Do UPREITs Suffer Tax-Timing Conflict of Interest?
Fred Wu
Doctoral candidate in Real Estate
Smeal College of Business
Penn State University
University Park, PA 16802
and
Abdullah Yavas
William Elliott Professor of Business Administration
Smeal College of Business
Penn State University
University Park, PA 16802
(This Draft: March 2005)
This research was supported by a research grant from the Real Estate Research Institute.
Abstract
The major cause of management’s breach of fiduciary duty is the misalignment of interest between management and shareholders. In the case of Umbrella Real Estate Investment Trusts (UPREITs), a frequently cited cause of the conflict of interest is tax timing. A tax timing conflict arises if management’ tax basis in the properties is different than that of REIT shareholders. Management may be reluctant to sell the properties in fear of triggering large personal capital gains tax (Sagalyn, 1996). In this research, we argue that this concern about tax timing conflict may be misplaced. We offer evidence that the property acquisition prices in the tax-deferred contributions are discounted, and therefore the agency problem has already been addressed by the market.
The major cause of management’s breach of fiduciary duty is the misalignment of interest between management and shareholders. In the case of Umbrella Real Estate Investment Trusts (UPREITs), a frequently cited cause of the conflict of interest is tax timing. UPREITs Tax timing conflict arises if management’s tax basis is lower than that of REIT shareholders. Management may be reluctant to sell the properties because of the negative personal tax implication that may result (Sagalyn, 1996). In this research, we argue that this concern on the tax timing conflict may be misplaced. We offer evidence that the property acquisition prices in tax-deferred contributions are discounted, and therefore the agency problem has already been addressed by the market.
1. Background on UPREITS and Literature
An UPREIT is a structure where a REIT owns substantially all of the assets of the REIT through an Operating Partnership (OP) composed of the REIT as general partner and the others as limited partners. Partnership interests in the OP are generally referred to as “OP Units.” The two-tiered structure of UPREIT has been adopted to take advantage of the tax deferred property transactions of partnership. Unlike selling real properties directly to a REIT which is usually a taxable event, selling properties to OP in exchange for OP Units is generally not a taxable event as the transaction can be structured as admitting a new partner by OP.[1] This is true even if OP units are exchangeable one-for-one into common stocks of REIT after a lock-up period. For real estate that has been held for a long time, the tax-deductible depreciation together with appreciation in property value can generate very large built-in capital gains. Consequently, UPREIT structure provides a distinct tax advantages to the property sellers over non-UPREITs by allowing them generally to sell the properties without adverse tax consequences. In principle, the sellers’ tax liability can be deferred indefinitely as long as OP continues to hold the contributed properties and the original sellers don’t convert OP units to other considerations such as cash or REIT common stocks. In the case any of these conditions is violated, the new OP partner (previous property seller) must recognize capital gains tax in the year of the event. Additionally, OP will receive a step-up in basis of its property with respect to any gain recognized by the new OP partner. Tax timing problems arise when the new OP partner also serves as management or board director. Since selling the contributed property triggers much larger share of capital gains tax for new OP partner, the optimal reservation price of the new OP partner (for selling) is much lower than that of REIT common shareholders. This creates the conflict of interest because a given property disposition transaction, while making no sense to the new OP partner, can be financially advantageous for REIT common shareholders.
In the current study, we argue that this frequently cited concern on the tax-timing conflict of interest has already been addressed by the market through discounting the property acquisition prices and by the contracting mechanism of lockup agreement. A delayed step-up in tax basis caused by the tax-deferred property contribution is detrimental to REIT shareholders’ wealth because it lowers the depreciation basis that OP may use initially to deduct taxable incomes.[2] Property sellers, expecting property prices to reflect on this, only defer their tax liability if their benefits from the tax deferral are larger than OP’s loss from it. The seller and OP then enter a lock-up agreement obligating OP not to sell properties, except in the tax-free or tax-deferred transactions, for a typical period of 5-12 years. The breach of the agreement calls for OP to compensate the property seller for the loss caused by the accelerated capital gains tax. As the new OP partner (previous property seller) is now made even by the lock-up agreement regardless the property disposition decision, his interest in tax-timing is no longer in conflict with that of REIT shareholders. Consequently, the tax-timing conflict of interest disappears. OP fully expecting the lock-up agreement will request compensation for the delayed tax depreciation benefits by asking for a lower property acquisition price. To examine how realistic our theory is, in this study we empirically investigate the hypothesis that the effect of the delayed tax benefits to OP is capitalized in property acquisition price.
A related study by Sinai and Gyourko (2000) examine how the stock price performance of, respectively, the regular REITs and REITs with UPREIT structure was affected by the capital gain tax rate reduction enacted in the Taxpayer Relief Act of 1997. They provide evidence that the tax change was substantially capitalized into share prices. In contrast, in the current study we seek the direct evidence that the tax effect is capitalized into the property acquisition price.[3] A recent work by Gentry, Kemsley, and Mayer (2003a) also examines how shareholder-level dividend taxes are incorporated into share prices. Their research finds strong evidence that shareholders value the tax benefits of depreciation deductions. Holding the market value of property constant, they find that the REITs with higher tax basis command higher stock market values. This result confirms that the stock market is efficient in incorporating tax benefits.
