SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References

Given global heterogeneity in climate-induced agricultural variability, Tanzania has the potential to substantially increase its maize exports to other countries. If global maize production is lower than usual owing to supply shocks in major exporting regions, Tanzania may be able to export more maize at higher prices, even if it also experiences below-trend productivity. Diverse destinations for exports can allow for enhanced trading opportunities when negative supply shocks affect the partners' usual import sources. Future climate predictions suggest that some of Tanzania's trading partners will experience severe dry conditions that may reduce agricultural production in years when Tanzania is only mildly affected. Tanzania could thus export grain to countries as climate change increases the likelihood of severe precipitation deficits in other countries while simultaneously decreasing the likelihood of severe precipitation deficits in Tanzania. Trade restrictions, like export bans, prevent Tanzania from taking advantage of these opportunities, foregoing significant economic benefits.


1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References

There is substantial evidence that the frequency and intensity of extreme climate events may change in the coming decades (Diffenbaugh et al., 2005; Easterling et al., 2000; IPCC, 2007), and countries, like Tanzania, are particularly sensitive to climate extremes owing to their reliance on rain-fed subsistence agriculture. Schlenker and Lobell (2010) estimate that average maize productivity in Sub-Saharan Africa may decline by 22% by mid-century. These projected declines have severe poverty implications given that maize is the most important staple food in Eastern Africa and the most widely traded agricultural commodity (World Bank, 2009). However, there is considerable heterogeneity in the impacts of climate change across countries, and farmers in countries that are less severely affected by particular weather outcomes may be able to sell excess supply to meet the excess demand from consumers in the more severely affected regions.

Despite the apparent benefits of greater openness to trade as a mechanism to reduce food supply variability and food price volatility, the trade policy response to climate volatility may in fact be one of greater international agricultural price insulation. Indeed, the food price crisis of 2007–2008 saw several countries erect export restrictions to enhance domestic food availability and price stability (Mitra and Josling, 2009). Tanzania was one such country, instituting a crisis-induced export ban on maize (FAO, 2009) that has only recently been lifted. Indeed, it can be argued that the export ban not only lowered exports, but also depressed producer prices, slowed agricultural growth, and created lost opportunities for farmers and consumers (World Bank, 2009).

If allowed to respond to market forces, international trade can play a role in mitigating the effects of climate-induced production shocks (Tobey et al., 1992; Reilly et al., 1994; Tsigas et al., 1997; Randhir and Hertel, 2000). However, most studies of this subject to date have focused only on decadal scale, mean climate change and do not consider the impact of changing climate volatility and the incidence and intensity of extreme events. Also, the scope of many of these studies is constrained by data limitations that prohibited analysis of individual African countries. Reimer and Li (2009) disaggregates many countries and examines of cereal and oilseeds trade, to conclude that world trade volumes will need to increase if yield variability increases. However, this analysis is unable to examine the complex impacts of factor incomes and prices on household welfare, as has been shown in studies of international price shocks following trade liberalization (e.g. Hertel et al., 2004, 2009; Hertel and Winters, 2006) and in analyses of climate change and poverty (Ahmed et al., 2009, 2011).

Tanzania is a country where grain production variability may increase because of changing climate volatility, with impacts on price and income volatility (Ahmed et al., 2011). Of particular interest are the potential inter-annual trading opportunities created by heterogeneous climate shocks, as well as the potential for trade to modulate the effects of climate-induced shocks on Tanzanian poverty.

In order to examine these potential trade effects, we first estimate the historical covariance of maize productivity shocks in Tanzania and her key trading partners. We then use an augmented version of the Global Trade Analysis Project (GTAP) global trade model and associated database to quantitatively examine the interaction between trade policy and the climate-induced maize production volatility, with emphasis on the resulting impacts on exports and poverty in Tanzania. We consider these impacts within the context of three historical case-study years in which we apply the historical productivity shocks under two different trade regimes to examine the sensitivity of the economic impacts of climate shocks to export restrictions. Finally, we use a suite of global climate model simulations to quantify the potential for changes in precipitation extremes in Tanzania and in key trading partners as greenhouse gas concentrations increase over the course of the 21st century.

2. Tanzanian Maize Trade in the Context of International Production Volatility

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References

Tanzanian's National Trade Policy emphasizes integration into the regional and multilateral trading systems (United Republic of Tanzania, 2003), which sets the goal of guiding the country from a supply-constrained economy to one with competitive, export-led growth. Export growth has been credited with an important role in raising the national growth rate from 2% per annum in 1990–95 to 6% in 2000–03 (Integrated Framework, 2005). The National Trade Policy, however, provides only a weak and superficial treatment of agriculture-enhancing policies, even though crop exports accounted for about 23% of total export value in 2001 (Dimaranan, 2006). There also appears to be little co-ordination of trade, agriculture, and poverty reduction strategies, despite the importance of sustaining export growth indicated in the National Strategy for Growth and Reduction of Poverty (United Republic of Tanzania, 2005).

Focusing on maize, Tanzania's trade policies have a history of rapid changes and great uncertainty (Chapoto and Jayne, 2010). For example, the government lifted a long-standing ban on maize exports around the same time that the East African Community was established in 1999. However, in 2003, the Ministry of Agriculture and Food Security imposed an export ban on maize by withdrawing export permits already issued to traders and suspending the issuance of new permits. In 2006, this ban was lifted for a month, and then re-imposed, before being lifted again in late 2010.

In principle, if Tanzanian maize productivity is above trend in a year when a major global maize trader like the USA has below-trend productivity, Tanzania may be able to export more maize at higher world prices. This point is best illustrated by Table 1 (column I), describing the correlations in maize production deviations from trend between 1971 and 2001 for Tanzania and the world's major producing regions. These deviations from trend can be conceptualized as being attributable to idiosyncratic shocks—including, most importantly, weather. There is positive correlation in production deviations from trend for Tanzania and many regional neighbors like Uganda and the Rest of Eastern Africa. This reflects the fact that when Tanzania experiences a given climate outcome that damages agricultural production in a given year (e.g. a drought) countries that are geographically close to it are likely to have similar experiences. Conversely, if Tanzania experiences climate that is conducive to good yields, and subsequently has above-average maize production, then its near neighbors are likely to experience the same positive climate. In contrast there is low, or negative, correlation between Tanzania's maize production volatility and the volatility in major maize traders like Argentina, the USA, and Western Europe.

Table 1. Tanzania's Maize Production Volatility Varying with Production Volatility in Other Regions and Percent Deviation of Maize Production from Trend
 Correlation (w.r.t. Tanzania) in percent deviations from trend of maize production, 1971–2001Deviations from trend production (percent of trend)
199519821983
IIIIIIIV
  1. Note: Global maize production time series data for the period 1971–2001 were obtained at the national level from FAOSTAT (2010), and then de-trended using a statistical model that explains physical production of maize as a function of time. The time series of estimated residuals, expressed as percentages of predicted production, is taken to represent the percentage deviations of production from trend.

