Preventive Veterinary Medicine 116 (2014) 26–36

Contents lists available at ScienceDirect

Preventive Veterinary Medicine

j ourna l ho me pa g e: www.elsev ier .com/ locate /prevetmed

Economic impact of foot and mouth disease outbreaks onsmallholder farmers in Ethiopia

W.T. Jemberua,b,∗, M.C.M. Mouritsa, T. Woldehannac, H. Hogeveena

a Business Economics Group, Wageningen University, Hollandseweg 1, 6706 KN Wageningen, The Netherlandsb Department of Veterinary Epidemiology and Public Health, Faculty of Veterinary Medicine, University of Gondar, P.O. Box 196, Gondar,Ethiopiac Department of Economics, College of Business and Economics, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia

a r t i c l e i n f o

Article history:Received 21 December 2013Received in revised form 23 April 2014Accepted 9 June 2014

Keywords:CattleEconomicEthiopiaFoot and mouth diseaseMorbiditySmallholder farming system

a b s t r a c t

Foot and mouth disease is endemic in Ethiopia with occurrences of several outbreaks everyyear. Quantitative information about the impact of the disease on smallholder farming sys-tems in the country is, however, scarce. This study presents a quantitative assessment ofthe clinical and direct economic impacts of foot and mouth disease outbreaks on house-hold level in smallholder livestock farming systems. Impacts were assessed based on dataobtained from case outbreaks in cattle in crop–livestock mixed and pastoral smallholderfarming systems that occurred in 2012 and 2013. Data were collected by using question-naires administered to 512 smallholder farmers in six districts within two administratezones that represent the two smallholder farming systems. Foot and mouth disease morbid-ity rates of 85.2% and 94.9% at herd level; and 74.3% and 60.8% at animal level in the affectedherds were determined for crop–livestock mixed system and pastoral system, respectively.The overall and calf specific mortality rates were 2.4% and 9.7% for the crop–livestock mixedsystem, and 0.7% and 2.6% for the pastoral system, respectively. Herd level morbidity ratewas statistically significantly higher in the pastoral system than in the crop–livestock mixedsystem (P < 0.001). The economic losses of foot and mouth disease outbreak due to milkloss, draft power loss and mortality were on average USD 76 per affected herd and USD 9.8per head of cattle in the affected herds in crop–livestock mixed system; and USD 174 peraffected herd and USD 5.3 per head of cattle in the affected herds in the pastoral system.The herd level economic losses were statistically significantly higher for the pastoral sys-tem than for the crop–livestock mixed system (P < 0.001). The major loss due to the disease

occurred as a result of milk losses and draft power losses whereas mortality losses wererelatively low. Although the presented estimates on the economic losses accounted only forthe visible direct impacts of the disease on herd level, these conservative estimates signifya potential socioeconomic gain from a control intervention.

© 2014 Elsevier B.V. All rights reserved.

∗ Corresponding author at: Business Economics Group, WageningenUniversity, Hollandseweg 1, 6706 KN Wageningen, The Netherlands.Tel.: +31 317484065; fax: +31 317482745.

E-mail addresses: wudu.jemberu@wur.nl,wudu.temesgen@uog.edu.et (W.T. Jemberu).

http://dx.doi.org/10.1016/j.prevetmed.2014.06.0040167-5877/© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Livestock production has a significant role to the

Ethiopian economy by contributing up to 45% of the agri-cultural GDP, 19% of total GDP and one fifth of the country’sexport (Behnke, 2011). Within the livestock populationof the country, cattle represent about 71% of the total

Veterina

4(mthparf8mpt(tnktsp

des2de

ieelwom2blTdaato

msTsfiovietSaFrb

W.T. Jemberu et al. / Preventive

2.2 million tropical livestock units (livestock biomass)MoARD, 2007). Cattle production in Ethiopia occurs

ainly in subsistence oriented small holder farming sys-ems. Within these systems, cattle play an essential role atousehold level by providing milk and meat for food, draftower, manure for soil fertilizer and cooking, cash income,nd other economic and social functions like financial secu-ity and matrimonial dowry. Crop–livestock mixed (CLM)arming is the dominant farming system in which about0–85% of the national cattle population is kept. The pri-ary purpose of cattle in this system is to provide draft

ower for crop production. In Ethiopia 80% of the requiredraction power for crop agriculture is provided by oxenMoARD, 2007). Pastoral farming is the second most impor-ant farming system, which represents about 15–20% of theational cattle population. Cattle in this system are mainlyept for milk production for household consumption. Thehird is market oriented farming system, which mainly con-ists of dairy and fattening, and represents only a smallroportion of the national cattle population.

Foot and mouth disease (FMD) is one of the endemiciseases in Ethiopia that occurs recurrently, causing sev-ral outbreaks every year (Ayelet et al., 2012). Serologicalurveys reported a sero-prevalence that ranges from 5% to5% at the animal level and up to 60% at the herd level inifferent parts of the country (Rufael et al., 2008; Megersat al., 2009; Bayissa et al., 2011).

Foot and mouth disease is considered as the mostmportant livestock disease in the world in terms of itsconomic impact (James and Rushton, 2002). The annualconomic impact of FMD in terms of visible productionosses and vaccination costs in endemic regions of the

orld is estimated between US$6.5 and 21 billion, whileutbreaks in FMD free countries and zones cause losses ofore than US$1.5 billion a year (Knight-Jones and Rushton,

013). The economic impact of FMD in endemic areas cane separated into two components: direct and indirect

osses (Rushton, 2009; Knight-Jones and Rushton, 2013).he direct losses of the disease consist of loss of milk pro-uction, loss of draft power, retardation of growth, abortionnd delayed breeding, and mortality especially in youngnimals. The indirect losses are related to market restric-ions, use of suboptimal production technologies and costsf control.

