INCENTIVES- DISINCENTIVES IN TRANSNATIONAL VALUE CHAINS
Lorenzo Giovanni Bellù (FAO-TCSP)
Piera Tortora (FAO consultant)
April 2010
1. Introduction 1
1. Value chain structures 1
2. The case of the Cotton Value Chain in Burkina Faso 5
3. Accounting for incentives or disincentives to producers in global value chains 9
4. Conclusions 10
5. References 11
1. Introduction
This note is the zero draft, to be further discussed and validated, of the chapter in the MAFAP methodology addressing the issue of transnational value chains, i.e. value chains covering different countries from the origin to the final destination of a commodity and/or related processed products, in which selected agents operate, directly or indirectly, in more than one country and dispose of market power at one or more stages of the chain. The section 1 presents a general discussion of concentration issues in value chains. In section 2, the cotton value chain in Burkina Faso is presented as a possible paradigmatic case of a global value chain where regulated local monopsonies at the processing stage exist, with the same agents acting to a good extent both at national and international level. Section 3 presents a simplified accounting framework highlighting relevant variables for policy monitoring and policy making in global value chain. Section 4 provides some concluding remarks for the issues addressed in the note.
1. Value chain structures
The form of vertical structure of a value chain impacts on both the price formation at every stage of the chain and on results of policies. On the one hand, concentration at one stage of the value chain alters the relative upstream-downstream bargaining power and the related terms of trade in ways that could either benefit or affect adversely the final consumers and/or the primary producers. On the other hand, depending on the vertical structure, policies that benefit firms at a certain stage of the vertical chain may either operate at the expense of firms at another (linked) stage of the same chain or benefit the entire chain. Moreover, alternative market structures entail very different net welfare and distributional effects of trade policy measures (see Messerlin 2001, Larsen 2002).
In monitoring policies we are interested, among other things, to the spread between a commodity price and some efficiency benchmark, as the result of market and/or policy failures. Such spread can be seen as an approximation of the incentive or disincentive relative to the efficiency benchmark.
The relevance of the incidence of market structure on price formation can be illustrated by a simplified example (we follow Gilbert, 2006). Let us focus on two stages of a given value chain: production and processing, and assume that the first operates in competitive markets, while the latter in monopoly. The fact that processing firms operate in monopoly implies that the processed commodity price is higher than it would be if firms operated in competitive markets ( ) and that the processed commodity supplied quantity is less ( ). In order to see what is the impact of this on the price and quantity supplied of the primary commodity, we assume linear demand and supply functions, and denote supply function of the primary commodity with Sr, while the supply function of the processed commodity with Sp. Processing costs and margins sum-up to a constant “c” and quantity losses in processing are zero. Figure 1 shows that the monopoly in the processing stage of the chain reflects negatively also on the upstream agent, i.e. the producer of the primary commodity, with a reduced raw commodity price compared with the one which would prevail under competitive markets( ) as well as reduced quantity traded ( ). This is due to the fact that the monopolist find its equilibrium not where the supply meets the demand schedule, as in point a, but where the supply meets the Marginal Revenue (MR), in point b.
Figure 1: Market structure and prices in a simplified value chain
Along agricultural value chains, imperfect competition is thought to be widespread in food processing and retailing, while more competitive markets usually characterize the upstream (primary commodities production) stages of the chain.
In figure 2, four theoretical cases of value chain structures are presented. The four cases are distinguished on the basis of market concentration and value added shares at each stage of the value chain. Case I illustrates market concentration in the processing and wholesale/retail stages and competitive markets at the production stage. Higher value added associated with downstream stages is assumed. Case II depicts the case of a monopoly at the processing stage which buys all the production available on the market from small scale producers and sells to wholesalers/retailers which operate with some degree of market power. The assumption of higher value added levels associated with downstream stages is retained. In Case III, the value chain is characterized by some degree of vertical integration between the production and processing stages, where agents operate in less then perfectly competitive markets. Higher concentration and higher value added are associated with the downstream stage. In Case IV, a key role in the value chain is played by middlemen, who exert monopsonistic power over small scale producers and avoid been squeezed by agents operating in downstream stages of the chain.
Figure 2: Typologies of value chain structures
Many value chains, including export-oriented ones, are also characterized by some forms of monopsonies on the output side and/or monopolies on the input side for primary producers. These are often organized on a geographical basis, in such a way that a single producer in a given zone cannot access alternative sources of inputs or sell its output through alternative channels. Unregulated monopolies increase the price of inputs paid by producers with respect to prevailing prices under competitive markets. On the other side, monopsonies reduce the price of the output as well as its optimal quantity. Figure 3 illustrates the latter case. Assuming “well-behaved” supply and production functions, the monopsonist (e.g. a sugar refinery) maximizes his profits when the marginal cost of the input (MC) (e.g. the sugarcane) equals the Marginal Revenue Product (MRP). Note however that, due to the fact that the monopsonist is the only-one who buys that input, he faces the whole supply curve of the sugarcane producers. This implies that the marginal cost curve (MC) for him is steeper than the supply curve S[1]. He will purchase the quantity Qe which is lower than the competitive equilibrium quantity Qe’. For that quantity he pays the Price Pe, which is lower than the competitive equilibrium price Pe’.
Figure 3. Monopsony on the output side of primary producers.
2. The case of the Cotton Value Chain in Burkina Faso
Cotton is the main export crop of Burkina Faso, covering between 50 to 60% of the export revenues, depending on the years.
.
The cotton Value Chain in Burkina Faso is characterized, a much as the cotton value chains in other countries, by three main stages: Primary production, (seed cotton), ginning (cotton fibre), bailing and trading. Primary production is essentially a relatively small-scale farm activity (350,000 producers with 8 hectares on average).
