Abstract:
The state of the current national agricultural capital input account is arguably misleading in measuring capital input use in South Africa. This is attributable to the problems associated with measurement of capital inputs, as well as changes in methodologies of different sources of capital data. This study questions the current state of national capital input accounts, with specific reference to machinery and implements, and a specific focus on tractors. It corrects the input series by revising the value of tractor sales and the machinery and implements series component of the capital input account for the South African agricultural sector, using more robust assumptions and underlying data as far as possible. This study shows that the current South African national capital input account is understated and resultantly has knock-on effects on service flows and agricultural productivity estimates for South Africa. Amongst other measurement problems in the capital formation series of South Africa, is the failure to incorporate quality adjustments in the valuation of capital. This measurement problem has been shown to lead to undercounting and effectively flawed productivity estimates. This study analyses the impact of incorporating quality changes in valuing tractors and using disaggregated data as far as possible. As such, this study focuses on tractor use in South Africa and analyses the tractor input by disaggregating the tractor categories and attributes. The tractor attributes and other drivers, such as net farming income and area planted in hectares, are used to trace the evolution of tractorisation in South Africa. As such, this study shows how tractorisation has changed over time and incorporates these quality changes by disaggregation, using underlying price and quality data for tractors in different classes and types. The assumptions used in past analyses have resulted in a failure to explain the evolution that has taken place in input use in the face of tractorisation innovation. The current Abstract estimates of the tractor price index are not cointegrated with the rand dollar exchange rate and net farm income, which is not in line with economic theory. In addition, current tractor trends fail to reflect the macro-economic trends that have taken place in South African agriculture. As such, the corrected capital input series will contribute to the understanding of economic trends in tractorisation and contribute to the labour debate in South Africa at an aggregate level. Thus, besides having a direct impact on service flow estimates and ultimately productivity analysis, this study has other agricultural policy impacts, such as on trade, mechanisation policy, and R&D policy, as well as on tax policies in the sector. For machinery manufacturers, this study is important in that it provides sound estimates of the relationship between tractor prices and other variables, which are useful for making informed business decisions. This study illustrates the effects of flawed measurement of inputs that impacts policy analysis and recommendations.
