WPS4601


Policy ReseaRch WoRking PaPeR                         4601




          Differential Adaptation Strategies by
    Agro-Ecological Zones in African Livestock
                             Management

                                   S. Niggol Seo
                                 Robert Mendelsohn
                                   Ariel Dinar
                              Pradeep Kurukulasuriya




The World Bank
Development Research Group
Sustainable Rural and Urban Development Team
April 2008

Policy ReseaRch WoRking PaPeR 4601


 Abstract

 This paper examines how farmers have adapted their                                    agro-ecological zones. Dairy cattle will decrease in semi-
 livestock operation to the current climate in each agro-                              arid regions, sheep will increase in the lowlands, and
 ecological zone in Africa. The authors examine how                                    chickens will increase at high elevations. With a mild
 climate has affected the farmer's choice to raise livestock                           and wet scenario, however, livestock adoption will fall
 or not and the choice of animal species. To measure                                   dramatically in lowland and high latitude moist agro-
 adaptation, the analysis regresses the farmer's choice on                             ecological zones. Beef cattle will increase and sheep will
 climate, soil, water flow, and socio-economic variables.                              fall in dry zones, dairy cattle will fall precipitously and
 The findings show that climate does in fact affect the                                goats will rise in moist zones, and chickens will increase
 farmer's decision about whether to raise livestock and the                            at high elevations but fall at mid elevations. Livestock
 species. The paper also simulates how future climates may                             adaptations depend on the climate scenario and will vary
 alter these decisions using forecasts from climate models                             across the landscape. Agro-ecological zones are a useful
 and the estimated model. With a hot dry scenario,                                     way to capture how these changes differ from place to
 livestock ownership will increase slightly across all of                              place.
 Africa, but especially in West Africa and high elevation



 This paper--a product of the Sustainable Rural and Urban DevelopmentTeam, Development Research Group--is part of
 a larger effort in the department to mainstream research on climate change. Policy Research Working Papers are also posted
 on the Web at http://econ.worldbank.org. The authors may be contacted at Niggol.seo@yale.edu, Robert.mendelsohn@
 yale.edu, Adinar@worldbank.org, and Pradeep.kurukulasuriya@undp.org.




          The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
          issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
          names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
          of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
          its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.


                                                          Produced by the Research Support Team

  Differential Adaptation Strategies by Agro-Ecological Zones in African Livestock

                                             Management1




    S. Niggol Seo2, Robert Mendelsohn3, Ariel Dinar4, and Pradeep Kurukulasuriya5.




1This paper is one of the product of a study "Measuring the Impact of and Adaptation to Climate Change
Using Agroecological Zones in Africa" funded by the KCP Trust Fund and conducted in DECRG at the
World Bank.. We benefited from comments by Christopher Delgado on an earlier draft. The views
expressed in this paper are the authors' alone.
2School of Forestry and Environmental Studies, Yale University; 230 Prospect St., New Haven, CT 06511,
USA; and a Consultant to the World Bank, email- Niggol.seo@yale.edu
3School of Forestry and Environmental Studies, Yale University; 230 Prospect St., New haven, CT 06511,
USA, and a Consultant to the World Bank, email- Robert.mendelsohn@yale.edu
4Development Research Group, World Bank, 1818 H St. NW, Washington DC 20433; phone 202-473-
0434; email adinar@worldbank.org .
5 Energy and Environment Group, Bureau of Development Policy, United Nations Development
Programme, New York; phone 212-217 2512; email: pradeep.kurukulasuriya@undp.org

1. Introduction

Past studies in developing countries on climate change impacts on agriculture revealed

that crops and especially grains are highly vulnerable to climate change (Rosenzweig and

Parry 1994, McCarthy et al. 2001). The main reasons for such high vulnerabilities are that

farmers in developing countries tend to be already located in a hot climate zone and that

these farmers may have less capacity to cope with climate risks. Researchers have argued

that farms in these areas should take adaptive measures in the coming decades to reduce

potential climate change impacts (Burton 1997, Smith 1997, Leary 1999, Mendelsohn

2000; Smit et al. 2000, Smit and Pilifosova 2001). Empirical work has revealed that

farmers can make several changes with relative ease to avoid large crop losses including

switching crop types and irrigation (Kurukulasuriya and Mendelsohn 2007; 2008a, Seo

and Mendelsohn 2008a). These adaptations are known to reduce but not eliminate the

damages from climate change (Kurukulasuriya et al. 2006; Kurukulasuriya and

Mendelsohn 2008b; Seo and Mendelsohn 2008b). Further, there are additional measures

that would require public (government) coordination such as the development of

irrigation potential and new breeds and varieties to cope with high temperatures. These

adaptations describe long-run behavior and do not capture some of the potential problems

associated with short-term adoption rates (Mendelsohn 2000; Kelly et al. 2005).

This paper examines long-term adaptations that farmers might make in their livestock

choices in response to climate.      We specifically examine whether adaptations vary

across Agro-Ecological Zones (AEZs). Livestock is an important topic since livestock

makes up over half of the total value of agricultural gross output in industrial countries,

and about a third of the total in developing countries, and this latter share is rising rapidly

(Nin, Ehui, and Benin 2007). Examining impacts across AEZs is important because there

is strong evidence to suggest that adaptations vary across the landscape. Of course,

climate is not the only variable that affects livestock choice. Farmers may invest in

livestock as part of a tribal custom or tradition. They may use livestock as an investment

device (wealth storage) in the absence of access to banking (Fafchamps et al. 1998).

Nonetheless, there is reason to believe that climate does affect livestock decisions. First,

just looking at the distribution of livestock across the planet, different species are more



                                              2

suited for different climates. Second, livestock may be used as a form of insurance

against crop loss in poor weather. Third, there is empirical evidence that climate affects

livestock decisions in Latin America and Africa (Seo and Mendelsohn 2007; 2008c,

2008d). Warmer temperatures increase the probability farmers will own livestock. In

addition, farmers change their portfolio of livestock species to match climate in their

location. However, past livestock research did not consider different adaptation

possibilities across AEZs.

Including information about AEZs helps in understanding the wide range of conditions

that farmers are currently facing and why each farmer may wish to adapt in a different

manner. The link between farm behavior and AEZs can also help extrapolate the results

from a limited sample area to all of Africa. We make use of the existing classification of

Agro-Ecological Zones in Africa by the Food and Agriculture Organization (FAO) and

quantify adaptation strategies by AEZ. Of course, AEZs were not developed to measure

livestock zones but rather crop zones.           Although attempts to develop suitable

classification zones for livestock have been undertaken, there is no final method that has

yet been developed. We consequently rely on the AEZ classification system in this paper.

This paper examines two important farm decisions by livestock owners in an effort to

adapt to climate change; livestock adoption and livestock species choice (Seo and

Mendelsohn 2007; 2008c, 2008d). We run binomial and multinomial choice models to

measure climate sensitivities of these choices and predict future choices based on the

estimated parameters. We then examine the link between these adaptive behaviors and

AEZs. We use this link to extrapolate behavior across Africa.

The paper proceeds as follows. The next section provides an economic theory of farm

decisions of livestock adoption and livestock species choice. The third section is devoted

to a detailed discussion of the data used in this study. Sections 4 and 5 present empirical

results of the models and predictions for the future, respectively. The paper concludes

with policy discussions and remaining issues.

2. Economic Theory

Although economic studies on climate change impacts on agriculture have focused on

crops, most farmers in Africa manage livestock in addition to crops. Some farmers


                                              3

manage only crops, but other farmers add livestock to their portfolio for various reasons

(Delgado 1999; Nin, Ehui, and Benin 2007). The first choice we examine is whether a

farmer adopts livestock because of different climate conditions. First of all, we assume
that farmers raise livestock if it is profitable to do so6. The second decision we examine is

which livestock species to own. In Africa, there were five major animals raised as

livestock: beef cattle, milk cattle, goats, sheep, and chickens. In our analysis, we will

consider only these five alternative livestock species. We specifically examine the

primary animal each farmer chooses, the animal that earns the greatest net revenue on

each farm. In practice, farmers can actually choose more than one species at a time. For

example, they can have beef cattle and chickens together. This paper assumes that the

farmer chooses only one species, the one that is most profitable (Train 2003). We

consequently focus only on the primary animal. In Africa, this seems reasonable, as

about 90% of total net revenue from livestock management is from the primary animal.

