EXECUTIVE SUMMARY Rural taxation policy is a major issue in many countries of Africa as they pursue more decentralized forms of governing and at the same time work to enhance the effectiveness, efficiency, and fairness of their tax systems. Tanzania has struggled with this issue since at least 1962, when it expanded countrywide the limited decentralization that had occurred under the colonial regime, then abolished LGAs in 1972 in favor of “Madaraka Mikoani,” only to reinstate them and enshrine them in the constitution in 1984. With wide powers to set tax policy and practice at local level, made possible by the Local Government Finance Act (LGFA) of 1982, Tanzania soon experienced a dizzying array of taxes and fees, with dramatically differing rates across LGAs. The situation became so extreme that some claimed that Tanzania by the late 1990s had “about 110 local authorities ... each with a different tax system” (Fjeldstad and Semboja 2000). A sustained effort at reform culminated in 2003, when the “head tax” and a series of “nuisance taxes” were abolished, and the produce cess was limited to a maximum of 5% (compared to rates as high as 20% in the past). Though the resulting system of local taxation is substantially less complex, less variable across LGAs, and less onerous than it was prior to these reforms, important problems remain, and stakeholder demands for further reform have been growing. Since the produce cess became the most important source of local revenue after 2003, much of the demand for reform has focused on it. In response to these concerns, GoT included a commitment to “reduce or abolish” produce cess when it signed the G8’s “New Alliance for Food Security and Nutrition” declaration. This study took advantage of a newly available database of LGA revenue and expenditure and complemented it with fieldwork in 27 LGAs with varying levels of reliance on the produce cess. Its overall purpose is to generate new empirical understanding that contributes to the on-going debate on produce cess and that informs the GoT on pros and cons of potential options for reform. Key new findings include: 1. Dependence on the produce cess varies widely among rural LGAs, from 0% of total locally generated revenue in Ngorongoro to 90% in Urambo; 2. Relative to the value of their marketed production, traditional export crops generate more than three times as much cess revenue as do food crops; 3. Much potential cess revenue goes uncollected: nationally, LGAs collect not more than one- quarter of the revenue potentially available from produce cess charges. This low level of collection reflects both limited human and institutional capacity at local level and widespread tax evasion, some of it likely featuring the collaboration of some local officials; 4. Because it is charged on the gross value of production, current cess rates can result in very high tax (even confiscatory) on net revenue among farmers that use a large amount of inputs but experience small net margins; Confirmed previous findings include: 1. With the reforms of 2003, local revenue fell sharply as a share of total LGA revenue, from 20% to a current level of 7%. Central government transfers provide the rest. Such a low share of locally generated revenue makes meaningful decentralization quite challenging. 2. Nationally, cess contributes only 1.8% of total LGA revenue, with other local taxes accounting for 5%; 3. Yet cess is the largest source of rural LGA own revenue, at 43%. Because this revenue is very flexible (it does not come with the spending dictates that accompany central government transfers), it is highly valued by local authorities, and is largely used for Councilor allowances and other “costs of doing business”; 4. Cess rates are highly variable across LGAs, varying by a factor of as much as four (Beans in Handeni at Tshs 1000/bag vs. Lushoto at Tshs 4000/bag); 5. Tax evasion is widespread and likely a more serious problem than tax avoidance; 6. But avoidance – farmers or traders or others changing their production and marketing behavior due to the tax (and especially due to the variation over space in tax rates) – can be a serious problem in particular instances. For example, some sugarcane growers in Mvomero are considering shifting their farming activities to Kilombero due to lower cess rates in the latter; and farmers and traders report that traders favor some districts over others in their food trade due to differences in cess rates; Reform options include: 1. Abolish cess in one step 2. Gradual phasing out of cess 3. Reduce the cess rate, broaden its base, and improve capacity for collection 4. Institute a differential cess for food- and non-food crops 5. Completely remove cess in food crops, leaving it only for traditional and other export crops. Simple simulations of option 3 combined with option 4 (3% for traditional cash crops, 2% for food crops) indicate that LGAs would need to improve their efficiency in collection (the share of potential cess that is actually collected) from the current estimated 28% to 41% to maintain revenue, and would increase revenue with further improvements. Complete elimination of cess on food crops (option 5) would make LGA’s jobs quite challenging, especially if rates were reduced on traditional export crops. Leaving the rate on these crops unchanged at 5%, LGAs would have to achieve nearly 60% efficiency in their collection to maintain their current revenues; dropping the cess on traditional export crops to 3% while eliminating it on food crops would require an almost certainly unattainable 83% efficiency. Based on the analysis in the paper, and in keeping with the view that improvement in tax systems is a long-term process featuring continuous, incremental improvement, the report suggests that option 3 combined with option 4 – reducing the rate of the cess (thereby reducing its variability over space), introducing a slight differential between food crops and traditional export crops, and broadening the cess collection base by working continuously to improve the human and institutional capacity of LGAs to collect taxes in efficient and fair fashion, is likely to be the best option for Tanzania. Piloting of technological and institutional innovations such as the use of mobile money for cess payment are proposed as one way to address both the inadequate local capacity and the scope for corruption in cess collection.