Foot and mouth disease is commonly considered asild in indigenous animals in the traditional productions

ystems (James and Rushton, 2002; Vosloo et al., 2002;homson and Bastos, 2005), implying a limited economicignificance of the disease for smallholder subsistencearmers who keep indigenous animals and do not partic-pate in the international trade. However, the importancef the disease for the smallholder farmers has been contro-ersial. For example, Perry et al. (2003) claimed a pro-poormpact of foot and mouth disease control in the south-rn African region through a national economic growthhat would create a suitable base for poverty reduction.coones and Wolmer (2006), however, commented that

lthough this study clearly showed that the investment inMD control primarily benefits the commercial sector, theesearchers inappropriately concluded that the poor wouldenefit indirectly from the national economic growth. In

ry Medicine 116 (2014) 26–36 27

a subsequent paper, Perry and Rich (2007) indicated thatFMD has diverse impacts across different farming systemsand that its control would constitute a pro-poor investmentin many developing countries. Meanwhile, several caseoutbreak studies that were conducted in different partsof the world reported a significant impact of FMD in thesmallholder settings (Barasa et al., 2008; Rast et al., 2010;Shankar et al., 2012; Young et al., 2013).

Quantitative information on the impact of FMD is essen-tial in order to make sound decisions about its control.Although it can be conjectured that losses from FMD inEthiopia could be significant due to the multiple functionsof cattle in the smallholder farming systems, empirical esti-mates about such losses are scarce. This study presents anempirical case study of the clinical and direct economicimpact of FMD outbreaks on household level in the small-holder livestock farming systems in Ethiopia.

2. Materials and methods

2.1. Study areas

The study was conducted in the administrative zones ofNorth Wollo and Borena (Fig. 1). These zones were selectedfor their representation of the two main smallholder farm-ing systems in Ethiopia and their recent FMD outbreakexperience.

North Wollo is located in northeast Ethiopia. The agro-ecological landscape within North Wollo ranges from drierlowlands at about 1000 m above sea level, through fertilemidlands, to cold highlands as high as 3700 m above sealevel (Edge and Adal, 2000). The main livelihood withinthis zone is crop–livestock mixed (CLM) farming. Cattleare primarily kept on smallholdings where they providedraft power for crop production, manure for soil fertilityand fuel, family diet, and cash income. The study withinthe North Wollo zone was concentrated on the districts ofKobo, Guba Lafto and Habru as these three districts wererecently affected by an FMD outbreak that occurred fromJune to November 2012.

Borena zone is located in southern part of the country.Except for a central mountain range and scattered volcaniccones and craters, the zone’s landscape is gently undulat-ing across an elevation of 1000–1600 m (Coppock, 1994).The zone is dominated by a semi-arid climate and mostparts of the zone are less endowed with moisture for reli-able crop based livelihoods. The main livelihood in this aridand semi-arid region is pastoralism. Three representativedistricts namely Dugda Dawa, Yabello and Dire were usedfor the study. These pastoral districts were selected basedon their recent history of experiencing an FMD outbreakthat occurred from December 2012 to February 2013 andon their accessibility for investigation.

2.2. Sampling and data collection

The necessary sample size was determined for each

zone (production system) independently. Based on theexpectation of a high herd level morbidity rate (80%), a95% confidence level and a 5% desired absolute precision, asample size 246 small holder farmers was determined for

28 W.T. Jemberu et al. / Preventive Veterinary Medicine 116 (2014) 26–36

wing th

Fig. 1. Map of Ethiopia sho

each zone (Thrusfield, 2005). A few additional farmers wereadded to make up for incomplete information. Accordingly,257 farmers from North Wollo (CLM system) and 255 farm-ers from Borena (pastoral system), were enrolled for thestudy from FMD outbreak affected kebeles1 in the studydistricts. Affected kebeles were identified from the out-break records of district veterinary offices. Sampling wasdone at market places, veterinary clinics and communitygatherings in the study districts. Farmers were selected atsome haphazard intervals from these sites (markets, veteri-nary clinics and meetings) and were included in the samplewhen they were from outbreak affected kebeles.

Data were collected using a questionnaire that wasadministered by interview. Before the interview, an oralconsent was obtained from each participating farmer afterreading a written explanation on the purpose of the study,the risks and benefits of participation in the study, the rightto refuse to participate in the study as well as the conditionsof confidentiality regarding the presentation of answers.

The interview was done at the end of FMD outbreaksin the respective zone by veterinarians using the struc-tured questionnaire (Appendix I). The questionnaire was

designed primarily to record morbidity and mortality rates,and production losses in different categories of cattle onindividual household herd level (referred simply as herd

1 Kebeles are administrative sub-divisions of a district in Ethiopia.

e study zones and districts.

here after). Cattle were grouped into six different cate-gories based upon their production status. These categorieswere (1) lactating cows, (2) pregnant cows; cows that werepregnant and not giving milk, (3) dry cows; cows that wereneither pregnant nor giving milk, (4) oxen; all adult maleanimals that were used for plowing, (5) young stock; bothmale and female animals older than one year and notinseminated or used for plowing, and (6) calves; youngcattle up to one year of age.

2.3. Estimation of morbidity and mortality rates

In the interview, farmers were asked to describe themain epidemiological and clinical features of FMD thatthey observed in their animals during the outbreaks. Iftheir description was approximately consistent with thekey words used in the literature description of the disease,they were considered to know the disease and the inter-view was continued. Then for each category of cattle, thenumber of animals at risk, affected, and killed by FMD dur-ing the outbreak was recorded to determine morbidity andmortality rates at animal and herd levels. The animal levelmorbidity rate was determined as the number of animalsinfected during the outbreak divided by the total number

of animals at risk, and the herd level morbidity rate wasdetermined as the number of positive herds (herds withone or more infected animals) divided by the total num-ber of herds. Mortality rate was determined as the number

Veterina

obead

2

salftbbattc

2

tyahieh

L

wllmtcs

tpcketw(toet

2

p

c0

tion only one third of the lactation yield is assumed to be

W.T. Jemberu et al. / Preventive

f animals which died of FMD during the outbreak dividedy the total number of animals at risk. Calves that appar-ntly died due to a shortage of milk as a consequence ofn infection in the dam were taken into account in theetermination of the FMD calf mortality.