However, we have also explored examining all combinations of species and the results

were quite similar, although not completely the same (Seo and Mendelsohn 2008c).

With the species choice, we also assume that farmers choose the species that is most

profitable. We hypothesize that the profitability and therefore livestock choices depend on

the AEZ in which the farm is located.

Let the profit associated with livestock farming in a specific AEZ (w) be written in the

following form:

1 =V1(Zw)+1 wherew =1,...,W.
   w                  w
0 =V0(Zw)+0                                                                                              (1)
   w                  w


where Z is a vector of exogenous characteristics of the farm and characteristics of the

farmer.     The subscript 1 refers to the ownership of livestock and 0 to no livestock. The

subscript w refers to the AEZs. The farmer will choose to raise livestock if:



6 The theory of profit maximization can be contested especially in Africa due to a fragile market system.
Some also argued that livestock in Africa are kept for the store of wealth (Singh et al. 1986; De Janvry et al.
1991; Bardhan and Urdy 1999; Moll 2005). We made the following two adjustments to address the issues
arising from special situations of African markets. First, we assume that if a farmer consumes his own
product, it is valued at market price.  Second, most farms depend on their own labor. Although it might
be reasonable to value own labor by market wages, empirical examinations did not support any specific
wage.


                                                     4

1 >0
  *      *  or if 0 - 1 < V1(Zw) -V0(Zw)                                                (2)

Assuming that the cumulative distribution of error term is a logistic function, the choice

of whether or not to raise livestock can be estimated with a standard logit model.

The choice of which species to select is slightly more difficult because there are more

choices. Let the profit from raising a specific livestock species (j) for a farm located in

a specific AEZ (w) be written in the following form:

 =V (Zjw)+ where j=1,...,Jand w =1,...,W.                                               (3)
  jw               jw


The vector Z could include climate, soils, water availability, access variables, electricity

provision, and education of the farmer. The subscript (j) refers to livestock species and

(w) refers to the AEZ where the farm is located. Note that the farmer chooses animal (j)

from the multiple alternatives, but he does not choose the AEZ (w). The profit function in

equation 1 is composed of two components: the observable component V and an error

term . The error term captures various errors such as measurement error, mis-

specification of the model, or lack of appropriately available data.

The decision of a farmer who is located in AEZ (w) is to choose the one species that gives

him the highest profit. Suppressing subscript w for convenience, the farmer will choose

species (j) over all other farm types if:

 >k fork  j. [orif k - <V (Zj)-V (Zk)for k  j]
   *     *                                                                            (4)
  j                                   j




The probability Pj for livestock (j) to be chosen is then

Pj = Pr[k -  < Vj -Vk ]  k  j where Vj = V (Zj)                                       (5)
               j


Assuming  follows an identical and independent Type I Extreme Value distribution

and the profit can be written linearly in the parameters, then the probability can be

calculated by successive integration of the density function as




                                             5

Pj =       eZj j+Z2jj
           J  eZkk+Z2jk                                                               (6)
           k=1


which gives the probability of livestock (j) to be chosen among (J) animals (McFadden

1981).

For each AEZ w, the marginal effect of a change in climate on the probability an animal j

is chosen can be obtained by differentiating Equation (6):

Pjw                           J
                                                                                     (7)
 zl   = Pjw[ + 2Z  ]- Pkw[kl + 2Zkwlkl] for w =1,2,...,W.
              jl     jwl jl  
                             k=1


The coefficients of the choice model  and  are not dependent on the AEZ. However,

the marginal impact of climate on the probability of selecting an animal depends on the

climate conditions in each AEZ and so will vary by AEZ.

3. Description of the Data

The FAO has developed a typology of Agro-Ecological Zones as a mechanism to classify

the growing potential of land (FAO 1978). The AEZs are defined using the length of the

growing season.      The growing season, in turn, is defined as the period where

precipitation and stored soil moisture is greater than half of the evapotranspiration. The

longer the growing season, the more crops can be planted (or in multiple seasons) and the

higher are the yields (Fischer and van Velthuizen 1996; Vortman et al. 1999). FAO has

classified land throughout Africa using this AEZ concept. Our study will use these FAO

defined AEZ classifications.

The economic data for this study were collected by national teams as part of the

GEF/World Bank project on climate change in Africa (Dinar et al 2008). The survey

asked detailed questions on crops and livestock operation during the agricultural period

of July 2002 to June 2003. The data were collected for each plot within a household and

household level data were constructed from plot level data. In each country, districts were

chosen to get a wide representation of farms across climate conditions in that country.

The districts were not representative of the distribution of farms in each country as there

are more farms in more productive locations. In each chosen district, a survey was



                                             6

conducted of randomly selected farms. The sampling was clustered in villages to reduce

sampling cost. A total of 9597 surveys were administered across the 11 countries in the

study: Burkina Faso, Cameroon, Egypt, Ethiopia, Ghana, Kenya, Niger, Senegal, South

Africa, Zambia, and Zimbabwe.

Climate data were gathered from two sources. We relied on satellite observations for

temperature and ground weather station observations for precipitation (Mendelsohn et al.

2007). The United States Department of Defense uses a set of polar orbiting satellites that

pass above each location on earth between 6am and 6pm every day. These satellites are

equipped with sensors that measure surface temperature by detecting microwaves that

pass through clouds (Weng & Grody 1998, Basist et al. 2001). The precipitation data

came from the Africa Rainfall and Temperature Evaluation System (ARTES) (World

Bank 2003). This dataset, created by the National Oceanic and Atmospheric Association's

Climate Prediction Center, is based on ground station measurements of precipitation.

The monthly data were organized into three month seasons. We define the winter in the

northern hemisphere as the average of November, December and January. February,

March and April are spring, May, June and July are summer, and August, September and

October are fall. The seasons in the southern hemispheres are assumed to be 6 months

apart from the northern hemisphere seasons. For example, the winter in the southern

hemisphere is May, June and July (Kurukulasuriya et al. 2006).

Soil data were obtained from the Food and Agriculture Organization's digital soil map of

the world (FAO 2003). The FAO data provide information about the major and minor

soils in each location as well as slope and texture. Data concerning the hydrology were

obtained from the results of an analysis of climate change impacts on African hydrology

(Strzepek and McCluskey 2006). Using a hydrological model for Africa, the authors

calculated flow and runoff for each district in the surveyed countries. Data on elevation at

the centroid of each district were obtained from the United States Geological Survey

(USGS 2004). The USGS data are derived from a global digital elevation model with a

horizontal grid spacing of 30 arc seconds (approximately one kilometer).

4. Empirical Results

Table 1 presents the fraction of the farms that have adopted livestock for Africa as a


                                              7

whole and by each AEZ. For Africa as a whole, almost 80% of farms own livestock7.

Table 2 presents the fraction of each livestock species chosen by farms with livestock.

Across Africa, 32% of livestock farms chose chickens, 20% chose goats, 20% chose

sheep, 20% chose dairy cattle, and 7% chose beef cattle. Within each species of animals,

there are different breeds of that species. For example, beef cattle which are origin to

India are believed to be more heat tolerant than the African native breeds of beef cattle

(Oklahoma State University 2007). The current data, however, did not have information
on the breeds within each species8.

Tables 1 and 2 also reveal how these percentages vary by AEZ. Livestock are chosen by

farms more often in high and mid elevation dry regions and less often in mid and high

elevation humid forest and low elevation moist savannah. In Africa, livestock is more

popular in drier locations. The choice of different livestock species also varies by AEZ.

Farmers choose beef cattle more often in deserts but less often in high elevation dry

regions. Dairy farms are chosen more often in desert, mid elevation, and high elevation

regions and less often in low elevation regions. Sheep are chosen less often in high

elevation and mid elevation regions that are moist and more often in low elevation

regions. Farms in mid and high elevations are less likely to choose goats whereas farms

in low elevations are more likely to pick goats. Chickens are chosen more often in wet

places in low and mid elevation regions.

Figure 1 maps the AEZs across Africa. Agro-Ecological Zones are classified by climate,

soils, and altitude. They are divided into five zones depending upon precipitation: semi-

arid, dry savannah, moist savannah, sub-humid, and humid forest. Each of these five

zones is again divided into three zones depending upon the elevation: lowland, mid-

elevation, and high elevation. The remaining AEZ is desert. The Sahara desert occupies a

vast amount of area in the north. There is also a desert in the south-western edge of the

continent. South of the Sahara desert is semi-arid zones followed by dryland savannah,

moist savannah, and humid forest. In central Africa around Cameroon, it is mostly

humid forest in low elevation with high rainfall. This low-elevation humid forest turns

7 It means the rest 20% of farms specialized in crops only.
8 Researchers also found that the transfer of a new technology to Africa has been much slower than the rest
of the developing countries (Evenson and Gollin 2003). Heat tolerant breeds of beef cattle largely had not
been adopted by African farmers.


