IMPACTS OF WORLD FUEL AND AGRICULTURAL PRICE CHANGES:

AN ECONOMYWIDE ANALYSIS OF TANZANIA

Channing Arndt

University of Copenhagen

1 Introduction

Like many low income countries, Tanzania is a structural importer of fuels and both an exporter and an importer of agricultural products. A considerable literature exists that analyzes development policy in the context of price volatility (Deaton 1999; Cuddington 1992; Combes and Guillaumont 2002). The recent rise in world commodity prices, combined with the proliferation of new tools for analysis that have come available since the commodity booms of the 1970s and early 1990s, has also produced a broad literature. Ivanic and Martin (2008) focused on food and concluded that poverty rates might rise substantially in low income countries. Arndt et al. (2008) examined both fuel and food prices and highlighted the particularly strong effect of world fuel price rises on economywide welfare and poverty in Mozambique, a low income country that is a reasonable analog to Tanzania. They found that the price shocks of 2008 increased poverty rates by about four percentage points with more than 80 percent of the increase in poverty attributable to the fuel price effect. Later work by Arndt et al. (2012) illustrated that the stagnation in poverty rates observed in Mozambique between 2002-03 and 2008-09 could be mainly attributed to a combination of rising world fuel and food prices and disappointing rates of technical advance in agriculture.

While predicting future commodity prices is inherently difficult, the commodity price rises observed since the beginning of the 21st century have been both profound and durable. Despite the "" that has enveloped the developed world in particular, world commodity prices have remained high and these high levels appear likely to persist and may even increase if global growth rates pick up. This paper considers the implications of world commodity price rises with these prices then persisting at historically high levels through the medium term. Because Tanzanian fuel imports are large (fuels alone accounted for between 17 and 25 percent of total imports in 2007 depending on the data source), particular attention is paid to the implications of rising fuel prices. Welfare and poverty implications are assessed using a dynamic economywide model of Tanzania that captures all of the major real economy impact channels resulting from changes in world prices of key import and export .

The remainder of this paper is structured as follows. In section 2, trends in world commodity prices are considered. This analysis focuses on agricultural commodity and crude oil prices. Section 3 presents the structure of the Tanzanian economy, and section 4 summarizes the DCGE model used to analyze the implications of world price changes. Section 5 details the simulations employed while section 6 presents results. A final section summarizes and concludes.

We find that changes in world prices for fuels and petroleum derived products are highly significant for welfare and poverty in Tanzania. The effects of permanent fuel price increases on welfare levels relative to a Baseline path tend to persist. However, once fuel prices stabilize at a higher level, macroeconomic trends, such as growth in absorption or reductions in consumption poverty, tend to strongly resemble Baseline trends. A temporary fuel price , with prices rising sharply and then returning to pre-shock levels, has few long run impacts with both levels and trends of key macroeconomic variables, including poverty rates, reverting close to the Baseline.

The welfare implications of changes in agricultural commodity prices are tempered by the tendency for prices of export and import commodities to rise and fall together meaning that price increases in some import commodities tend to be offset by similar increases in some export commodities, which tempers the changes in overall terms of trade (the ratio of export to import prices). If high prices for agricultural products persist (and investors react to these prices shifts), changes in production structure through time can allow the Tanzanian economy to profit from export commodity prices rises and reduce the implications of import commodity price rises. These production shifts can, in principle, lead to fairly substantial economywide welfare gains and reductions in poverty, particularly in rural areas.

2 World price changes

2.1 Fuels Figure 1 illustrates trends in real crude oil prices since 1950 using the United States Consumer Price Index as a deflator. The prices are presented in real 2012 USD. Oil prices are characterized by a long period of stability up to the 1970s. Prices then rose throughout the 1970s with a real peak of 113 USD per barrel in early 1980 (measured in 2012 USD). From this peak, a substantial decline in the crude oil price occurred from 1980 to 1986 followed by a second prolonged period of relative stability from 1986 to about 2000. A marked upward trend has characterized the first dozen years of the new millennium. Prices in the new millennium have also been characterized by increased volatility. As of mid 2012, world oil prices have more than tripled since the beginning of the new millennium on January 1 2000.

[Insert Figure 1]

Viewed in long historical series and in real terms, current prices are high corresponding roughly to the peak values experienced in the late 1970s and early 1980s. From 1979 to 1983, real oil prices, inflated to current dollars, were above 70 USD per barrel for 46 months. Since 2005, real oil prices have been above 70 USD per barrel for 72 months and remain above that level today. While predicting future oil prices in inherently tricky, futures markets indicate crude oil prices at roughly current market levels out to about 2020.

2.2 Global agricultural markets With the exception of occasional price spikes, the prices for agricultural commodities received by farmers and paid by consumers declined broadly from the initiation of the to the end of the 20th century. Figure 2 depicts an index of prices received by farmers for 48 commodities in the United States divided by the general level of prices as measured by the United States consumer price index in order to account for inflation beginning in 1954.1 During the latter half of the 20th century, prices received by farmers declined nearly continuously in real terms. This general decline in prices was interrupted in the early 1970s when prices received by farmers briefly rose above the levels of 1954, the first year for the depicted price series. Nevertheless, this rise was short lived. By the end of 20th century, prices had declined by about 60 percent from the levels observed in mid-century.

