|
Article Excerpt
The possible existence of investor clientele groups has received little attention in the real estate finance literature. In this paper we develop a clientele model, which in equilibrium produces a clustering of investors by tax characteristics. Low-tax-bracket investors are concentrated in low-value rental housing that attracts rents which are high in relation to property values. On the other hand, only high-tax-bracket investors will be observed in high-value rental housing, and they charge rents that are low in relation to property values. An empirical model is specified and estimated using a cross section of investors in Australian private rental housing markets. Investor clienteles are detected among property investors, though there is a weak diversification effect indicating that clientele effects may be stronger among single property investors.
**********
This paper investigates whether investor clientele effects characterize real estate investments in rental housing. In their seminal paper concerning the valuation of shares, Miller and Modigliani (1961) acknowledge that tax-driven clientele effects could result in low-dividend-yield shares selling at a premium as returns are "packaged" in the form of capital gains that are tax advantaged. But Miller and Modigliani question the empirical significance of such effects for investors in tax brackets other than the very top bracket. A plethora of empirical studies have explored whether the separation of shareholders by dividend yield is as exact as the raw tax differentials might seem to suggest. A parallel literature has sought evidence of tax-disadvantaged dividends leaving a detectable track in the prices of shares. (1)
In physical asset markets such as real estate, where again capital gains are typically given favorable tax treatment, the possible existence of clientele effects has received relatively little attention. This is despite evidence suggesting that rates of house price appreciation differ between locations (Kiel and Carson 1990) and between higher and lower priced homes (Seward, Delaney and Smith 1992, Haurin, Hendershott and Wachter 1996, Smith and Ho 1996). (2) Mills and Kaiser (1999) find that over the 1930-1976 period the rent-to-value ratio has varied inversely and consistently with the inflation rate. They conclude that "the reason for this relationship is that landlords require a roughly constant total return to induce them to hold rental housing" (Mills and Kaiser 1999, p. 84). Since capital gains are a large part of investor returns, submarkets where expected house price appreciation is high will feature competition between investors from different tax brackets that should push down current rent-to-value ratios. In submarkets where expected house price appreciation is low, the exit of investors will force current rent-to-value ratios higher.
Steele (1993) recognizes tax-driven clientele effects in Canadian rental markets where (in the early 1990s) only 50% of investors' capital gains were taxable. Thus, the "larger the capital gain component relative to the net rental income component of rental real estate, the more attractive is investment in rental real estate to high-bracket investors relative to low-bracket investors" (Steele 1993, p. 114). Implications can be drawn for ownership of the stock. If corporations are typically subject to lower marginal rates of tax, then they will be uncompetitive with high income individual investors and displaced by the latter, particularly during inflationary periods.
Both Steele (1993) and Titman (1982) recognize that during inflationary periods rents will not keep pace with inflation because landlords' capital gains receive tax advantaged treatment. Kiefer (1978), in simulating the cost of housing services for renters and buyers at different income levels, also finds that accelerating rates of inflation benefit investors at higher income levels due to the asymmetric tax treatment of income and capital gains. It can then be advantageous for tenants to locate in submarkets where high-income landlords are concentrated (Keifer 1978, p. 138). This tax arbitrage view of tenure choice recognizes that tax expenditures can allow high-tax-bracket investors to lease rental housing services to low-tax-bracket individuals, at a rent which leaves the latter financially better off renting rather than owning (Anstie, Findlay and Harper 1983, Gordon, Hines and Summers 1987, Follain and Ling 1988, Hendershott 1988, Wood 2001). But these models typically assume that marginal investors belong to the very top bracket, a view for which Narwold (1992) provides both theoretical and empirical support. In contrast, Wood and Watson (2001) find that investors from the top tax bracket are a minority of all investors in Australia. Furthermore, they find concentrations of high-tax-bracket investors in high-value rental housing and concentrations of low-tax-bracket investors in low-value rental housing.
In this paper we develop an investor clientele model that offers an explanation for these patterns. We utilize a micro database to detect investor clientele effects in Australia, where capital gains are given favorable tax treatment because only real capital gains are taxed on realization. (3) In the next section the model is presented. It is followed by an explanation of the empirical specification used for detecting investor clientele effects as well as the econometric issues that we address. The data source is then described and our findings are presented. A concluding section summarizes the main points and offers some thoughts on policy implications. The principal finding of the research is that marginal tax rates affect the gross and net rental yields of investors' rental property portfolios, even after controlling for risk, operating costs and market segmentation.
Analysis
Consider an investor who has acquired a residential property investment in a housing market where house prices and rents appreciate at the constant uniform rate [[pi].sub.h]. Given perfect capital markets, zero transaction costs and assuming that preferential tax treatment of capital gains takes the form of tax exemption, the investor's user cost expression can be written as (4)
R = i + v - [[pi].sub.h]/(1 - [tau]) (1)
where R is the reservation rental rate, i is the market rate of interest, v is operating cost (per dollar of capital value) and [tau] is the investor's marginal income tax rate. (5) With capital gains tax exempt, the final term in Equation (1) can be interpreted as the pretax rental rate equivalent of the nontaxable component of the landlord's rate of return (Litzenberger and Sosin 1978). If the expected rate of capital appreciation falls by one percentage point, R must rise by enough to cover not only the loss of capital gains, but also by enough to pay taxes on the compensating increase in rents, that is, by 1/(1 - [tau]) percentage points.
With an exogenously determined market rental rate and a homogenous housing stock, Equation (1) implies that landlord's excess returns will be an increasing, convex function of their marginal tax rate. (6) The intuition behind this conclusion becomes apparent on rewriting Equation (1) as
(1 - [tau])(R - [i + v]) = -[[pi].sub.h]. (2)
The left-hand side is an after-tax net rental rate measure, defined inclusive of the investor's nominal cost of capital. At positive rates of capital gain, the after-tax net rental rate must be in deficit when the investor secures a real rate of return equal to the real rate of interest. An increase in [tau] will reduce this after-tax deficit; ceteris paribus, it raises the investor's real rate of return above the real interest rate. (7)
Let us now relax the assumption of a uniform housing market. Consider a housing market comprised of two submarkets X and Y offering rental-housing services only. Housing in these two submarkets is offered at the same rent and is identical in all respects other than location. However, expected rates of house price appreciation are higher in submarket X ([[pi].sub.x]) than submarket Y ([[pi].sub.y]).
This assumption is consistent with numerous empirical studies. (8) There are two income tax brackets: high ([[tau].sub.H]) and low ([[tau].sub.L]). Rent levels are sticky (9) and the stock of housing is fixed in the short run. The reservation rental rates of potential investors from the two tax brackets are given by
[R.sub.k](j) = i + v - [[pi].sub.j]/[1 - [[tau].sub.k]] k = H, L; j = X, Y. (3)
Investors from both tax brackets will be prepared to pay a price premium to acquire rental housing in submarket X, but high-tax-bracket investors will be willing to offer the higher premium because the benefit derived from tax-exempt capital gains is greater. In the short run, market values in submarket X are then driven up and rental housing in Y must sell at a discount to attract investors. Thus, expected capital gains tax differentials are capitalized into market values. As a consequence, relative market rental rates in submarket X decline. The higher the marginal tax rate in the highest tax bracket, the more the market rental rate in X can be reduced before this submarket's tax-advantaged package of returns is offset. (10)
A pooling equilibrium (11) at a uniform market rental rate cannot be sustained because posttax rates of return will be higher in submarket X and high-tax-bracket investors holding properties...
|