                                                     8

into mid-elevation and then into dry savannah as it stretches east toward Kenya. South of

the humid forest is moist savannah followed by dry savannah. The AEZs of South Africa

are mostly moist savannah in the east, dry savannah in the center, and desert in the west.

Because each AEZ faces a different climate and agricultural condition, we anticipate that

farms in each AEZ adopt different agricultural practices. This is confirmed in Tables 1

and 2. We now wish to quantify this relationship more carefully with an empirical

model. We first examine whether farms in each AEZ are more or less likely to adopt

livestock operations. Table 3 presents the estimated results of a binary choice model of

livestock adoption using a logit regression. The dummy variable for livestock ownership

is run against linear and squared climate variables, climate interaction variables, soils,

and socio-economic variables following the literature in this area (Mendelsohn et al. 1994,

Adams et al. 1999). The regression uses summer and winter temperature and precipitation.

The regression includes temperature and precipitation interaction terms for both summer

and winter. Although models of crops have used all four seasons of the year, none of the

livestock models estimated to date have found all four seasons to be significant in either

South America or Africa (Seo and Mendelsohn 2007; 2008c; 2008d).

The regression confirms that the livestock ownership decision is highly dependent on the

climate in which the farm is located. Both temperature and precipitation are significant

especially in the summer. The summer temperature and precipitation interaction variable

is negative and significant. This negative estimate of the summer interaction variable

implies that farms are less likely to adopt livestock in warmer locations if the area is also

wet. The interaction variables may be picking up problems with livestock diseases in

hot wet locations (such as Trypanosomiasis (Nagana), Theileriasis (East Coast Fever),

and Rift Valley Fever) (Ford and Katondo 1977; University of Georgia 2007).

Some of the control variables are also significant. When there is electricity in the farm,

they are less likely to have livestock at all. Livestock is chosen more often in West Africa,

the Sahel, and high elevation AEZs in which fewer farms have electricity compared to the
other parts of Africa9. Most of the soil and flow variables are not significant. When water



9 These results seem to be related to small livestock farms that own sheep, goats and chickens but do not
have electricity. Farms with electricity have more often chosen cattle.


                                                      9

flow is high in a district, the district is more likely to have livestock, but the estimate is

not significant. When the head of farm works at the farm, the farm is less likely to have

livestock, but the estimate is not significant.

Due to its nonlinear functional form, it is difficult to interpret the estimated parameters in

Table 3. We consequently calculate the marginal effect of temperature and precipitation

on the probability to have livestock evaluated at the annual mean of each climate variable.

The results presented in Table 4 reveal that, for Africa as a whole, livestock ownership

increases with warming. However, as precipitation rises, farmers choose to own livestock

less frequently. Both these results are reasonable given that a farmer's main choices are

between crops and livestock. As it gets warmer, crops become much less profitable,

making livestock more attractive. As it gets wetter, however, crops become much more

productive, making livestock relatively less attractive. Livestock in Africa is also very

susceptible to various livestock diseases which become prevalent in wet places (Ford and

Katondo 1977, University of Georgia 2007).

To examine the choice of primary livestock species, we estimate a multinomial logit

model. Table 5 shows a set of regressions, one for each animal, leaving out chickens.

The independent variables include a set of climate variables, social variables, and own

and cross prices. The coefficients on the seasonal climate variables reveal that species

choice is highly sensitive to climate. Every species choice has at least some significant

climate coefficients. Among the control variables, electricity and water availability are

significant for cattle ownership. Farms with electricity are more likely to have dairy cattle,

but less likely to have beef cattle. Electricity is needed in milking and cooling of milk.

Farms in districts with more water flow are less likely to choose beef cattle but more

likely to own dairy cattle. Beef cattle is concentrated in South Africa and high elevation

farms in Kenya in which climate is dry and have lower water flow. Some but not all price

variables are significant. The own price has a positive effect on choosing dairy cattle as

expected but an unexpected negative effect on choosing beef cattle. The results reflect the

fact that beef cattle can be sold only after 4-5 years of raising on the commons, but milk

can be sold immediately. Hence higher milk price is incentive for the farmers whereas

higher beef price is a disincentive to the farmers. Own price terms are not significant for

goats and sheep, but cross price terms are highly significant for sheep ownership. When


                                              10

the prices of other animals are high, farmers tend to switch to other animals from sheep.

To understand the nonlinear climate coefficients, we calculate the marginal effects of

temperature and precipitation on the selection of the above five species evaluated at the

mean climate. Table 6 shows that farmers would change their portfolio of livestock as

the current climate is disturbed. As temperature increases, farmers will switch from cattle

and chickens to goats and sheep. Goats and sheep are more heat tolerant so they can
endure the warmer temperatures10.                Changing rainfall also shifts African farmer's

choices. As rainfall increases, fewer farms choose dairy cattle and sheep while more
farms raise goats and chickens11.

One of the insights of this paper is that farmers should adapt to climate change differently

depending upon the agricultural economic conditions of where the farm is located. This

can be seen clearly from the Tables 1 and 2 in which farms in different agro-ecological

zones show different preferences for livestock ownership and livestock species choice.

Hence, as climate changes and agro-ecological conditions change for a farmer in a

specific AEZ, the farmer is likely to change the portfolio of livestock.

Based on the parameter estimates in Table 3, we map in Figure 2 the current probability

of owning livestock in each district. The leftmost figure in Figure 2 shows the current

probability of owning livestock across Africa. The lowest livestock levels are in the

moist lowland forests of central Africa. It is also clear that farms have less livestock in

the desert areas of the Sahara and southwest and eastern Africa. Semi arid and savannah

regions, however, have more livestock as can be seen in the Sahel and near Mozambique.

The current probability of owning dairy cattle and sheep are mapped in the leftmost

figures in Figure 3 and Figure 4, respectively. The other livestock species are shown in

the Appendix. Dairy cattle are chosen widely across Africa except in the Sahel and

along the eastern edge of Africa. Note that dairy cattle are raised even in the driest and

hottest places in Africa. In contrast, farmers avoid sheep in the dry and hot parts of Africa

and also in lowland humid forests. Farmers are much more likely to select sheep in

West Africa and mid elevation dry zones.

10In contrast to beef cattle, goats and sheep provide a shorter cycle returns.
11Sheep are known to be much more vulnerable to the parasites that spread in wet conditions (Delgado
1999).


                                                    11

The maps for goats and chickens are placed in the Appendix. The maps for goats are

similar to the maps for sheep. Goats are chosen more often in West Africa and in the

eastern countries such as Kenya and Mozambique. However, in contrast to sheep, goats

are owned by many farms in lowland humid zones around Cameroon. Chicken ownership

resembles the results for dairy cattle. Chickens are owned widely across Africa, but are

favored in wetter places below the Sahel. We do not present a beef cattle map. Beef

cattle are highly concentrated in two places: South Africa and the highlands of East

Africa.

5. Forecasting Livestock Adaptations

As climate change unfolds over the coming century, farmers are likely to adapt to the

changes by adding or subtracting livestock operations and by switching livestock species

to minimize the damage and take best advantage of new climate conditions. In this

section, we provide an analysis of how farmers would make such changes in the next 100

years. We assume that the cross sectional patterns of behavior we observe today can be

used to predict how farmers will adapt over the long run. In making these predictions,

we assume that it is only climate that changes over time. Obviously, there will be many

other changes as Africa develops including income, technology, and land use. Because

we do not take these factors into account the projections should not be interpreted as

predictions of what would happen in the future. These many factors need to be taken into

account to predict future outcomes. The projections in this study are merely trying to see

how farm choices might vary with climate. It is also important to understand that these

projections are intended to represent long run adaptations, not changes that farmers make
from day to day12.