[Insert Figure 2]

Since 2002, this decline has ceased and been replaced by a consistently rising trend in the index. Rapid visual inspection of the level of the index makes the price increases observed since 2002 appear to be rather small. However, in proportional terms--the more economically relevant measure-- they are considerable. Relative to the 20th century trend, they are even more pronounced. By this index, prices relative to the 20th century trend are proportionally higher than at any time in the past 50 plus years. As of mid 2012, prices were 74 percent higher than the 20th century (linear) trend would dictate. By this measure, the current rise in agricultural prices is larger than the price shock of the 1970s by nearly a factor of two (74% in 2012 versus 39% in 1973).2 The recent rise in agricultural commodity prices may not be another spike as in the 1970s. A recent comprehensive review undertaken by the Foresight project concludes that "there is a strong likelihood that food prices will rise significantly over the next 40 years" (Foresight, 2011, p. 64). In sum, the recent increases in agricultural prices either represents an extraordinary deviation from the trend of decreasing agricultural prices, implying an upcoming dramatic decline in commodity prices just to return to the levels observed in 2002, or an end to the trend.

1 Data are from the National Agricultural Statistics Service of the United States Department of Agriculture. The United States is chosen because it offers a long price time series. The United States is also a major agricultural producer, importer, and exporter making it a reasonable global price bellwether. The application of a US price index to Tanzania is limited by substantial differences in the commodity composition of output and the influence of US price policies. Nevertheless, the US index is adequate for the general purposes considered here.

2 The trend line is determined based on a linear regression of the index values versus time from 1954 to 2002. Even if one uses the full sample (1954-2012), the departure from linear trend in recent years is unprecedented. 3 Structure of the Tanzanian economy

Table 1 shows the structure of the Tanzanian economy in 2007, which is the base year for our analysis using the economywide model. Note that this year is convenient in that it allows us to consider the implications of the considerable world price movements that occurred over the period 2008-2011 from the most recent possible base. Agriculture accounts for nearly one third of gross domestic product (GDP) and more than four-fifths of employment. Within agriculture, most farmers are smallholders with average land holdings of 1.6 hectares. They produce most of the country's food, which dominates the agricultural and downstream manufacturing sectors. Tanzania also imports foods (mainly cereals), which account for about five percent of total imports. This dependence on food imports stems in part from smallholders’ low crop yields and a reliance on traditional rain-fed farming technologies. Larger-scale commercial farmers are more engaged in nonfood export crops, such as coffee, tobacco and tea, which together account for nearly a third of total merchandize exports.

[Insert Table 1]

Commodities strongly influenced by the world price for crude oil, including liquid fuels, fertilizer, rubber, and other chemicals together account for 26 percent of total imports. Of these four, liquid fuels represent the single largest import share at nearly 17 percent of imports.3 Unlike agriculture, these commodities account from very small shares of value added or exports. Hence, increases in the prices of fuels and derived products represent almost purely a terms of trade loss.

4 Description of the economywide model

Econonywide models, also called computable general equilibrium (CGE) models, are often used to examine external shocks and policies in low income countries. Their strength is their ability to measure linkages between producers, households and the government, while also accounting for resource constraints and their role in determining product and factor prices. These models are, however, limited by their underlying assumptions and the quality of the data used to calibrate them. Tanzania's initial economic structure, as presented in Table 1 and discussed in section 3, will strongly influence model results.

Our model belongs to the neoclassical class of CGE models.4 Economic decision-making in the model is the outcome of decentralized optimization by producers and consumers within a coherent economywide framework. The model identifies 58 sectors (i.e., 26 in agriculture, 22

3 Data from UN Comtrade point to even higher import levels for fuels in 2007.

4 The model’s mathematical specification is discussed in Diao and Thurlow (2012). industries and 10 services). Based on the 2000/01 Household Budget Survey (HBS) (NBS, 2002), labor markets are segmented across three skill groups: (1) workers with less than primary education; (2) workers with primary and possibly some secondary schooling; and (3) workers who have completed secondary or tertiary schooling. Agricultural land is divided across small- and large-scale farms using the 2002/03 Agricultural Sample Survey (MINAG, 2004).

Substitution possibilities exist between production for domestic and foreign markets. Profit maximization drives producers to sell in markets where they achieve the highest returns based on domestic and export prices. Further substitution possibilities exist between imported and domestic goods. Households, firms, and government minimize costs in sourcing goods from domestic and foreign markets while accounting for potential differences in the characteristics of domestic and foreign commodities. Under the small-country assumption, world demand and supply is perfectly elastic at fixed world prices, with the final ratio of traded to domestic goods determined by the endogenous interaction of relative prices. Production and trade elasticities that govern these substitution possibilities are drawn from Dimaranan (2006).

The model distinguishes between 15 representative households (rural farm, rural nonfarm and urban nonfarm groups by per capita expenditure quintiles). Households receive income in payment for producers’ use of their factors of production, and then pay direct taxes, save and make foreign transfers (all at fixed rates). Households use their remaining disposable income to consume commodities with the allocations sensitive to movements in prices and income. In order to estimate the implications of world price shocks for poverty rates, the CGE model is linked to a micro-simulation module. In particular, each of the more than 22,000 respondents in the HBS is linked to their corresponding representative household in the CGE model. Changes in commodity prices and households’ consumption spending are passed down from the CGE model to the survey, where total per capita consumption and poverty measures are recalculated.