.4. Estimation of economic losses

The framework for economic impact assessment of live-tock disease provided by Rushton (2009) and Knight-Jonesnd Rushton (2013) as direct impacts (visible and invisibleosses) and indirect impacts (additional costs and revenuesoregone) was used in assessing the economic losses ofhe FMD outbreaks. In this study we considered the visi-le direct economic impact of FMD; and it was determinedy an estimation of the most visible production losses suchs milk loss, draft power loss and mortality loss. Becausehe variables used for estimation of losses were based onhe responses of the individual farmers, economic loss cal-ulations were done at herd level.

.4.1. Milk lossMilk loss due to acute FMD can arise from two situa-

ions: (1) when a lactating cow is affected and her milkield decreases or stops during the duration of the illnessnd (2) when a lactating cow dries off because of the loss ofer calf. The impact of the latter situation was considered

n the losses calculation due to calf mortality for expedi-ncy. Economic losses due to milk loss per FMD affectederd were calculated as

milkij = Ncowi ∗ Qi ∗ Tmilki ∗ Pmilkj

here Lmilkij represents the economic losses due to milkoss for herd i in farming system j; Ncowi the number ofactating cows affected in herd i; Qi the average quantity of

ilk lost in liters per affected cow per day in herd i; Tmilkihe average duration of illness in days of affected lactatingows in herd i, Pmilkj the price of milk per liter for farmingystem j.

It was difficult to determine the price of milk due tohe subsistence nature of the farming systems. Whereasastoralists sell raw milk that is extra to their householdonsumption at markets in nearby towns, such type of mar-ets were nonexistent in the CLM system. In the latter casextra milk was sold only in the form of butter. In the pas-oral system a seasonal average milk price of USD2 0.66/las determined from a survey of four small town markets

Finchewa, Surpha, Bekie, Dubluk) in the study districts. Inhe CLM system a proxy price of USD 0.49/l was used basedn the milk price obtained by the market oriented farm-rs in the surrounding urban centers (Woldya and Koboowns).

.4.2. Draft power lossThe losses associated with a reduction in available draft

ower were calculated only for the CLM system as draft

2 Prices were originally collected in Birr (Ethiopian currency) andonverted into USD using the study period’s average exchange rate of.0547 USD/Birr.

ry Medicine 116 (2014) 26–36 29

power was not commonly used in the pastoral system. Eco-nomic losses due to draft loss per herd equaled

Ldrafti = Noxeni ∗ (Tdrafti ∗ adj) ∗ Pdraft

where Ldrafti represents the economic loss due to draftpower loss for herd i; Noxeni the number oxen affected inherd i, Tdrafti the average duration of illness in days of anaffected ox in herd i, adj an adjustment factor, and Pdraftthe price of draft power rent of an ox per day.

Draft power for crop production (plowing and thresh-ing) is not needed throughout the year because ofseasonality in crop production. Moreover not all days inthe planting season are effective working days. Accordingto Goe (1987) draft oxen in smallholder farms of Ethiopiawork only for about 65 days a year. The probability that aday on which an ox is ill coincides with an effective workingday equals, therefore, to 65/365 (0.178). This ratio is usedas an adjustment factor (adj) to change the days of illnessto actual working days lost. There was no market for draftpower rent except in some peri-urban localities. In mostcases farmers whose oxen are unable to work borrow oxenfrom their relatives to work their land. The rent price of anox of USD 2.74/day as quoted in peri-urban areas was usedto estimate the costs related to the draft loss.

2.4.3. Mortality lossThe losses due to mortality equaled the market price of

the animal that died. Therefore, the economic loss due tomortality per herd was calculated as

Lmortij = (Nmortcalfi ∗ (Pcalf

+ (1/3 average milk yield per lactation ∗ Pmilkj)))

+ (Nmortyoungi ∗ Pyoung) +(Nmortadulti ∗ Padult)

where Lmortij represents the economic losses due to mor-tality for a herd i in farming system j; Nmortcalfi is thenumber of calves died in herd i; Pcalf is the price of acalf; Pmilkj is price of milk per liter in framing system j;Nmortyoungi is the number of young stock died in herd i;Pyoung is the price of a head of young stock, Nmortadulti isthe number of adult cattle died in herd i, and Padult is theprice of a head of adult cattle.

When a calf dies, the dam stops lactating for the rest ofthe lactation period. The economic loss due to a calf deathis, therefore, based on the value of the calf as well as thevalue of the milk lost due to early drying off the dam. Ascalf deaths did not always occur at the beginning of lacta-

lost due to calf death. The total lactation yield of local cowsas estimated from several studies in rural areas of Ethiopiais on average 448 l (Behnke, 2010). The average prices ofUSD 164.10 and USD 273.50 respectively per young andadult animal (cows and oxen), were derived from the farm-ers response in both farming systems. No price for calveswas quoted by the farmers and a price of USD 27.35 wasestimated by the authors.

Veterinary Medicine 116 (2014) 26–36

Table 1Cattle herd size and structure by farming system.

