Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Can Linear Predictability Models Time Bull and Bear Real Estate Markets? Out-of-Sample Evidence from REIT Portfolios


A recent literature has shown that REIT returns contain strong evidence of bull and bear dynamic regimes that may be best captured using nonlinear econometric models of the Markov switching type. In fact, REIT returns would display regime shifts that are more abrupt and persistent than in the case of other asset classes. In this paper we ask whether and how simple linear predictability models of the vector autoregressive (VAR) type may be extended to capture the bull and bear patterns typical of many asset classes, including REITs. We find that nonlinearities are so deep that it is impossibile for a large family of VAR models to either produce similar portfolio weights or to yield realized, ex-post out-of-sample long-horizon portfolio performances that may compete with those typical of bull and bear models. A typical investor with intermediate risk aversion and a 5-year horizon ought to be ready to pay an annual fee of up to 5.7 % to have access to forecasts of REIT returns that take their bull and bear dynamics into account instead of simpler, linear forecast.

This is a preview of subscription content, log in to check access.

We’re sorry, something doesn't seem to be working properly.

Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5


  1. 1.

    Two concurrent explanations exist for the existence of predictability in asset returns. First, numerous asset pricing papers show that predictability results from business cycle movements and changes in investors’ perceptions of risk that are reflected in time-varying risk premia. Second, other research shows that predictability may reflect an inefficient market populated with overreacting and irrational investors. Ling et al. (2000) contain references to this debate. In our paper, we take the existence of both linear and nonlinear predictability patterns as an empirical fact and investigate whether linearizing such patterns may produce the same improvement in portfolio performance (if any) as nonlinear ones would.

  2. 2.

    However Campbell and Thompson (2008) defend the finding of predictability in stock returns and show that it is possible to beat the historical average by imposing restrictions on the signs of regression coefficients and return forecasts.

  3. 3.

    In the United States, the REIT market offers investors a way to invest in real estate without problems of illiquidity, intense management, and large lot size/high unit cost (see Ciochetti et al. 2002).

  4. 4.

    As typical of the literature since Nelling and Gyourko (1998), our VARs also allow the possibility that lagged asset returns forecast future asset returns, also reflecting cross-asset patterns (i.e., it is allowed that past stock or bond returns may forecast subsequent REIT returns, etc.).

  5. 5.

    However, in the short-run investment horizon case, the a large set of VAR models provide a superior realized, out-of-sample CER than a three-state MRSM does.

  6. 6.

    Fugazza et al. (2009) report that for long-horizon investors that ignore predictability, ignoring real estate assets results in an approximate CER loss of 200 basis point, which climbs up to 400 basis point under linear predictability. See also welfare loss estimates of omitting real estate from the asset menu in MacKinnon and Al Zaman (2009).

  7. 7.

    Other recent papers that have featured either nonlinearities or breaks in patterns of REIT return predictability are Guirguis et al. (2005) and Serrano and Hoesli (2007). In the finance literature, Park (2010) argues that in a sub-sample of US data that includes the 1990s, the predictive power of the dividend yield disappears. Henkel et al. (2011) find that stock return predictability occurs only during economic contractions but disappears during expansions.

  8. 8.

    The 3 derives from the fact that we study the case of \(p=1,\) 2, and 4; the addition of one to \(2^{M}\) depends on the fact that our models always include as a default p lags of past asset returns.

  9. 9.

    Notably, (3) can be interpreted as a MRSM under a specific restriction on the transition probabilities, i.e., when an absorbing state in the transition matrix is imposed, such that \(k=1\) obtains.

  10. 10.

    Notice that the exact value to which \(\delta \) is calibrated will only affect the intertemporal trade-off between current and future consumption and not the resulting optimal portfolio weights, which are instead our main focus.

  11. 11.

    Detailed evidence and cross-correlogram plots are available upon request.

  12. 12.

    Davies (1977) has derived an upper bound for the significance level of the standard likelihood ratio test under nuisance parameters:

    \(\Pr \left( LRT>x\right) \leq \Pr \left( \chi _{1}^{2}>x\right) +\sqrt{2x} \exp \left( -\frac{x}{2}\right) \left[ \Gamma \left( \frac{1}{2}\right) \right] ^{-1}. \)

    This bound holds if the log-likelihood has a single peak.

