Regional Econometric Housing Start Forecast Accuracy in Florida

Abstract

Regional econometric forecasting accuracy assessment has traditionally received less attention than its national macroeconometric counterpart. While evidence is available that state and local employment forecasts perform well relative to standard benchmarks, little is known about the historical performance of widely cited regional housing start forecasts. This paper attempts to fill partially that gap in the literature by assessing the accuracy of previously published residential construction forecasts for Florida and its six largest metropolitan areas. Results indicate that regional econometric housing start forecasts are less reliable than regional employment structural model forecasts.

Determinants of Construction Sector Activity in Turkey: A Vector Autoregression Approach

“…In order to cognize the factors affecting the economic fluctuations and exposing the economic situation of a country, the analysis of construction sector and especially housing sector are required since changes in the volume of construction and housing sector and real economic activity are usually parallel to and also interacted with each other. In this context, Fullerton and West (1998) studied the residential construction activity for Florida and its six largest metropolitan areas between 1985 and 1995 with quarterly data. Autoregressive Integrated Moving Average (ARIMA) and random walk prediction rule were used in their study.…”

Specification of a Borderplex Econometric Forecasting Model

“…Given the degree of freedom constraints imposed on regional estimation efforts for both sides of the border, it probably makes sense to employ the structural approach discussed above for large-scale forecasting efforts. Extrapolation "sanity checks" could be fairly easily and reliably conducted via the time-honored random-walk comparisons long employed by practitioners (Fullerton and West 1998). The slightly modified structural framework does prove attractive from the vantage point of offering a comprehensive mechanism for simultaneously examining multiple segments of border region economies and at least partially accounting for international business cycle developments.…”

“…Their flexibility has allowed them to be applied to a wide range of regional taxation issues in the U.S. (Fullerton 1987) and international monetary topics in Latin America (Fullerton and Araki 1996). Limited applications of these techniques to border region issues have also started to emerge in the academic literature (Fullerton 1998).…”

“…Because annual frequency data are used, it is not surprising that higher order serial correlation is not encountered. That nearly one-third of all the border model stochastic specifications require autocorrelation correction in part reflects widespread data constraints that also affect traditional regional econometric modeling efforts (Fullerton and West 1998) as well as standard Latin American forecasting models (Fullerton and Araki 1996).…”

“…An eleven-equation subsystem is used to characterize residential construction and real estate in El Paso, reflective of the overall importance of this segment of any regional economy (Zandi 1999). Given the typical difficulties associated with housing start forecasts (Fullerton and West 1998), broader coverage has been included to augment the overall usefulness of this portion of the model. Variables modeled in this block include local housing affordability, average monthly housing payments, median prices for existing and new homes, single-unit and multiunit starts, single-family and multifamily home sales, and single-unit and multiunit stock levels.…”