Farming system

CLMa Pastoral

Number of herds 257 255Herd size

Mean 7 36Percentiles (5th, 50th, 95th) 2, 6, 16 9, 28, 95

Cattle category (%)Lactating cow 18.87 22.22Pregnant cow 8.97 14.01Dry cow 7.04 10.22Oxen 27.02 11.28Young stock 20.17 20.27

30 W.T. Jemberu et al. / Preventive

2.4.4. Total economic lossesThe total economic losses per individual herd were

aggregated as the sum of all losses arising from milk loss,draft loss and mortality:

TELij = Milkij + Drafti + Mortij

where TELij represents the total economic losses for herd iin farming system j, Milkij the economic losses due to milkloss in herd i in farming system j, Drafti the economic lossesdue to draft loss for herd i; and Mortij the economic lossesdue to mortality in herd i in farming sytem j.

The average economic loss per individual head of cattlepresent in the affected herds was determined by dividingthe total economic losses in the herd by the total numberof cattle in the herd.

2.5. Statistical analysis

Statistical analyses were conducted to test the signifi-cance of differences in herd level FMD morbidity rates andeconomic losses between farming systems and betweendistricts within the farming systems. A Chi-square testwas used to evaluate the differences in herd level morbid-ity rates between farming systems and between districtswithin the farming systems; and Fisher’s exact test wasused for those comparisons where the expected counts ofcells in the contingency tables were less than five. An inde-pendent samples t-test was used to evaluate differences inherd level economic losses between farming systems; anda one-way ANOVA to test the differences between districtswithin the farming systems. A least significant differencetest was conducted for multiple comparisons of meanswhen they appeared significant in the one-way ANOVAtest. The data on the herd level economic losses were logtransformed to correct for the skewness observed in theoriginal data. All statistical analyses were performed usingIBM SPSS statistical package (IBM Corp. Released 2012. IBMSPSS Statistics for Windows, Version 21.0. Armonk, NY: IBMCorp).

3. Results

3.1. Cattle herd size and structure

A total of 512 smallholder farmers were surveyed in thesix districts that represent the crop livestock mixed (CLM)and pastoral farming systems. The herd size and structurefor both farming systems are presented in Table 1. Themean household cattle herd size in the pastoral system (36heads of cattle per herd) was higher than in the CLM system(7 heads of cattle per herd). Cattle herd sizes were variableand skewed to the right. Whereas oxen represented thehighest percentage of the herd in CLM system, lactatingcows and calves represented the highest percentage in thepastoral system.

3.2. Morbidity and mortality

Several outbreaks of FMD were reported in the studyareas during the period of 2012–2013. The outbreaks wereconfirmed and the causal virus identified as serotype O by

Calf 17.93 22.00

a CLM, crop–livestock mixed.

antigen detection ELISA at the national animal health diag-nostic and investigation center (Sebeta, Ethiopia). All thefarmers surveyed were able to describe the disease in termsof its key symptoms and epidemiologic features. The pas-toral districts were more widely affected by outbreaks interms of affected kebeles (78.2%) and herds (94.9%) thanthe districts in the CLM in which 49.0% of the kebeles and85.2% of the herds were affected (Table 2). However, animallevel morbidities and mortalities in the affected herds werehigher for the CLM system (Table 3). Herd level morbidityrates were significantly higher in the pastoral system thanin the CLM system (P < 0.001). Further analysis of differ-ences of herd level morbidity rates between districts withineach farming system showed no statistically significant dif-ferences (P > 0.05) (Table 2).

Based on farmers’ diagnosis, the animal level morbidityand mortality rates due to FMD outbreaks in the affectedherds were, respectively, 74.3% and 2.4% in the CLM systemand 60.8% and 0.7% in the pastoral system (Table 3). Mor-bidity and mortality rates in different categories of cattle inthe affected herds are presented in Table 4. High morbiditywas observed in lactating cows and oxen in the CLM systemwith the similar pattern in all districts. No specific patternof morbidity in the cattle categories was observed in thepastoral system. Adult mortality was generally low. Rela-tively higher mortality rates were observed in calves, viz.9.7% for CLM and 2.6% for pastoral system (Table 4). Animallevel morbidity rates in the affected herds range from 13%to 100% CLM system and from 7% to 100% in the pastoralsystem.

3.3. Economic losses

Economic losses from milk reduction, draft power loss,and mortality per affected individual animal are presentedin Tables 5–7, respectively. The mean daily milk loss percow was 1.8 liters (ranging 0–4 liters) and 1.8 liters (ranging0.5–3 liters) for the CLM and pastoral system, respectively.The losses were roughly the same in the two systems andwere on average more than 75% of the daily milk yield. The

mean duration of illness in cows that led to milk reductionwas 33.6 days (ranging 7–90 days) for the CLM system and23 days (ranging 7–35 days) for the pastoral system. Themean number of effective working days lost by an affected

W.T. Jemberu et al. / Preventive Veterinary Medicine 116 (2014) 26–36 31

Table 2Foot and mouth disease infection rates at kebele and herd level by farming system and district.

Farming system District No. of kebeles Kebeles affected (%) No. of herds surveyed Herds affected (%)*

CLMa

Kobo 40 82.5 80 88.8Guba Lafto 32 21.9 75 78.7Habru 30 33.3 102 87.3Overall 102 49.0 257 85.2**

Pastoral

Dugda Dawa 16 81.3 81 98.8Dire 16 100 80 93.8Yabello 23 60.9 94 92.6Overall 55 78.2 255 94.9***

a CLM, crop–livestock mixed.* �2(1) = 13.4, P < 0.001 for differences in herd level morbidity between farming systems.

** �2(2) = 3.68, P = 0.159 for differences in herd level morbidity between CLM districts.*** Fishers exact P = 0.122 for differences in herd level morbidity between pastoral districts.

Table 3Foot and mouth disease morbidity and mortality rates in affected herds by farming system and district.

