Introduction

Input costs for farming in India have increased significantly over the last few years (Raghavan 2008; Srivastava, Chand, and Singh 2017). Neither the minimum support price (MSP) offered by the government nor the “free” market price in the market (mandi) of the Agricultural Produce Market Committee (APMC) has kept pace with the rise in input prices (Srivastava, Chand, and Singh 2017).

An important demand of agrarian movements in the country is a minimum support price that is at least 50 per cent higher than the cost of production, as estimated by the Commission for Agricultural Costs and Prices (CACP). This procedure for fixing the MSP was recommended by the National Commission on Farmers (NCF) headed by Professor M. S. Swaminathan in 2006. In the budget speech for 2018–19, the Union Finance Minister announced that the government would ensure that the MSP would be at least 50 per cent higher than the cost of production. This announcement initiated a debate about which cost of production should be taken into consideration while deciding upon the MSP: the A2+FL cost (i.e. paid-out costs plus imputed cost of family labour), or the C2 cost (i.e. the sum of paid-out costs, imputed value of family labour, interest on the value of owned capital assets, and the rental value of owned land).

This note is concerned with the potential underestimation of costs by the CACP (and not with whether the A2+FL cost or the C2 cost is a suitable measure of the cost of production). The methods and procedures followed by the CACP, it is argued, are likely to underestimate the actual cost of production for two reasons. The first possible reason has to do with problems of the methodological framework, an issue that has been discussed by different scholars (Sen and Bhatia 2004; Surjit 2008; Nawn 2013). The second reason relates to lags in the availability of data. This has been mentioned in CACP (2012), but has not adequately been discussed in the literature. In this note, we demonstrate how lags in availability of data, especially on input prices, can lead to underestimation of costs of production.

CACP Data and Methodology

The methodology used to arrive at the cost of production for a particular crop is described in detail in the Annual Price Policy report published by the CACP. According to these reports, the CACP uses cost estimates generated by the Directorate of Economics and Statistics (DES) of the Union Ministry of Agriculture and Farmer Welfare under the “Comprehensive Scheme for Studying the Cost of Cultivation of Principal Crops in India” (henceforth, Comprehensive Scheme). However, Comprehensive Scheme data are usually available only after a lag of two to three years; for instance, cost of production data available to the CACP in 2018–19 are for the year 2015–16. To arrive at a projected cost of production for the current year, then, CACP uses the “actual estimates” available for the most recent three years. Thus, the projected cost of cultivation (CoC) (in rupees per hectare) for the crop season of 2018–19 is based on “actual estimates” of the crop seasons of 2013–14, 2014–15, and 2015–16.

These “actual estimates” show the changes in input costs over these three years. To assess future changes in individual input costs, the CACP constructs a composite input price index (CIPI) based on the latest prices of different inputs, including human labour, bullock labour, machine labour, manure, fertilizer, seed, pesticides, and irrigation, using data from the Labour Bureau, Ministry of Labour and Employment, State Governments, and the Office of the Economic Adviser, Ministry of Commerce and Industry. The CACP combines CIPI with “actual estimates” to arrive at crop-wise, State-wise projected costs of cultivation. Crop-wise, State-wise costs of production (CoP) (in rupees per quintal) are derived from the costs of cultivation using projected yields. The all-India cost of production is arrived at by calculating a weighted average of State-level costs of production, with the weights being the shares of States in all-India production for the most recent year for which production estimates are available. For the crop season 2018–19, production estimates of 2016–17 are used. An all-India crop-wise weighted average input price index for all inputs is created by the CACP using State-level CIPIs, with the weights being the relative shares of States in the national area under the crop during the latest crop year for which production estimates are available. These indices are then used to compute an all-India weighted average composite input price index for each crop, with the weights being relative shares of crops in total production at the all-India level during the latest crop year for which production estimates are available.

Underestimation of Input Costs

Problems in Methodological Framework

The Special Expert Committee on Cost of Production Estimates (chaired by S. R. Sen), constituted in 1979–80, and the Expert Committee for Review of Methodology of the Cost of Production of Crops (chaired by C. H. Hanumantha Rao), constituted in 1990, raised a number of issues related to the methodology used by the CACP. Both these committees studied the sampling framework of the scheme. As a result of the recommendations of these committees, the CACP made some changes to its methodology; for instance, it has shifted from a single-crop approach to a crop-complex approach in order to improve its estimates.

