Corporate Governance and Market Valuation of Publicly Traded Real Estate Companies: Evidence from Europe

Tobin’s Q represents the relation between market value and replacement costs and serves as a popular valuation measure in financial literature.

Including more sophisticated empirical studies, such as the ones of queryAgrawal et al. (1996), Beiner et al. (2006) and Black et al. (2006). Other important finance literature on the topic includes Gompers et al. (2003), Bebchuk and Cohen (2005), Cremers and Nair (2005), Bebchuk et al. (2005), Core et al. (2006) and Faleye (2007).

Other real estate studies investigating wealth effects of corporate governance use alternative measures of firm performance, such as ROA and ROE. Analyzing a sample of 38 US REITs between 1972 and 1981, Solt and Miller (1985) find evidence that the payment of incentive fees is generally positively related to financial performance, as measured by ROA and ROE. In turn, Ghosh and Sirmans (2003) provide results indicating that greater board independence enhances ROE-levels of US REITs whereas greater CEO stock ownership appears to have an adverse effect on firm performance. In addition, Feng et al. (2005) find support for the view that board structure plays an important role in financial performance of US REITs. Their findings indicate that a smaller and more independent board of directors is associated with higher levels of ROA.

Similarly, Ghosh and Sirmans (2003) and Feng et al. (2005) report a significantly positive association between greater outsider representation on REIT boards and performance, as measured by ROE and ROA.

For an excellent review of econometric problems in empirical studies on corporate governance see Börsch-Supan and Köke (2002).

Empirical studies that find evidence for endogeneity of corporate governance mechanisms include Agrawal and Knoeber (1996), Chung and Pruitt (1996), Loderer and Martin (1997), Eisenberg et al. (1998), Bhagat and Black (2002), Ghosh and Sirmans (2003), Klapper and Love (2004), Beiner et al. (2006), Black et al. (2006) and Ghosh and Sirmans (2006).

Simultaneous equations models are models in which two or more endogenous variables are determined simultaneously; see Hausman (1983); Davidson and MacKinnon (1993).

As Davidson and MacKinnon (1993) indicate there are numerous variants of instrumental variable estimation methods in econometrics, e.g. two-stage least squares (2SLS), three-stage least squares (3SLS) and the generalized method of moments (GMM).

These aggregate ratings are likely to be inadequate proxies for corporate governance for three basic reasons. First, important governance characteristics might not be considered. Second, the construction of the index is necessarily biased to the extent that weights are more or less arbitrarily assigned to certain governance indicators; see also Bruno and Claessens (2006). And third, corporate governance indices in terms of aggregate measures per se do not allow capturing the effect of dynamic interactions among governance mechanisms in an empirical model. Sonnenfeld (2004) criticizes the quality of the governance metrics published by professional corporate governance data and service providers. He argues that they look merely at public records to score firms on their governance effectiveness by using simplistic checklists of standards or metrics based heavily upon myths, rather than on genuine research. He further points out that they also may cross the line from being independent raters to becoming active consultants for the firms they study leading to doubts about their objective credibility. Finally and most importantly, he holds the opinion that their methods do not deliver reliable and accurate governance ratings.

According to Bebchuk et al. (2005) governance quality could well be measured by focusing on a few provisions that matter instead of using a very broad index that includes numerous less important measures which only serve to introduce noise.

The equations of the system can be considered interdependent. Estimating each equation in isolation would not be sufficient to determine the actual meaning of the statistical relationships between the variables; see Hausman (1983).

3SLS was originally introduced by Zellner and Theil (1962). This estimation method accounts for joint endogeneity and basically relies on instrumental variables as well as the least squares method. 3SLS is a so-called full system estimation method, indicating that all equations of the system are estimated jointly; see Schmidt (1976); Davidson and MacKinnon (1993); Greene (2003).

The USD 50 million threshold corresponds to a common criterion of admittance to recognized real estate indices such as GPR. For reasons of comparability and data availability companies with a smaller market capitalization were excluded from the sample.

The underlying concept goes back to Brainard and Tobin (1968).

Studies using Tobin’s Q implicitly assume that capital markets are efficient and correctly reflect the value of companies. Even though this might not correspond to the truth, Tobin’s Q might be the best performance measure available according to Börsch-Supan and Köke (2002).

Companies reporting in accordance with IFRS basically have the option to state their investment properties at cost or at fair value. Nevertheless, those companies deciding to state their investment properties at cost are required disclose the fair value of these properties in the notes of the balance sheet.

For a similar line of argumentation see Capozza and Seguin (2000), Benveniste et al. (2001), Gentry and Mayer (2003), Hartzell et al. (2004).

For a general and intuitive description of the Hausman test for endogeneity see Pindyck and Rubinfeld (1998) and Wooldridge (2000).

According to Cronqvist et al. (2001) gains from specialization with respect to particular property types arise from the specific knowledge of the management team on the individual properties, how to value them and about potential buyers and sellers in the market.

Among others, La Porta et al. (2002), Doidge et al. (2004), Klapper and Love (2004), Durnev and Kim (2005) and Agrawal et al. (2007) emphasize the importance of country-level effects in predicting corporate governance. Therefore, the COUNTRY indicator variables will be utilized in equations 1 to 7.

For a similar argumentation see Agrawal and Knoeber (1996).

Numerous empirical studies, such as Titman and Wessels (1988) and Fama and French (2002), were able to find evidence for an inverse relationship between profitability and leverage, eventually supporting the pecking order theory.

For a more detailed description of the single test statistics see Kohl (2009).