To introduce climate change, we examine a set of climate change scenarios predicted by

Atmospheric-Oceanic Global Circulation Models (AOGCM's). We rely on two climate

scenarios that bracket the range of outcomes predicted in the most recent IPCC

(Intergovernmental Panel on Climate Change) report (IPCC 2007): CCC (Canadian


12 Individual producer livestock holdings in Africa have been typically explained by a multi-year risk mitigation where
the risks are driven by droughts and diseases (Delgado 1999). This analysis, however, examines livestock holdings as a
long term adaptation to existing climate. Hence, if the climate in the year 2002 that our study is based on was
substantially different from the normal climate, the estimates in the study suffer from the biases introduced by such
discrepancies. However, there is no evidence of such a peculiarity in the year 2002 weather.



                                                         12

Climate Centre) (Boer et al. 2000) and PCM (Parallel Climate Model) (Washington et al.

2000).

Table 7 presents the mean temperature and rainfall predicted by the two models for the

years 2020 and 2100. In Africa in 2100, PCM predicts a 2°C increase and CCC a 6°C

increase in temperature. Rainfall predictions vary. PCM predicts a 10% increase in

rainfall in Africa and CCC a 10% decrease by 2100. Even though the mean rainfall in

Africa is predicted to increase/decrease depending on the scenario, there is also

substantial variation in rainfall across countries. Examining the path of climate change

over time reveals that temperatures are predicted to increase over time for all models.

Precipitation predictions, however, vary across time for Africa: CCC predicts declining

precipitation whereas PCM predicts a slight increase. However, it should be noted that

predicted changes vary slightly for individual countries and regions.

5.1 Analysis for Africa

We simulate climate change impacts on the choice of livestock ownership and livestock

species based on the parameter estimates in the previous section conditional on each of

the above two climate scenarios. Table 8a indicates that more farmers in Africa will own

livestock by 2020 under the hot CCC scenario and that much fewer farms will own

livestock under the wet PCM scenario. Table 8b indicates that by 2100, the effects of

each climate scenario will intensify. There will be a small reduction in livestock

ownership with the PCM scenario and an increase under the CCC scenario. These
results are due to a combination of seasonal changes in precipitation and temperature13.

The results for each livestock species reveal that by 2020, beef cattle, dairy cattle, and

chickens will begin to decline, but goats and sheep will increase. The results, however,

depend on the climate scenarios. For example, although sheep selection will increase

under the CCC scenario, it will decrease with the PCM scenario. Beef cattle will be

chosen more often with the PCM. These trends intensify through 2100.




13 These results may not hold if there will be a great deal of changes in the demand for livestock in the
future (Delgado 1999). Population changes and urbanization in the coming decades will affect the overall
results described in this paper. Although it is desirable to model these additional changes, it is difficult
without a good general equilibrium model of the world economy which does not exist at the moment.


                                                    13

5.2 Analysis by Agro-Ecological Zones

The livestock choices in each AEZ are not identical to the average choices for the

continent as shown in Tables 8a and 8b. For example, in 2100, continental livestock

ownership is expected to decrease in the PCM scenario.            Yet, in some AEZs, the

probability to adopt livestock increases. For example, in the PCM scenario, livestock

increases in the desert and lowland semi-arid regions. Although livestock ownership

increases in every AEZ under the CCC 2100 scenario, ownership rises less in the lowland

moist regions and much more in the mid and high elevation regions.

Table 8a also shows how the choice of each species would change by 2020. With the

PCM scenario, chickens are chosen less often in most AEZs, but more often in the mid

elevation semi-arid AEZ. Dairy cattle choice also decreases across Africa except for the

desert areas. With the CCC scenario, goats are chosen more frequently across all the

AEZs except for the lowland dry savannah and lowland semi-arid zones. Chickens also

decrease overall, but they increase in the highland AEZs.

Table 8b shows the simulation results for 2100. Beef, dairy, and chickens had similar

reactions across the two climate scenarios. Beef will likely fall in the desert but increase

in the low and mid elevation moist regions. Dairy will fall in mid and high elevations

but increase in low semi-arid regions. Chickens will increase in high elevation but

decrease in low elevation regions. However, sheep and goats had different effects

across AEZs depending on the climate scenario.           In the PCM scenario, goats will

increase in the high elevation regions. In the PCM scenario, goats will also increase in the

mid elevation regions. Goats will fall in the mid elevation regions and in some lowland

regions in the CCC scenario. Sheep will increase across the board in the CCC scenario.

In the PCM scenario, sheep will fall in the high and mid elevation regions.

The change in the probability of livestock is presented in Figure 2 for the PCM (right

figure) and CCC (center figure) scenarios. The PCM scenario shows that livestock

ownership will decrease precipitously in lowland humid AEZs, mid elevation and high

elevation AEZs. Only deserts will see an increase in livestock ownership in the PCM

scenario. The CCC scenario shows that livestock ownership will increase across Africa

except for the desert areas but it will increase more in the highland zones and West Africa.



                                              14

The change in the probability of adopting dairy cattle and sheep are shown in Figures 3

and 4 respectively. The effect of the 2100 PCM scenario is shown in the right figure

and the effect of the 2100 CCC scenario is shown in the middle figure. With the PCM

scenario, dairy cattle ownership decreases in general but it increases in the areas close to

the desert.   The additional precipitation makes dairy cattle relatively more attractive in

these dry areas. With the CCC scenario, dairy cattle ownership also decreases across

the landscape, but especially in the high elevation AEZs.

Figure 4 shows the results for sheep. With the PCM scenario, sheep ownership decreases

in desert, high elevation, and lowland humid regions. However, sheep ownership

increases in West Africa and mid elevation areas in Southern Africa. With the CCC

scenario, sheep ownership increases in all the AEZs, but especially in West Africa and

mid elevation AEZs.

The figures for the distributions of goats and chickens in 2100 are shown in the Appendix.

In general, goats and sheep show a similar change in their spatial distribution except that

goats prefer wetter places while sheep prefer dryer places. Dairy cattle, beef cattle, and

chickens follow a similar pattern as well in that they are all very vulnerable to warming.

But chickens do well in the high elevation farms even if climate turns dry. When climate

turns wet under PCM scenario, farms close to the deserts or in high mountains increase

chicken ownership. Beef cattle ownership in the temperate regions of South Africa and

Kenya will fall with warming. The effect is more severe, the warmer the temperature

becomes.

6. Conclusion and Policy Implications

This paper examines how farmers currently adapt to the climate in each of their AEZs by

choosing to raise livestock and choosing different species of livestock. In the first

analysis, we ran a binary logit model of livestock adoption, and in the second analysis,

we used a multinomial logit model across livestock species. The empirical estimates were

used to compare climate change impacts in 2100 for each of the 16 AEZs of Africa.

Our results indicate that farmers will adopt more livestock as temperature increases.

However, farmers decrease their ownership of livestock if rainfall increases. Livestock is

chosen more often when climate is dry. They will also switch livestock species. As


                                             15

temperature rises, they tend to move away from beef cattle, dairy cattle, and chickens and

towards goats and sheep. As precipitation increases, they tend to shift away from beef

cattle, dairy cattle, and sheep towards goats and chickens.

These results, however, vary across Agro-Ecological Zones. Some adaptation strategies

suggested by the Africa-wide results are not appropriate for a specific AEZ. To see the

AEZ specific results, we simulated how climate change impacts would affect the above

two choices for each AEZ. The results suggest that livestock ownership will increase

across Africa except for the desert areas. The largest increase in livestock adoption will

happen in the high elevation AEZs under the CCC scenario.

Livestock species choice will also differ by the AEZs. Dairy cattle, beef cattle, and

chickens have similar responses across all two climate scenarios.             Dairy cattle

ownership will decrease across Africa, but especially in the high elevation AEZs. Beef

cattle will fall in the desert but increase in low and mid elevation moist regions.

Chickens will increase in high elevation regions but fall in low elevation regions. Sheep

and goats, however, had different responses across the AEZs depending on the climate

scenario. Under the CCC scenario, sheep ownership increases in all the AEZs, but

especially in West Africa. Under the PCM scenario, desert, high elevation, and lowland

humid regions all see a decrease of sheep ownership but sheep ownership increases in

West Africa and mid elevation areas in the Southern Africa. In general, goats and sheep

show a similar response to temperature changes, but goats prefer wetter places while

sheep prefer drier places. Thus goats increase in low and mid elevation regions in the wet

PCM scenario but decrease in these same regions in the CCC scenario.