The government receives revenues from direct and indirect taxes, and makes transfers to households and the rest of the world. The government purchases consumption goods and services, and remaining revenues are saved (budget deficits are negative savings). All private, public and foreign savings are collected in a savings pool from which investment is financed.

The model includes three macroeconomic accounts: government, current account, and savings- investment. To balance these macro-accounts, it is necessary to specify a set of “macro-closure” rules that provide a mechanism through which macroeconomic balance is maintained. A savings- driven closure is assumed in order to balance the savings-investment account. This means that households’ marginal propensities to save are fixed, and investment adjusts to income changes to ensure that the level of investment and savings are equal in equilibrium. For the current account, it is assumed that a flexible exchange rate adjusts in order to maintain a fixed level of foreign savings. In other words, the external balance is held fixed in foreign currency terms. For the government account, direct tax rate rates are fixed and the fiscal deficit adjusts to equate total revenues and expenditures. Finally, the producer price index is chosen as the model’s numéraire, and so all product and factor price movements are relative to this fixed price index.

The description of the model in the above paragraphs ignores the time dimension. The model, as described above, is static in that it adjusts to shocks by restoring equilibrium in all markets subject to macroeconomic accounting identities and the macro-closure rules applied. The model is rendered dynamic by solving a series of static equilibriums. Unlike full inter-temporal models, which include forward-looking expectations, the recursive dynamic model used in this paper adopts a simpler set of adaptive rules, under which investors essentially expect prevailing price ratios to persist indefinitely. Under this specification, the investment levels of the previous year are used to augment sectoral capital stocks, net of depreciation. The model adopts a “putty-clay” formulation, whereby each new investment can be directed to any sector in response to differential rates of return to capital, but installed equipment must remain in the same sector. Unlike capital, growth in labor and land supply is determined exogenously. In addition, labor and land can be reallocated across sectors in response to changing economic conditions. Sectoral productivity growth is also exogenous, but may vary by factor. Using these simple relationships to update key variables, we can generate a series of growth paths based on different world price scenarios.

5 Simulations

Table 2 summarizes the simulations imposed on the model in order to investigate the implications of changes in world prices for fuels and derived products and agricultural commodities for welfare and poverty in Tanzania over the period 2008-2015. The details of each simulation are discussed below.

[Insert Table 2]

The purpose of the Baseline scenario is to generate a reasonable counterfactual against which the other scenarios can be compared. World prices are constant at 2007 levels in the Baseline. Subsequent scenarios modify world prices but leave all other aspects of the Baseline scenario constant. Differences between the Baseline and other scenarios can then be attributed to the world price shocks. The Baseline scenario is not a forecast; rather it is a reasonable benchmark against which other scenarios containing world price shifts can be measured. In the Baseline scenario, population growth is set at 2.5 percent per year during 2007-2015. Total labor supply grows at 2.1 percent per year in all scenarios, reflecting our assumption of full employment (perfectly inelastic labor supply). Skilled labor supply grows faster than unskilled labor reflecting gradual improvements in educational attainment. Livestock stocks and agricultural land expand at one percent each year, capturing rising population density, especially in rural areas. Total factor productivity (TFP) growth is set at one percent per annum in agriculture and three percent per annum in non-agriculture. These assumptions combine to generate a 4.7 percent per annum GDP growth rate over the period 2007-2015. The Historical simulation applies world price shocks as observed between 2007 and 2011 (see Table 3). World price levels observed in 2011 are applied to the 2012-15 period (e.g., real prices remain constant from 2011). It is important to highlight that, while the Historical simulation applies observed world price changes over the period 2007-2011, the simulation is not an attempt to replicate history. First, where credible international price data are not available, no shocks are applied. Second, there is no attempt to track other salient events, such as droughts or particularly good agricultural seasons, which may impact welfare and poverty rates over the period. Instead, the Historical scenario simply relies on available world price data to generate an interesting series of world price shocks. The model is then used as a simulation laboratory to consider the implications of these price shocks for the Tanzanian economy.

[Insert Table 3]

The next three scenarios-- H-Fuel, H-Agriculture, and H-AgXsugar-- examine subsets of the historical shocks in order to gain greater insight into the implications of world price changes. The H-Fuel scenario only considers world prices changes for petroleum, fertilizer, other chemicals and rubber. All other prices, including those for agricultural products, remain at baseline 2007 levels. The H-Agriculture scenario only considers changes for agricultural products leaving all other prices at base levels.5 The H-AgXsugar is the same as H-Agriculture except for the export price for sugar, which is maintained at baseline 2007 levels. This scenario is run because Tanzanian sugar exports are mostly directed towards the European Union where they have enjoyed preferential access and high prices. It is quite possible that the prices received by Tanzanian processed sugar exporters since 2007 have remained constant or declined, even while world market prices have risen, due to preference erosion on exports to the European Union.

The final two scenarios are hypothetical scenarios and focus on import prices for fuels and derived products. In both scenarios, prices for petroleum, fertilizer, other chemicals, and rubber rise by 75, 35, 25, and 25 percent respectively in 2008. In 2009, prices of all four commodities rise a further 10 percent. In the FuelPerm scenario, these prices remain at the 2009 level for the period 2010-15. In the FuelTemp scenario, the prices drop back to the levels observed in 2008 in the year 2010 and fall back to baseline 2007 levels in 2011. These baseline levels then persist over the 2012-15 period.