However, according to Surjit (2008), the sampling framework continues to have major shortcomings. First, the scheme does not operate in all the States and covers only 25 crops. The North-Eastern States (except Assam) and Jammu & Kashmir are not included in the sample. The scheme is limited mainly to seasonal and annual crops (coconut and sugarcane being exceptions), and does not take into consideration changing cropping patterns in many States. Secondly, the scheme fails to adequately accommodate various institutional arrangements, such as tenancy, in the sample. Tenant farmers are under-represented in the sample.¹

Problems with the scheme also lie in the methods used to calculate imputed costs, such as the cost of time spent at managerial tasks, rental value of owned land, and interest rates charged for fixed capital and working capital. The Hanumantha Rao Committee recommended that 10 per cent of cost A2 should be added as management cost to the total cost of cultivation. The government accepted the recommendation but decided to add 10 per cent of C2 cost (rather than A2 cost), and created a new cost concept, C3. Such a method would typically double the cost to be accounted for management while at the same time keeping management costs out of A2+FL as well as C2 costs. According to Sen and Bhatia (2004), the CACP has conveyed its disagreement in this regard to the government.

Sen and Bhatia (2004) also point out that the rental value of owned land is still calculated on the basis of the share of rent in the gross value of output, even though the two review committees had recommended more comprehensive methods to compute this variable. Similarly, the interest cost on owned fixed capital is estimated at 10 per cent per annum and the interest cost on owned working capital is estimated at 12.5 per cent for half the period of a crop. This remains the practice despite the fact that the two committees recommended that the two interest rates be calculated by taking the weighted average of the actual interest rates canvassed from sample cultivators (which is likely to be higher than the assumed rate of interest).

There are also problems with respect to the collection, processing, and analysis of data, and of the quality of data. These include problems with respect to the classification of farms into different size-classes for sampling purposes, problems related to the FARMAP software, which results in substantial inaccuracies in the estimation of gross cropped area and net sown area, and problems arising from incorrect coding (Surjit 2008). The exclusion of transport costs also contributes to underestimating the costs of cultivation.

Lags in the Availability of Data

In this section, we examine the impact of lags in the availability of data on estimates of costs of production. To this end, we constructed input price indices using the “actual” Comprehensive Scheme plot-level data for kharif as well as rabi crops for a particular year (obtained after a lag), and compared these with the “projected” input price indices constructed by the CACP in their price policy reports for that particular year. We also computed the “actual” A2+FL and C2 costs of production of major crops using the Comprehensive Scheme plot-level summary data for a particular year, and compared it with the respective “projected” costs estimated by CACP for that particular year using lagged data. To construct the input price indices using original data, we used the methodology described in the Manual on Cost of Cultivation Surveys released by the Central Statistical Office (CSO). The most recent plot-level data made available by the DES are for 2013–14. So, the input price indices were compared for the period from 2004–05 to 2013–14. The plot-level data do not provide information about all inputs. So, the indices constructed using plot-level data cover the costs of fertilizer, human labour, animal labour, and machinery.

In this note, the extent of underestimation is determined by comparing the costs “projected” by CACP with costs estimated by using cost of cultivation/production-related data for a particular year, as obtained from the DES website. Since the latest cost of cultivation/production-related data available are for 2015–16, we have compared the price levels from 2004–05 to 2015–16 for six major crops, namely paddy, cotton, wheat, maize, groundnut, and soybean. To arrive at State-level, crop-level A2+FL and C2 costs from cultivation/production-related data, the methodology described in the Manual on Cost of Cultivation Survey was used. To arrive at a weighted average of A2+FL and C2 costs at the crop level, the production share of a particular State (for a particular crop) in that particular year was used as the weight. This is different from the CACP methodology, which uses lagged production shares as weights. For example, the production estimates of 2016–17 are used by CACP to calculate prices in 2018–19.

We have also calculated adjusted MSP for each crop, showing the impact of underestimation of production/cultivation cost on the support price that farmers get. We assumed that MSP is decided as a direct mark-up on A2+FL cost, i. e., that the ratio of MSP to A2+FL cost remains the same for a crop in a particular year, and calculated the adjusted MSP based on the actual cost of production.

Discussion and Results

A comparison of the two input price indices for the four inputs – namely, fertilizer, human labour, animal labour, and machinery – shows an underestimation of input costs in the CACP procedure. This underestimation has, in fact, increased with time. In the case of fertilizer prices, the CACP estimated that the price index with base 2004–05= 00 was 152.7 in 2013–14 (see Figure 1). The same index for the same year computed from plot-level data was 205.3. In other words, CACP estimated that prices of fertilizers increased by 1.5 times in the period between 2004–05 and 2013–14, whereas, according to the plot-level data, the prices had almost doubled. Similarly, the cost of machinery (see Figure 2) and the cost of manual labour (see Figure 3) have been underestimated. Among all the inputs we examined, underestimation was greatest for machinery. The only instance where the two estimates matched was in the case of animal labour (see Figure 4).