In conclusion, farmers can make some adaptations to climate change by adding or

removing livestock from their portfolio and by switching species. These adaptations

vary across Agro-Ecological Zones. It is important that policies designed to facilitate

adaptation avoid a uniform Africa-wide approach but rather be tailored to each Agro-

Ecological Zone. Of course, there are many other determinants to livestock ownership

and species choice that this study did not take into account. Future policies will have to

take into account changes in global prices for livestock, and local changes in population,

income, and development. This study provides just a preliminary perspective on what



                                             16

adaptations are appropriate for each future climate scenario.




                                            17

References

Basist, A., Williams Jr., C., Grody, N., Ross, T.E., Shen, S., Chang, A., Ferraro, R.,
       Menne, M. J.: 2001. "Using the Special Sensor Microwave Imager to monitor
       surface wetness". J. of Hydrometeorology 2: 297-308.

Bardhan, P. and C. Udry, 1999. Development Economics. Oxford University Press,
       Oxford.

Boer, G., G. Flato, and D. Ramsden (2000), "A transient climate change simulation with
       greenhouse gas and aerosol forcing: projected climate for the 21st century",
       Climate Dynamics 16, 427-450.

Burton, I. 1997. "Vulnerability and adaptive response in the context of climate and
       climate change", Climatic Change 36: 185-196.

Delgado, Christopher. Livestock to 2020: The Next Food Revolution. International Food
       Policy Research Institute. 1999. pp 79.

Evenson, R., and D. Gollin. 2003. Assessing the impact of the Green Revolution, 1960-
       2000. Science 300:758-762.

De Janvry, A., M. Fafchamps, and E. Sadoulet, " Peasant Household Behavior with
       Missi-Ng markets: Some Paradoxes Explained." Economic Journal 101, No.409
       (1991), 1400-1417.

Dinar, A., R. Hassan, R. Mendelsohn, and J. Benhin (eds), 2008. Climate Change and
       Agriculture in Africa: Impact Assessment and Adaptation Strategies, London:
       EarthScan (forthcoming).

Fafchamps, M., Udry, C., and Czukas, K., ``Drought and Saving in West Africa: Are
Livestock a Buffer Stock?,'' Journal of Development Economics, 55: 273-305,
1998.

FAO. 1978. "Report on Agro-Ecological Zones; Volume 1: Methodology and Results for
       Africa" Rome.

Fischer, G. and H. van Velthuizen 1996. "Climate Change and Global Agricultural
       Potential: A Case of Kenya" IIASA Working Paper WP-96-071.

Ford, J. and K. Katondo, The Distribution of Tsetse flies in Africa (Organization of
       African Unity: Nairobi, Kenya, 1997).

IPCC 2007. (Intergovernmental Panel on Climate Change) Climate Change 2007: The
       Physical Science Basis. Contribution of Working Group I to the Fourth
       Assessment Report [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B.
       Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press,
       Cambridge, United Kingdom

Kelly, D.L., Kolstad, C. D., and G.T. Mitchell, 2005. Adjustment costs from
       environmental change. Journal of Environmental Economics and Management
       50: 468-495.

Kurukulasuriya, P., R. Mendelsohn, R. Hassan, J. Benhin, M. Diop, H. M. Eid, K.Y. Fosu,



                                           18

       G. Gbetibouo, S. Jain, A. Mahamadou, S. El-Marsafawy, S. Ouda, M. Ouedraogo,
       I. Sène, D. Maddision, N. Seo and A. Dinar. 2006. "Will African Agriculture
       Survive Climate Change?" World Bank Economic Review 20(3): 367-388.

Kurukulasuriya, P. and R. Mendelsohn. 2007.. "Modeling Endogenous Irrigation: The
       Impact Of Climate Change On Farmers In Africa" World Bank Policy Research
       Working Paper 4278, Washington DC.

Kurukulasuriya, P. and R. Mendelsohn, 2008a. "Crop Switching as an Adaptation
       Strategy to Climate Change" African Journal Agriculture and Resource
       Economics (forthcoming).
Kurukulasuriya, P. and R. Mendelsohn, 2008b. "A Ricardian Analysis of The Impact
       of Climate Change on African Cropland" African Journal Agriculture and
       Resource Economics (forthcoming).
Leary, N. A. 1999, "A framework for Benefit-Cost Analysis of Adaptation to Climate
       Change and Climate Variability", Mitigation and Adaptation Strategies for Global
       Change 4: 307-318.

McFadden, D. L. 1981. "Econometric Models of Probabilistic Choice", in D. McFadden,
       Structural Analysis of Discrete Data and Econometric Applications MIT Press,
       Cambridge.

McCarthy, J., O. Canziani, N. Leary, D. Dokken, and K. White (eds.) 2001, Climate
       Change 2001: Impacts, Adaptation, and Vulnerability, Intergovernmental Panel
       on Climate Change Cambridge University Press: Cambridge.

Mendelsohn, R. "Efficient Adaptation to Climate Change", Climatic Change 45 (2000)
       583-600.

Mendelsohn, R., A. Dinar and A. Sanghi. "The Effect of Development on the Climate
       Sensitivity of Agriculture", Environment and Development Economics 6 (2001)
       85-101.

Mendelsohn, Robert, Pradeep Kurukulasuriya, Alan Basist, Felix Kogan, and Claude
       Williams. 2007. "Measuring Climate Change Impacts with Satellite versus
       Weather Station Data.", Climatic Change 81: 71-83.

Mendelsohn, R. and S.N. Seo 2007, "An Integrated Farm Model of Crops and Livestock:
       Modeling Latin American Agricultural Impacts and Adaptations to Climate
       Change", World Bank Policy Research Series Working Paper4161, Washington
       DC, USA.

Moll, H. A.J., "Costs and Benefits of Livestock Systems and the Role of Market and
       Nonmarket Relationships", Agricultural Economics 32, (2005),181-193.

Nin, A., S. Ehui, and S. Benin 2007, "Livestock Productivity in Developing Countries:
       An Assessment", in R. Evenson and P. Pingali (eds), Handbook of Agricultural
       Economics, Volume 3, North Holland, Oxford, UK. pp 2467-2532.

Oklahoma      State  University,    2007.  Breeds    of   Livestock,    Available    at
       htttp://www.ansi.okstate.edu/breeds.


                                          19

Reilly, J., et al. 1996. "Agriculture in a Changing Climate: Impacts and Adaptations" in
        IPCC (Intergovernmental Panel on Climate Change), Watson, R., M. Zinyowera,
        R. Moss, and D. Dokken (eds.) Climate Change 1995: Impacts, Adaptations, and
        Mitigation of Climate Change: Scientific-Technical Analyses, Cambridge
        University Press: Cambridge p427-468.

Seo, N. and R. Mendelsohn. 2007. "An Analysis of Livestock Choice: Adapting to
        Climate Change in Latin American Farms" World Bank Policy Research Working
        Paper 4164.
Seo, S. N.. and R. Mendelsohn 2008a, "The Analysis of Adaptation of Farmers in Latin
        America: Crop Switching", Ecological Economics (forthcoming).

Seo, S. N. and R. Mendelsohn 2008b, "A Ricardian Analysis of the Impact of Climate
        Change on Latin American Farms", Agricultural Technica (forthcoming)

Seo, N. and R. Mendelsohn. 2008c. "Climate Change Impacts and Adaptations on Animal
        Husbandry in Africa" African Journal Agriculture and Resource Economics
        (forthcoming).

Seo, S. N. and R. Mendelsohn 2008d, "Measuring Impacts and Adaptations to Climate
        Change: A Structural Ricardian Model of African Livestock Management",
        Agricultural Economics (forthcoming).

Singh, I., L. Squire, and J. Strauss, "A survey of Agricultural Household Models: Recent
        Findings and Policy Implications", World Bank Economic Review 1, No. 1(1986),
        149-179.

Smit, Barry, Ian Burton, R. J. T. Klein, and J. Wandel. 2000. "An anatomy of adaptation
        to climate change and variability", Climatic Change 45(1): 223-251.

Smit, Barry and Olga Pilifosova. 2001. "Adaptation to Climate Change in the Context of
        Sustainable Development and Equity." In J.J. McCarthy, O.F. Canzianni, N.A.
        Leary, D.J.

Dokken, and K.S. White, eds., Climate Change 2001: Impacts, Adaptation, and
        vulnerability - Contribution of Working Group II to the Third Assessment Report
        of the Intergovernmental Panel on Climate Change. Cambridge, U.K.: Cambridge
        University Press.