6 Results

We begin by focusing on the Baseline scenario and then compare the historical prices simulations (H-Fuel, H-Agriculture, and H-AgXsugar) with the Baseline. Principal macroeconomic results are presented for these simulations in Table 4. For the Baseline scenario,

5 The set of world price shocks in H-Fuel combined with the set of world price shocks in H-Agriculture yields the complete set of world price shocks imposed in the Historical simulation. real absorption, defined as C+I+G=GDP+M-X, grows alongside real GDP at a rate of about 4.7 percent per annum.6 This Baseline growth path brings about a reduction in the national poverty rate of 5.5 percentage points between 2007 and 2015.

Because world prices are assumed constant in the Baseline, the terms of trade remain constant. However, the real exchange rate depreciates mildly over the period 2007-15. The real exchange rate is an important measure for the analysis of world price shocks. It is defined as the ratio of the prices of traded to non-traded goods. A rise in the real exchange rate implies a real depreciation of the currency. This real depreciation provides incentives for producers to increase production of traded goods (such as maize), which can either compete with imports or add to exports, at the expense of non-traded goods (such as services like haircuts). Changes in the real exchange rate represent the principal macroeconomic mechanism through which the real economy adjusts to changes in the terms of trade.

In a static analysis, a decline in the terms of trade would indicate that the world prices of imported commodities had tended to rise relative to the prices of exported commodities. As a consequence, the same volume of exports generates insufficient foreign currency to finance the same volume of imports. Absent an increase in foreign exchange from other sources (e.g., an increment to aid, an increase in remittances, or a drawdown on foreign reserves of the central bank), the economy must apply some combination of increased exports and reduced imports in order to reestablish external balance. When the terms of trade decline is driven by an increase in the price of a critical imported input such as fuel, the adjustment mechanism in developing countries is almost invariably accomplished through a depreciation in the real exchange rate and associated changes in the structure of production towards tradeable goods. Note that increased exports and/or reduced imports imply a reduction in the quantity of goods available within the economy and hence a reduction in real absorption.

[Insert Table 4]

The historical simulation provides an interesting contrast. Using 2007 trade shares to weight the commodity price shocks, terms of trade are shown to mostly decline over the 2008-15 period (the exception is 2009 where low oil prices provide a slight terms of trade gain). In keeping with this decline in terms of trade, total real absorption, which is a good measure of economywide welfare, declines relative to the Baseline in 2008 by 2.1 percent. This decline is erased in 2009 when a slight terms of trade improvement is experienced. With the rebound in oil prices, terms of trade decline once again in 2010 and then register a constant 3.9 percent loss for the period 2011- 15. Nevertheless, despite this measured terms of trade loss and contrary to the static analysis discussed above, economywide welfare, as measured by absorption, is actually nearly three

6 Foreign savings, which contain aid and foreign investment, are assumed to increase at a rate of five percent per annum. Foreign savings determines the value of exports less imports (X-M). percent higher in 2015 compared with the Baseline. In addition, the real exchange rate is appreciated and the national poverty rate is lower in the Historical scenario relative to baseline. In short, in the near term (2008), the model responds to a terms of trade loss in a manner consistent with normal static analysis (reduction in absorption, depreciation of the real exchange rate, and an increase in poverty relative to the Baseline). However, over time, the dynamics yield positive outcomes.

This can be further investigated through the additional scenarios. The H-Fuel scenario illustrates the powerful effects that fuel prices exert on economies such as Tanzania. The fuel, fertilizer, chemical, and rubber price increases illustrated in Table 3 are sufficient to decrease terms of trade by about 10 percent for the period 2011-2015. During this period, economywide welfare, as measured by absorption, is about 3.3 percent lower. National poverty rates are also about 2.4 percentage points higher. Increases in world prices of these critical commodities, particularly fuel, can be likened to a removal of goods from the Tanzanian economy. Because it is very difficult to economize on fuel imports particularly in the short run, the quantity of imports remains relatively constant despite the higher price (fuel imports in the H-Fuel scenario in 2008 are 92 percent of baseline levels). To pay for these imports, Tanzania must export more and import less, implying reduced quantities of goods available for consumption, investment, and government spending, which is reflected in reduced absorption. Less consumption implies higher poverty (unless it is somehow possible to confine the consumption reduction to the non-poor).