Figure 1

Comparison of price indices for fertilizer, actual and projected, 2004–05 to 2013–14

Figure 2

Comparison of price indices for machinery, actual and projected, 2004–05 to 2013–14

Figure 3

Comparison of price indices for manual labour, actual and projected, 2004–05 to 2013–14

Figure 4

Comparison of price indices for animal labour, actual and projected, 2004–05 to 2013–14

For rice (see Table 1), the CACP underestimated C2 costs every year except 2012–13 and 2013–14. For A2+FL costs, the CACP method gave underestimates for five years and overestimates for seven years. However, in the case of cotton (see Table 2), the CACP consistently underestimated the costs every year after 2008–09. This underestimation was greatest in 2012–13, when the CACP’s projected A2+FL cost was 32 per cent lower than the actual A2+FL cost. If we assume that MSP is decided as a direct mark-up on A2+FL cost, i. e., that the ratio of MSP to A2+FL cost remains the same for a particular year, then the MSP of cotton from 2008–09 onwards should have been far higher than the announced levels (see Table 2). For example, in 2014–15, the MSP should have been Rs 4,848 and Rs 5,236 for medium-staple and long-staple cotton respectively, rather than Rs 3,750 and Rs 4,050, if the A2+FL cost had been projected accurately. In other words, the MSP should have been about 20–30 per cent higher than the announced price.

Similarly, soybean and maize farmers were receiving lower MSPs than warranted because costs of production were underestimated, particularly after 2013–14 (see Table 3 and Table 4, respectively). For instance, in 2015–16, CACP projected the A2+FL cost for soybean at Rs 1,770 per quintal, whereas the A2+FL cost from the plot-level data amounted to Rs 4,242 per quintal – more than twice the cost as estimated by the CACP. The costs of cultivation of wheat and groundnut as estimated by CACP were more accurate than the estimates for other crops (see Table 5 and Table 6).

CACP is aware of the underestimation of costs of production. In the Kharif Price Policy Report for 2012–13, the CACP stated:

The assumption of holding constant fixed cost components in cost projection for two to three years ahead does not stand the test of time. As far as kharif crops are concerned, such correction for underestimation/overestimation for different States of earlier projected cost compared to actuals has been effected in their likely projected costs for 2012–13.

And according to the Kharif Price Policy Report for 2013–14:

Since 2012, Commission also introduced a correction factor (CF) based on the difference between actual and projected costs for three years, for which latest information is available. Continuing with a similar practice, in pursuit of improvising projections, the Commission looks into the changes in the CF and adjusts its projected costs accordingly.

However, as our analysis shows, even after the introduction of the correction factor (CF), the CACP has continued to underestimate projected cost. In this regard, CACP itself recommended some changes in its 2012–13 report:

Greater credibility has to be built in the methodology of collection, compilation and generation of cost estimates, as has been followed for quite a long time, by bringing in more transparency to bridge the trust deficit on cost estimates thrown up by the DES. In this context, the Commission recommends the following: (i) there is need for switching over from the old manual mode of data collection to data collection on real time basis by providing palm tops to field investigators interacting with farmers and canvassing information from them on day to day basis; (ii) cost estimates being crucial to the formulation of price policy the cell phone numbers of the sample farmers covered under the Comprehensive Scheme (CS) be forwarded to the Commission to enable a preliminary cross check and reassessment of information being collected in the field on real factors of production. The Commission strongly feels that to begin with, change has to be initiated not only to cut down on time lag in the generation of estimates but also to strengthen monitoring mechanism in data collection.

Nevertheless, the problem of data lag persists.

Conclusion

Our analysis shows that the Commission for Agricultural Costs and Prices (CACP) underestimates the price indices of inputs. The underestimation is highest for machinery and fertilizers. The underestimation has increased with time even after the Commission introduced a correction factor (CF) to rectify the methodology. This leads to underestimates of the costs of production (both A2+FL cost and C2 cost) of major crops, including cotton, soybean, and maize. The underestimation is higher for C2 costs as compared to A2+FL costs. If these costs are accurately estimated, the minimum support price (MSP) for these crops should rise by at least 20 to 30 per cent, as MSP is linked to production cost. At the same time, underestimation due to non-inclusion of management cost, transportation cost, and insurance premium, and incorrect definition of interest rate on working capital, also depress input costs. In short, actual production costs are higher than the CACP estimates for several crops. We argue that there is urgent need to reduce the time-lag between the collection and release of data on cost of cultivation. For administrators, this may be a minor statistical or logistical problem, but for farmers, the impact it has on their lives is significant.