Argentina−0.22−6.9415.987.43
Brazil0.1217.908.45−10.02
China0.295.68−7.150.03
E. Asia0.18−6.89−14.76−0.31
E. Europe & Former USSR0.06−9.6620.358.29
Latin America & Caribbean0.372.75−6.83−11.72
Malawi0.110.2314.579.83
Mexico−0.199.32−11.4512.02
Middle East & North Africa0.00−14.90−7.28−5.20
Mozambique−0.02−2.2124.2814.98
Oceania0.13−11.9514.12−17.12
Rest of E. Africa0.28−1.904.7010.07
Rest of N. America0.27−3.2317.082.56
Rest of Sub-Saharan Africa0.1313.43−32.15−34.30
S. Asia0.15−2.41−3.768.89
Southern Africa−0.11−43.460.04−49.22
Tanzania1.0018.60−13.10−16.69
Uganda0.2016.84−3.22−0.07
USA−0.01−14.9619.26−40.63
W. Europe0.01−8.698.223.12
Zambia0.29−29.00−37.99−22.64
Zimbabwe−0.13−52.311.90−48.75

In order to explore the potential for trade to buffer or intensify the effects of climate-change-induced grain production volatility, we select from our time series of production deviations from trends, three individual historical years in which Tanzania and/or her major trading partners experienced large deviations from their respective maize production trends. The historical test-case approach is motivated by the importance of climate for grain production (Lobell et al., 2008), the anticipated damages to Sub-Saharan African agriculture owing to climate change (Schlenker and Lobell, 2010), and the negative effects of grain productivity shocks on poverty in developing countries (Ahmed et al., 2009, 2011). Given this chain of influence linking climate, grains production, and poverty, the case-study years allow us to quantitatively test the potential for different trade regimes to moderate the impacts of climate change on poverty, within the context of known productivity shocks in Tanzania and key trading partners.

Let us begin with the case in which Tanzania experiences a positive deviation from trend, as reported in column II of Table 1. In 1995, Tanzania's production was 19% above trend, while major exporters like the USA and Argentina had sub-par production (recall the negative correlation of these regions with Tanzanian maize deviations in column I). Maize production was also below trend in many Southern and East African countries. Uganda, one of Tanzania's neighbors, also had substantially higher production that year. Tanzania had below average production in 1982 and 1983 (columns III and IV). In 1982, major maize exporters had above-average production, putting downward pressure on prices. In contrast, in 1983, some of those major exporters had below-average production pushing up global maize prices to the advantage of smaller exporters that had good harvests in that year.

An export ban on grains would have made it impossible for Tanzania to take advantage of higher world prices, especially in a year like 1995, when it had above-average production while other African countries and major global traders had low production. Analysis of FAOSTAT (2010) maize trade shows that there were 14 years when Tanzania did not officially export any maize. In six of these years—1976, 1977, 1981, 1986, 1996, and 1996—Tanzania had a positive maize production deviation from trend. In some of these cases, the world price of maize was also above trend, suggesting that the country was missing an opportunity to take advantage of the higher commodity prices. While import restrictions in partner countries surely played a role in limiting Tanzania's exports, Tanzania has had a range of different initiatives in place that have also contributed to low levels of exports (Chapoto and Jayne, 2010). For example, in 1995, when Tanzania had higher-than-usual maize production, the world price of maize was 38% above trend.1 However, there was an export ban in place which limited the country's ability to capitalize on this market development. This was finally lifted at the end of 1996, allowing Tanzania to export at greater-than-trend world maize prices till 1998, after which maize prices dropped below trend again.

3. Methodology

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References

Tanzania's response to the type of global maize production heterogeneity delineated in the case-study years must be understood through mechanisms that can control for the substantial year-to-year changes in the global economy, and the frequent changes in Tanzanian maize trade policy that have occurred over the 1971–2001 period. In order to understand the poverty implications of Tanzania's responses, the mechanism must also account for the changes in prices and factor incomes that will occur across multiple markets, and other general equilibrium (GE) effects.

A computable GE simulation approach is thus necessary, and we employ a modified version of the GTAP simulation model. In addition to allowing us to estimate the changes in consumer prices and earnings stemming from changes to agricultural productivity caused by climate effects, this approach also allows us to examine the additional sensitivity of Tanzania's economic responses to the historical maize production heterogeneity under alternative trade regimes, without including the additional uncertainties of a forecasting approach.

Model Description

We begin with the GTAP Database Version 6 (Dimaranan, 2006) and a modified version of the standard GTAP model (Hertel, 1997). Maize is a component of the “Coarse Grains” composite commodity group in the database. However, in Tanzania, maize accounts for more than 78% of the total output of “Coarse Grains” and more than 98% of “Coarse Grains” trade. As such, we consider the market and production structure of the Tanzanian “Coarse Grains” sector to be a reasonable approximation of the Tanzanian maize sector. The global database has the additional advantage of reconciling the global input–output and trade data from a range of sources, and benchmarking them to a single representative year; 2001 in our particular case. The analysis in this paper focuses on the sensitivity of this “representative” Tanzanian economy2 to a very specific set of stressors, where the comparability of the effects of the different realizations of the stressors (i.e. the climate shocks) will be maintained as long as the same benchmark representative dataset is used in every realization.

We retain the empirically robust assumptions of constant returns to scale and perfect competition, and introduce factor market segmentation, following Keeney and Hertel (2005). Farm and non-farm mobility of factors are restricted by specifying a constant elasticity of transformation function which allows “transformation” of farm employed labor and capital into non-farm uses and vice-versa. This allows for persistent wage differences between the farm and non-farm sectors, and is the foundation of the inter-sectoral distributional analysis (Ahmed et al., 2011). Land is disaggregated by agro-ecological zones, based on the data of Lee et al. (2009) and Monfreda et al. (2009). We assume a constant aggregate level of land, labor, and capital employment reflecting the belief that the aggregate employment of factors is unaffected by the climate shocks that are affecting grain production. The model is then calibrated such that simulations of estimated historical productivity volatility of coarse grains replicate observed historical price volatility, following Ahmed et al. (2011).

International trade is modeled using the Armington assumption, which is particularly appropriate for agricultural markets, where products do not tend to be differentiated by firm. The key import demand elasticities are estimated in Hertel et al. (2007) by exploiting cross-sectional variation in delivered prices owing to variation in international transport costs.

