Faculty Sponsor(s)
Jeff Summers
Location
Jereld R. Nicholson Library: Grand Avenue
Subject Area
Economics
Description
This research analyzes the short and long run relationships between new housing starts (HOUS), real gross domestic product (RGDP), and average 30-year fixed mortgage rates (AFM) over the period 1972:Q2 to 2018:Q3 in order to forecast housing starts for the fourth quarter of 2018. After concluding all three series exhibit unit roots, a cointegrating regression was used to test for a long run relationship between the variables. The hypothesis that HOUS, RGDP, and AFM are not cointegrated was rejected at the 5% level, indicating they maintain a long run relationship. Next, a Vector Error Correction Model (VECM) was estimated and used to forecast HOUS for the fourth quarter of 2018. The forecasted value of 1,333,000 new housing units compares favorably to the actual value of 1,151,000 new units with a forecast error of less than 15%.
Recommended Citation
Fowler, Lindsay, "Forecasting New Housing Starts Using Real GDP and Average 30-Year Fixed Mortgage Rates" (2019). Linfield University Student Symposium: A Celebration of Scholarship and Creative Achievement. Event. Submission 40.
https://digitalcommons.linfield.edu/symposium/2019/all/40
Forecasting New Housing Starts Using Real GDP and Average 30-Year Fixed Mortgage Rates
Jereld R. Nicholson Library: Grand Avenue
This research analyzes the short and long run relationships between new housing starts (HOUS), real gross domestic product (RGDP), and average 30-year fixed mortgage rates (AFM) over the period 1972:Q2 to 2018:Q3 in order to forecast housing starts for the fourth quarter of 2018. After concluding all three series exhibit unit roots, a cointegrating regression was used to test for a long run relationship between the variables. The hypothesis that HOUS, RGDP, and AFM are not cointegrated was rejected at the 5% level, indicating they maintain a long run relationship. Next, a Vector Error Correction Model (VECM) was estimated and used to forecast HOUS for the fourth quarter of 2018. The forecasted value of 1,333,000 new housing units compares favorably to the actual value of 1,151,000 new units with a forecast error of less than 15%.
