This study used panel data in analyzing the effects of the cost of machines, tools and equiptments; labor costs; quantity of fertilizer; cost of seeds/planting materials; cost of crop protection products; and irrigation fees on the yield per hectare of rice. Panel data were utilized using fifteen (15) regions of the Philippines namely Bicol, Cagayan Valley, Cordillera Administrative Region (CAR), Central Luzon, Ilocos, Southern Tagalog, Central Visayas, Eastern Visayas , Western Visayas, Northern Mindanao, Southern Mindanao, Western Mindanao, Central Mindanao, CARAGA and ARMM over the 12 periods, 1991 to 2002. Regression analysis was applied to the irrigated and non-irrigated areas. One hundred eighty (180) samples points were utilized. The study ran four (4) models, each for irrigated and non-irrigated rice areas. These four models consisted of different combinations of the explanatory variables to test the consistency of the factors that contribute significantly to the yield per hectare of palay. All models were of the unconstrained Cobb-Douglas type production function. All explanatory variables, except the fertilizer which was in kilograms, were deflated using the GDP deflator. The result of the analysis revealed that in the irrigated rice area, the three explanatory variables, namely: fertilizer, crop protection, and irrigation fees had regression coefficients which were highly significant in all models. It should be noted that fertilizer was in kilograms while crop protection products (agri-chemicals) and irrigation fees were deflated peso values of these variables. The remaining three (3) explanatory variables namely, labor, machines, and seeds gave regression coefficients that were not significant. In other words, the marginal product of these variables or factors of production were statistically equal to zero. While in the non-irrigated model, results of the regression analysis showed that three (3) variables namely, fertilizer, machines, and crop protection products gave significant regression coefficients in all regression runs. Two variables (labor and seeds) were not significant but seeds gave varying results. The significant explanatory variables of the models could explain about two-thirds of the total variation of the yield per hectare of palay in both irrigated and non-irrigated rice areas