  13. 13.

    For instance, a three-state MSI leads to a maximized log-likelihood of 5026.22 to be contrasted with a single-state maximized log-likelihood of 4725.92. The increment of 300.30 log-likelihood “points” originates an LRT statistic of 600.6 that however corresponds to a loss of 16 degree of freedom because a MSIH(3) implies the estimation of 28 parameters against the 14 parameters required by a simple Gaussian IID model. Moreover, one has to take into account the nuisance parameter problems related to the fact that when the Markov chain becomes degenerate, a few parametes under the MSIH framework cannot be identified and the log-likelihood is unbounded (see Hamilton 1994). Even taking these issues into account, the p-value associated with the Davies’ bound is 3.4e-06.

  14. 14.

    The estimated transition matrix reveals that if in month t the markets are in a bear state, in month \(t+1\) there is an almost 96 % probability of still being a bear regime.

  15. 15.

    The reason is that under a Gaussian IID model, the dynamics of portfolio weights does not come from movements in the value of current and past predictors, but simply from the fact that the vector of expected returns and the covariance matrix need to be recursively estimated over time.

  16. 16.

    The 90 % empirical range equals the difference between the two values from the time series of weights that leave the 5 % of the recursive holdings in each of the two tails. This measure is computed and reported as a proxy of dispersion that is especially useful when the weights have non-normal distributions.

  17. 17.

    Of course, such lack of important differences between average weights over our OOS, 1995–2009 period does not imply that there cannot be any differences in the time-varying path of the weights, as commented with reference to Figs. 3, 4 and 5. Some minor differences in the average weights assigned to long-term bonds can be found between the Gaussian IID model and the VAR frameworks, with the latter implying higher average weights than the former does.

  18. 18.

    We have also tabulated summary statistics similar to those in Table 4 for the cases of \(\gamma =2\) and 10, reaching qualitatively similar conclusions. Detailed results are available upon request.

  19. 19.

    Here there are also some important differences within the VAR class, as for instance a full VAR(1) generates larger and more unstable demands than the best performing VAR does. We conjecture that this latent instability in the relationship between long- and short-term demands is responsible for the disappointing OOS performance of the full VAR when compared to a more parsimonious VAR.

  20. 20.

    In Tables 5 and 6, when these are tabled among the best ranked models, we have boldfaced the highest Sharpe ratios, mean realized returns, and skewness as well as as the lowest realized standard deviation of realized returns and excess kurtosis among all the models entertained in the paper.

  21. 21.

    These 95 % confidence intervals have been computed by applying a block bootstrap to the time series of OOS recursive, realized performance statistics, such as realized utilities, and returns.

  22. 22.

    In fact, realized power utility depends on the entire realized density of portfolio returns and as such on all realized moments, not only mean, standard deviation, skewness, and kurtosis (see Campbell and Viceira 2002).

  23. 23.

    In more detail, it is easy to show that

    \(\kappa _{0,T}(\gamma )\equiv v_{T}^{1-\gamma }\left[ (1-\gamma )^{-1}-1-\frac{1}{2}\gamma -\frac{1}{6}\gamma (\gamma +1)-\frac{1}{24}\gamma (\gamma+1)(\gamma +2)\right]\)

    \(\begin{array}{rll}\kappa _{1,T}(\gamma ) &\equiv& \frac{1}{6}v_{T}^{-\gamma }\left[ 6+6\gamma +3\gamma (\gamma +1)+\gamma (\gamma +1)(\gamma +2)\right] >0\quad \kappa _{2,T}(\gamma )\equiv -\frac{1}{4}\gamma v_{T}^{-(1+\gamma )} \\&& \qquad\qquad\qquad\quad \left[2+2(\gamma +1)+(\gamma +1)(\gamma +2)\right] <0 \\ \kappa _{3,T}(\gamma ) &\equiv &\frac{1}{6}\gamma (\gamma +1)(\gamma+3)v_{T}^{-(2+\gamma )}>0\qquad \kappa _{4,T}(\gamma )\equiv -\frac{1}{24} \gamma (\gamma +1)(\gamma +2)v_{T}^{-(3+\gamma )}<0.\end{array}\)

    This approach is only heuristic because our utility function in (4) links welfare to consumption flows and not directly to realized portfolio returns.

