Technical efficiency and economies of scale: A non-parametric analysis of REIT operating efficiency

Abstract

This study measures technical efficiency and economies of scale for real estate investment trusts (REITs) by employing data envelopment analysis (DEA), a linear-programming technique. Using data from the National Association of Real Estate Investment Trusts (NAREITs) for the years 1992–1996, we find that REITs are technically inefficient, and the inefficiencies are a result of both poor input utilization and failure to operate at constant returns to scale. With respect to scale inefficiency, most REITs are operating at increasing returns to scale, suggesting that REITs could improve performance through expansion. Moreover, we employ regression analysis to determine what characteristics influence the efficiency measures obtained. The results show that internal REIT management is positively related to all measures of efficiency. Increasing leverage is negatively related to REIT input utilization. Finally, increasing REIT diversification across property types enhances scale efficiency (SE) but reduces input usage efficiency.

Introduction

Real estate investment trust (REIT) managers, security analysts, investors, and academicians are interested in measuring the operating efficiency of REITs and the impact of various REIT characteristics on operating efficiency. Such information can prove useful in improving REIT operating efficiency and ultimately increasing long-run performance.

In classical economic theory, efficiency is generally determined by estimating a long-run average cost curve and determining whether firms experience economies of scale. A firm is scale efficient if it operates at the bottom of the assumed U-shaped average cost curve. That is, a firm does not decrease its average costs by increasing or decreasing output levels. For REITs, if the outputs are measured in terms of total assets, scale efficiency (SE) exists when total assets cannot be altered without increasing average costs.

Several studies have indirectly tested for the existence of scale economies in REITs. These studies have found mixed results. Allen and Sirmans (1987) examine the price effects of REITs that have merged. They find positive price effects that are attributed to better asset utilization and that suggests potential scale economies for REITs. Several studies on the wealth effects of property acquisition by REITs show no compelling evidence for the existence of scale economies (see Corgel et al., 1995, for a review). Both McIntosh et al., 1991, McIntosh et al., 1995 find evidence against the existence of scale economies in REITs. The prior study finds that smaller firms earn higher risk-adjusted rates of return. The latter shows positive price impacts for REIT sell-offs, indicating benefits associated with reductions in size.

While these studies have scale implications for REITs, the academic literature at this time includes only one study that directly measures economies of scale for REITs. That study (Bers and Springer, 1997) estimates economies of scale for REITs using data from the National Association of Real Estate Investment Trusts (NAREITs) for the years 1992, 1993, and 1994 (NAREIT, 1993, NAREIT, 1994, NAREIT, 1995). The authors estimate a classical translog cost function and examine the elasticity of average cost with respect to output to determine if scale economies are present. The results show the existence of significant scale economies, suggesting that gains in productivity can be realized through expansion.¹ Also, the authors find that REIT characteristics affect the measured efficiency levels.

While informative, additional information is needed. First, Bers and Springer provide no inefficiency estimates. Their traditional regression-based (non-efficient frontier) technique has a single error term, and assumes that each REIT is operating on its efficient cost frontier, where the scale economies are correctly defined (Berger et al., 1993). Studies in banking, insurance, and real estate have revealed that firms operate off of the efficient cost frontier, or are “X-inefficient” (see Anderson et al., 1998). In fact, the losses in efficiency from failure to be X-efficient have been shown to be more substantial than losses resulting from failure to be scale efficient (Berger et al., 1993). Moreover, Springer and Bers note that their results may be biased due to errors in the functional form. Thus, to further examine the operating efficiency of REITs and to test for the presence of economies of scale, it is appropriate to examine X-inefficiencies using an approach that does not require choosing a functional form.

While several techniques are available for analyzing X-efficiencies, this study uses data envelopment analysis (DEA). DEA is a linear-programming technique used to construct an efficient cost frontier.² This particular application of DEA allows for a measurement of overall technical efficiency (OTE). A firm has technical efficiency when input usage (costs) cannot be decreased without decreasing output. In the context of X-inefficiency, OTE is the product of the deviation from the efficient cost frontier due to inefficient input utilization (pure technical inefficiency) and the deviation from failure to operate at constant returns to scale (scale inefficiency).

This estimation technique has both advantages and disadvantages relative to parametric efficient frontier techniques such as the stochastic frontier approach. The main advantage is that DEA allows efficiency and economies of scale estimations without specifying a functional form, while being able to handle a multiple input multiple output production process. The most cited disadvantage of the DEA technique is that it does not allow for deviations from the efficient frontier to be a function of random error. As such, DEA can produce results that are sensitive to outliers, model specification, and data errors. DEA is employed here as prior research indicates problems in identifying and estimating various functional forms such as the translog and Cobb–Douglas cost functions.

Once the various efficiency measures have been estimated using DEA, a regression analysis is used to determine the impact of various REIT characteristics on the DEA efficiency measures.

The results show that X-inefficiencies exist in the REIT industry and that poor utilization of inputs is the dominant factor contributing to these inefficiencies. Scale economies are shown to exist, thus further corroborating the results of Bers and Springer, 1997, Bers and Springer, 1998. Also, larger REITs are more efficient than smaller REITs for all efficiency measures for most of the years tested. The regression results suggest that internal REIT management enhances all measures of efficiency, while increasing leverage decreases pure technical efficiency (PTE). Finally, increasing REIT diversification is positively related to SE, but negatively related to PTE.

The remainder of the paper is organized as follows: Section 2 discusses the DEA technique; Section 3 describes the data; Section 4 is a discussion of the efficiency results estimated by DEA; Section 5 discusses the impacts of various REIT characteristics on the estimated efficiency measures; and finally, Section 6 summarizes the results and concludes the paper.