Farming system District No. of herds No. of cattle Morbidity (%) Mortality (%)

CLMa

Kobo 71 713 78.4 3.2Guba Lafto 59 231 67.1 2.2Habru 89 662 71.8 1.5Overall 219 1606 74.3 2.4

Pastoral

Dugda Dawa 80 2882 61.0 0.4Dire 75 2994 62.2 0.6Yabello 87 2726 59.2 1.0Overall 242 8602 60.8 0.7

a CLM, crop–livestock mixed.

Table 4Foot and mouth disease morbidity and mortality rates in different categories of cattle in the affected herds by farming system.

Farming system Category of cattle No. of cattle Morbidity (%) Mortality (%)

CLMa

Lactating cow 303 83.8 0.3Pregnant cow 144 75.7 0.0Dry cow 113 70.8 0.0Oxen 434 85.7 0.7Young stock 324 71.0 1.9Calf 288 50.0 9.7

Pastoral

Lactating cow 1912 63.9 0.2Pregnant cow 1204 57.8 0.2Dry cow 879 68.2 0.3Oxen 970 63.5 0.3Young stock 1744 60.2 0.1Calf 1893 55.5 2.6

a CLM, crop–livestock mixed.

Table 5Economic losses due to milk loss per FMD affected lactating cow by farming system and district.

Farming system District Daily milk yield (l)Mean (range)

Daily milk loss dueto FMD (l)Mean (range)

Duration of FMDillness (days)Mean (range)

Quantity of milklost (l)Mean (range)

Economic losses(USD)Mean (range)

CLMa

Kobo 2.3 (1–5) 1.8 (1–4) 40 (7–90) 72 (7–360) 35 (3–176)Guba Lafto 2.5 (1–3) 1.6(0–2.3) 35 (14–60) 56 (0–138) 27 (0–68)Habru 2.3 (1–4) 1.8 (1–4) 25 (7–60) 45 (7–360) 22 (3–176)Overall 2.3 (1–4) 1.8 (0–4) 33 (7–90) 59.4 (0–360) 29 (0–176)

Pastoral

Dugda Dawa 2.1 (0.7–4) 1.6 (0.7–3) 23 (10–35) 37 (7–105) 24 (5–69)Dire 2.3 (1–4) 1.8 (1–3) 21 (7–30) 38 (7–90) 24 (5–59)Yabello 2.2 (0.5–4) 1.8 (0.5–3) 26 (10–30) 47 (5–90) 31(5–59)Overall 2.2 (0.5–4) 1.8 (0.5–3) 23 (7–35) 41 (3.5 –105) 26(5–69)

a CLM, crop–livestock mixed.

32 W.T. Jemberu et al. / Preventive Veterinary Medicine 116 (2014) 26–36

Table 6Economic losses due to draft power loss per affected ox in the crop–livestock mixed districts.

Districts Duration of FMD illness(days)Mean (range)

Effective working dayslosta

Mean (range)

Economic loss (USD)Mean (range)

Kobo 37 (14–84) 7 (2–15) 18 (7–41)Guba Lafto 28 (7–56) 5 (1–10) 13 (3–27)Habru 25 (7–56) 5 (1–10) 12 (3–27)Overall 30 (7–84) 5 (1–15) 15 (3–41)

a An adjustment factor of 0.178 was used to change duration of illness into effective working days lost.

Table 7Economic losses due to mortality per head of cattle by farming system and district.

Farming system District No. of adult cattledied

No. of youngstock died

No. of calvesdied

Average economic lossper animal died (USD)

CLMa

Kobo 4 6 13 148Guba Lafto 0 0 5 101Habru 0 0 10 101Overall 4 6 28 129

Pastoral

Dugda dawa 3 1 11 160Dire 1 0 18 134Yabello 6 0 20 160

1

Overall 10

a CLM, crop–livestock mixed.

ox was 5 days (ranging 1–15 days). The highest mean eco-nomic loss at the individual affected animal level was dueto mortality (Table 7) which was on average USD 129 peranimal died in the CLM and USD 151 per animal died inthe pastoral system, followed by the losses resulting frommilk reduction which was on an average USD 29 and USD26 per affected cow in the CLM and the pastoral systems,respectively (Table 5).

The mean total economic losses of FMD outbreak at herdlevel were USD 76 (5th–95th percentiles: USD 7–271) forthe CLM system and USD 174 (5th–95th percentiles: USD10–648) for the pastoral system. These herd level losseswere variable and highly skewed to the right in both farm-ing systems (Table 8). The mean total herd level economiclosses in the CLM system were statistically significantlylower than in the pastoral system (P < 0.001). Comparisonof mean total herd level losses between districts withineach farming system using a one-way ANOVA showed a sta-tistically significant difference between the CLM districts(P < 0.001) but not in the pastoral districts (P > 0.05). Posthoc comparison of districts in the CLM system using theleast significant difference technique revealed significantdifferences between all pairs of districts with the lossesbeing highest in Kobo and lowest in Guba Lafto districts.At herd level, the largest component of the economic losswas due to milk loss, and in CLM system it was followed bydraft loss. Mortality was the least contributor of herd levellosses in both systems (Table 8).

The average economic losses per animal present in anaffected herd were USD 9.8 in CLM system and USD 5.3 inthe pastoral system.

4. Discussion

The mean cattle herd sizes recorded in this study (7heads for crop–livestock mixed (CLM) and 36 heads for

49 151

pastoral) were higher than the national averages reportedpreviously for the respective systems (4 heads for CLMand 13 heads for pastoral) (Negassa and Jabbar, 2008). Oneexplanation for these differences could be the fact that inthe present study only those farmers who owned cattlewere included, whereas in the national estimates cattleherd sizes were averaged over all smallholders keepinglivestock, independent of whether they owned cattle or not.Negassa and Jabbar (2008) reported that only 83% farmersin the CLM system and 78% of the farmers in the pastoralsystem owned cattle.