In order to understand how climate shocks can affect household welfare and poverty, we use the household micro-simulation model from Ahmed et al. (2011). That approach involves augmenting the simulation framework with the household model of Hertel et al. (2004), to estimate changes in income and consumption of households in the neighborhood of the poverty line. For poverty analysis, the utility of the household at the poverty line is then defined as the poverty level of utility. If an adverse climate shock pushes households' utility below this level, they enter poverty. The poverty module is calibrated using Tanzania's Household Budget Survey 2000/2001 and households are stratified into seven groups based on earnings sources.3

Experimental Design

The patterns of historical production deviations from trend for each of the three case study years highlighted in the previous section—scenarios 1995, 1982, and 1983—are reproduced in the model via an appropriate combination of international productivity shocks.4 In order to mimic the short run/transient nature of these productivity shocks, these simulations are conducted under the assumption that agricultural land and capital are immobile across sectors. That is, farmers are limited in their ability to adapt to these stochastic climate outcomes. We then consider the impacts on trade and poverty in the context of a range of different trade scenarios to explore the interplay between trade regimes and the poverty impacts of climate shocks.5

The production deviations for the case study years are simulated under two trade regimes:

  • 1
    Baseline—The trade regime prevailing in the 2001 world economy.
  • 2
    Export restriction—As in (1) but with Tanzania imposing an export restriction on maize, preventing maize exports from rising above their 2001 levels.

As noted above, Tanzania introduced grain export bans in response to the 2007/08 food price crisis. Implementation of the export restriction is treated via a complementary slackness condition, with the 2001 benchmark year's export level taken as the quota level of exports. Thus, when changes in economic conditions push Tanzania's maize exports to increase, they are prevented from rising above the 2001 benchmark year's level by means of an endogenous export tax-equivalent of the ban. Maize trade flows, however, are permitted to fall below 2001 benchmark year levels if economic conditions dictate a decline in maize export demand. Finally, since producers may adjust their behavior ex ante in anticipation of future export restrictions, we estimate factor-tax equivalents that shift production accordingly to reflect these producer expectations for the simulations under the export restriction trade regime.

4. Analysis

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References

Additional Impact of Export Restriction

Table 2 describes the marginal changes in maize exports from Tanzania caused by maize productivity shocks in the case study years under the two policy regimes. Under the baseline regime, Tanzanian maize exports expand in scenarios 1995, when the country experiences positive deviations from its production trend (columns I). This is driven by the large declines in the domestic price of maize in Tanzania, as reported in Table 3. The large decreases in Tanzanian maize prices relative to her competitors, dominate any reduction in aggregate import demand, thereby increasing maize exports from Tanzania for all countries.6

Table 2. Percentage Change in Maize Exports from Tanzania Caused by Historical Production Deviations from Trend under Baseline and Export Restriction
RegionBaselineTanzania export restriction
199519821983199519821983
IIIIIIIVVVI
  1. Source: Authors' simulations.

Argentina194−66−70−64−2
Brazil131−65−200−63−15
China182−64−180−62−13
East Asia153−59−10−570
Eastern Europe & Former USSR205−69−630−68−60
Latin America & Caribbean179−61310−590
Malawi118−68−610−67−58
Mexico222−661070−640
Middle East & North Africa216−63−260−61−22
Mozambique107−51−430−49−40
Oceania225−6770−650
Rest of East Africa194−68−370−66−33
Rest of North America206−68470−660
Rest of Sub-Saharan Africa115−55−410−53−38
South Asia120−42−570−39−54
Southern Africa137−58−300−56−26
Uganda62−51−360−46−29
USA180−6550−48−33
Western Europe204−65−510−630
Zambia43−45−500−63−47
Zimbabwe142−57−530−42−47
Total175−62−280−60−29
Table 3. Impact of Export Ban on Maize Prices and as a Result of Global Maize Production Deviations from Trend
YearChange under baselineAdditional change owing to export restriction (difference between baseline and export restriction)
Maize pricePoverty change 1000s of people, by effectMaize pricePoverty change 1000s of people, by effect
Cost of living effectEarnings effectTotalCost of living effectEarnings effectTotal
IIIIIIIVVVIVIIVIII
  1. Source: Authors' simulations.

1995−23.03−149.6−5.2−154.8−2.902.701.704.40
198228.54135.49.0144.40.200.00−0.20−0.20
198343.79173.725.0198.7−0.070.100.000.10

Tanzania's maize exports more than double for some trading partners. The results in Table 2 illustrate the important role of diverse trade partners in mitigating potential food security crises in the wake of a major supply shock. In 1995, maize production in Uganda was more than 16% below trend. Given that 95% of its maize is for domestic consumption, this represents a major maize supply contraction. The high maize demand in Uganda, coupled with the higher supply in Tanzania allows Tanzania to substantially increase its exports, alongside those of a few other select countries.

The case of baseline scenario 1982 is the converse of scenario 1995. In this case, Tanzanian maize production was 13% below trend, pushing up domestic prices by more than 28%. The supply shocks in other countries were such that the supply price of Tanzanian maize was relatively higher than the price of maize from other countries, and Tanzanian exports fell across the board.

The case of baseline scenario 1983 is more complicated. Even though total Tanzanian maize exports decrease by about 28%, Tanzania is able to increase its exports to some countries that experienced even more negative deviations in their production from trend. For example, Tanzanian exports of maize to the USA are predicted to increase by 5%. The USA accounts for 15% of Tanzania's maize exports and experienced production 40% below-trend in the baseline scenario 1983. This drives up the average price of maize imports around the world, and results in a drop in imports in most markets.

Let us now turn to the impacts of the Export Restriction Regime—reported in the second group of columns in Table 2 as well as the price effects reported in Table 3 (column I and V). In the year when production is above trend and an export restriction is in place, the maize price drops below the price that would have prevailed if there were no restriction. In the case of scenario 1982, when Tanzanian maize production was below-trend, maize exports decreased for all countries (since the modeling of the restriction allows for decreases from 2001 benchmark flows). The effect of the export ban was negligible.

When considering the poverty impacts of production deviations from trend caused by climate under trade baseline and export restriction regimes, we must account for the impacts on prices both, and factor incomes. From Table 3 (columns II–IV, VI–VIII) we can see that poverty decreases in the year that Tanzania has a positive production deviation from trend (scenario 1995), with most of the poverty reduction owing to reductions in the cost of living. The price of maize falls substantially, and poor households are able to purchase more food at lower prices. Improvements in earnings also contribute to the poverty reduction. The higher demand for workers in expanding downstream sectors like Other Food and Beverages, Farm Livestock, and Processed Livestock leads to improvements to factor returns.

Tanzania's negative deviations from maize production trends in 1982 and 1983 lead to increases in poverty in the case-study simulations. There were sharp increases in food prices arising from the negative production shocks to maize, leading to major increases in poverty through higher costs of living.

Aside from generally being poverty-increasing (or dampening poverty reduction), the export restriction also has distributional implications. In the scenarios where the export restriction increased poverty (1983), the poverty increases were among households that depended on agriculture as the primary source of income, relied on transfers, or had diverse sources of income. In contrast, the export ban actually had a slight poverty reducing effect among households that were not involved in agriculture, or relied on market wages as their main source of income.

