferent dvdend yelds and attrbute the clams of hgher dvdend yeld on non-traded REITs to dfference n dvdend yeld computaton. I also fnd that non-traded REITs would have lower market-to-book multples f they were publcly traded. Consderng all ths evdence there are two clear patterns that may make non-traded REITs a more problematc or rsky nvestment than traded REITs. Frst, ther ablty to make dvdend payments and mantan such hgh payout ratos may be lmted f operatons are not as proftable as traded REITs. If the hgh payout rato s supported by greater use of outsde captal, mostly n the form of equty, then nontraded nvestors are lkely to suffer severe losses snce ths structure would not be sustanable. Ths necesstates a mandate on the source of dvdend payments that dvdends should be supported by operatng profts and not by external captal. Such a mandate on how dvdends are fnanced would make non-traded REITs even more credble to nvestors. Second, non-traded REIT portfolos may be concentrated n specfc local markets. If these local markets were to experence a contracton, non-traded REIT nvestors would suffer sgnfcant losses. Ths may not be as much of a problem for traded REITs snce they hold relatvely larger property portfolos and I would expect that ther concentraton n partcular local markets s lmted. Thrd, t s not clear what justfes hgh transacton costs assocate wth non-traded REITs. It does not appear to be superor ROIC generated by non-traded REITs based on the avalable evdence. 58 Journal of Accountng and Fnance vol. 12(1) 2012

12 ENDNOTES 1 Randyl Drummer, Talkng non-trade REITs wth Nck Schorsch of Amercan Realty Captal, September 29, REFERENCES Chrstopher Farrell, Gettng a slce of the commercal market, Bloomberg Busnessweek, February 13, 2006, accessed September 18, 2010, Janet Morrssey, Investors are seekng refuge n non-traded REITs, February 1, 2009 accessed November 05, 2010, Jason Mattox, Know your REITs, Commercal Investment Real Estate, July-August 2006, p Joanna Randell, Non-traded REITs slated to rase $10 bllon, REIT Cafe, accessed February 18, 2011, 10_Bllon.html. Kevn Hogan, Non-traded REITs can dversfy nvestment portfolo whle reducng volatlty, September 24, 2010, accessed November 5, 2010, REITs-Can-Dversfy-Investment-Portfolos-Whle-Reducng-Volatlty/ REIT Café, REIT Talk Epsode 29, Prvate vs. publc what nvestors need to know, July 07, 2010, accessed July 12, 2010, Vee Kmbrell and Stacy Chtty, Are non-traded REITs gong manstream, Natonal Real Estate Investor, May 1, 2010, accessed November 5, 2010, Journal of Accountng and Fnance vol. 12(1)