Monetary Policies and European Office Markets Dynamics

Abstract

The purpose of this paper is to study the consequences of unconventional monetary policies on Europe’s commercial real estate market. To investigate the role of money supply on real estate markets dynamics, we use panel modelling. Our main objective is to analyze the relationships between office price indexes and monetary variables. Our panel analysis focuses on price dynamics across 16 main office markets in Europe between 2009 and 2019. We have constructed for each market a monetary index suitable for commercial real estate. Our robust results, corrected for the presence of errors-in-variables, report that a positive relationship exists between the global money supply and office prices. Moreover, the largest markets seem more affected by the massive injections of liquidity from the central banks, especially those from the European Central Bank (ECB). We highlight significant differences among office markets in Europe.

Keywords:

• Real estate
• office prices dynamics
• Monetary supply
• European markets
• panel estimation

Acknowledgments

We would like to thank an anonymous referee and the editor William Hardin for helpful comments. We also thank Raphael Languillon for comments and suggestions on a previous and preliminary version of this study. The usual disclaimer applies.

Notes

1 Estimation based on market data from BNP Paribas Real Estate.

2 The European debt crisis also had a strong influence on the level of international investors during this period.

3 In this approach, the explanatory variables are supposed to have the same impact on the endogenous variable, while the heterogeneity between the markets is captured by the constants (the pooling test is rejected).

4 Brounen and Jennen (2009) find a marginal difference between output and employment in their study on the European office market.

5 As countries have various definitions of broad money, a simple sum of monetary aggregates would under-represent countries with narrow definitions of M2.

6 M2 is a money aggregate that measures the money supply. It is generally defined as the sum of currency and coins, demand deposits, money markets and savings deposits.

7 Between 3.5 and 5 months for the execution (Devaney and Scofield (2014)), plus the period between the change in monetary policy and the investment decision.

8 For more details, see Cragg (1994) and Carmichael and Coën (2008).

9 To detect the presence of EIV, we can also proceed in two steps using artificial regression as proposed by Davidson and MacKinnon (2004). First, we compute estimates of EIV, W, as the residuals of k OLS regressions with observed ⁁ variables, X as dependent variables and the instruments as regressors (higher moments of X such as X − X⁁ = W⁁, with X, estimates of the true variables. Second, we add estimates of EIV as additional regressors (see the appendix ⁁ of Carmichael and Coën, 2008, for more details). The standard errors are calculated following a standard approach taking into account instruments: the same as the approach described by Carmichael and Coën (2008) (for more details see their appendix) and Davidson and MacKinnon (2004) (artificial regression techniques) and then applying the GMM procedure introduced by Hansen (1982). Thus, we obtain robust statistics with a correction for EIV in a GMM framework.

10 Because of data availability we use annual data for the post Global Financial Crisis period (2009-2019). As mentioned in the Summary statistics (Table 3), we have 11 observations for each market and each variable. Therefore, we use panel modelling and correct for EIV in a GMM procedure. Nevertheless, the results could be slightly sensitive to data availability for small samples.

11 This group is composed of London, Paris and Lyon. London and Paris are the two biggest markets in Europe, both in terms of stock and of investment volume. If we consider only the French markets (Paris and Lyon), we obtain the same conclusions. The estimates of MI-Global is highly significant at 1% (0.479; t-stat:3.708) while MI-Euro is not (0.583; t-stat: 0.922). Detailed results are available upon request.

12 As the Central Bank policies are different for these two countries, the lack of significance from the European index might have been expected. However, the results have also been confirmed when separating the UK and France. These results are available on request.