33172 Environmentally and Socially Sustainable Development 251 June 2005 Findings reports on ongoing operational, economic, and sector work carried out by the World Bank and its member governments in the Africa Region. It is published periodically by the Knowledge and Learning Center on behalf of the Region. The views expressed in Findings are those of the author/s and should not be attributed to the World Bank Group. Reform Experience with the Tanzanian Cotton Sector by John Baffes Cotton, Tanzania's second largest gional cooperatives licensed to crop after coffee, was introduced at trade cotton. In 1994/95, some 22 the turn of the century by German private companies started trading settlers as a plantation crop but later cotton, and 8 new private ginneries efforts focused on smallholder produc- were constructed. tion. Output rose considerably with By 1996/97, private businesses the releases of new varieties, along were purchasing almost half of all with better organization of the sec- cotton. Private traders and tor following establishment of the ginneries were able to capture a Tanganyika Lint and Seed Market- considerable share of the market ing Board in 1956. By 1966, produc- because they offered higher prices tion reached 80,000 tons, or 0.75 per- than cooperatives and paid cent of world production of 10.7 mil- promptly. Some private ginneries lion tons. A turning point came in the also engaged in contract farming, 1960s following the government's providing inputs to producers who shift to central planning thinking, agreed to supply cotton in return. whereby several cooperatives were The ginneries and producers usu- formed and began building ginneries, ally established a minimum price training staff, and taking over enter- at planting time, but the price prises from foreign owners. could be adjusted if the market By 1968, the cooperatives and the price was higher during the har- Cotton Board had gained a monopoly vest. of virtually all commercial activities of the sector. They soon became large Assessing the reforms bureaucracies and failed to pay ad- equate attention to the needs of the A principal argument for policy sector (see box), in turn rendering reform was the large gap between reforms the only feasible alternative. world prices and the prices re- The first step toward reforms was ceived by cotton growers. During taken in 1989/90, when the govern- the six seasons prior to the re- ment launched the Agricultural Ad- forms, the average grower's share justment Program. More reforms was 41 percent of the cotton export came with the Cotton Act of 1994, price. In the six seasons after the Findings when the government formally elimi- reform, the share was 51 percent. nated the monopoly held by the Board Payments were made more and the cooperatives and allowed promptly as well. Before the re- competition in marketing and gin- forms, growers often had to wait as ning. At the time there were 14 re- long as two years for payment. With inflation running at 20­30 percent a year, the value of their ward as causes of the decline in quality--the open- payments was halved by the delay. The reforms also meant ing of new ginneries and the mixing of grades--are that input prices increased considerably, primarily reflect- more ambiguous. There are also two factors pointing ing the removal of input subsidies. The average cost of pes- to an increase in quality. First, prior to the reforms, ticides, for instance, rose from 1,600 Tanzanian shillings cotton had to wait in storage for as long as a year (Tsh) a kilogram in 1993/94 to Tsh 5,000 in 1998/99. before it was sold. Since the reforms, cotton sales have Lack of reliable data makes it difficult to analyze the im- been expedited, as evidenced by the shorter period of pact of reforms on supply, but a simple comparison before trading in Northern Europe which fell from 36 weeks and after the reforms using International Cotton Advisory between 1984/85 and 1993/94 to 14 weeks after the Committee data indicates that cotton output averaged reforms. Second, prior to the reforms a large share of 64,800 tons in the five seasons before the reforms and 66,400 cotton went to the domestic textile industry. Although tons in the five seasons after the reforms, a very moderate not documented, it is likely that lower quality cotton increase. But other data sources throw even this modest was consumed domestically, while higher quality cot- increase into doubt. Three government sources show no ton was exported. Since the reforms, most cotton is supply response after the 1994/95 policy changes, which is exported. So, although the quality of exported cotton consistent with the large reduction in input use that took may have declined on average, it does not necessar- place after the reforms. ily follow that the quality of all cotton has declined. Once input supply and credit for purchasing inputs were One measure of cotton quality is the premium that no longer integrated into a single cotton marketing chan- Tanzanian cotton commands over the A Index (or nel, use of inputs declined sharply. Loss of the single mar- "world" price of cotton). In the seven seasons prior to keting channel pushed up the costs of marketing chemi- the reforms, Tanzania's premia averaged 10.0 per- cals and led to a collapse in supply and distribution. And cent over the A Index. In the seven seasons after the most farmers could not purchase chemicals at market