Draft oxen constitute the highest proportion of cattlein the herds within the CLM system and their proportionincreases with decreasing herd size. This indicates the pri-ority given to oxen by farmers, emphasizing the importanceof cattle as source of draft power in this type of farmingsystem. This corresponds to the general objective of cattleraising in the CLM farming system where cattle are primar-ily kept for provision of draft power and other inputs forcrop production (Negassa et al., 2011). The highest propor-tion of cattle in the pastoral systems represents lactatingcows on which the pastoral households depend on for foodsubsistence.

A high morbidity of FMD was reported both at herdlevel (85% for CLM system and 95% for pastoral systems)and animal level (74% for CLM and 61% for pastoral). Theherd level morbidity rate was statistically higher in the pas-toral herds as compared to CLM herds but the differencesin morbidity rates between districts within each farmingsystem were not statistically significant. This indicates theimportance of the farming system in the epidemiology ofthe disease. The difference might be due to the high con-

tact rates among household herds in the pastoral farmingsystem that is related to their mobility.

Comparable animal level morbidity rates rangingbetween 60 and 100% were reported previously in the CLM

W.T. Jemberu et al. / Preventive Veterinary Medicine 116 (2014) 26–36 33

Table 8Economic losses of FMD per herd by farming system and district in USD indicated by mean and percentiles.

Farming system Districts Milk lossesMean (5th, 50th, 95th)

Draft lossesMean (5th, 50th, 95th)

Mortality lossesMean (5th, 50th, 95th)

Total economic losses*

Mean (5th, 50th, 95th)

CLMa

Kobo 56 (0, 44, 212) 36 (0, 27, 82) 48 (0, 0, 285) 140 (20, 89, 47)Guba Lafto 2 (0, 0, 5) 19 (0, 14, 55) 9 (0, 0, 101) 29 (0, 14, 128)Habru 23 (0, 15, 88) 21 (0, 14, 55) 11 (0, 0, 101) 55 (0, 34, 176)Overall 28 (0, 7, 123) 25 (0, 27, 55) 22 (0, 0, 164) 76 (7, 35, 271)**

Pastoral

Dugda dawa 128 (0, 89, 394) – 33 (0, 0, 273) 161 (0, 116, 613)Dire 136 (17, 87, 576) – 33 (0, 0, 256) 169 (17, 94, 749)Yabello 143 (0, 109, 426) – 48 (0, 0, 391) 190 (0, 119, 706)Overall 136 (9, 97, 415) – 39 (0, 0, 274) 174 (10, 110, 648)***

a CLM = crop–livestock mixed.* t(459) = −7.05, P < 0.001 for differences in total economic losses between farming systems.

ween CLeen pas

sRsathpopmooioterpt

iowtTndorstrdptoreysttlf

** F(2, 216) = 35.85, P < 0.001 for differences in total economic losses bet*** F(2, 239) = 0.57, P = 0.564 for differences in total economic losses betw

ystem in eastern parts of Ethiopia (Mersie et al., 1992;oeder et al., 1994). The mortality rates of 2.4% in the CLMystem and 0.7% in the pastoral system are generally lownd in line with common knowledge that FMD is charac-erized by low mortality in adult animals (OIE, 2009). Aigher morbidity was seen in oxen and lactating cows, theroducing members of the herds, than the other categoriesf cattle in the CLM system. This could be associated withroduction stress and/or high contact with humans duringilking and plowing, which might facilitate transmission

f infection among these groups of animals. During FMDutbreak oxen from non-infected herds often moved tonfected areas to assist in plowing lands of farmers whosexen are affected. This may result in a wider dissemina-ion of an outbreak. This indicates the importance of specialmphasis on oxen in control efforts like vaccination toeduce both farmers’ loss of draft power at critical time oflanting, and the disease spread in the cattle population inhe CLM farming system.

Assessment of economic losses due to the outbreaksndicated that the largest losses at the animal levelbviously occurs when the animal dies. Mortality lossesere on average USD 129 per animal died in the CLM sys-

em and USD 151 per animal died in the pastoral system.he differences are merely a reflection of the difference inumber of mortality cases per category of animals in whicheath of more adult animals means higher losses. The sec-nd largest losses at animal level were due to milk loss,epresenting USD 29 per affected lactating cow for the CLMystem and USD 26 per affected lactating cow for the pas-oral system. These losses vary wildly as indicated by theanges in Tables 5 and 6. For example the economic lossesue to milk loss in the CLM varied from nil to USD 176er affected lactating cow depending on the yield level ofhe cow, the severity of milk reduction and the durationf the illness. In this study it was assumed that milk yieldeturned to pre-clinical FMD levels after the clinical recov-ry of the cows. There is, however, an indication that milkield may not return to normal after clinical recovery. Atudy in dairy farms in Pakistan reported a milk yield reduc-

ion of up to 30% two months after infection compared tohe pre-clinical FMD yield (Ferrari et al., 2013). Neverthe-ess, as management and physiological (e.g. lactation stage)actors were not controlled, it is difficult to account the

M districts.toral districts.

remaining yield reduction primarily as an effect of the dis-ease. We, therefore, chose to remain cautionary with ourcalculations and assumed a return of the milk productionto pre-clinical levels so that we did not overestimate thelosses due to FMD.

At herd level the most important losses arise from milkloss followed by draft loss in case of the CLM system. Mor-tality losses are relatively low. The importance of milk anddraft losses in the economic impact estimate is consis-tent with the primary objectives of keeping cattle in thesmallholder subsistence systems, which are aimed at theproduction of subsistence milk and draft power for thepastoral and CLM system, respectively. The mean total eco-nomic losses per herd of USD 76 for the CLM system wasstatistically significantly lower than the average mean lossof USD 174 for the pastoral system. The higher losses perherd in the pastoral system could mainly be explained byits larger herd sizes. The herd level losses were variableand highly skewed to the right in both farming systems.This was the result of high losses in a few herds which hadmortality cases, and the skewed distribution of the herdsizes.

