Price-volume Correlation in the Housing Market: Causality and Co-movements (P.20)

estimation results in Tables 2, 3, 4 and 5: the benchmark specification in Table 2, the
asymmetric specification in Table 3, and subsample estimation for high supply
elasticity markets (Table 4) and low supply elasticity markets (Table 5). The table
also reports t-statistics of testing two-sided hypotheses that the correlations follow a
distribution with 0 mean.
We have a few interesting findings. First, we find evidence of the statistically
significant positive price–volume correlation. The raw correlation is 0.048 (0.072 in
markets with high elasticity and 0.025 in markets with low elasticity) and significant
at the1% level (insignificant for markets with low elasticity). Second, we find strong
evidence of positive correlations between “fitted” prices and volume. The “fitted”
price–volume correlations are much higher than raw price–volume correlations.
They are 0.168 and 0.165 for the benchmark and asymmetric specifications, and
0.214 and 0.108 for sub-samples of MSAs with high and low supply elasticity. All
“fitted” price–volume correlations are significant at the 1% level. Finally, the
correlations between residuals are always lower than the raw price–volume
correlations, and are insignificant. Panel A seems to indicate that our model
captures the price–volume correlation well.

Next Page 21



Leave a Reply

You must be logged in to post a comment.