  24. 24.

    This is consistent with the fact that most of the predictors often included in the best performing VARs express forecasting power for REIT returns, such as the REIT cap rate, the rate of growth in housing starts, or mortgage rates.


  1. Ang, A., & Bekaert, G. (2002). International asset allocation with regime shifts. Review of Financial Studies, 15(4), 1137–1187.

  2. Ang, A., & Bekaert, G. (2007). Stock return predictability: is it there? Review of Financial Studies, 20(3), 651–707.

  3. Ang, A., Bekaert, G., Liu, J. (2005). Why stocks may disappoint? Journal of Financial Economics, 76(3), 471–508.

  4. Brennan, M., Schwartz, M., Lagnado, R. (1997). Strategic asset allocation. Journal of Economic Dynamics and Control, 21(8–9), 1377–1403.

  5. Campbell, J., & Thompson, S. (2008). Predicting excess stock returns out of sample: can anything beat the historical average? Review of Financial Studies, 21(4), 1509–1531.

  6. Campbell, J.Y., & Viceira, L. (2002). Strategic asset allocation: portfolio choice for long-term investors. New York: Oxford University Press.

  7. Campbell, J.Y., Chan, Y.L., Viceira, L.M. (2003). A multivariate model of strategic asset allocation. Journal of Financial Economics, 67, 41–80.

  8. Case, B., Guidolin, M., Yildirim, Y. (2011). Markov switching dynamics in REIT returns: univariate and multivariate evidence on forecasting performance. Real Estate Economics. doi:10.2139/ssrn.2138256.

  9. Cauley, S., Pavlov, A., Schwartz, E. (2007). Homeownership as a constraint on asset allocation. Journal of Real Estate Finance and Economics, 34(3), 283–311.

  10. Chandrashekaran, V. (1999). Time-series properties and diversification benefits of REIT returns. Journal of Real Estate Research, 17(1), 91–112.

  11. Chong, L., Miffre, J., Stevenson, S. (2009). Conditional correlations and real estate investment trusts. Journal of Real Estate Portfolio Management, 15(2), 173–184.

  12. Chun, G., Sa-Aadu, J., Shilling, J. (2004). The role of real estate in an institutional’s investor portfolio. Journal of Real Estate Finance and Economics, 29(3), 295–320.

  13. Ciochetti, B., Craft, T., Shilling, J. (2002). Istitutional investors’ preferences for REIT stocks. Real Estate Economics, 30(4), 567–593.

  14. Clayton, J., & MacKinnon, G. (2001). The time-varying nature of the link between REIT, real estate and financial asset returns. Journal of Real Estate Portfolio Management, 7(1), 43–54.

  15. Cotter, J., & Stevenson, S. (2006). Multivariate modeling of daily REIT volatility. Journal of Real Estate Finance and Economics, 32(3), 305–325.

  16. Cotter, J., & Stevenson, S. (2007). Uncovering volatility dynamics in daily REIT returns. Journal of Real Estate Portfolio Management, 13(2), 119–128.

  17. Crawford, G., & Fratantoni, M. (2003). Assessing the forecasting performance of regime switching, ARIMA and GARCH models of house prices. Real Estate Economics, 31(2), 223–243.

  18. Davies, R. (1977). Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika, 74(1), 33–43.

  19. Feldman, B. (2003). Investment policy for securitized and direct real estate. Journal of Portfolio Management (Special Real Estate Issue), 29(5), 112–121.

  20. Fugazza, C., Guidolin, M., Nicodano, G. (2007). Investing for the long-run in european real-estate. The Journal of Real Estate Finance and Economics, 34(1), 35–80.

  21. Fugazza, C., Guidolin, M., Nicodano, G. (2009). Time and risk diversification in real estate investments: assessing the ex-post economic values. Real Estate Economics, 37(3), 341–381.