The H-Agriculture scenario provides the difference that helps to explain the eventual welfare gains in the Historical simulation. For every year except 2009, the agriculture shocks yield terms of trade improvements. For the period 2011-15, the terms of trade gain is significant at nearly five percent. Even though this terms of trade gain is constant over the 2011-15 period (when measured using 2007 trade shares), welfare measures such as total real absorption and the national poverty rate improve relative to the Baseline. These welfare gains relative to the baseline occur because the price shocks effect both imported and exported agricultural commodities. With the incentives embedded in the shifts in international prices, the production structure of Tanzanian agriculture evolves. In particular, the agricultural sector produces less of commodities whose world prices have remained constant or declined and more of commodities whose prices have risen. This shift in structure allows Tanzania to export more of the items whose prices have risen substantially, such as coffee and processed sugar, and import less of items whose prices have risen, such as maize and milled rice. As coffee and processed sugar were relatively small sectors to begin with and because the price rise is particularly steep, export growth is considerable in percentage terms. Real export totals increase by a factor of about nine for coffee and more than 20 for processed sugar relative to the Baseline in the year 2015. Imports of maize and rice, which were relatively large in the Baseline (see Cereals in Table 1), contract by about nine and 16 percent respectively relative to the Baseline in the year 2015. Table 5 illustrates changes in poverty by urban and rural zone. Not surprisingly, increases in agricultural prices for both imports and exports (scenario H-Agriculture) favor the rural sector and generate greater reductions in rural poverty than in the baseline. Relative to the urban sector, the rural sector experiences substantially larger percentage point reductions in poverty. Relative to the Baseline in 2015, the rural poverty rate falls by 4.5 percentage points while the urban rate falls by only 1.4 percentage points. It is also unsurprising that both rural and urban consumption poverty rates rise due to increased fuel prices (scenario H-Fuel). The incidence is much more evenly distributed in this scenario with the rural poverty rate rising by 2.5 percentage points relative to the baseline in 2015 and urban rate rising by 2.2 percentage points. The slightly larger rise in rural rates is largely explained by the larger concentration of the rural population near the poverty line making the rural poverty rate somewhat more sensitive to shifts in welfare. In addition, rubber exports enjoy a world price improvement with benefits largely accruing in urban areas. Overall, the results are indicative of the way that the negative effects of oil price rises permeate through the economy.

[Insert Table 5]

Returning to agriculture, we have seen that if the shifts in relative international prices for agricultural commodities are perceived to be permanent by investors and then actually persist, the economy, at least as modeled, is capable of taking advantage of the price shifts by shifting resources towards the production of goods whose relative prices have increased. This is, of course, easier to accomplish in a model than in reality. In addition, the price rises may interact with existing distortions. For example, as mentioned, Tanzanian sugar enjoys preferential access to the tightly regulated market in the European Union. As a result, increases in international sugar prices may not be particularly meaningful to Tanzanian sugar producers.

The implications of removing one of the sources of growth in the historical simulations are considered with a focus on sugar. The scenario H-AgXsugar simply maintains sugar export prices at Baseline levels. The measured terms of trade effect for the period 2011-15 shifts from about a five percent gain in the H-Agriculture scenario to a 4.7 percent loss in H-AgXsugar scenario. Welfare indicators are commensurately worse across the board relative to H- Agriculture; however, relative to the Baseline, the economy is able to generate an overall welfare gain (a 0.8 percent increase in absorption) by 2015. The national poverty rate is slightly higher though this impact is confined to urban zones. Rural poverty rates decline by about 0.4 percentage points in 2015. These gains are dependent upon strong production response in coffee and other export products, which may or may not be realistic within the time frame envisioned. Nevertheless, overall, the simulations do illustrate that Tanzania has the possibility to profit from a generalized and sustained rise in agricultural commodity prices even if the price rise generates an initial terms of trade loss. We turn now to considering in greater depth the implications of temporary versus permanent rises in the world prices of fuels and petroleum derived products holding all other prices constant. Table 6 illustrates the principal results. In both the FuelTemp and FuelPerm scenarios, world prices for petroleum, fertilizer, other chemicals, and rubber rise by 75, 35, 25, and 25 percent respectively in 2008. In 2009, world prices of all four commodities rise a further 10 percent. Consequently, the results in 2008 and 2009 are exactly the same for the two scenarios. Results differ beginning in 2010. In the FuelPerm scenario, world prices for fuels and derived products remain at the 2009 level for the period 2010-15. In the FuelTemp scenario, world prices drop back to the levels observed in 2008 in the year 2010 and fall back to baseline 2007 levels in 2011. These baseline levels then persist over the 2012-15 period.

[Insert Table 6]

The implications of the shocks in 2008 and 2009 are substantial. Terms of trade decline by 13 percentage points in 2008 and a further three percentage points in 2009. Economywide welfare in 2009 is more than six percent lower and national poverty rates are nearly five percentage points higher compared with Baseline levels. These are large numbers. Under the FuelTemp scenario, import prices return to Baseline levels by 2011 providing a close to symmetric positive shock to the economy. Welfare, as measured by absorption and the poverty rates, returns to essentially Baseline levels (see Table 6).

A residual impact of the world price shocks is observable in the real exchange rate in the FuelTemp scenario. In response to the price shocks of 2008 and 2009, the real exchange rate depreciates in order to encourage production of tradeable goods so that imports of fuel and petroleum derived products can be financed. The increase in the price of tradeable goods causes resources to be allocated towards the tradeable sectors. This includes increases in investment directed towards tradeable sectors, which expands the stock of capital in those sectors. When the price shock is removed, the putty-clay capital stock accumulation rules imply that the accumulated capital in tradeable sectors cannot be removed straight away. As a result of this legacy of investment in 2008 and 2009, the economy is more oriented towards the production of tradeables and away from the production of non-tradeables than in the Baseline. To address this capital stock legacy, the exchange rate appreciates somewhat relative to the Baseline (see, for example, 2011), which encourages investment in non-tradeable sectors. Over time, the real exchange rate in the FuelTemp scneario trends towards the levels observed in the Baseline.

