Table 4
Granger causality test results for industrial property
Δ tbi |
Δ reit |
Explanatory variable |
ΔD |
Adj. RP2P | ||
ΔIRN |
Δ S |
Δ SE | ||||
Pairwise model | ||||||
■g ω Δ ind_tbi .52 |
.00 |
.54 | ||||
■8 5 Δ ind reit .07 eg — о Й |
.45 |
.56 | ||||
Multiple variable model | ||||||
Δ indtbi .20 |
.00 |
.26 |
.99 |
.97 |
.91 |
.54 |
Δ ind reit .19 |
.34 |
.80 |
.53 |
.07 |
.11 |
.61 |
The table shows the p-values in the tests. The null hypothesis is that of no Granger causality between the
variables. The models include one lag and two dummy variables that take the value one in a single period
(1995Q2, 2008Q4) and are zero otherwise.
Table 5 Granger causality test results for offices
Δ tbi |
Δ reit |
Explanatory variable |
ΔD |
Adj. RP2P | |||
ΔIRN |
Δ S |
Δ SE | |||||
Pairwise model | |||||||
й O <υ əa й S |
Δ of_tbi .00 |
.03 |
.22 | ||||
Δ ofreit .92 |
.00 |
.32 | |||||
Multiple variable model Δ of_tbi .14 |
.00 |
.02 |
.22 |
.58 |
.03 |
.28 | |
Δ ofreit .70 |
.40 |
.70 |
.50 |
.07 |
.12 |
36 |
The table shows the p-values in the tests. The null hypothesis is that of no Granger causality between the
variables. The models include one lag and one dummy variable that takes the value one in a single period
(2008Q4) and is zero otherwise.
Table 6 Granger causality test results for retail property
Explanatory variable |
Adj RP2P | |||||||||
Δ tbi |
Δ reit |
INF |
ΔIR |
Δ S |
Δ SE |
ΔD |
Δ GDP | |||
O ð S Q |
Pairwise model | |||||||||
Δ retbi Δ re reit |
.15 .96 |
.07 .00 |
.41 .34 | |||||||
Multiple variable model | ||||||||||
Δ retbi |
.28 |
.04 |
.08 |
.64 |
.21 |
.04 |
.31 |
.01 |
.47 | |
Δ re reit |
.39 |
.50 |
.14 |
.58 |
.19 |
.51 |
.01 |
.09 |
.46 |
The table shows the p-values in the tests. The null hypothesis is that of no Granger causality between the
variables. The models include one lag and two dummy variables that take the value one in a single period
(1995Q2, 2008Q4) and are zero otherwise. Due to heteroscedasticity of the residuals in the model for Δrereit,
the p-values for Δ rereit in the multiple variable model are based on a covariance matrix that is computed
allowing for heteroscedasticity as in White (1980).
28