  22. Fugazza, C., Guidolin, M., Nicodano, G. (2011). 1/N and long run optimal portfolios: results for mixed asset menus. Federal Reserve Bank of St. Louis working paper No. 2009-001.

  23. Ghysels, E., Plazzi, A., Valkanov, R. (2007). Valuation in US commercial real estate. European Financial Management, 13(3), 472–497.

  24. Glascock, J. (1991). Market conditions, risk, and real estate portfolio returns: some empirical evidence. Journal of Real Estate Finance and Economics, 4(4), 367–373.

  25. Guidolin, M., & Hyde, S. (2012). Simple VARs cannot approximate Markov switching asset allocation decisions: an out-of-sample assessment. Computational Statistics and Data Analysis, 56(11), 3546–3566.

  26. Guidolin, M., & Timmermann, A. (2007). Asset allocation under multivariate regime switching. Journal of Economic Dynamics and Control, 31(11), 3502–3544.

  27. Guidolin, M., & Timmermann, A. (2008). International asset allocation under regime switching, skew and kurtosis preferences. Review of Financial Studies, 21(2), 889–935.

  28. Guirguis, H., Giannikos, C., Anderson, R. (2005). The us housing market: asset pricing forecasts using time varying coefficients. Journal of Real Estate Finance and Economics, 30(1), 33–53.

  29. Hamilton, J.D. (1994). Time series analysis. Princeton: Princeton University Press.

  30. Henkel, S., Martin, J., Nardari, F. (2011). Time-varying short-horizon predictability. Journal of Financial Economics, 99(3), 560–580.

  31. Hoesli, M., Lizieri, C., MacGregor, B. (2008). The inflation hedging characteristics of US and UK investments: a multi-factor error correction approach. Journal of Real Estate Finance and Economics, 36(2), 183–206.

  32. Hudson-Wilson, S., Fabozzi, F., Gordon, J. (2005). Why real estate? An expanding role for institutional investors. The Journal of Portfolio Management (Special Real Estate Issue), 29(5), 12–25.

  33. Hung, K., Onayev, Z., Tu, C. (2008). Time-varying diversification effect of real estate in institutional portfolio. Journal of Real Estate Portfolio Management, 14(4), 241–261.

  34. Hutson, E., & Stevenson, S. (2008). Asymmetry in REIT returns. Journal of Real Estate Portfolio Management, 14(2), 105–124.

  35. Ingersoll, J. (1987). Theory of financial decision making. Totowa: Rowman & Littlefield.

  36. Jirasakuldech, B., Campbell, R., Emekter, R. (2009). Conditional volatility of equity real estate investment trust returns: a pre- and post-1993 comparison. Journal of Real Estate Finance and Economics, 38(2), 137–154.

  37. Kallberg, J., Crocker, H., Wylie, G. (1996). The role of real estate in the portfolio allocation process. Real Estate Economics, 24, 359–377.

  38. Lee, S., & Stevenson, S. (2005). The case for REITs in the mixed-asset portfolio in the short and long run. Journal of Real Estate Portfolio Management, 11(1), 55–80.

  39. Ling, D., & Naranjio, A. (1997). Economic risk factors and commercial real estate returns. Journal of Real Estate Finance and Economics, 14(3), 283–301.

  40. Ling, D., Naranjio, A., Ryngaert, M. (2000). The predictability of equity reit returns: time variation and economic significance. Journal of Real Estate Finance and Economics, 20(2), 117–136.

  41. Liow, K., Chen, Z., Liu, J. (2011). Multiple regimes and volatility transmission in securitized real estate markets. Journal of Real Estate Finance and Economics, 42(3), 295–328.

  42. Liow, K., Ho, H., Ibrahim, M., Chen, Z. (2009). Correlation and volatility dynamics in international real estate securities markets. Journal of Real Estate Finance and Economics, 39(2), 202–223.

  43. Liow, K., Zhu, H., Ho, D., Addae-Dapaah, K. (2005). Regime changes in international securitized property markets. Journal of Real Estate Portfolio Management, 11(2), 147–165.