The quantitative assessment of the direct economicimpact of FMD in this study did not include all conceivableimpacts of the disease. Only those losses that are of cru-cial importance to the farming systems under study andthat were relatively easy to quantify were considered. Assuch the economic losses estimated should be seen as aconservative estimate of the potential losses. Besides theevaluated losses, loss of condition could be an importantFMD impact as well. Some studies in similar smallholdersettings in other countries used the loss of condition as themain basis for the estimation of the economic impact ofthe disease (Rast et al., 2010; Young et al., 2013). Subsis-tence smallholder farmers in Ethiopia have, however, lowmarket off-take rates and low market participation level(Negassa and Jabbar, 2008). They sell animals only at timesof an acute cash shortage like during a bad crop harvest anddrought. So the impact of FMD in terms of condition losswill only be significant at such selling occasions. In normal

times, condition loss will be regained slowly after recov-ery and its impact will not be that high. The impacts of thedisease related to poor growth and reproduction perfor-mance occur in very subtle way and may halve long term

Veterina

34 W.T. Jemberu et al. / Preventive

effects in the herd dynamics and, therefore, need not tobe overlooked in the decision making about the controlof the disease. In addition to the production impacts, wehave attempted to include costs related to the treatment ofsick animals and the disposal of dead animals during theoutbreaks. Expenditures on modern treatment of sick ani-mals were, however, made only by a few farmers in thepastoral system. Most farmers used traditional practices totreat sick animals which were difficult to translate into amonetary value. Dead animals were disposed in the openfield for scavengers; farmers, therefore, did not experienceany disposal costs.

Even for the visible direct FMD impacts quantified inthis study, it was difficult and probably improper to con-vert the impacts into financial loss. As the farming systemsare not market oriented, the impact on the goods and ser-vices of cattle affected by the disease have in most casesno direct cash effect. The milk produced in these systems isused for own family consumption. The actual loss of milk is,therefore, not reflected in terms of cash income alone butalso has negative effect in the food security and nutritionof the family, especially in essential micronutrients in milkdiet that are crucial for children health and growth. Simi-larly the draft power loss is reflected as a reduction in theproductivity of crop production rather than an immediatefinancial loss. The loss estimated here can be, therefore,considered as a proxy quantitative measure that is usedto uncover the disease’s otherwise covert impacts on thefarmers’ welfare.

The estimated losses due to the disease outbreak rep-resent a significant burden for the affected households. Forexample, the losses of USD 76 per household in the CLM sys-tem constitutes about 6.8% of the total annual expenditure(a proxy of income) of an average household in this sys-tem which is estimated to be about USD 1122 (CSA, 2012).Taking a loss of more than 10% of income as a bench markfor a catastrophic shock (Shankar et al., 2012), the 6.8% lossof income for CLM households is a significant burden. Thispercentage obviously would be much higher and the con-sequences more sever for pastoral households because oftheir total dependency on livestock for their livelihood.

To the best knowledge of the authors no quantita-tive assessment of production losses has been previouslyreported for smallholder farmers in Ethiopia for compar-ison. Similar studies in countries of South and SoutheastAsia where cattle and buffalo are extensively used in small-holder rice-livestock production have reported a muchhigher loss due to FMD (Rast et al., 2010; Shankar et al.,2012; Young et al., 2013). However these economic lossestimates vary widely among the reports. For example aloss of USD 52.4 per affected head of cattle was reportedform Laos (Rast et al., 2010) while this reached USD216.32–370.54 per affected cattle in Southern Cambodia(Young et al., 2013). Even in the same country Cambo-dia, widely variable estimates are reported like USD 45per affected household by Shankar et al. (2012) and USD216.32–370.54 per affected cattle by Young et al. (2013).

These differences, in addition to the intrinsic variationsof outbreaks used for the case studies that affect someof the parameters (e.g. duration illness, level of produc-tion loss etc.), are also results of differences in the range

ry Medicine 116 (2014) 26–36

and type of losses and/or costs quantified, as well as theprices. This wide variation in estimation of FMD losses maybe a problem in evaluating the cost benefit ratio of con-trol interventions and calls for a standard methodologicalframework for estimating the impact in smallholder farm-ing system in developing countries where cattle are usedfor multiple functions.

This study excluded the indirect effects of the diseaselike its effect on international trade, costs of surveillanceand the disease’s impact on other species like small rumi-nants, all of which are important in Ethiopia. These factors,added to more subtle direct effects that were not quantified(like long term effect on herd structure), make the loss ofthe disease much more substantial than the one estimatedin this study. This demonstrates a potential economic gainfrom a control intervention against the disease in theselivestock systems. For example, the estimated losses ofUSD10 for the CLM system, and USD 5 in the pastoral sys-tem per animal indicate that a biannual vaccination witha cost of two USD per vaccination could be economicallyprofitable, if the vaccine is fully effective and the risk ofFMD is at least once every 5 years for the CLM system and2.5 years for pastoral system. It must be noted, however,that the impact of the disease, as mentioned elsewherein this report, did not occur directly as financial loss andits main impact would be on the social wellbeing of thefarmers in terms of food security and nutrition. In somesituations, the disease’s impact even goes beyond the eco-nomic domain. For example, in Borena pastoral area it wasreported that cattle affected by FMD cannot be used fordowry which is one the basic functions of cattle in thatarea (Jibat et al., 2013). Generally the decision related tointervention against the disease is complex and needs con-sideration of all facets of the disease’s impact.