The results of these case study simulations under the baseline and export restriction highlight a number of important points.7 First, Tanzania has the potential to substantially increase its maize exports to other countries, and not only when its production is above trend. If global maize production is lower than usual owing to supply shocks in major exporters, Tanzania can export more maize at higher prices, even if it also experiences below-trend production. Secondly, diversified sources of imports can help mitigate the effects of a negative supply shock in a major source, as shown in the case of Uganda in 1995. Conversely, having diverse destinations for exports can allow for substantial trade when negative supply shocks affect the partners' usual sources. For example, Tanzania could export to Latin America and the Caribbean when US production was below-trend in scenario 1983. Finally, export bans suppress maize prices, either by making price declines greater, or price increases less positive. However, they are an ineffective tool for altering the poverty impact of the underlying climate/productivity shocks, and come at the cost of significant reductions in exports, GDP and long run credibility as a supplier of agricultural products.

Results—Projected Changes in Climate Extremes in Tanzania and Trading Partners

Given the sensitivity of poverty outcomes in Tanzania to productivity shocks in key trading partners, we seek to quantify the potential changes in country-level climate extremes as greenhouse gas concentrations rise in the 21st century. To do so, we draw on a suite of climate model experiments to project 21st century changes in (i) the occurrence of extreme dry years in Tanzania and her key trading partners; (ii) how often Tanzania experiences a dry year in the same year that her key trading partners do not experience dry years; and (iii) how often each of Tanzania's key trading partners experiences a dry year in the same year that Tanzania does not experience a dry year.

Models, Scenarios, and Climate Analyses

We analyze climate model results from the Coupled Model Intercomparison Project (CMIP3) (Meehl et al., 2007a,b). CMIP3 has archived results from multiple general circulation models (GCMs) developed at climate modeling centers around the world. The archive has been extensively analyzed by the international community, and the multi-model output served as the backbone of much of the Working Group I contribution to the Fourth Assessment Report (“AR4”) of Intergovernmental Panel on Climate Change (IPCC, 2007).

The CMIP3 archive contains GCM simulations of the pre-industrial period, the 20th century, and various scenarios of the 21st century, including a number of the 21st century “SRES scenarios” produced by the IPCC in its Special Report on Emissions Scenarios (Nakicenovic and Swart, 2000). Varying numbers of models have archived varying subsets of results for these different simulations. Together, these various climate model simulations create an “ensemble of opportunity” that has been used to analyze a wide variety of phenomena in the climate system, including detection and attribution of historical climate change (e.g. Santer et al., 2009), the sensitivity of global mean temperature to elevated greenhouse forcing (e.g. Knutti et al., 2008), the response of extreme events to global warming (e.g. Tebaldi et al., 2006; Knutson et al., 2008; Diffenbaugh and Scherer, 2011), and the potential impacts of climate change on natural and human systems (e.g. Williams et al., 2007; Ahmed et al., 2009; Battisti and Naylor, 2009; Loariel et al., 2009).

In the present study, we focus on the CMIP3 simulations that have been forced by the IPCC A1B scenario (Nakicenovic and Swart, 2000). The cumulative CO2 emissions and global mean temperature change are very similar in the IPCC illustrative scenario suite over the first half of the 21st century (Meehl et al., 2007b). The A1B scenario falls near the middle of the illustrative suite in the second half of the 21st century, with global atmospheric carbon dioxide concentrations reaching between 600 and 800 parts per million (ppm) by the end of the 21st century, and global mean temperature warming by between 2 and 4 degC (Meehl et al., 2007b). For our analyses, we select “run 1” from each of the 22 climate models that archived surface temperature and precipitation data for the A1B 21st century scenario.

Given the importance of adequate precipitation for maize production in Tanzania and the world (e.g. Ahmed et al., 2011; Lobell et al., 2008), we focus our analyses on the occurrence of dry years in Tanzania and key (or potential) trading partners during the 20th and 21st century CMIP3 simulations. For the purposes of this study, we take a dry year to be any year in which the annual precipitation is equal to or less than the historical 1-in-10-year dry event. We take the historical period to be 1951–2000, meaning that for each trading partner, the 1-in-10-year dry event is the fifth driest year in that partner during the 1951–2000 period, and a dry year is any year in which the annual precipitation is less than or equal to the precipitation value of the fifth driest year of the 1951–2000 period.8

We use the CMIP3 climate model experiments to quantify the likelihood of different trading partners experiencing dry years, including the likelihood that these dry years co-occur in Tanzania and the trading partners. To do so, we first calculate the annual precipitation in each partner-country in each year from 1951 to 2000 in each of the CMIP3 historical simulations. We then calculate the magnitude of the 1-in-10-year dry event for each trading partner for the 1951–2000 period, which we may call the “dry year threshold”.

Having obtained the dry year threshold for each partner for the 1951–2000 period, we calculate three metrics of dry year occurrence in each decade of the CMIP3 A1B 21st century projections. The first metric is the expected number of dry years in each decade of the 21st century in Tanzania and her key trading partners. Secondly, we calculate how often Tanzania experiences a dry year in the same year that her key trading partners do not experience dry years; this metric indicates the likelihood that Tanzania's trading partners will experience non-adverse climate conditions in the same year that Tanzania experiences adverse conditions, therefore offering the potential for Tanzania to ameliorate adverse conditions in Tanzania through trade. For each trading partner, we report the percentage of Tanzania dry years in which that country does not also experience a dry year. We report these percentages for each decade of the 21st century. The third metric involves calculating how often each of Tanzania's trading partners experiences a dry year in the same year that Tanzania does not experience a dry year. This metric indicates the likelihood that Tanzania will experience non-adverse climate conditions in the same year as her trading partners experience adverse conditions, therefore offering the potential for Tanzania to benefit through trade. For each trading partner, we report the percentage of dry years in that partner-country in which Tanzania does not also experience a dry year over each each decade of the 21st century.

We combine the 22 GCM realizations from the CMIP3 archive by first performing the above calculations for each GCM realization individually, and then calculating the mean across the ensemble of GCM realizations.

Severe Dry Events in Tanzania

Table 4 shows the number of dry years in Tanzania in each decade of the 21st century, along with the percentage of Tanzania dry years in which each trading partner does not experience a dry year. The decadal occurrences are reported as the mean of the decadal occurrences of the individual climate model realizations. With this measure, a dry year occurrence of 0.5, as reported for the decade from 2000 to 2009, indicates that across all of the GCM realizations in that decade, the occurrence of dry years in Tanzania was half as common as in the historical period (which was one year on average per decade given the way the dry year has been defined). In addition, the number of years in which Tanzania experiences a dry year but another trading partner does not is reported as a percentage of the Tanzania dry years in that decade. For example, in the case of Canada, the value of 100% indicates that in all of the years in which Tanzania experienced a dry year, Canada did not simultaneously experience a dry year.

Table 4. Severe Dry Events in Tanzania
Country2000s2010s2020s2030s2040s2050s2060s2070s2080s
  1. Source: Authors' estimates from Meehl et al. (2007a).

  2. Notes: The decadal occurrence of 1-in-10-year dry events in Tanzania is reported as the mean of the decadal occurrences of the individual climate model realizations. The number of years in which Tanzania experiences a 1-in-10-year dry event but another country does not is reported as a percentage of the decadal occurrences in Tanzania.