  44. MacKinnon, G., & Al Zaman, A. (2009). Real estate for the long term: the effect of return predictability on long-horizon allocations. Real Estate Economics, 37(1), 117–153.

  45. Mei, J., & Liu, C. (1994). The predictability of real estate returns and market timing. Journal of Real Estate Finance and Economics, 8(2), 115–135.

  46. Myer, N., & Webb, J. (1994). Statistical properties of returns: financial assets versus commercial real estate. Journal of Real Estate Finance and Economics, 8(3), 267–282.

  47. Nelling, E., & Gyourko, J. (1998). The predictability of equity REIT returns. Journal of Real Estate Research, 20(3), 117–136.

  48. Nicodano, G., & Guidolin, M. (2009). Small Caps in international equity portfolios: the effects of variance risk. Annals of Finance, 5(1), 15–48.

  49. Okunev, J., Wilson, P., Zurbruegg, R. (2000). The casual relationship between real estate and stock markets. Journal of Real Estate Finance and Economics, 21(3), 251–261.

  50. Park, C. (2010). When does the dividend-price ratio predict stock returns? Journal of Empirical Finance, 17(1), 81–101.

  51. Peng, L., & Schulz, R. (2012). Does the diversification potential of securitized real estate vary over time and should investors care? Journal of Real Estate Finance and Economics, 1–31. doi:10.1007/s11146-011-9357-5.

  52. Pesaran, H., & Timmermann, A. (1995). Predictability of stock returns: robustness and economic significance. Journal of Finance, 50(4), 1201–1228.

  53. Rerhing, C. (2012). Real estate in a mixed-asset portfolio: the role of the investment horizon. Real Estate Economics, 40(1), 65–95.

  54. Sa-Aadu, J., Shilling, J., Tiwari, A. (2010). On the portfolio properties of real estate in good times and bad times. Real Estate Economics, 38(3), 529–565.

  55. Serrano, C., & Hoesli, M. (2007). Forecasting REIT returns. Journal of Real Estate Portfolio Management, 13(4), 293–309.

  56. Serrano, C., & Hoesli M. (2010). Are securitized real estate returns more predictable than stock returns?Journal of Real Estate Finance and Economics, 41(2), 170–192.

  57. Swanson, Z., Theis, J., Casey, K. (2002). Reit risk premium sensitivity and interest rates. Journal of Real Estate Finance and Economics, 24(3), 319–330.

  58. Welch, I., & Goyal, A. (2008). A comprehensive look at the empirical performance of equity premium prediction. Review of Financial Studies, 21(4), 1455–1508.

  59. Zhou, J., & Anderson, I. (2011). Extreme risk measures for international REIT markets. Journal of Real Estate Finance and Economics, 45(1), 152–170.

Download references

Author information

Correspondence to Massimo Guidolin.

Appendix: Solution of Dynamic Asset Allocation Problems by Monte Carlo Methods

Appendix: Solution of Dynamic Asset Allocation Problems by Monte Carlo Methods

Markov Switching Model

Given the optimization problem is solved backwards at each time t (since the portfolio can be rebalanced every month), such that \(a\left ( \mathbf {\pi }_{t+1}^{i},t+1\right )\) is known for all values of \(i=1,2,\ldots ,Q\) on a discretization grid. Here \(a(.)\) is a function of the regime probabilities \(\pi _{t+1}\). Computing a Monte Carlo approximation of

$$E_{t}\left[ \left \{ \mathbf{\omega }_{t}\mathbf{R}_{t+1,g}\right \} ^{1-\gamma }a\left( \mathbf{\pi }_{t+1}^{i},t+1\right) \right] $$

requires drawing G random samples of asset returns \(\left \{R_{t+1,g}\left (\pi _{t+1}^{i}\right )\right \}_{g=1}^{G}\) from the \(t+1\) one-step joint density conditional on the period-t parameter estimates \(\hat {\mathbf {\theta }}_{t}=\left (\left \{ \hat {\mathbf {\mu }}_{S},\hat {\Omega }_{S}\right \}_{S=1}^{k},\hat {\mathbf {P}}\right )\) assuming that, at each point \(\pi _{t}^{i}\) is updated to \(\pi _{t+1}\left (\pi _{t}^{i}\right )\). The algorithm consists of the following steps:

  1. 1.