The implications of a permanent increase in fuel and petroleum derived products are largely attributable to the initial price shocks. Once import prices stabilize at higher levels, trends in welfare, as measured by absorption and poverty, are largely, but not completely, restored. As shown in Table 6, the absorption loss relative to the Baseline expands from 6.1 percent in 2011 to 6.3 percent in 2015. Similarly, the national poverty rate falls by 3.4 percentage points in the Baseline over the period 2011 to 2015 but by only 3.1 percentage points in the FuelPerm scenario. As modeled, the movement of productive resources, in the FuelPerm scenario, into agriculture relative to the Baseline contributes to these dynamics. The share of agriculture in GDP in 2015 is 28.3 percent in FuelPerm versus 27.9 percent in the Baseline. Because TFP growth is assumed to be one and three percent per annum in agriculture and non-agriculture respectively, economywide TFP growth is somewhat slower in FuelPerm.

7 Conclusions

Low income countries, such as Tanzania, are often structural importers of fuels and petroleum derived products. These products often comprise very significant shares in total imports. In addition, fuels are critical inputs into a vast array of production processes. Economizing on fuel use, particularly in the short run, is extremely difficult. As a result of these factors, Tanzania appears to be quite vulnerable to world fuel price increases. Based on our modeling, a cumulative oil price shock of about 92 percent from 2007 levels (with all other world prices held constant) reduces absorption, a measure of economywide welfare, by 6.1 percent and increases poverty rates by about five percentage points.

With a temporary world fuel price shock, the return of world prices to baseline levels mainly results in a return to Baseline levels and trends with respect to welfare indicators such as real absorption and poverty rates. With permanent fuel price shocks, the initial impact of the shock on welfare levels persists with time (e.g., poverty rates remain about five percentage point higher). However, once world prices have stabilized at a higher level, the economy essentially returns to Baseline welfare trends (e.g., the annual reduction in poverty rates is quite similar to the Baseline).

Agricultural price shocks differ substantially from fuel price shocks both in terms of immediate impact and in terms of dynamics. With respect to immediate impact, welfare impacts depend upon the import and export composition of the products whose prices have changed. As agricultural product prices tend to be at least loosely linked in global markets, price increases tend to occur for both imported and exported commodities. Implications for terms of trade may be positive or negative. In addition, even if the terms of trade effect of the initial increase is negative, dynamic shifts in the structure of production can allow the economy to benefit from the higher prices by exporting more and importing less of the commodities whose prices have increased. The stimulus to agriculture from higher world prices for agricultural commodities also tends to favor poverty reduction particularly in rural areas.

Based on available data for world prices for basic commodities that are important import and export items for Tanzania, world price trends since 2007 have led to terms of trade deteriorations when terms of trade are measured using trade shares from 2007. Increases in fuel prices largely drive the terms of trade deterioration. The CGE model demonstrates the possibility that Tanzania can reverse static terms of trade losses by shifting resources towards production in sectors that have experienced world price increases. Should structural or policy barriers impede this reallocation of resources, these possibilities will be compromised.

The right level of investment into agricultural commodity sectors with historically volatile prices has always been a difficult question. As discussed in section 2, there are good reasons to believe that the trend of declining prices for agricultural commodities that characterized the 20th century may have ended. In addition, fuel prices are currently at historically high levels and seem likely to persist. These observations support investments in import competing and export oriented agricultural production. Further support for investment in agriculture is gained from the relatively large reductions in poverty that occur as a result of agricultural growth.

Figure 1: Real crude oil prices ( expressed in 2012 USD).

160

140

120

100

80 $/barrel 60

40

20

0

7/1/1964 2/1/1991 1/1/1950 6/1/1952 4/1/1957 9/1/1959 2/1/1962 5/1/1969 3/1/1974 8/1/1976 1/1/1979 6/1/1981 4/1/1986 9/1/1988 7/1/1993 5/1/1998 3/1/2003 8/1/2005 1/1/2008 6/1/2010

12/1/1966 10/1/1971 11/1/1983 12/1/1995 10/1/2000 11/1/1954

Source: Economagic.com.

Figure 2: Index of prices received by farmers in the United States.

200 180 160 140 120 Index 100 20th Century Trend 80 Index/Trend Ratio 60 40 20

0

1957 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 1954

Source: United States Department of Agriculture. Table 1: Structure of the Tanzanian economy in 2007.

Shares Ratios Exports/ Imports/ Value Added Production Employment Exports Imports Output Demand GDP 100.0 100.0 100.0 100.0 100.0 11.5 24.4 Agriculture 32.0 23.0 82.3 31.8 5.3 18.3 9.2 Food crops 19.2 13.2 40.1 2.2 5.1 2.1 13.4 Cereals 8.4 6.3 15.0 0.0 4.8 0.0 20.6 Export crops 3.2 2.8 12.1 18.6 0.2 67.5 7.4 Livestock 5.7 4.1 13.3 1.4 0.0 4.2 0.0 Other agriculture 3.9 2.9 16.9 9.7 0.0 43.5 0.0 Industry 23.3 31.3 2.7 32.5 93.4 11.7 48.6 Mining 3.8 3.0 0.2 21.5 3.7 74.7 59.2 Manufactures 9.0 14.9 1.5 11.0 89.7 9.1 66.0 Other chemicals 0.1 0.1 0.0 0.0 6.5 0.0 94.3 Fertilizers 0.0 0.0 0.0 0.0 1.6 0.0 95.0 Petroleum 0.1 0.1 0.0 0.0 16.8 0.0 98.1 Rubber 0.2 0.3 0.0 0.2 1.2 6.6 52.7 Other industry 10.6 13.3 1.0 0.0 0.0 0.0 0.0 Services 44.7 45.7 15.0 35.6 1.3 8.7 0.7 Private Services 31.9 32.4 13.5 35.6 1.3 12.4 1.0 Government Services 12.8 13.3 1.5 0.0 0.0 0.0 0.0 Source: 2007 Tanzania SAM.