Number of years in decade that benchmark 1-in-10 dry event threshold is exceeded in Tanzania
Tanzania0.50.590.770.450.640.550.360.360.32
Number of years in decade that benchmark 1-in-10 dry event threshold is exceeded in Tanzania
Australia82.00100.0088.31100.0085.9474.5588.8963.8956.25
Canada100.00100.00100.00100.00100.00100.00100.00100.00100.00
China100.0061.0288.3191.11100.00100.00100.00100.00100.00
Ethiopia64.0061.0276.62100.00100.0081.8288.8975.00100.00
India90.0093.2288.31100.0092.19100.00100.00100.00100.00
Kenya36.0084.7553.2560.0078.1374.5563.89100.0084.38
Madagascar100.0076.2783.1291.1178.1365.4575.0063.8984.38
Malawi90.0069.4983.1271.1164.0658.1888.8950.0071.88
Mexico90.0076.2783.1291.1178.1349.0988.8975.0071.88
Mozambique90.0069.4971.4380.0050.0041.8288.8963.8956.25
Russian Federation100.00100.00100.00100.00100.00100.00100.00100.00100.00
South Africa72.0093.2288.3180.0064.0674.55100.0088.89100.00
Sudan100.0084.75100.00100.0085.9481.8288.8975.00100.00
Uganda64.0076.2764.9491.1178.1365.4563.8988.8984.38
UK64.0084.7594.81100.0085.9481.8288.8988.8971.88
USA90.00100.0076.6291.1192.19100.00100.00100.00100.00
Zambia82.0069.4964.9471.1156.2558.1888.8975.0056.25
Zimbabwe82.0093.2283.1280.0056.2574.55100.0063.8971.88

The CMIP3 GCM ensemble projects mean-annual precipitation to increase in Tanzania over the course of the 21st century in response to increasing greenhouse gas concentrations (e.g. Christensen et al. 2007; Meehl et al. 2007a). In accordance with this increase in mean-annual precipitation, we find that the occurrence of dry years is substantially reduced in Tanzania (Table 4). The ensemble-mean dry year occurrences range from 0.45 to 0.77 from the early 2000s through the 2050s, and from 0.32 to 0.36 in the 2060s, 2070s and 2080s.

In almost all cases, more than 50% of Tanzania's dry years coincide with non-dry years in Tanzania's selected African trading partners (Table 4). Exceptions include 36.00% co-occurrence of dry conditions in Tanzania and non-dry conditions in Kenya in the 2000s, and 41.82% co-occurrence of dry conditions in Tanzania and non-dry conditions in Mozambique in the 2050s. This means that extreme events in these three trading partners are expected to coincide more frequently. Of the African trading partners, Sudan exhibits the highest co-occurrence of non-dry years with Tanzania dry years, with at least 75% of Tanzania dry years co-occurring with a non-dry year in Sudan in all decades of the 2000–2089 period. Conversely, Kenya, Mozambique and Zambia exhibit the lowest co-occurrence of non-dry years with Tanzania dry years, with each partner-country exhibiting four decades in the 2000–2089 period in which less than 65% of Tanzania dry years co-occur with a non-dry year in that trading partner. The 2040s exhibit the lowest co-occurrence of non-dry years with Tanzania dry years across the African trading partners, with five of the 10 trading partners experiencing less than 65% co-occurrence of non-dry years with Tanzania dry years.

As might be expected simply from geographic proximity, Tanzania's key trading partners outside of Africa more consistently exhibit co-occurrence of non-dry years with Tanzania's dry years than do Tanzania's key trading partners that are within Africa. Of those partners outside of Africa, Canada and the Russian Federation exhibit the highest co-occurrence of non-dry years with Tanzania's dry years, with 100.00% of Tanzania's dry years co-occurring with a non-dry year in those trading partners in all decades of the 2000–2089 period. China, India, and the USA also exhibit high co-occurrence of non-dry years with Tanzania's dry years, with at least 88% of Tanzania's dry years co-occurring with a non-dry year in those countries in nine decades of the 2000–2089 period. Conversely, Australia and Mexico both exhibit less than 76% co-occurrence of non-dry years with Tanzania's dry years in the 2050s, 2070s, and 2080s.

Severe Dry Events in Tanzania's Key Trading Partners

Table 5 shows the number of dry years in each of Tanzania's key trading partners in each decade of the 21st century, along with the percentage of trading partners' dry years in which Tanzania does not experience a dry year. The decadal occurrences are reported as the mean of the decadal occurrences of the individual climate model realizations. Therefore, as in the preceding table, a dry year occurrence of 0.5 indicates that across all of the GCM realizations in that decade, the occurrence of dry years was half as common as in the historical period, and an occurrence of 2.0 indicates that the occurrence of dry years was twice as common as in the historical period. In addition, the number of years in which each trading partner experiences a dry year but Tanzania does not is reported as a percentage of the trading partners' dry years in that decade. Under this metric, for a given trading partner, a value of 50% indicates that in half of the years in which that country experienced a dry year Tanzania is not expected to simultaneously experience a dry year.

Table 5. Severe Dry Events in Tanzania's Key Trading Partners
CountryMetric2000s2010s2020s2030s2040s2050s2060s2070s2080s
  1. Source: Authors' estimates from Meehl et al. (2007a).

  2. Notes: The decadal occurrence of 1-in-10-year dry events in key trading partners is reported as the mean of the decadal occurrences of the individual climate model realizations. The number of years in which a partner experiences a 1-in-10-year dry event but Tanzania does not is reported as a percentage of the respective decadal occurrences in each partner.

AustraliaDry years0.81.00.81.01.31.31.51.21.6
Dry but TZA not89.0100.088.3100.093.289.896.788.691.2
CanadaDry years0.20.40.10.10.00.00.00.00.0
Dry but TZA not100.0100.0100.0100.0n/an/an/an/an/a
ChinaDry years0.81.20.90.60.40.20.10.10.1
Dry but TZA not100.080.589.590.9100.0100.0100.0100.0100.0
EthiopiaDry years1.21.11.01.20.81.10.60.50.6
Dry but TZA not84.778.182.0100.0100.090.593.280.0100.0
IndiaDry years1.11.11.11.00.70.50.50.60.7
Dry but TZA not96.395.290.5100.093.2100.0100.0100.0100.0
KenyaDry years0.70.60.70.70.70.80.40.10.2
Dry but TZA not56.281.847.175.380.982.963.9100.077.8
MadagascarDry years0.91.41.10.91.41.51.61.61.7
Dry but TZA not100.090.187.794.590.487.693.591.097.6
MalawiDry years0.91.10.80.91.10.90.61.20.8
Dry but TZA not95.383.383.184.979.874.793.285.489.0
MexicoDry years1.72.62.02.63.13.33.94.64.6
Dry but TZA not97.694.593.398.595.591.799.098.197.8
MozambiqueDry years1.01.31.11.21.71.91.11.71.6
Dry but TZA not95.086.479.892.481.083.395.292.391.5
Russian FederationDry years0.40.10.10.10.00.00.00.00.0
Dry but TZA not100.0100.0100.0100.0n/an/an/an/an/a
South africaDry years0.90.91.11.51.81.92.32.11.8
Dry but TZA not84.695.392.194.087.693.0100.097.7100.0
SudanDry years1.00.91.10.81.11.00.91.21.2
Dry but TZA not100.090.1100.0100.092.191.094.592.4100.0
UgandaDry years0.60.60.90.50.80.60.50.40.4
Dry but TZA not70.374.568.690.082.969.572.088.988.9
UKDry years1.00.80.50.70.50.60.60.40.8
Dry but TZA not82.088.390.0100.082.085.990.988.989.0
USADry years1.10.70.80.70.90.90.60.70.5
Dry but TZA not95.2100.076.693.295.3100.0100.0100.0100.0
ZambiaDry years1.11.01.41.31.61.61.11.30.9
Dry but TZA not90.582.080.189.881.985.595.293.284.6
ZimbabweDry years1.51.41.11.51.82.21.81.91.4
Dry but TZA not93.896.587.293.885.293.7100.093.093.6