    For each possible value of the current regime \(S_{t}\) simulate G returns \(\left \{ \mathbf {R}_{t+1,g}(S_{t+1}) \right \} _{g=1}^{G}\) in calendar time from the regime switching model:

    $$\mathbf{R}_{t+1,g}\left( S_{t}\right) =\mathbf{\mu }_{S_{t+1}}+\mathbf{\varepsilon }_{t+1,g}\qquad \mathbf{\varepsilon }_{t+1,g}\sim N\left(\mathbf{0},\Omega_{S_{t+1}}\right). $$

    The simulation enables switching as governed by the transition probability matrix \(\hat {\mathbf {P}}_{t}\). For example, starting in state 1, the probability of switching to state 2 between t and \(t+1\) is \(\hat {p}_{12}\equiv e_{1}^{\prime }\hat {\mathbf {P}}_{t}e_{2}\), while the probability of remaining in state 1 is \(\hat {p}_{11}\equiv e_{1}^{\prime }\hat {\mathbf {P}}_{t}e_{1}\). Hence, at each point in time, \(\hat {\mathbf {P}}_{t}\) governs possible state transitions.

    Combine the simulated returns \(\left \{ \mathbf {R}_{t+1,g}\right \} _{g=1}^{G}\) into a sample of size G, using the probability weights \(\pi _{t}^{j}\):

    $$\mathbf{R}_{t+1,g}\left(\mathbf{\pi }_{t}^{i}\right)=\sum \nolimits_{j=1}^{k}\left(\mathbf{\pi}_{t}^{i}\mathbf{e}_{j}\right)\mathbf{R}_{t+1,g}\left( S_{t}=j\right) $$
  2. 2.

    Update the future regime probabilities perceived by the investor using the Hamilton-Kim filtering formula

    $$\mathbf{\pi }_{t+1}\left( \mathbf{\pi }_{t}^{i}\right) =\frac{\left( \mathbf{\pi }_{t}^{i}\right) ^{\prime }\hat{\mathbf{P}}\odot \mathbf{\eta }\left(\mathbf{R}_{t+1,g}(\mathbf{\pi }_{t}^{i});\hat{\mathbf{\theta }}_{t}\right) }{\left( \left( \mathbf{\pi }_{t}^{i}\right) ^{\prime }\hat{\mathbf{P}}\odot \mathbf{\eta }\left( \mathbf{R}_{t+1,g}(\mathbf{\pi }_{t}^{i});\hat{\mathbf{\theta }}_{t}\right) \right) \mathbf{\iota}_{k}}. $$

    This gives an \(G\times k\) matrix \(\left \{\mathbf {\pi }_{t+1}\left ( \mathbf {\pi }_{t}^{i}\right ) \right \} _{g=1}^{G}\), whose rows correspond to simulated vectors of perceived regime probabilities at time \(t+1\).

  3. 3.

    For all \(g=1,2,\ldots ,G\) calculate the value \(\tilde {\mathbf {\pi }}_{t+1,g}^{i}\) on the discretization grid \((i=1,2,\ldots ,Q)\) closest to \(\pi _{t+1,g}\left (\pi _{t}^{i}\right )\) using the distance measure \(\sum _{j=1}^{k-1}\left \vert \mathbf {\pi }_{t+1}^{i}\mathbf {e}_{j}-\mathbf {\pi }_{t+1,g}\mathbf {e}_{j}\right \vert \), i.e.

    $$\tilde{\mathbf{\pi}}_{t+1,g}^{i}\left(\mathbf{\pi }_{t}^{i}\right)\equiv \arg \min \sum \limits_{j=1}^{k-1}\left \vert \mathbf{x}\mathbf{e}_{j}-\mathbf{\pi}_{t+1,g}\mathbf{e}_{j}\right \vert . $$