Smith, Joel. 1997. "Setting Priorities for Adaptation to Climate Change." Global
        Environmental Change 7(3): 251-264.

Train, K., Discrete Choice Methods with Simulation (Cambridge, U.K.: Cambridge
        University Press, 2003).

University of Georgia, College of Veterinary Medicine, 2007. Foreign Animal Diseases:
        The Gray Book, Available at http://www.vet.uga.edu/vpp/gray_book02.

USGS (United States Geological Survey), 2004. Global 30 Arc Second Elevation Data,
        USGS         National     Mapping       Division,     EROS      Data     Centre.
        http://edcdaac.usgs.gov/gtopo30/gtopo30.asp

Voortman, R., B. Sonnedfeld, J. Langeweld, G. Fischer, H. Van Veldhuizen. 1999,


                                             20

      "Climate Change and Global Agricultural Potential: A Case of Nigeria" Centre for
      World Food Studies, Vrije Universiteit, Amsterdam.

Washington,W., J. Weatherly, G. Meehl, A. Semtner, T. Bettge, A. Craig,W. Strand, J.
      Arblaster, V. Wayland, R. James, and Y. Zhang, "Parallel Climate Model (PCM):
      Control and Transient Scenarios" Climate Dynamics 16(2003), 755-774.

World Bank. 2003. Africa Rainfall and Temperature Evaluation System (ARTES). World
      Bank, Washington DC.




                                         21

Table 1: Number of Livestock Farms by AEZs


       AEZ            Description                   Number Percentage
       Africa                                         8427       79.7

       AEZ 1          Desert                           959       82.1

       AEZ 2          High elevation dry savanna       117       91.5

       AEZ 3          High elevation humid forest      941       74.2

       AEZ 4          High elevation moist savannah    362       80.1

       AEZ 5          High elevation semi-arid          73       87.7

       AEZ 6          High elevation sub-humid         800       82.8

       AEZ 7          Lowland dry savannah            2798       79.6

       AEZ 8          Lowland humid forest            1242       83.6

       AEZ 9          Lowland moist Savannah          2152       73.1

       AEZ 10         Lowland semi-arid                699       81.8

       AEZ 11         Lowland sub-humid               1286       82.3

       AEZ 12         Mid-elevation dry savannah       907       81.5

       AEZ 13         Mid-elevation humid forest       984       75.2

       AEZ 14         Mid-elevation moist savannah    2014       78.6

       AEZ 15         Mid-elevation semi-arid          134       88.8

       AEZ 16         Mid-elevation sub-humid         1039       79.2




                                        22

Table 2: Number of Livestock Species by AEZ
(1) Africa
                           Percentage
 Africa
 Beef cattle                                                            7.2

 Dairy cattle                                                         21.7

 Goats                                                                19.1

 Sheep                                                                20.0

 Chickens                                                             32.0

(2) By AEZs.
 AEZ                       Percentage       AEZ           Percentage
                                                            Lowland moist
 AEZ 1                               Desert AEZ 9                 Savannah
 Beef cattle                           18.2Beef cattle                  4.1

 Dairy cattle                          51.6Dairy cattle               16.3

 Goats                                  4.5Goats                      22.9

 Sheep                                  6.8Sheep                      23.2

 Chickens                              19.0Chickens                   33.5

                             High elevation                 Lowland semi-
 AEZ 2                         dry savanna AEZ 10                      arid
 Beef cattle                            0.0Beef cattle                  5.3

 Dairy cattle                          53.9Dairy cattle               40.2

 Goats                                  2.6Goats                      18.7

 Sheep                                 16.7Sheep                      20.4

 Chickens                              26.9Chickens                   15.4

                             High elevation                   Lowland sub-
 AEZ 3                        humid forest AEZ 11                    humid
 Beef cattle                            4.0Beef cattle                  9.8

 Dairy cattle                          60.5Dairy cattle               12.8

 Goats                                  7.5Goats                      21.4

 Sheep                                 11.7Sheep                      20.9

 Chickens                              16.4Chickens                   35.0

                             High elevation               Mid-elevation dry
 AEZ 4                      moist savannah AEZ 12                 savannah
 Beef cattle                            7.5Beef cattle                 11.5

 Dairy cattle                          43.7Dairy cattle               33.5



                                         23

Goats                      6.0 Goats                      9.4

Sheep                    23.6  Sheep                     14.2

Chickens                 19.1  Chickens                  31.6

                High elevation                 Mid-elevation
AEZ 5                semi-arid AEZ 13           humid forest
Beef cattle                0.0 Beef cattle                4.5

Dairy cattle             67.4  Dairy cattle              47.0

Goats                    14.3  Goats                     12.1

Sheep                      2.0 Sheep                     12.1

Chickens                 16.3  Chickens                  24.4

                High elevation                 Mid-elevation
AEZ 6              sub-humid AEZ 14            moist savannah
Beef cattle                7.9 Beef cattle               10.2

Dairy cattle             46.5  Dairy cattle              30.6

Goats                    12.2  Goats                     11.0

Sheep                    14.3  Sheep                      8.3

Chickens                 19.2  Chickens                  39.9

                  Lowland dry                  Mid-elevation
AEZ 7                savannah AEZ 15                semi-arid
Beef cattle                3.8 Beef cattle                5.1

Dairy cattle             16.5  Dairy cattle              47.5

Goats                    22.7  Goats                     14.1

Sheep                    28.2  Sheep                     16.2

Chickens                 28.9  Chickens                  17.2

                Lowland humid                  Mid-elevation
AEZ 8                   forest AEZ 16             sub-humid
Beef cattle                4.8 Beef cattle                5.9

Dairy cattle               7.5 Dairy cattle              51.2

Goats                    28.0  Goats                     10.9

Sheep                    20.4  Sheep                     13.2

Chickens                 39.3  Chickens                  18.8




                            24

Table 3: Logit Model of Livestock Adoption


Variable                   Estimate          Chisq Statistic   P value
Intercept                             -1.1147              0.95          0.33

Summer Temperature                     0.1837              4.58          0.03
Summer Temperature2                   -0.0032              4.00          0.05

Summer Precipitation                   0.0123              5.95          0.01
Summer Precipitation2               -4.43E-06              0.37          0.54

Winter Temperature                    -0.0249              0.13          0.72
Winter Temperature2                  0.00158               0.66          0.42

Winter Precipitation                   0.0038              0.32          0.57
Winter Precipitation2                -0.00006              8.35          0.00

Summer Temp * Prec                   -0.00057             10.81          0.00

Winter Temp * Prec                   -0.00002              0.00          0.96

Flow                                 0.00263               0.15          0.70

Head Farm                             -0.0541              0.63          0.43

Electricity                           -0.0836              4.93          0.03

Soil Ferralsols                         1.196              0.45          0.50

Soil Luvisols                          0.8625              3.04          0.08

Soil Vertisols                        -0.4351              0.29          0.59

N=8113
Likelihood Ratio Test= 1640.70 (P value< 0.0001)




                                         25

Table 4: Marginal Effects on Livestock Adoption (%)


                                                  High       High
                                      High dry humid         moist      High
            Africa       Desert       savannah    forest     savannah   semi-arid
Livestock        78.47        78.41        79.35       77.55     79.21      79.39
T                 0.16         0.56         0.30        0.54      0.46       0.53
P               -0.063       -0.004       -0.017      -0.038    -0.004     -0.004


                                      Low         Low        Low        Low sub-
            High sub- Low       dry humid         moist      semi-arid  humid
            humid        savannah     forest      savannah
Livestock        79.57        79.69        77.87       77.10     80.21      79.26
T                 0.28         0.02         0.06        0.05      0.11       0.03
P                0.012       -0.070       -0.075      -0.119    -0.043     -0.055


            Mid       dry Mid      humid Mid      moist Mid    semi- Mid     sub-
            savannah        forest         savannah      arid         humid
Livestock          78.49            77.76         77.26        79.19        78.52
T                    0.42            0.19          0.40         0.37         0.26
P                 -0.051           -0.040        -0.084       -0.014       -0.034




                                          26

Table 5: Multinomial Logit Model of Livestock Species Choice


                           Beef cattle              Dairy Cattle
                           Est         Chi-sq       Est          Chi-sq
Intercept                     4.5920          1.35      24.0842       121.41
Summer Temperature            0.5554          3.32      -1.8232       128.46
Summer Temperature2          -0.0099          3.02       0.0323       113.83
Summer Precipitation          0.0148          0.81      -0.0555        39.12
Summer Precipitation2          0.0000         2.88       0.0001        33.27
Winter Temperature           -1.4005        88.81        0.2586         4.36
Winter Temperature2            0.0320       49.72       -0.0101         8.73
Winter Precipitation         -0.0395          5.19      -0.1186        94.31
Winter Precipitation2        -0.0001          0.75      -0.0002        23.16
Summer Temp * Prec           -0.0002          0.08       0.0013        15.52
Winter Temp * Prec             0.0030       10.16        0.0067       101.52
Flow                         -0.0480          5.23       0.0353         7.70
Head Farm                    -0.3760          2.60       0.0432         0.14
Electricity                  -0.4324        17.83        0.1552         4.14
Beef price                   -0.0020          4.80      -0.0008         1.58
Milk price                   -0.0009          0.91       0.0013         4.34
Goat price                   -0.0106          1.58      -0.0021         0.14
Sheep price                    0.0090         2.26       0.0063         2.61


Note: N=4379, Likelihood Ratio Test=4497.60 (P value<0.0001)




                                         27

Table 5: Continued.