We find that a number of Tanzania's key trading partners experience increases in the occurrence of dry years as greenhouse gas concentrations rise in the 21st century GCM simulations (Table 5). For example, within Africa, Madagascar, Mozambique, South Africa, and Zimbabwe experience an average of at least 1.5 dry years per decade in at least four decades of the 2040–2089 period. Conversely, Ethiopia, Kenya, and Uganda all exhibit decreased occurrence of dry years in the late 21st century, including mean occurrence of less than 0.7 years per decade in the 2060s, 2070s, and 2080s. Kenya exhibits the lowest occurrence of dry years of any of Tanzania's key African trading partners, including lower dry year occurrence than Tanzania in the 2070s and 2080s (Tables 4 and 5).

The 21st century climate model experiments suggest that Tanzania is likely to experience non-dry years in most of the years in which her key African trading partners experience dry years (Table 5). For example, although South Africa exhibits increasing occurrence of dry years throughout the 21st century period, a minimum average of 84.6% of those dry years in a given decade co-occur with non-dry years in Tanzania, including greater than 97% in the 2060s, 2070s, and 2080s. Likewise, between 87% and 100% of the dry years in Madagascar and Zimbabwe co-occur with non-dry years in Tanzania, as do between 79% and 85% of the dry years in Mozambique. Kenya exhibits the lowest co-occurrence of dry years with Tanzania non-dry years, including the three lowest mean decadal co-occurrences of any of Tanzania's key trading partners (47.1%, 56.2%, and 63.9%).

Outside of Africa, Mexico experiences the most substantial intensification of dry year occurrence of Tanzania's key trading partners, including a minimum average of 1.68 dry years per decade in the 2000s and a maximum of 4.64 dry years per decade in the 2070s (Table 5). Australia likewise exhibits at least 1.2 dry years per decade in each decade of the 2040–2089 period, including an average of 1.6 dry years in the 2080s. Conversely, Canada, China, India, and the Russian Federation all exhibit decreases in dry year occurrence over the 21st century, including near-zero occurrence in Canada and the Russian Federation beginning in the 2020s, and in China after the 2050s.

Despite the very high occurrence of dry years in Mexico, the mean co-occurrence of those dry years with non-dry years in Tanzania is at least 90% in all decades of the 2000–2089 period, including greater than 97% in the 2060–2089 period, in which Mexico experiences the highest occurrence of dry years (Table 5). Australia, which also experiences increasing dry year occurrence in the 21st century, exhibits lower co-occurrence of dry years with Tanzania non-dry years than does Mexico, although the mean decadal co-occurrence is greater than 88% in all decades of the 2000–2089 period. In contrast, the mean co-occurrence of dry years with non-dry years in Tanzania is 100% for Canada and the Russian Federation throughout the 21st century, 100% for China after the 2030s, and 100% for India and the USA after the 2040s.

Discussion

The CMIP3 suite of global climate model projections for the 21st century suggests that further global warming is likely to both increase the mean seasonal precipitation (Christensen et al., 2007; Meehl et al., 2007a) and decrease the occurrence of dry years in Tanzania (Table 4). These results suggest that Tanzania could experience decreased agricultural stress from precipitation deficits in the future.

However, dry years do persist in Tanzania through the 21st century (Table 4). Quantifying the co-occurrence of these dry years with non-dry years in Tanzania's key trading partners indicates the likelihood that Tanzania's trading partners will experience non-adverse climate conditions in the same year that Tanzania experiences adverse conditions, thereby offering the potential for Tanzania to ameliorate adverse conditions in Tanzania through increased imports. Analysis of the CMIP3 archive of global climate model simulations indicates that those dry years that do occur in Tanzania in the 21st century will often coincide with non-dry years in Tanzania's key trading partners both within and outside of Africa (Table 5). These results suggest that importing grains from other trading partners could help to alleviate negative effects of severe dry conditions going forward in the 21st century, particularly for a mix of trading partners that can help to hedge against the coincidence of severe dry years both within and outside of Africa. In particular, our analysis identifies Sudan as the trading partner within Africa that is most likely to consistently experience non-dry conditions in the years in which Tanzania experiences dry conditions, and Kenya, Mozambique, and Zambia as the least likely. Likewise, our analysis identifies Canada, China, the Russian Federation, and the USA as the trading partners outside of Africa that are most likely to consistently experience non-dry conditions in the years in which Tanzania experiences dry conditions, and Mexico and Australia as the least likely.

The CMIP3 suite of global climate model projections also suggests that further global warming is likely to increase the occurrence of dry conditions in many of Tanzania's trading partners, and to alter the co-occurrence of dry years in Tanzania's trading partners with non-dry years in Tanzania as Tanzania's dry year occurrence decreases through the 21st century (Table 5). Quantifying the co-occurrence of dry years in Tanzania's trading partners with non-dry years in Tanzania indicates the likelihood that Tanzania will experience non-adverse climate conditions in the same year as her trading partners experience adverse conditions, therefore offering the potential for Tanzania to benefit from the non-adverse conditions through increased exports. Tanzania is likely to experience non-dry years in most of the years in which key trading partners experience severe dry conditions in the 21st century. These results suggest that Tanzania might benefit from exporting grains to trading partners within and outside of Africa as climate change increases the likelihood of severe precipitation deficits in other partner-countries while simultaneously decreasing the likelihood of severe precipitation deficits in Tanzania. In particular, our analysis identifies Madagascar, Mozambique, South Africa, and Zimbabwe as the potential trading partners within Africa that are most likely to experience substantial increases in dry years, with Tanzania experiencing non-dry years in the vast majority of those trading partners' dry years. Our analysis also identifies Mexico and Australia as the key trading partners outside of Africa that are most likely to experience substantial increases in dry years over the course of the 21st century, but also suggests that Tanzania is likely to experience non-dry years in most of the years in the 21st century in which her key trading partners outside of Africa experience dry years. This could present Tanzania with some export opportunities in the future.