reforms, the premia fell to 8.8 percent. Thus, while prices either because they could not afford credit (due to the quality of cotton may have declined marginally, it high interest rates) or because credit was denied by the has by no means deteriorated, as most reports show. Banks (due to high probability of default). Two credit provi- The sector's ginning capacity has increased con- sion schemes were initiated to boost input use: the siderably while marketing has improved. Before the government's Agricultural Input Trust Fund in 1995 and reforms there were 34 ginneries in Tanzania. Dur- the Board's Cotton Development in 1999. The first scheme ing the first three years of the reforms 17 new pri- ended in failure. The second, still in force, faces numerous vate ginneries were built mainly in response to coop- problems; interviews with industry, for example, allege that eratives' refusal to allow private traders to gin their corruption is pervasive in the distribution chain and that cotton on a contract basis. The new private ginneries substantial quantities of chemicals are diverted to the free added 17,000 tons of monthly capacity to the existing market. 19,000 tons. On the marketing side, during the sec- Similar problems plagued seed distribution. Until 1997, ond year of reform the private sector took over more seed distribution was handled by the cooperatives, which than a quarter of cotton marketing. As of 1997/98, were required to retain one-tenth of the seed for free distri- the Board withdrew completely from cotton market- bution (the rest was milled for oil). Once private traders took ing. Currently, most cotton is marketed by the pri- over, the quality of seeds deteriorated and growers received vate sector. far less than the required 10 percent of seeds. After the Finally, considerable achievements have been made reforms, at least eight new oil mills were built, increasing in research. The cotton research station has devel- milling capacity by almost 50 percent. This intensified com- oped a new cotton variety which, according to re- petition for seed, thus diverting it to oil mills. search data and interviews with researchers, is su- Quality deterioration has been cited often as another perior to the existing varieties. However, achieving negative outcome of the reforms. Among a host of reasons, the higher yields promised by the new variety requires the following four appear to have affected quality: decline in multiplication and the release of enough quantities input use; abandonment of zoning; mixing of cotton variet- to replace the older varieties and to forestall mixing ies; and El Niņo and La Niņa for the 1997/98 and 1998/99 with existing ones. That, in turn, will require a con- seasons. The effects of two other factors sometimes put for- certed effort by the various institutions including the Board and the Ministry of Agriculture. contrary, if the market price is lower than the in- dicative price, farmers may view the indicative price Constraints as a guaranteed price and refuse to sell their cotton at the prevailing price. Despite the achievements, several issues need to The power of the Board also appears to be excessive be addressed, including taxation, power of the Board and goes far beyond what a regulatory agency should and the Ministries, and reliability of statistics. Cot- do. According to the 2001 Cotton Industry Act, the ton, like all other export crops, is subject to a host of Board is entitled to "... do anything or enter into any taxes, levies, and fees administered at both district transaction which in the opinion of the Board is cal- and central government levels. One study found that culated to facilitate the proper and efficient carrying the tax burden on cotton was almost 15 percent of the out of its activities and the proper exercise of its func- producer's gross receipts. Another government report tions under the provision of this Act." found that nominal protection rates in the sector in- Another concern is the poor quality of statistics. A creased from ­17.9 for 1986-89 to ­69.0 for 1990-93 recent government report contained such discrepan- and ­67.7 for 1994-99 (the negative sign indicates cies in its statistics of cotton production that attempt- taxation). Moreover, numerous reports have shown ing to gauge the degree of supply response by com- that there is considerable tax avoidance and evasion, paring pre- and post-1994 averages would yield out- a not surprising outcome in such a high taxation comes ranging from a 15 percent reduction to an 8 environment. The recommendation regarding taxa- percent increase in cotton output. Improving collec- tion is simple: the tax code should be simplified and tion and dissemination of statistics not only will guide taxes should be reduced. policy makers in making the proper policy decisions One of the Board's responsibilities is quality inspec- but will also help the private sector in making the tion at buying posts. However, a lack of adequate re- proper investment choices. sources severely limits these inspections. In prac- tice, ginneries send samples to the board from each Cooperative Unions and the Private Sector--Percep- bale of cotton ginned, and the Board simply informs tions and Reality ginneries if samples are deficient. In order to be more effective, the Board hired a private company which Members of cooperatives along with government offi- placed inspectors at every