                          Goats                    Sheep
Intercept                 Est         Chi-sq       Est         Chi-sq
Summer Temperature           -3.6110         1.78        4.2311      2.86
Summer Temperature2          -0.1826         1.13       -0.1708      1.03
Summer Precipitation          0.0053         3.11       0.0023       0.60
Summer Precipitation2          0.0058        0.30       -0.0009      0.01
Winter Temperature             0.0000        1.67        0.0000       0.28
Winter Temperature2            0.3197        3.14       -0.2345       2.55
Winter Precipitation         -0.0058         1.69        0.0097      6.70
Winter Precipitation2        -0.0207         2.23       -0.0212      1.48
Summer Temp * Prec             0.0001        4.41        0.0000       0.10
Winter Temp * Prec           -0.0003         0.67       -0.0002       0.26
Flow                           0.0006        0.82        0.0004       0.22
Head Farm                    -0.0107         0.20       -0.0322       2.27
Electricity                  -0.1011         0.70       -0.0084       0.01
Beef price                   -0.1184         2.72       -0.2674      14.98
Milk price                   -0.0001         0.00       -0.0018      5.40
Goat price                   -0.0014         2.54       -0.0037      15.69
Sheep price                    0.0022        0.10        0.0064      0.95


Note: N=4379, Likelihood Ratio Test=4497.60 (P value<0.0001).




                                       28

Table 6: Marginal Effects by AEZs (%)


                                              High      High
                                     High dry humid     moist     High
            Africa        Desert     savannah forest    savannah  semi-arid
 Beef              3.07        4.48       1.97      1.47      3.81      1.80
 T                -0.12       -1.46       0.04      0.31     -0.27      0.04
 P                 0.02        0.02       0.02      0.02      0.03      0.02
 Dairy            24.11       59.99      21.47     48.78    38.73     30.27
 T                -2.08       -1.27      -5.80     -6.92     -7.57     -8.07
 P                -0.15        -0.11     -0.15     -0.46     -0.30     -0.24
 Goats            17.16        4.29      13.14      9.98      9.48      9.54
 T                 1.13        0.83       1.72      1.88      1.93      1.91
 P                 0.13        0.01       0.11      0.20      0.12      0.06
 Sheep            22.46        6.59      28.51     11.17    17.96     27.05
 T                 1.92        0.37       2.88      1.95      2.74      3.66
 P                -0.13       -0.01      -0.13     -0.07     -0.07     -0.02
 Chickens         33.20       24.64      34.91     28.61    30.02     31.33
 T                -0.84        1.53       1.15      2.78      3.17      2.46
 P                 0.12        0.10       0.15      0.31      0.22      0.18




                                        29

Table 6 continued.


                                Low
            High sub- Low   dry humid     Low moist Low semi- Low sub-
            humid     savannah  forest    Savannah  arid      humid
Beef              3.95     2.06       2.48     2.78      1.60      3.16
T                -0.41     0.08       0.40     0.25     -0.03      0.48
P                 0.02     0.02       0.02     0.03      0.01      0.03
Dairy           43.76     13.20      24.45    13.04     24.37     16.50
T                -8.55    -1.02      -2.52    -0.69     -1.17     -2.34
P                -0.26    -0.05      -0.18    -0.11     -0.06     -0.05
Goats           10.19     21.47      18.28    22.16     18.13     18.70
T                 2.30     0.54       1.40     1.18      0.25      1.28
P                 0.17     0.07       0.25     0.17      0.04      0.17
Sheep           14.23     34.38      16.09    25.29     30.51     22.86
T                 2.58     2.78       1.36     1.98      2.50      1.99
P                -0.08     -0.11     -0.18    -0.19     -0.09     -0.24
Chicken         27.86     28.89      38.70    36.73     25.39     38.79
T                 4.09    -2.38      -0.65    -2.72     -1.54     -1.41
P                 0.15     0.07       0.09     0.10      0.10      0.09




                                     30

Table 6 continued

              Mid     dry Mid humid Mid     moist Mid  semi- Mid   sub-
              savannah    forest      savannah    arid       humid
Beef cattle          3.75        1.67        5.31       7.75       2.86
T                   -0.19        0.31       -0.32      -1.65       0.06
P                    0.04        0.02        0.04       0.03       0.02
Dairy cattle        27.80       42.63       30.61      22.41      39.20
T                   -4.17       -5.29       -3.45      -5.29      -6.37
P                   -0.26       -0.43       -0.30      -0.17      -0.24
Goats               10.73       13.00       12.18      10.84      13.71
T                    1.58        1.82        1.66       1.67       1.97
P                    0.11        0.24        0.19       0.07       0.23
Sheep               18.23       10.71       12.00      26.56      14.03
T                    2.12        1.35        1.34       3.13       2.15
P                   -0.07       -0.11       -0.10      -0.07      -0.13
Chickens            39.48       31.98       39.89      32.45      30.21
T                    0.67        1.80        0.77       2.14       2.19
P                    0.19        0.28        0.17       0.14       0.13




                                     31

Table 7: AOGCM Scenarios


                              Current    2020     2100

Summer Temperature (°C )
CCC                             25.7     +1.4     +6.0
PCM                             25.7     +0.7     +2.2
Winter Temperature (°C )
CCC                             22.4     +2.2     +7.3
PCM                             22.4     +1.1     +3.1
Summer Rainfall (mm/month)
CCC                            149.8      -4.6    -33.7
PCM                            149.8      -4.7     -4.7
Winter Rainfall (mm/month)
CCC                             12.8     +1.1     +3.5
PCM                             12.8     +18.8    +21.6




                              32

Table 8a: Climate Change Impacts by AEZs by 2020 (%)


                                               High       High
                                    High dry humid        moist     High
            Africa     Desert       savannah   forest     savannah  semi-arid
Livestock        78.47      78.41        79.35       77.55     79.21     79.39

CCC                0.98      0.73         0.99        1.45      1.12      1.20

PCM               -4.13      0.01        -2.64     -24.68     -10.16     -4.32

Beef

cattle             3.92      4.48         1.97        1.47      3.81      1.80

CCC               -0.72     -2.32       +0.12       +0.46      -0.55     +0.06

PCM              +1.24      -0.60       +6.18        -0.03     +2.38     +6.02

Dairy

cattle           24.11      59.99        21.47       48.78     38.73     30.27

CCC               -1.90     +1.24        -6.99       -6.46     -8.18     -9.67

PCM               -5.37     +2.15        -2.86     -31.61     -21.88     -9.76

Goats

                 17.16       4.29        13.14        9.98      9.48      9.54

CCC               -0.04     +0.79       +1.45       +1.28      +1.82     +2.25

PCM              +6.58      +1.48       +2.43      +42.57     +17.37     +5.00

Sheep

                 22.46       6.59        28.51       11.17     17.96     27.05

CCC              +5.76      +0.77       +5.20       +2.57      +4.14     +5.20

PCM               -0.84     -1.65       -15.67       -7.53    -10.41    -16.47

Chickens

                 33.20      24.64        34.91       28.61     30.02     31.33

CCC               -3.86     -0.48       +0.23       +2.14      +2.77     +2.16

PCM               -2.13     -1.37       +9.92        -3.40   +12.54    +15.20




                                        33

Table 8a continued.