Finally, although we have focused our analyses on dry years as defined by the historical 1-in-10-year dry event, we note that temperature is also an important determinant of grains production. Maize, in particular, can be sensitive to high temperatures (e.g. Schlenker and Roberts, 2009). We have repeated the above analyses for the 1-in-10-year hot event, and we find that most of our focus partner-countries are projected to regularly experience annual temperatures that exceed the baseline 1-in-10-year hot threshold relatively early in the 21st century (not shown). Indeed, most are also projected to regularly experience summers that are hotter than the historical maximum by the late 21st century in the A1B scenario (Battisti and Naylor, 2009; Diffenbaugh and Scherer, 2011). The effects of such temperature rises much also be taken into account in order to obtain a more complete picture of the interplay between future climate and Tanzania's trade potential.

5. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References

Focusing on case-study years representing a range of production shocks in Tanzania and key trading partners, it is found that Tanzania has the potential to substantially increase its maize exports to other countries, and not only when its production is above trend. If global maize production is lower than usual owing to supply shocks in major exporters, Tanzania can export more maize at higher prices, even if it also experiences below-trend production. As expected, diversified sources of imports can help mitigate the effects of a negative supply shock in a given country. Conversely, having diverse destinations for exports can allow for export increases when negative supply shocks affect the partners' dominant sources. Tanzanian export restrictions are found to suppress maize price responses, either by making price declines larger, or price increases less positive. However, the marginal impact of the maize export restriction on poverty is small.

Considering climate change, severe dry conditions in Tanzania will most often coincide with non-dry conditions in Tanzania's key trading partners within Africa in the future. However, there may be decades in the 21st century when some countries frequently experience coinciding severe dry conditions. Based on those future predictions, Tanzania's key trading partners will also experience increases in the occurrence of severe dry conditions as greenhouse gas concentrations rise in the 21st century. Tanzania is likely to experience non-dry years in most of the years in which key trading partners experience severe dry conditions in the 21st century. These results suggest that Tanzania could benefit from exporting grains to countries within and outside of Africa as climate change increases the likelihood of severe precipitation deficits in other countries while simultaneously decreasing the likelihood of severe precipitation deficits in Tanzania.