Two other studies are also related. Holmes and Slade (2001) examine the impact of tax-deferred Section 1031 exchanges on apartment transaction prices in Phoenix area. They find that, while selling prices of relinquished properties are not significantly impacted, the purchase prices of replacement property in a Section 1031 exchange hold price premiums. They argue that the purchase price premium is consistent with the reverse tax capitalization hypothesis where the replacement property acquirer bids up the property price due to the pressure to acquire a replacement property within the required time frame. Additionally, Campbell et al. (2001) find that the status of being an UPREIT seems to hold incremental explanation power (for the abnormal returns) over the ratio of the value of OP units offered (by the acquirer) to total acquisition price. They argue that the UPREIT status captures more than just the tax benefits. But as we argue later in this paper, the ratio of the value of OP units offered is not a good indicator of the size of tax effect. So the additional explanation power of the UPREIT status in Campbell et al. (2001) might just pick up the remaining tax effect not captured by the ratio of the value of OP units.
2. Discussions and Hypothesis
The benefits of the tax-deferred property contribution to partnerships as a way to dispose real estate assets are widely discussed in both the academic and the professional literature. What are not clear is how the market prices of the contributed properties reflect tax effect and how the net tax benefits are shared among property buyers and sellers. To develop a better understanding of these issues, we first gauge the potential benefit that a tax-deferred property contribution (to a partnership) may have on both the buyer and the seller. Consider a commercial property sold for cash at , of which 100% is depreciable due to a ground lease. Suppose an investor expects to hold this property for its statutory depreciation period of 39-years[4] and sells it for at least what she paid for it.[5] Let the investor’s marginal income tax rate to be and the depreciation recapture capital gains tax rate to be .[6] The investor’s net benefit, as of the time of purchase, from depreciation tax shield can be expressed as following:
,(1)
where is the after-tax discount rate. In contrast, if the same property with zero tax basis is purchased in a tax-deferred contribution to OP, the buyer (OP) will not be able to take tax-deductible depreciation until 5-12 years later when the lock-up agreement is expired. The buyer’s net loss due to the delayed tax benefit (Lb) is given by:
,(2)
where is the term of the lock-up agreement.
On the other hand, the seller’s benefit from deferring the capital gains tax of the zero tax basis property is:
,(3)
where is the capital gains tax rate and P is the seller’s historical cost for the property.To simplify the analysis, we assume , then Equation (3) becomes
(4)
It is interesting to point out that it is not always better to defer the capital gains tax. When the capital gains tax rate is lower relative to the individual’s marginal income tax rate, it becomes attractive to realize the tax gains immediately so that the buyer can take depreciations at the higher individual marginal income tax rate. Additionally, when the interest rate is lower, deferring capital gains tax also becomes less attractive because the time value of deferring tax to a later date is getting smaller while the loss from delaying the future depreciation tax shields is discounted at a lower rate.
To see the above analysis in an example, assume the property is 100% depreciable, , , and years, and let to vary from 0% to 28%. We can obtain how Lb/MV and Ds/MV change with respect to . In Figure 1, we illustrate that when capital gains tax rate falls to 15%, it is no longer optimal to defer the capital gains tax through property contribution to a partnership. This is consistent with the observations that the outstanding numbers of OP units in many UPREITs are dramatically reduced in recent years after capital gains tax rate is reduced from 28% to 20% and then to 15%.
Depending on the relative bargaining power of the buyer and seller, the transaction price in a tax-deferred property contribution will have a price discount that splits the total gains of between the buyer and seller. In Figure 1, for , BS/MV=9.5% and LB/MV=6.5%; while for , BS/MV=13.3% and LB/MV=6.0%. So it is reasonable to expect the price discount in a tax-deferred property contribution to be in the range of 6.5% to 13.3%. These numbers assume property is 100% depreciable. If land accounts for 25% of the property value, then observed price discount should be in the range of 4.9% to 10%. If the contributed property still has some depreciable tax basis instead of zero, we expect the observed price discount to be even lower than 4.9% to 10%.
Figure 1: Benefit of Tax-Deferred Contribution of Property vs. Capital Gains Tax Rate
Since the lower tax basis results in a reduced tax-deductible depreciation for the buyer, we expect, holding all other factors constant, the acquisition price to be lower for properties with a lower tax basis. If we can observe the level of tax basis in a property, our null hypothesis can be expressed as following:
H1: Holding all other factors constant, the property acquisition price is negatively associated with that property’s tax basis.