    Knowledge of thevector \(\left \{\tilde {\mathbf {\pi }}_{t+1,g}^{i}\left (\mathbf {\pi }_{t}^{i}\right ) \right \} _{g=1}^{G}\)allows us to build \(\left \{a\left (\mathbf {\pi }_{t\hspace *{-1.5pt}+\hspace *{-1.5pt}1}^{(i,g)},t\hspace *{-1.5pt}+\hspace *{-1.5pt}1 \right )\right \} _{g=1}^{G}\), where \(\pi _{t+1}^{(i,g)}\equiv \tilde {\mathbf {\pi }} _{t+1,g}^{i}\left (\pi _{t}^{i}\right )\) is a function of the assumed, initial vector of regime probabilities \(\pi _{t}^{i}\).

  4. 4.

    Solve the program

    $$\max \limits_{\mathbf{\omega }_{t}\left(\mathbf{\pi }_{t}^{i}\right)}G^{-1}\sum \limits_{g=1}^{G}\left \{\left[\mathbf{\omega }_{t}\mathbf{R}_{t+1,g}\right]^{1-\gamma }a\left( \mathbf{\pi }_{t+1}^{(i,g)},t+1\right) \right \} $$

    For large values of G this provides a precise Monte Carlo approximation to \(E\left [\left \{ \mathbf {\omega }_{t}\mathbf {R}_{t+1,g}\right \} ^{1-\gamma } a\left ( \mathbf {\pi } _{t+1}^{i},t+1\right )\right ]\). The value function evaluated at the optimal weights \(\hat {\mathbf {\omega }}_{t}\left (\pi _{t}^{i}\right )\) gives \( a\left (\pi _{t}^{i},t\right )\) for the ith point on the initial grid. We also check whether \(\omega _{t}R_{t+1,g}\) is negative and reject all corresponding sample paths.

The algorithm is applied to all possible values \(\pi _{t}^{i}\) on the discretization grid until all values of \(a\left (\pi _{t}^{i},t\right )\) are obtained for \(i=1,2,\ldots ,Q\). It is then iterated backwards. We take \(a\left (\pi _{t+1}^{i},t+1\right )\) as given and use the actual vector of smoothed probabilities \(\pi _{t}\). The resultant vector \(\hat {\mathbf {\omega }}_{t}\) gives the optimal portfolio allocation at time t, while \(a(\pi _{t},t)\) is the optimal value function. In our application, Q is selected as 5\( ^{2}=25\) which fits the standard formula \(5^{k-1}\) as in Guidolin and Timmermann (2008) and the number of Monte Carlo simulations is 30,000.

VAR Model

Again the optimization problem is solved by backward iteration for each point t so that \(a\left ( \mathbf {Z}_{t+1},t+1\right )\). A Monte Carlo approximation of the expectation

$$E_{t}\left[ \left \{ \mathbf{\omega }_{t}\mathbf{R}_{t+1,g}\right \} ^{1-\gamma }a\left( \mathbf{Z}_{t+1}^{i},t+1\right) \right] $$

now requires drawing G random samples of the state variables \(\left \{ \mathbf {Z}_{t+1,g}\right \} _{g=1}^{G}\) from the \(t+1\) one-step joint density conditional on the period-t parameter estimates \(\hat {\mathbf {\theta }}_{t}=(\hat {\mathbf {\mu }},\hat {\mathbf {A}},\hat {\Omega })\). The algorithm is similar but much simpler than for the MRSM. The G returns \(\left \{\mathbf {R}_{t+1}(\mathbf {Z}_{t}^{i})\right \} _{g=1}^{G}\) need to be simulated from the VAR model. In this case \(Q=20\) delivers quite accurate results (because of the linearity of the prediction framework) and we set again \(G=30{,}000\).

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Bianchi, D., Guidolin, M. Can Linear Predictability Models Time Bull and Bear Real Estate Markets? Out-of-Sample Evidence from REIT Portfolios. J Real Estate Finan Econ 49, 116–164 (2014).

Download citation


  • REIT returns
  • Predictability
  • Strategic asset allocation
  • Markov switching
  • Vector autoregressive models
  • Out-of-sample performance

JEL Classifications

  • G11
  • C53