The Tanzanian economy thus has the potential to capitalize on the increasing heterogeneity of climate impacts on agriculture in the future, but only through removal of export restrictions or movement to a rules-based policy mechanism as advocated by Chapoto and Jayne (2010), removing both the policy uncertainty and the resulting price instability.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Tanzanian Maize Trade in the Context of International Production Volatility
  5. 3. Methodology
  6. 4. Analysis
  7. 5. Conclusion
  8. References
  • Ahmed, S. A., N. S. Diffenbaugh, and T. W. Hertel, “Climate Volatility Deepens Poverty Vulnerability in Developing Countries,” Environmental Research Letters 4 (2009):8pp.
  • Ahmed, S. A., N. S. Diffenbaugh, T. W. Hertel, D. Lobell, N. Ramankutty, A. R. Rios, and P. Rowhani, “Climate Volatility and Poverty Vulnerability in Tanzania,” Global Environmental Change 21 (2011):4655.
  • Battisti, D. S. and R. L. Naylor, “Historical Warnings of Future Food Insecurity with Unprecedented Seasonal Heat,” Science 323 (2009):24044.
  • Chapoto, A. and T. S. Jayne, “Maize Price Instability in Eastern and Southern Africa: The Impact of Trade Barriers and Market Interventions,” Paper presented at the COMESA Policy Seminar on Variation in Staple Food Process: Causes, Consequences, and Policy Options, Maputo, Mozambique, 25–26 January (2010).
  • Christensen, J. H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R. K. Kolli, W.-T. Kwon, R. Laprise, V. Magaña Rueda, L. Mearns, C. G. Menéndez, J. Räisänen, A. Rinke, A. Sarr, and P. Whetton, “Regional Climate Projections. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I,” in S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds), Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge: Cambridge University Press (2007).
  • Diffenbaugh, N. S. and M. Scherer, “Observational and Model Evidence of Global Emergence of Permanent, Unprecedented Heat in the 20th and 21st Centuries,” Climatic Change 107 (2011):61524.
  • Diffenbaugh, N. S., J. S. Pal, R. J. Trapp, and F. Giorgi, “Fine-scale Processes Regulate the Response of Extreme Events to Global Climate Change,” Proceedings of the National Academy of Sciences 102 (2005):157748.
  • Dimaranan, B. V. (ed.), Global Trade, Assistance, and Production: The GTAP 6 Data Base, Center for Global Trade Analysis, West Lafayette, IN: Purdue University (2006).
  • Easterling D. R., G. A. Meehl, C. Parmesan, S. A. Changnon, T. R. Karl, and L. O. Mearns, “Climate Extremes: Observations, Modeling, and Impacts,” Science 289 (2000):206874.
  • FAO, The State of Agricultural Commodity Markets 2009, Rome: Food and Agricultural Organization (2009).
  • FAO, FAOSTAT Database, Food and Agriculture Organization, Rome (2010), available at: faostat.fao.org
  • Hertel, T. W. (ed.), Global Trade Analysis: Models and Applications, Cambridge: Cambridge University Press (1997).
  • Hertel, T. W. and L. A. Winters, Poverty and the WTO: Impacts of the Doha Development Agenda, New York: Palgrave McMillan (2006).
  • Hertel, T. W., D. Hummels, M. Ivanic, and R. Keeney, “How Confident Can We Be of CGE-Based Assessments of Free Trade Agreements?” Economic Modelling 24 (2007):61135.
  • Hertel, T. W., M. Ivanic, P. Preckel, and J. A. L. Cranfield, “The Earnings Effects of Multilateral Trade Liberalizations: Implications for Poverty,” World Bank Economic Review 18 (2004):20536.
  • Hertel, T. W., R. Keeney, M. Ivanic, and L. A. Winters, “Why isn't the Doha Development Agenda More Poverty Friendly?” Review of Development Economics 13 (2009):54359.
  • Integrated Framework, Tanzania: Diagnostic Trade Integration Study, (1) (2005), available at: http://www.enhancedif.org/documents/DTIS%20english%20documents/english/Tanzania_DTIS_Vol1_Nov05.pdf.
  • IPCC, “Climate Change 2007: The Physical Science Basis. Contribution of Working Group I,” in S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds), Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge: Cambridge University Press (2007).
  • Keeney, R. and T. W. Hertel, “GTAP-AGR: A Framework for Assessing the Implications Multilateral Changes in Agricultural Policies,” GTAP technical paper 24, Center for Global Trade Analysis, Purdue University (2005).
  • Knutson, T. R., J. J. Sirutis, S. T. Garner, G. A. Vecchi, and I. M. Held, “Simulated Reduction in Atlantic Hurricane Frequency under Twenty-first-century Warming Conditions,” Nature Geoscience 1 (2008):35964.
  • Knutti, R., M. R. Allen, P. Friedlingstein, J. M. Gregory, G. C. Hegerl, G. A. Meehl, M. Meinshausen, J. M. Murphy, G.-K. Plattner, S. C. B. Raper, T. F. Stocker, P. A. Stott, H. Teng, and T. M. L. Wigley, “A Review of Uncertainties in Global Temperature Projections Over The Twenty-first Century,” Journal of Climate 21 (2008):265163.
  • Lee, H. L., T. Hertel, S. Rose, and M. Avetisyan, “An Integrated Land Use Database for CGE Analysis and Climate Policy Options,” in T. W. Hertel, S. Rose, and R. Tol (eds), Economic Analysis of Land Use in Global Climate Change Policy, Abingdon, Oxon: Routledge (2009): ch. 4.
  • Loariel, S. R., P. B. Duffy, H. Hamilton, G. P. Asner, C. B. Field, and D. D. Ackerly, “The Velocity of Climate Change,” Nature 462 (2009):105255.
  • Lobell, D. B., M. B. Burke, C. Tebaldi, M. D. Mastrandrea, W. P. Falcon, and R. L. Naylor, “Prioritizing Climate Change Adaptation Needs for Food Security in 2030,” Science 319 (2008):60710.
  • Meehl, G.A., C. Covey, K.E. Taylor, T. Delworth, R.J. Stouffer, M. Lotif, B. McAvaney, and J. F. B. Mitchell, “The WCRP CMIP3 Multimodel Dataset: A New Era in Climate Change Research,” Bulletin of the American Meteorological Society 88 (2007a):138394.
  • Meehl, G. A., T. F. Stocker, W. D. Collins, P. Friedlingstein, A. T. Gaye, J. M. Gregory, A. Kitoh, R. Knutti, J. M. Murphy, A. Noda, S. C. B. Raper, I. G. Watterson, A. J. Weaver and Z.-C. Zhao, Global Climate Projections, in S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds), Fourth Assessment Report of the Intergovernmental Panel on Climate Change, ch. 10, Cambridge: Cambridge University Press (2007b):747845.
  • Mitra, S. and T. Josling, “Agricultural Export Restrictions: Welfare Implications and Trade Disciplines,” IPC position paper, Washington, DC: International Food & Agricultural Trade Policy Council (2009), available at: http://www.agritrade.org/GlobalExpRestrictions.html
  • Monfreda, C., N. Ramankutty, and T. W. Hertel, “Global Agricultural Land Use Data for Climate Change Analysis,” in T. W. Hertel, S. Rose, and R. Tol (eds), Economic Analysis of Land Use in Global Climate Change Policy, Abingdon, Oxon: Routledge (2009).
  • Nakicenovic, N. and R. Swart (eds), Special Report on Emissions Scenarios, Cambridge: Cambridge University Press (2000).
  • Randhir, T. and T. W. Hertel, “Trade Liberalization as a Vehicle for Adapting to Global Warming,” Agricultural and Resource Economics Review 29 (2000):15972.
  • Reilly, J., N. Hohmann, and S. Jane, “Climate Change and Agricultural Trade: Who Benefits, Who Loses?” Global Environmental Change 4 (1994):2436.
  • Reimer, J. J. and Man Li, “Yield Variability and Agricultural Trade,” Agricultural and Resource Economics Review 38 (2009): 25870.
  • Santer, B. D., K. E. Taylor, P. J. Gleckler, C. Bonfils, T. P. Barnett, D. W. Pierce, T. M. L. Wigley, C. Mears, F. J. Wentz, W. Brüggemann, N. P. Gillett, S. A. Klein, S. Solomon, P. A. Stott, and M. F. Wehner, “Incorporating Model Quality Information in Climate Change Detection and Attribution Studies,” Proceedings of the National Academy of Sciences 106 (2009):1477883.
  • Schlenker, W. and D. B. Lobell, “Robust Negative Impacts of Climate Change on African Agriculture,” Environmental Research Letters 5 (2010): 8pp.
  • Schlenker, W. and M. J. Roberts, “Nonlinear Temperature Effects Indicate Severe Damages to U.S. Crop Yields under Climate Change,” Proceedings of the National Academy of Sciences 106 (2009):1559498.
  • Tebaldi, C., K. Hayhoe, J. M. Arblaster, and G. A. Meehl “Going to the Extremes,” Climatic Change 79 (2006):185211.
  • Tobey, J., J. Reilly, and S. Jane, “Economic Implications of Global Climate Change for World Agriculture,” Journal of Agricultural and Resource Economics 17 (1992):195204.
  • Tsigas, M. E., G. B. Friswold, and B. Kuhn, “Global Climate Change and Agriculture,” in T. W. Hertel (ed.), Global Trade Analysis: Modeling and Applications, New York: Cambridge University Press (1997).
  • United Republic of Tanzania, “National Trade Policy,” Ministry of Industry and Trade, Dar es Saalam (2003), available at: http://www.tanzania.go.tz/pdf/tradepolicy.pdf
  • United Republic of Tanzania, “National Strategy for Growth and Reduction of Poverty (MKUKUTA),” Vice President's Office, Tanzania (2005).
  • Williams, J. W., S. T. Jackson, and J. E. Kutzbach, “Projected Distributions of Novel and Disappearing Climates by 2100AD,” Proceedings of the National Academy of Sciences 104 (2007):573842.
  • World Bank, “Eastern Africa: A Study of the Regional Maize Market and Marketing Costs,” Report 49831-AFR, World Bank, Washington, DC (2009).
Notes
  • 1

    Details on the estimation of percentage deviations from trend are available in the appendices of the World Bank working paper version.

  • 2

    The data for domestic production and consumption in Tanzania is based on a 1992 input–output table that was updated to be consistent with 2001 macro-economic and trade data. Details of the process and validity are documented in Dimaranan (2006).

  • 3

    The poverty model is fully documented in Ahmed et al. (2011), including parameter estimates for the AIDADS demand system used.

  • 4

    To obtain the appropriate scale for the productivity shocks, we perform a pre-simulation in which output is exogenous and productivity endogenous.

  • 5

    The production deviations for Mozambique, Rest of Sub-Saharan Africa, Southern Africa, Zambia, and Zimbabwe were too large for the model to converge on a solution. Production in these countries was thus assumed to be at trend for any given year simulated, and no deviations were simulated.

  • 6

    Becaue of the importance of the value of the Armington elasticity in this analysis, the value was varied by a factor of 50%. The impact on the results was small. Details are available in the appendices of the working paper version.

  • 7

    The sensitivity of these findings with respect to key parameter settings is explored in detail in the appendices of the working paper version. Suffice it to say that these conclusions are robust to such variation.

  • 8

    We note that although sub-annual variability is certainly important for agriculture, use of annual-scale aggregations allows us to compare countries from different hemispheres using a single metric.