Unfortunately, as in many other tax studies, the tax basis of property is not directly observable from our data. So, as a first step, we need to develop a reasonable proxy for the property’s tax basis. One candidate proxy is the payment method of transaction (partnership units, cash, debt assumption, stocks, etc.). Since selling property for OP units does not trigger capital gains taxes like other forms of payment do, we expect properties with the adjusted tax basis below the property’s market value to self-select into using OP units as the payment method. In the meantime, we expect properties with the adjusted tax basis equal or above the property’s market value to self-select into using other forms of payment method that cause the transactions to be taxable. It shall be noted here that if only a portion of total acquisition price is paid by OP units, it is possible that the transaction may be partially taxable. For example, when the contributing partner’s net debt shifts to OP (through debt assumption by OP) as a result of property contribution is in excess of the contributed property’s adjusted tax basis plus the value of OP units, the excess will still be recognized as gains at the year of contribution, and thus is partially taxable. However, even in this case, any gains associated with the consideration paid in OP units will still be tax-deferred and thus it is expected that the property’s adjusted tax basis in a partially tax-deferred transaction will also be below the property’s market value. So we believe that the dummy variable (OP Dummy), which equals to one if OP units are used and zero otherwise, can be a good proxy for the relative position of the property’s adjusted tax basis with respect to the property’s market value. When OP units are used to pay for the entire or part of the consideration, the property’s tax basis is always lower than its full market value. Our testable null hypothesis is the following:
H2: Holding all other factors constant, the property acquisition price is negatively associated with the use of partnership units as the payment method.
A few additional considerations may complicate our analysis and deserve further discussion. First, the tax basis of OP immediately after receiving the contributed property increases by the contributed property’s inherent tax basis at the time of contribution plus any gains realized in the transaction. In a completely tax-deferred transaction, since no gain is realized, the partnership’s tax basis increases only by the property’s inherent tax basis, which can be very low (relative to property’s market price) depending on the amount of depreciation has been taken previously. What complicates our analysis further is the government rule [Internal Revenue Code Sec. 704 (c)] that requires, to the extent possible (subject to the ceiling rule), the noncontributing partners be put in the same position they would have been in had the property had tax basis equal to value. Put differently, to the extent possible, the noncontributing partner must be allocated depreciation deductions based on the property’s market value.[7] The important implication of this government rule is that property price in an acquisition involving partnership units may be not discounted as much or at all, which would decrease the likelihood of obtaining our hypothesized price effect.
Second, property sellers may use other tax deferral strategies involving no partnership units, for example, an installment sale (with or without wrap-around mortgage) and the Section 1031 exchange, to achieve the tax deferral benefit. An installment sale permits a taxpayer to defer gains recognition until the taxpayer actually receives the installment payments, while Section 1031 exchange allows property sellers to defer gains as long as a qualified replacement property is purchased within a certain period. However, unlike in the tax-deferred property contribution to partnership, the buyer’s tax basis in these strategies is not adversely impacted by the seller’s tax-deferral. To the contrary, in the case of a Section 1031 exchange, replacement property buyers may even share some of his tax deferral benefits by bidding up the property acquisition price as demonstrated by Holmes and Slade (2001). Thus, the inclusion of the observations involving these strategies can bias our analysis. However, in this study we take steps to exclude those transactions by identifying them through either SEC filings or available commercial data. So the final sample is believed to be relatively clean of these alternative tax-deferred transactions.
Third, in addition to the first type of self-selection based on the property’s tax basis relative to the property’s market value, there exists another type of self-selection based on the over- or under-valuation of REIT common stocks relative to the underlying assets owned by OP. Since each share of OP unit is exchangeable one-for-one to REIT common stock after a lock-up period, the overvalued OP will have incentives to use the overvalued OP units as currency to pay for the acquisition. The seller is likely to recognize this and request more shares of OP units. The final result will be a higher acquisition price if the acquisition price is paid by overvalued OP units. This result will bias our analysis. In this analysis we address this problem by explicitly incorporating a variable that measures over and under valuation of partnership units, details of which we describe later in the paper.
Alternatively, one can use the ratio of the value of partnership units offered by the acquirer to total acquisition price to proxy for inherent tax basis of property at the time of transaction. One common tendency is to assume that the more partnership units are used in a property transaction, the lower the adjusted tax basis relative to the contributed property’s market value is. The situation, however, is more convoluted. We use a simplified example to illustrate our point. The owner of an unencumbered building with zero tax basis but a market value of $100 can cash out his investment by first taking $99 out of a mortgage loan and then sell the remaining equity in the building for $1 of partnership units. Provided some requirements are satisfied, for example, if his post- contribution liability is not reduced by shifting the mortgage debt to OP,[8] the property seller (or new OP partner) can sell the property and defer all of the capital gains of $100, while still utilizing very little partnership units, i.e., only 1% of property’s market value. There are also additional rules and considerations in structuring the tax-deferred property sales, so that the ratio of the value of OP units may be a very noisy indicator of the level of tax basis. Our conjecture is confirmed by our conversations with the acquisition professionals employed by a major office REIT. Consequently, we expect using the ratio of the value of partnership, instead of the OP_dummy, to proxy for the unobserved tax basis, may yield less robust results.