This study utilised empirical specifications by Muchapondwa (2008), Leaver (2004) and Townsend and Thirtle (1997). The study closely followed
the ARDL model of the adapted Nerlove (1958) log-linear supply function by Leaver (2004). Minor modifications were made by using a different time period and dropping some variables due to unavailability of data. The adoption of the logarithmic function allows coefficients to be directly interpreted as elasticities and also reduces problem of non-normality in residuals. The tobacco supply response econometric model is presented as follows:
11
ln(TOUT
t
)=β
0
+β
1
ln(TOUT
t-1
)+β
2
ln(PT
t-1
)+β
3
ln(MP
t-1
)+β
4
ln(ATG
t
)+β
5
SQ
t-
1
+β
6
T+β
7
T
2
+µ
t
…………………………………………………………………………….....(3.2)
where; β
0
= regression intercept to β
7
= regression coefficients ln(TOUTt)= natural log of total tobacco output in tonnes ln(TOUT
t-1
)= log of lagged total tobacco output ln(PT
t-1
)= natural log of lagged tobacco price ln(MP
t-1
)= natural
log of lagged maize price, expressed in US dollar terms per tonne ln(ATG
t
) = natural log of annual population
of active tobacco growers SQt-1
= dummy variable for the years with sales quota taking the value of 1 in years in which sales quota applies and 0, otherwise T simple time trend (t for 1980 tot for 2015); T quadratic time trend (t for 1980 tot fort random error term with zero mean and constant variance.
3.2. Definition and Justification of Variables3.2.1 The Dependent VariableTotal tobacco output (lnTOUTt)Total tobacco output refers to the actual total amount of all unprocessed tobacco leaf in million tonnes supplied and delivered to the market. There is a great disagreement in existing literature on the correct measure of crop supply response and the
available options are acreage, productivity and total output. Askari and Cummings (1977) stated that agricultural supply response functions are best expressed in terms of acreage assuming that farmers can only increase output in response to price changes by utilising more land. However, Leaver (2004) argued that use of actual output is also a best measure because this acknowledges that farmers may respond to changes in prices by farming their land more intensively. Therefore, from the above argument and also
guided by empirical literature, this study used total output as the dependent variable.
Share with your friends: