 |
|
SECTOR
Home | What's New
| About SECTOR | Meetings
| Publications | Discussion
Forum | Contact Us
| Links | Site
Map
|
|
|
SAP 2.74/WP.128
Agrarian transition in Viet Nam
Accounting for agricultural growth
The approach followed to identify sources of agricultural growth rests on the work by Kawagoe, Hayami and Ruttan (KHR) (1985), which in turn is based on earlier work by Hayami and Ruttan. Lau and Yotopoulos (1989) and Carter and Bin Zhang (1994) follow a similar method.
A log-linear Cobb-Douglas production function is proposed in which total output is a function of land (A), labour (L), livestock (S), fertilizer (F) and machines (M). Time-series analysis of the production function is difficult for several reasons (see Lau and Yotopoulos, 1989). For some variables there is a limited range of variation, e.g. land area is fixed and labour changes only slowly. Moreover, there tends to be high multi-collinearity as the ratio between some factors of production tends to be rather rigid. And, in time-series, it may be difficult to separate the effects of productivity growth from scale effects. Because of these difficulties, studies have concentrated on using cross-country data to estimate the production function. The implicit assumption is that all countries have access to the same technology. KHR (1985) find that the sum of the input coefficients for the sample of developing countries is not significantly different from one, suggesting the absence of economies of scale. They argue that this is intuitively acceptable since the shortage of land and the abundance of labour results in the intensive cultivation of small plots which will not have the scale effects that more mechanized cultivation of large plots might have.
The results of the various studies have been used to derive the factor shares for the various inputs:
Data for the period 1976-95 have been collected.
Output: gross output of crops, measured in tons. This ignores the output from non-crop activities, such as animal husbandry and aquaculture, which were small in earlier years but have grown rapidly recently. The advantage of using volume measures rather than value terms is that no prices are needed to deflate nominal values. A problem, however, is that the total gross output combines the output volumes of different crops with different "bulkiness", e.g. one hectare of land may produce 3.7 tons of paddy or 45.2 tons of sugar cane. This is problematic because the area under sugar cane has fluctuated rather strongly over the years. For this reason sugar-cane output has been excluded from total output
Land: the sown area of land in hectares
Labour: labour registered as employed in agriculture
Fertilizer: kg of nitrogen, phosphorous and potash
Livestock: number of working animals
Machines: number of tractors