The seed cotton (primary commodity), is non-tradable internationally. It needs to be processed (ginned, cleaned and bailed) before selling it to foreign buyers.
Collection and ginning activities are organized as local monopsonies (figure 4). Since 2004, a national law attributed to the former national monopsonist SOFITEX, a joint venture of the government Burkinabè with the state-owned French company DAGRIS[2], the local monopsonies in the west provinces of the country. Two additional ginning companies were created and were attributed two local monopolies: the FASOCOTON and the SOCOMA, covering the central and eastern provinces respectively.
FASOCOTON is controlled by the Swiss multinational company REINHART, one of the major cotton traders world-wide, while SOCOMA is controlled by DAGRIS. The partial privatization of DAGRIS by the French government in 2008 shifted the control of DAGRIS, as well as SOCOMA, to the French holding GÉOCOTON. This company is controlled by ADVENS (51%), a French multi-national corporation which owns interests in agro-industrial, logistics and transport activities in Senegal and Mali, and participated by CMA-CGM (49%), the world’s third-largest container shipping French company (APE 2008)[3]. These local monopsonies export their cotton mostly through REINHART, and DUNAVANT[4], the American largest private-owned cotton merchandiser in the world.
Figure 4: Monopsony zones in Burkina Faso.
REINHART, DUNAVANT as well as GEOCOTON are, directly or indirectly (via controlled companies), both cotton processors in various African countries and cotton traders.
At the beginning of the season, the monopsonists announce a floor price for the seed cotton (the same for all the three monopsonists), which is the 95% of the “Pivot price”, a reference price based on the average international price of the cotton fibre of a given quality in the last five years (Cotlook A index), taking into account a technical conversion factor between seed cotton and cotton fibre (around 3:1) and some standard processing costs. During the campaign, farmers are paid the floor price at the delivery of the seed cotton, net of the cost of the inputs they received at the beginning of the campaign and related interest. At the end of the campaign, the “ex-post” price of seed cotton is calculated using the same criteria applied for the pivot price but this time, the past five-year average Cotlook A index is replaced by the average “Cotlook A index” of the campaign. If the “Ex-post” price is lower than the floor price, the monopsonists receive a compensating payment from a stabilization fund (“fond de lissage”). If the “ex-post” price of the seed cotton is comprised between the “floor price” and 101% of the “pivot” price, the monopsonists pay to the farmers the difference between the “Ex-post” price and the “floor” price. If the “Ex-post price exceeds 101% of the “pivot” price, the part up to the 101% goes to farmers, while the part exceeding the 101% goes partly to the “stabilization” fund, partly to the monopsonists and partly to the farmers, according to an algorithm which considers the level of the surplus and the needs of the fund.
Note that at the beginning of the season, the cotton companies act under uncertainty. They assume that, on average, a given output price (cotton fibre) will prevail in the markets where they operate. With reference to figure 5, the marginal revenue product curve MRP will be an expected-one. They will therefore announce a floor price for seed cotton close to Pe, in order to stimulate farmers to produce the quantity Qe.
Figure 5. Structure of cotton value chain in Burkina and ownership of cotton companies
Three issues related to pricing mechanisms in the Burkina Faso cotton value chain must be noted:
1) the cotton company do not pay the producers on the basis of the price they actually receive by the international traders, but on the basis of an index that is, by construction, a systematic underestimate of the actual market prices[5]; This could configure a systematic “disincentive” to cotton seed producers, not captured by the gap between the “official” international price (the Cotlook index) and the price received by the producers.
2) This “disincentive”, as long as the producers’ supply curve is rigid, might translate almost exclusively in a reduction of their income, rather than also in a reduction of output, which would affect also the cotton companies. This rigidity could be the result of institutional factors, such as the influence of cotton companies on farmers via agricultural input availability, “informal” social pressure etc.
3) The price actually received by the cotton company might not be the same price received by the international traders (net of a “normal” trade commission) due to the fact that the sellers are directly or indirectly controlled by the traders[6]. They are indeed to a good extent two sides of the same economic subject, due to the substantial (if not formal) vertical integration between them. This vertical integration may constitute a constraint regarding the choice of the international trader[7].
The latter point, on conceptual grounds, presents some analogies with the classical “Principal-Agent theory” (Spence, A.M and Zeckhauser R. (1991) [8]. Here the principal (the cotton company) relies on the services of the agent (the international trader). Clearly, the cotton company has limited instruments to verify the performances of the international trader. However, if the cotton company and the trader were completely different economic subjects, the first should be free to choose among traders, the one(s) which maximises the company’s revenue, i.e. who pays the highest price for cotton. On the contrary, in situations where the trader itself has a say on the choice of the trader by means of its control on the company itself, the company may not be free to maximize its revenue. This collusion between part of the cotton companies and the international traders is represented in figure 5 by the dashed rectangle surrounding both the cotton companies and the international traders.
By lowering the prices paid to domestic companies below those actually received on international markets, foreign investors (i.e. the international traders) generate profits downstream, by shifting losses upstream. This is a well known mechanism to inflate and expatriate profits used by transnational companies (see Brealey and Myers, 2007)[9]. By the point of view of the producers, with reference to figure 3, the collusion between cotton companies and international traders results in lower seed cotton prices, as, in absence of collusion, the marginal revenue product curve would be MRP”, generating the equilibrium price Pe” > Pe.
All this translates in a further disincentive to producers which is not captured by the gap between the “official” international price and the price actually received by the producers.
3. Accounting for incentives or disincentives to producers in global value chains
In general terms, accounting for rents in global value chains implies looking at both the segment in the producing and exporting country as well as the segment in the importing one. Assuming that the producing country is a small country characterized by some inefficiency and market imperfections in processing and transport technologies, the producer price of the exported commodity can be represented as: