Table 6: Regression Results with Total Output dln(TOUT

21
4.4 Interpretation and Discussion of Regression Results
As shown in Table 6, the coefficient of determination, R of approximately 0.55 is higher than
0.5 indicating that the model is relatively of good fit. This value of 0.55 indicate that about 55% of the sample variance in the dependent variable is being jointly explained within the model by lagged tobacco output, lagged tobacco price, lagged maize price, population of tobacco growers and the quadratic time trend. In addition, the whole model is significant at 1% level since the probability value of the F-statistic is 0.001858 meaning that the whole model is valid and significant. With reference to Table 6, a positive coefficient of the lag of real tobacco price was found to be
0.117526 and significant at 5% level implying that producers price expectations influence tobacco supply. A higher price for tobacco in one period is followed by an increase in output supplied in the next period. The coefficient is the short-run price elasticity of supply implying that a 10% increase in current price would result in approximately 1.18% increase in tobacco output in the subsequent period. In conjunction with this, the long-run price elasticity of tobacco was found to be 0.16 (Appendix 5). Both elasticities are less than 1, showing that supply of tobacco is price inelastic. Hence, tobacco output is relatively unresponsive to price changes both in the short-run and the long-run. These findings are also comparable to short-run price elasticities of 0.34 by Leaver (2004) and 0.28 by Townsend and Thirtle (1997). Similarly, the long-run price elasticity of 0.16 is similar to that of 0.18 calculated by Muchapondwa (2008). Lagged tobacco output (dlnTTOUT
t-1
) has a coefficient of -0.266044 which is negative and statistically significant at 10% level as shown in Table 6. These findings were expected. This implies that a 10% increase in tobacco supply in one period is accompanied by a 2.6% fall in output supplied in the subsequent period. Although these results are different from those of
Muchapondwa (2008) and Leaver (2004) who found a positive relationship, the findings are consistent with the Cobweb theory. The negative impact of lagged output on tobacco output supply indicates that if tobacco output is currently high, prices would fall due to excess supply and in the next season farmers would reduce tobacco production as predicted by the Cobweb theory. Lagged price of maize was found significant at all conventional levels with a negative coefficient of -0.355482. This was expected and consistent with the findings of Townsend and Thirtle

22
(1997) who found a negative relationship. The coefficient is the short-run cross price elasticity of tobacco to price of maize in Zimbabwe. This implies that an expected 10% increase in relative price of maize in one year causes a fall in tobacco supply of 3.6% the following year. In addition the cross-price elasticity is less than 1, therefore, supply of tobacco is inelastic to changes in prices of its substitutes in production as also noted by Townsend and Thirtle (1997). As expected, supply of tobacco was found positively related to population of tobacco farmers
(lnATG
t
). A positive coefficient of 0.265710 which is significant at 10% level was estimated. The results do concur with the predictions of the micro-economic theory of supply which posits that as new firms enter the market, quantity supplied increases. This coefficient indicate that an annual 10% increase in the number of tobacco growers will cause about 2.6% rightward shift of the tobacco supply function. These findings are also comparable to those of Dean (1966) who also found a positive relationship in Malawi. A negative coefficient for the quadratic time trend variable was expected but a positive significant coefficient of 0.001246 was found. This indicates that in the long-run, the tobacco supply function exhibits increasing returns to advances in technology. This positive sign contradict with the findings of Leaver (2004) who found a negative relationship between the quadratic time trend and tobacco output. A possible explanation for this is that there is a slow rate of technology adoption in Zimbabwe. The coefficients for the simple time trend and sales quota dummy variables were unexpectedly found to be statistically insignificant at all conventional levels.

Download 0.76 Mb.

Share with your friends: