# 8.3 Comparison of Average Profit Efficiency Across Farm Classes

Chapter 6 showed that small-scale farmers across the country studies typically earned higher profits per unit than large-scale farmers if family labor is not costed, which suggests that they at least have a chance to survive. It was also clear that charging family labor at market rates largely eliminated this profit advantage.

The previous chapters and first section of this chapter now yield the data and methodology necessary to formally test hypotheses 2 and 7 from Chapter 1: "Large-scale producers are more efficient users of farm resources to secure profits, other things equal" and "Contract farmers are more profit efficient than independent farmers for comparable scales of operation".

The methodology and model are discussed in Chapter 4, and the estimated equations leading to the end result are discussed further in the next section. The results in this section are the "bottom line" estimates of a combination of technical and allocative efficiency, referred to here as "relative profit efficiency". In terms of the methodology of Chapter 4, each farm has a specific measure of relative profit efficiency that ranges from 0 (worst) to 1 (best). In Figure 4.1 in Chapter 4, this is (Yi/Y*), or the ratio of actual profit performance for the farm in question divided by estimated ideal performance for a farm endowed with the same bundle of fixed resources and facing the same set of prices. Farms with a score of 1 are on the estimated profit efficiency frontier. Farm-specific score are averaged over farms in the class in question to get an estimate of relative profit efficiency for that class of farm. Computationally these measures are estimated from a function of the residuals of the stochastic profit frontiers of Chapter 4, or the first stage of our model.

Results for a sample of commodities and countries are shown in Figures 8.1 to 8.10. Family labor is not costed. Results for other categories of farm and with costing of family labor are shown in the country annexes. Figures 8.1 to 8.3 show a diversity of results for broilers, and Figure 8.4 shows the case for layers in India. When family labor is not costed in the Philippines (Figure 8.1), smallholders as a group actually do better than larger-scale (commercial) operations as a group, but this is due to the composition of the smallholder sample, which deliberately includes a large number of contract farmers. In the Philippines, contract farmers as a group does much better (84 percent of maximum profit efficiency) than independents and price contractors (who are otherwise comparable to independents, except for having a steady forward outlet), who manage only 51 percent of maximum profit efficiency. Costing family labor and repeating the same exercise shows that commercial farmers as a whole do about the same as smallholders, but does not change the divide between contract and independent farmers.

In Thailand (Figure 8.2), the prevalence of contract farming for broilers in both the sample and nationally is much greater than is the case in the Philippines, with most broiler farming being done under price contracts. The pooled (price contract/fee contract/independent) results for Thai broilers are reported in the figure, in ascending order of size of farm. The largest farms are the most profit efficient (85 percent of maximum), compared to the smallest (78 percent of maximum). The bump in the 5,000 to 10,000 bird category is probably due to the prevalence of fee contract farmers in this size of operation, which is likely to increase technical and allocative profit efficiency (if not profits) over independent and price contracts at the same scale.

In Brazil (Figure 8.3), profit efficiency levels are substantially higher for larger farms for both broilers and layers; all farms used for this figure are independent. The Indian layer farms in Figure 8.4 are also all independent. Here the trend across increasing scales is not clear (at least with statistical confidence in the difference of averages across quintiles of the size distribution of farms). Subject to the latter caveat, it seems that efficiency increases slightly up to the 25,000-50,000 egg per annum production level and may even decline slightly above that.

Insights for the relative profit efficiency of swine production are found in Figures 8.5 to 8.7. In the Philippines, mean efficiency slightly declines at first with increasing size, as the unit advantage of not costing family labor becomes less important with increasing scale, and then increases. In Brazil (Figure 8.6), large producers are seen to be clearly more efficient (78 percent of maximum profit) than small producers (71 percent of maximum). In Thailand (Figure 8.7), mean profit efficiency in swine finishing increases marginally in going from backyard operations to small-scale commercial, and then levels off with further increases in scale. For weaners, profit efficiency increases sharply in going from backyard to small-scale commercial, but then falls again with increases in scale, being about the same at the very large-scale level as in backyards if family labor is not costed. The higher efficiency of small-scale commercial over backyards is hardly surprising; the fall in efficiency for larger scale units for weaners is probably due to the spreading of owner supervision too thin at larger scales in a business where piglet mortality is a major cost factor.

Figure 8.1 Relative profit efficiency for Philippines broiler farm by size and integration

Source: Costales, A., et. al., Annex I.

Figure 8.2 Relative profit efficiency of Thai broiler operations by size of farm

Source: Poapongsakorn, N., et. al., Annex IV.

Figure 8.3 Relative profit efficiency by farm size for Brazilian poultry operations

Mean Efficiency by Scale (Broilers)

Source: Camargo Barros, G.S., et. al., Annex V.

Mean Efficiency by Scale (Layers)

Source: Camargo Barros, G.S., et. al., Annex V.

Figure 8.4 Relative profit efficiency for Indian layer farms by farm size

Source: Mehta, R., et. al., Annex II

Figure 8.5 Relative profit efficiency by farm size for Philippines swine farms

Source: Costales, A., et. al., Annex I.

Figure 8.6 Relative profit efficiency by farm size for Brazilian swine

Source: Camargo Barros, G.S., et. al., Annex V.

Figure 8.7 Relative profit efficiency by size of two types of Thai swine farms

Source: Poapongsakorn, N., et. al., Annex IV.

Unlike broilers, layers and swine, dairy production is not usually thought to be subject to large technical economies of scale in production. Results from the three countries were dairy was studied are found in Figures 8.8 to 8.10. A different pattern is observed in Western India (Gujarat) and Northern India (Haryana). In Gujarat, where the cooperative movement is very active and has done much to increase the profit efficiency of smallholders (like contract farming for broilers in the Philippines), mean efficiency declines with increasing production up to the 80-150 liter/day production level (a herd of 20-30 animals, depending on composition). In large part this decline reflects the effect of not costing family labor. Above 150 liters a day, a different kind of dairying enterprise is involved, and mean efficiency increases. In Haryana, where more larger scale dairy farmers are found, scale related differences in mean efficiency are less pronounced: mean efficiency does fall slightly at first, but then levels off.

Results for Brazilian dairy (Figure 8.9) show decreases in mean profit efficiency up to 50-70 cows, and increases thereafter. This is a similar trend to the India results, for the same reasons: the efficiency advantage to smallholders of not costing family labor is less with increasing scale and the largest scales start bringing in new technologies and practices. It is noteworthy that in both the India and Brazil cases, smallholder dairy farmers are more efficient than very large-scale, as long as family labor is not costed. On Thai dairy farms, (Figure 8.10), mean profit efficiency is fairly stable up to 50 cows, and declines thereafter.

The overall result from comparing relative profit efficiency across countries, commodities, and degree of vertical integration are that small farms are not less efficient at securing profits per unit of output when family labor and environmental externalities are not costed. Following the argument in Chapter 1, smallholders at least have a chance, although large-scale producers may still drive them out by driving unit margins very low in absolute terms. Hypothesis 2 is therefore not supported by this study, with the possible exception of the largest producers of broilers and finished hogs. It is clearly not the case for dairy.

Figure 8.8 Relative profit efficiency of India dairy farms by size and region

Source: Sharma, V.P., et. al., Annex III.

Figure 8.9 Relative profit efficiency by farm size for Brazilian dairy

Source: Camargo Barros, G.S., et. al., Annex V.

Figure 8.10 Relative profit efficiency of Thai dairy farms

Source: Sharma, V.P., et. al., Annex IV.

Second, the efficiency advantage of smallholders increases when going from the backyard to the smallholder commercial model, but disappears fairly quickly with increasing size of operation, as the unit cost advantages of a "free" stock of family labor become less important. The smallholder commercial model, which is really an improved version of the backyard model, would seem to be a viable target for technology and institutional development.

Dairy production clearly is most efficient at small (not tiny) scales, consistent with 20 to 30 cow herds. Dairy clearly is a smallholder activity, and there is potential for keeping smallholders involved in poultry and swine, particularly with vertical coordination. All this abstracts from possible economies of scale in collection, processing and distribution of products such as milk and poultry, where transaction costs on the marketing (not production) side are major. The next section will look in detail at what factors explain why specific farms were more or less profit efficient.

Third, vertical coordination such as contract farming and dairy cooperatives clearly improve the relative profit efficiency of smallholder farmers, even if in some case (e.g. India broilers in Chapter 6), unit profits were lower for contract farmers. Hypothesis 7 is supported by the results above, although there is more work to do in this area. Contract farming works to improve efficiency (and thus competitiveness) by reducing transaction costs. The next section gives insights on which transaction costs and other factors explain why specific farms are profit inefficient.