Academe Research Journals

Journal of Agricultural Economics and Development

Full Length Research Paper

Model misspecification and factors affecting land use decisions using linear probability model (LPM)

Ochalibe A. I.*, Ogah, M. O. and Obo, V.

Department of Agricultural Economics, University of Agriculture Makurdi, Benue State, Nigeria.

*Corresponding author. E-mail: alexee2007@yahoo.com. Tel: +2348101342902, +2348053362445.

Accepted 18 November, 2015

Abstract

This study examined linear probability approach and misspecification of econometric model on land fallow decisions. It utilised primary data with sample size of 4187 agricultural households. The data were analyzed using linear probability model (LPM). Specification error test, robustness, goodness of fit and significance test were performed after estimation. Ramsey RESET test using powers of fitted values is significant at 5% (p>f=0.000), indicating that there is no serious signs of misspecification. The inktest (P=<0.05) is not significant indicating normality. In the same vein, diagnostic tests further indicated that Breusch-pagan/Cook-weisberg test for heteroskedasticity with the null hypothesis of constant variance was rejected (p>chi2=0.000). From the results, LPM passed only some of the tests suggesting that LPM may be less suited when dependent variable is dichotomous. The study revealed that distance and transaction costs along with socioeconomic and institutional factors significantly (p=>0.000) affect the decision to leave land fallow. It was concluded that the model was correctly specified, and the relevant variables in the economic relationship of land use decision were clearly identified with well defined pattern of relationship. It was recommended that a combination of econometric techniques like binary choice model as well as policy measures to ensure accurate estimates and reduce possibility of leaving agricultural land unutilised is required.

Key words: Linear probability model (LPM), distance, transaction costs, land use.