ginnery in order to moni- cials are skeptical of the private sector's involvement tor the quality of cotton. These inspectors were given in marketing and trade activities in Tanzania. Their the right to "reject" cotton. Not only is it not clear views reflect, in part, 30 years of experimentation what "rejection" means, but the quality control with central planning. Consider the following inter- scheme appears not to be functional at all. Corrup- view with a high ranking official of a cooperative. tion also appears to have been a major problem. In- When asked for his views on private sector involve- terviews with ginning staff indicated that quality in- ment in cotton marketing, he replied: "The problem spectors would "accept" or "reject" cotton on the basis with the private traders is that you cannot control of whether a side-payment of $US 30 per truck load them! The World Bank should support cotton to be had been made. Quality of cotton is something that produced. It should support the cooperative union and should concern the private sector participants, not forget about the private sector." This thinking ech- the Board. oed the Arusha Declaration some 34 years earlier: By statute, the Board is supposed to ensure free "The way to build and maintain socialism is to en- competition, fair trade, and to set and monitor indica- sure that the major means of production are under tive prices as established by market forces. For ex- the control and ownership of the Peasants and Work- ample, in 2002/03 the Board announced an "indica- ers themselves through their Government and their tive" price of 140 Tsh per kilogram. However, this is Cooperatives." a dangerous practice that should be discontinued. If When asked what kind of assistance would be the market price is above the "indicative" price, the needed to overcome the difficulties the cooperative announcement motivates traders to collude and union faced, he replied: "I need finance. If I am given refuse to pay more than the indicative price. On the enough finance, I could do better." Though unstated, implicit in this response was the assumption that small quantity attributed to each of the two major co- the loan would not be at market rates or, if a history operative unions. Purchases by SHIRECU and NCU of debt accumulation by the unions and eventual for- are estimated at 2,648 and 1,348 tonnes, respectively. giveness is any guide, would not be repaid at all. Most A local press report has referred to the huge debts cooperative unions have difficulty obtaining bank with which NCU is burdened. The article mentions a loans because they lack creditworthiness, not because figure of over seven billion shillings (roughly US banks are unwilling to lend. Since banking sector lib- $7,000,000), of which 4.5 billion (US $4,500,000) are eralization, obtaining finance has not been a prob- owed to the National Bank of Commerce." lem for creditworthy borrowers. The manager of a union-owned cotton oil mill, currently leased to an This article was prepared by John Baffes, Senior Econo- entrepreneur, noted that there had been no difficulty mist, Development Prospects Group. It is an extract of a obtaining a $340,000 loan to purchase the factory and full document, Tanzania's Cotton Sector: Con- another $300,000 loan to rehabilitate it--a 5-year straints and Challenges in A Global Environment, loan at 12 percent interest from a private bank. Africa Working Paper no. 42, December 2002, When asked about the cooperative's achievements www.worldbank.org/afr/wps/wp42/htm (also pub- prior to the reforms, the official responded: "We built lished in Development Policy Review, vol. 22, no. 1, all this infrastructure!" "All this infrastructure" is now January 2004, pp. 75-96). For more information, please insolvent and up for liquidation (a cotton oil mill), run- email jbaffes@worldbank.org. down or underutilized (several ginneries), or aban- doned (two 2,000-hectare cotton farms). The problems with the cooperatives were vividly described as follows in a review of the sector: "Inter- nal corruption, inefficiency, maintenance of loss- making non-core activities, ... over-manning, under- qualified management and outdated and poorly main- tained plant ... remain largely uncorrected." This de- scription is similar to that of a government review which reported: "A study ... on the quality of ginnery manpower working in ginneries owned by the coop- erative unions found that: most of the staff were over 45 years of age; 90 percent of them were unqualified for their posts; most ginnery managers had no formal management training; above 80 percent of the ginnery workers did not meet the required minimum qualifications and their promotions were based on experience without undergoing any further formal training." The Netherlands Development Cooperation reached remarkably similar conclusions: "Studies in 1989 and 1992 judged that financial management of the [unions] was very weak, and revealed that both the Nyanza Cooperative Union and Shirecu were in- solvent. They listed serious shortcomings in accoun- tancy practices, revealed high marketing costs, and unrecoverable bank overdrafts. In addition they men- tioned that the virtual absence of members' equity had made the unions become wholly financed by gov- ernment, chiefly through bank overdrafts." Cotton Outlook (October 28, 2002) reported: "A note- worthy of the [2002/2003 production] figures is the