                                Low
           High sub- Low    dry humid      Low moist Low semi- Low sub-
           humid      savannah  forest     savannah  arid      humid
Livestock        79.57     79.69      77.87     77.10    80.21     79.26

CCC               0.62      1.38      -0.60      1.85     1.13      0.12

PCM             -27.33     -1.06      -7.44     -3.87    -7.76     -9.09

Beef

cattle            3.95      2.06       2.48      2.78     1.60      3.16

CCC              -0.55    +0.34      +1.18       0.44     0.46      1.72

PCM              +0.80    +2.47      +1.16       1.96     5.57      2.55

Dairy

cattle           43.76     13.20      24.45     13.04    24.37     16.50

CCC             -10.64     -2.13      -2.61     -0.11    -1.39     -2.91

PCM             -33.26     -0.91      -7.41     -2.93    -0.93     -2.35

Goats

                 10.19     21.47      18.28     22.16    18.13     18.70

CCC              +2.19     -2.21     +3.11      -0.89    -2.56      1.34

PCM             +41.12    +0.54     +13.16       5.71     9.81     11.80

Sheep

                 14.23     34.38      16.09     25.29    30.51     22.86

CCC              +3.52    +9.95      +1.78       7.53     8.18      3.99

PCM             -10.51    +2.41       -4.63      0.80    -4.46     -5.82

Chickens

                 27.86     28.89      38.70     36.73    25.39     38.79

CCC              +5.47     -5.94      -3.46     -6.97    -4.68     -4.15

PCM              +1.85     -4.50      -2.28     -5.54    -9.99     -6.18




                                     34

Table 8a continued.


           Mid       dry Mid   humid Mid   moist Mid  semi- Mid   sub-
           savannah      forest      savannah    arid       humid
Livestock          78.49        77.76       77.26      79.19      78.52

CCC                  1.62        0.67        1.94       0.99       0.73

PCM                 -8.42      -20.51       -8.13      -3.29     -28.92

Beef

cattle               3.75        1.67        5.31       7.75       2.86

CCC                 -0.61        0.54       -1.11      -2.55       0.25

PCM                  6.10       -0.11        3.84       5.86       2.68

Dairy

cattle             27.80        42.63       30.61      22.41      39.20

CCC                 -3.02       -4.21       -1.21      -6.08      -7.16

PCM                 -7.20      -26.76      -11.70      -1.90     -28.18

Goats

                   10.73        13.00       12.18      10.84      13.71

CCC                  1.30        1.30        0.61       1.78       1.36

PCM                10.56        38.58       13.03       3.06      42.25

Sheep

                   18.23        10.71       12.00      26.56      14.03

CCC                  3.50        2.01        2.99       4.97       3.25

PCM                -11.27       -6.31       -5.34     -15.55     -10.29

Chickens

                   39.48        31.98       39.89      32.45      30.21

CCC                 -1.18        0.37       -1.28       1.88       2.30

PCM                 1.80        -5.41        0.16       8.53      -6.46




                                     35

Table 8b: Climate Change Impacts by AEZs by 2100


                                              High       High
                                    High  dry humid      moist     High
           Africa      Desert       savannah  forest     savannah  semi-arid
Livestoc

k               78.47       78.41        79.35      77.55     79.21     79.39

CCC               2.99       2.96         4.31       3.97      3.98      4.47

PCM              -3.51       1.52        -0.15     -21.29     -6.63     -1.02

Beef

cattle            3.92       4.48         1.97       1.47      3.81      1.80

CCC               1.67      -3.89         0.16       2.57     -0.76     -0.08

PCM               2.39      -1.80         6.68       1.95      3.82      5.90

Dairy

cattle           24.11      59.99        21.47      48.78     38.73     30.27

CCC              -5.76      -3.24       -16.93     -23.58    -25.36    -23.51

PCM              -5.57      -0.24         2.32     -33.95    -19.46     -0.31

Goats

                17.16        4.29        13.14       9.98      9.48      9.54

CCC              -5.05       5.15         0.79       2.65      3.42      4.62

PCM               5.85       2.63         1.93      44.07     16.50      3.70

Sheep

                22.46        6.59        28.51      11.17     17.96     27.05

CCC              22.10       3.00        27.99      17.34     23.80     25.71

PCM               3.11      -0.75       -12.26      -8.43     -8.82    -14.17

Chickens

                33.20       24.64        34.91      28.61     30.02     31.33

CCC             -15.50      -1.04       -12.02       1.02     -1.10     -6.74

PCM              -6.90       0.16         1.32      -3.63      7.95      4.87




                                        36

Table 8b continued.


                                Low
           High sub- Low    dry humid      Low moist Low semi- Low sub-
           humid      savannah  forest     savannah  arid      humid
Livestock        79.57     79.69      77.87     77.10     80.21    79.26

CCC               2.70      3.07       0.62      4.70      2.58     1.76

PCM             -20.98      0.76     -11.84     -2.12     -4.31   -10.24

Beef              3.95      2.06       2.48      2.78      1.60     3.16

CCC              -0.66      2.98      12.86      4.61      1.92    14.94

PCM               2.37      3.45       5.13      5.38      6.29     8.37

Dairy            43.76     13.20      24.45     13.04     24.37    16.50

CCC             -28.81     -4.08      -0.01     -3.03     -2.50    -3.06

PCM             -29.68     -0.42      -8.63     -3.06      2.15    -0.48

Goats            10.19     21.47      18.28     22.16     18.13    18.70

CCC               3.68    -10.09      -4.96     -9.40     -8.78    -7.82

PCM              36.66     -3.58      17.19      2.83      2.75    10.77

Sheep            14.23     34.38      16.09     25.29     30.51    22.86

CCC              21.97     29.28      12.20     32.37    23.13     19.82

PCM             -10.06     10.72      -4.53      6.62      2.25    -5.22

Chickens         27.86     28.89      38.70     36.73     25.39    38.79

CCC               3.81    -18.10    -20.08     -24.55    -13.77   -23.88

PCM               0.71     -10.1      -9.16    -11.77    -13.44   -13.43




                                     37

Table 8b continued.


           Mid       dry Mid   humid Mid   moist Mid  semi- Mid   sub-
           savannah      forest      savannah    arid       humid
Livestock           78.49       77.76       77.26      79.19      78.52

CCC                  4.72        2.26        5.09       3.97       3.11

PCM                 -5.31      -19.53       -5.72      -0.66     -23.75

Beef                 3.75        1.67        5.31       7.75       2.86

CCC                 -0.16        4.25       -0.89      -5.47       2.41

PCM                  5.55        1.69        5.22       3.68       7.08

Dairy               27.80       42.63       30.61      22.41      39.20

CCC                -14.14      -16.61      -12.50     -14.50     -21.81

PCM                 -2.80      -30.35      -10.00       5.42     -25.12

Goats               10.73       13.00       12.18      10.84      13.71

CCC                  4.23        0.76        3.20       3.13       0.27

PCM                  9.71       42.04       11.10       2.10      36.32

Sheep               18.23       10.71       12.00      26.56      14.03

CCC                 20.46       14.97       19.14     24.44      22.27

PCM                -10.59       -7.42       -5.04     -12.85     -10.41

Chickens            39.48       31.98       39.89      32.45      30.21

CCC                -10.39       -3.37       -8.96      -7.61      -3.14

PCM                 -1.86       -5.95       -1.28       1.65      -7.87




                                     38

Fig 1: Agro-Ecological Zones of Africa

Figure 2: Estimated Probability to Own Livestock (Left), Change in Probability under CCC 2100 (Middle), and Change in Probability
under PCM 2100 (Right).




                                                               40

Figure 3: Estimated Probability to Choose Dairy Cattle (Left), Change in Probability under CCC 2100 (Middle), and Change in
Probability under PCM 2100 (Right).




                                                              41

Figure 4: Estimated Probability to Choose Sheep (Left), Change in Probability under CCC 2100 (Middle), and Change in Probability
under PCM 2100 (Right).




                                                               42

Appendix


Appendix 1: Estimated Probability to Choose Goats (Left), Change in Probability under CCC 2100 (Middle), and Change in
Probability under PCM 2100 (Right).




                                                          43

Appendix 2: Estimated Probability to Choose Chickens (Left), Change in Probability under CCC 2100 (Middle), and Change in
Probability under PCM 2100 (Right).




                                                            44