Table 2: Simulations run for the dynamic CGE model.

Number Label Description 1 Baseline Baseline path with constant world prices. 2 Historical World prices adjusted to reflect world market changes for 15 commodities from 2008-2011. 3 H-Fuel World prices adjusted for fuels and derived commodities only from 2008-2011. 4 H-Agriculture World prices adjusted for agricultural commodities only from 2008-2011. 5 H-AgXsugar World prices adjusted for agricultural commodities, except sugar exports, from 2008-2011. 6 FuelTemp Temporary increase in fuel prices in 2008 and 2009 with return to 2007 levels in 2011. 7 FuelPerm Increases in fuel prices in 2008 and 2009 with these levels retained through 2015.

Table 3: Indices of real prices for the historical simulations (2007=100).

2008 2009 2010 2011 Export Commodities Vegetable oils 117.8 84.8 91.1 123.2 Coffee 110.5 96.3 156.5 229.4 Cotton 106.1 103.6 99.3 136.5 Leaf tea 124.4 143.8 143.0 153.1 Fish 137.9 82.2 95.7 126.7 Import and Export Commodities Forestry & wood products 95.2 87.7 89.5 92.9 Processed sugar 122.4 176.9 200.7 247.0 Rubber 95.7 70.3 101.2 116.0 Import Commodities Maize 133.8 98.5 109.0 167.5 Wheat 125.2 85.1 83.9 116.3 Milled rice 206.1 172.2 150.0 155.5 Chemicals 111.5 98.6 106.7 118.4 Fertilizers 117.2 96.9 108.8 126.7 Petroleum 134.4 87.8 109.4 145.4

Source: International Monetary Fund commodity price data combined with the US GDP deflator. Table 4: Macroeconomic indicators for historical simulations.

2007 2008 2009 2010 2011 2012 2013 2014 2015 Absorption (C+I+G) in Baseline and percent change relative to the Baseline

Baseline 24,474 25,561 26,716 27,946 29,255 30,648 32,131 33,710 35,391 Historical 0.0 -2.1 1.7 0.8 0.5 1.2 1.9 2.4 2.9 H-Fuel 0.0 -2.3 1.0 -0.8 -3.3 -3.3 -3.3 -3.3 -3.4 H-Agriculture 0.0 0.2 0.7 1.5 3.5 4.1 4.7 5.3 5.8 H-AgXsugar 0.0 -2.2 1.4 0.1 -0.7 -0.1 0.3 0.6 0.8 Terms of Trade using 2007 trade shares (Baseline = 100) Historical 100.0 94.6 100.9 98.8 96.1 96.1 96.1 96.1 96.1 H-Fuel 100.0 93.7 102.6 97.9 91.4 91.4 91.4 91.4 91.4 H-Agriculture 100.0 100.7 98.4 100.9 104.9 104.9 104.9 104.9 104.9 H-AgXsugar 100.0 94.4 100.4 98.2 95.3 95.3 95.3 95.3 95.3 Real Exchange Rate Index Baseline 100.0 101.2 102.3 103.2 104.0 104.7 105.2 105.5 105.6 Historical 100.0 108.8 95.0 99.0 96.7 93.9 91.5 89.5 87.6 H-Fuel 100.0 110.2 98.8 106.0 116.5 116.7 116.8 116.6 116.3 H-Agriculture 100.0 99.8 98.3 96.5 87.5 85.7 84.0 82.6 81.2 H-AgXsugar 100.0 109.0 95.9 100.9 100.3 97.9 96.1 94.7 93.6 National Consumption Poverty Headcount Baseline 40.0 39.6 39.0 38.6 37.9 37.0 36.2 35.5 34.5 Historical 40.0 41.6 38.5 38.6 38.1 36.4 35.4 33.9 32.4 H-Fuel 40.0 41.2 38.5 39.2 40.2 39.6 38.7 37.9 37.0 H-Agriculture 40.0 39.8 39.1 38.0 35.7 34.7 33.1 31.9 30.7 H-AgXsugar 40.0 41.6 38.8 39.2 39.2 38.0 36.6 35.6 34.6

Source: Tanzania linked CGE/microsimulation model results. Table 5: Rural and urban consumption poverty headcounts.

2007 2008 2009 2010 2011 2012 2013 2014 2015 Rural Poverty Headcount Baseline 44.7 44.2 43.6 43.2 42.5 41.5 40.6 39.8 38.8 Historical 44.7 46.2 43.0 43.0 42.3 40.5 39.5 37.8 36.1 H-Fuel 44.7 45.9 43.2 43.9 44.8 44.1 43.2 42.3 41.3 H-Agriculture 44.7 44.4 43.6 42.4 39.9 38.8 37.0 35.8 34.4 H-AgXsugar 44.7 46.3 43.3 43.6 43.4 42.1 40.5 39.5 38.4 Urban Poverty Headcount Baseline 20.2 19.9 19.5 19.2 18.8 18.4 18.0 17.5 16.6 Historical 20.2 22.3 19.6 20.0 20.3 19.3 18.5 17.4 16.7 H-Fuel 20.2 21.6 19.0 19.7 21.1 20.7 19.8 19.3 18.8 H-Agriculture 20.2 20.3 20.2 19.4 18.2 17.3 16.5 15.7 15.2 H-AgXsugar 20.2 22.3 19.8 20.7 21.6 20.9 20.0 19.2 18.7

Source: Tanzania linked CGE/microsimulation model results.

