Examining a Non-Linear Relationship Between Housing Price Shocks and Credit Growth in a Threshold Vector Autoregressive (TVAR) Model*

This study examines a non-linear relationship between housing prices, interest rates, and credit growth, depending on the state of economic activity. Using the industrial production index, inflation rate, CD rate, the housing price index, and total household debt, I estimate a structural threshold vector autoregression model, wherein the threshold variable is set to the real housing price growth. The two different regimes are identified based on the threshold value, which is estimated at 2.13 percent. I find that the non-linear relationship between the real housing price growth, interest rate, and credit growth across the two different regimes is statistically significant. In the regime where the housing price growth is above the threshold value, housing price shocks do not yield significant responses for credit growth. By contrast, in the regime where the housing price growth is below the threshold value, housing price shocks generate large and significant responses for credit growth. Regarding interest rate shocks that significantly reduce credit growth, the impact of credit growth ceases in 6 months in the above-threshold regime, whereas it lasts for 12 months in the below-threshold regime.