Table 6: Principal results for fuel and petroleum derived products import price shocks.

2007 2008 2009 2010 2011 2012 2013 2014 2015 Absorption (C+I+G) in Baseline and percent change relative to the Baseline

Baseline 24,474 25,561 26,716 27,946 29,255 30,648 32,131 33,710 35,391 FuelTemp 0.0 -4.9 -6.1 -4.9 -0.1 -0.1 -0.1 -0.1 -0.1 FuelPerm 0.0 -4.9 -6.1 -6.1 -6.1 -6.1 -6.2 -6.2 -6.3 Terms of Trade using 2007 trade shares (Baseline = 100) FuelTemp 100.0 86.9 83.9 86.9 100.0 100.0 100.0 100.0 100.0 FuelPerm 100.0 86.9 83.9 83.9 83.9 83.9 83.9 83.9 83.9 Real Exchange Rate Index Baseline 100.0 101.2 102.3 103.2 104.0 104.7 105.2 105.5 105.6 FuelTemp 100.0 120.8 126.3 121.4 102.2 103.1 103.8 104.3 104.5 FuelPerm 100.0 120.8 126.3 126.5 126.5 126.4 126.1 125.7 125.0 National Consumption Poverty Headcount Baseline 40.0 39.6 39.0 38.6 37.9 37.0 36.2 35.5 34.5 FuelTemp 40.0 43.6 43.9 41.9 37.9 36.9 36.2 35.4 34.5 FuelPerm 40.0 43.6 43.9 43.0 42.2 41.4 41.0 39.8 39.1 Rural Consumption Poverty Headcount Baseline 44.7 44.2 43.6 43.2 42.5 41.5 40.6 39.8 38.8 FuelTemp 44.7 48.3 48.5 46.4 42.4 41.3 40.5 39.6 38.8 FuelPerm 44.7 48.3 48.5 47.6 46.7 45.9 45.4 44.2 43.4 Urban Consumption Poverty Headcount Baseline 20.2 19.9 19.5 19.2 18.8 18.4 18.0 17.5 16.6 FuelTemp 20.2 23.9 24.7 22.7 18.8 18.4 18.0 17.5 16.6 FuelPerm 20.2 23.9 24.7 23.7 23.1 22.6 22.2 21.5 20.8

Source: Tanzania linked CGE/microsimulation model results. 8 References

Arndt, C., R. Benfica, N. Maximiano, A. Nucifora, and J. Thurlow (2008). “Higher Fuel and Food Prices: Impacts and Responses for Mozambique.” Agricultural Economics. 39: 497-511.

Arndt, C., M.A. Hussain, E.S. Jones, V. Nhate, F. Tarp, and J. Thurlow (2012). “Explaining the Evolution of Poverty: The Case of Mozambique.” American Journal of Agricultural Economics. 94: 854-872. doi: 10.1093/ajae/aas022.

Combes, J.-L., and P. Guillaumont (2002). “Commodity Price Volatility, Vulnerability and Development.” Development Policy Review. 20: 25-39.

Cuddington, J., (1992). “Long-run trends in 26 primary commodity prices: A disaggregated look at the Prebisch-Singer hypothesis.” Journal of Development Economics. 39: 207-227.

Deaton, A. (1999). “Commodity Prices and Growth in Africa.” Journal of Economic Perspectives. 13: 24- 40.

Diao, X., and J. Thurlow (2012). A Recursive Dynamic Computable General Equilibrium Model. In Diao, X., J. Thurlow, S. Benin and S. Fan (eds.) Strategies and Priorities for African Agriculture: Economywide Perspectives from Country Studies, International Food Policy Research Institute, Washington, DC.

Dimaranan, B. (ed.) (2006). Global Trade, Assistance, and Production: The GTAP 6 Data Base, Center for Global Trade Analysis, Purdue University, Indiana.

Economagic.com. (2012) Electronic data accessed on November 11 2012. http://www.economagic.com/em-cgi/data.exe/var/west-texas-crude-long.

Foresight (2011). The Future of Food and Farming. Final Project Report. The Government Office for Science, London.

Ivanic, M., and W. Martin (2008). “Implications of Higher Global Food Prices for Poverty in Low- Income Countries.” Policy Research Working Paper WPS4594, World Bank, Washington, DC.

McMillan, M., and D. Rodrik (2012). “Globalization, Structural Change, and Productivity Growth.” IFPRI Discussion Paper 01160, International Food Policy Research Institute, Washington, DC.

MINAG (2004). National Sample Census of Agriculture, 2002/2003, Ministry of Agriculture, Food Security and Cooperatives, Dar es Salaam, Tanzania.

NBS (2002). Household Budget Survey 2000/01. National Bureau of Statistics, Dar es Salaam, Tanzania.

UN Comtrade (2012). United Nations Commodity Trade Statistics Database. Available from http://comtrade.un.org/db/.