5. Conclusion

Foot and mouth disease outbreak in the cattle popu-lation of smallholder farmers causes high morbidity andassociated economic losses that represent significant partof smallholders’ income. The losses are variable amonghouseholds and are higher for pastoral households thanfor households within the crop–livestock mixed system.Although the presented estimates on the economic lossesaccounted for only the visible direct impacts of the diseaseon herd level, these conservative results already signifya potential socioeconomic gain from a control interven-tion given the current frequency of FMD outbreak in thecountry. Formal and comprehensive evaluation of all costsand benefits of the disease and potential control optionsis, however, required before a recommendation for controlintervention.

Acknowledgements

The authors would like to thank Nuffic (Netherlandsorganization for international cooperation in higher edu-

cation) for funding the study. We would like also to extendour acknowledgments to the national animal health diag-nostic investigation center in Sebeta, Ethiopia and theanimal health personnel in the study districts for their

Veterina

ap

A

1

1

R

A

W.T. Jemberu et al. / Preventive

ssistance during data collection, and to the farmers whoarticipated and provided information for the study.

ppendix I.

FMD outbreak impact assessment questionnaire(Translated from Amharic to English)

1. Name of the herd owner (optional)2. Herd location:

villagekebeledistrict

3. Have your ever heard of FMD? Yes/NoIf yes, can you describe its clinical signs (Continue

with questionnaire if one or more of the following signsare described)

i. Lameness due to foot lesions and salivation due tomouth lesions in cattle and/or small ruminants.

ii. Foot lesions in cattle and/or small ruminants that iscontagious.

iii. Mouth lesions in cattle and/or small ruminants thatis contagious.

iv. Lameness or mouth lesions, and blisters (sores) onteats of cows

4. Did the recent FMD outbreak affect your herd? Yes/No.5. If yes, can you indicate the period at which the outbreak

occurred?startedended

6. Morbidity and associated consequences of the FMDoutbreak per category of cattle within the herd.

Category Total no.present

No.affectedby FMD

Averageduration(weeks)

No.aborted

No.died

Averagetreatmentcost (Birr)

Lactatingcows

Pregnantcows

Dry cowsDraft oxenYoung

stocksCalves

7. Average daily milk production before the onset of out-break ---------- (liter/day)

8. Average daily milk yield production during the FMDillness period ----- (liter/day)

9. How long did the affected lactating cows produce less?-----------

0. How long did the affected oxen stay out of work duringthe illness -------

1. What did you do with animals that died of FMD? Meatconsumed or thrown away? If thrown away, what werethe costs for disposal? --------

eferences

yelet, G., Gelaye, E., Negussie, H., Asmare, K., 2012. Study on the epi-demiology of foot and mouth disease in Ethiopia. Rev. Sci. Tech. 31,789–798.

ry Medicine 116 (2014) 26–36 35

Barasa, M., Catley, A., Machuchu, D., Laqua, H., Puot, E., Tap Kot, D.,Ikiror, D., 2008. Foot-and-mouth disease vaccination in South Sudan:Benefit–cost analysis and livelihoods impact. Transbound. Emerg. Dis.55, 339–351.

Bayissa, B., Ayelet, G., Kyule, M., Jibril, Y., Gelaye, E., 2011. Studyon seroprevalence, risk factors, and economic impact of foot-and-mouth disease in Borena pastoral and agro-pastoral system, southernEthiopia. Trop. Anim. Health Prod. 43, 759–766.

Behnke, R., 2010. The Contribution of Livestock to the Economies of IGADMember States. Study Findings, Application of the Methodology inEthiopia and Recommendations for Further Work. IGAD Livestock Pol-icy Initiative, pp. 12, Working Paper No. 02-10.

Behnke, R., 2011. The Contribution of Livestock to the Ethiopian Economy– Part II., pp. 7–8, IGAD LPI Working Paper No. 02-11.

Coppock, D.L. (Ed.), 1994. The Borana Plateau of Southern Ethiopia: Syn-thesis of Pastoral Research, Development and Change, 1980–91. ILCA(International Livestock Centre for Africa), Addis Ababa, Ethiopia, p.39.

CSA (Central Statistical Agency), 2012. The Federal Democratic Republicof Ethiopia Central Statistical Agency. Household Consumption andExpenditure (HCE) Survey 2010/11. Analytical report. Addis Ababa,Ethiopia., pp. 47–69.

Edge, S., Adal, Y., 2000. Farming Assets in North Wollo. Statistics, Maps andImpression from Travel to North Wollo. Department of Social Anthro-pology, Norwegian University of Science and Technology, Trondhiem,Norway, pp. 1.

Ferrari, G., Tasciotti, L., Khan, E., Kiani, A., 2013. Foot-and-mouthdisease and its effect on milk yield: an economic analysis on live-stock holders in Pakistan. Transbound. Emerg. Dis., http://dx.doi.org/10.1111/tbed.12072.

Goe, M.R., (Ph.D. dissertation) 1987. Animal Traction on SmallholderFarms in the Ethiopian Highlands. Department of Animal Science,Cornell University, Ithaca, NY, USA.

James, A., Rushton, J., 2002. The economics of foot and mouth disease. Rev.Sci. Tech. 21, 637–644.

Jibat, T., Admassu, B., Rufael, T., Baumann, M.P.O., Pötzsch, C.J., 2013.Impacts of foot-and-mouth disease on livelihoods in the BorenaPlateau of Ethiopia. Pastoralism: Res. Policy Pract. 3, 1–11.

Knight-Jones, T.D.J., Rushton, J., 2013. The economic impacts of foot andmouth disease – what are they, how big are they and where do theyoccur? Prev. Vet. Med. 112, 161–173.

Megersa, B., Beyene, B., Abunna, F., Regassa, A., Amenu, K., Rufael, T., 2009.Risk factors for foot and mouth disease seroprevalence in indigenouscattle in Southern Ethiopia: the effect of production system. Trop.Anim. Health Prod. 41, 891–898.

Mersie, A., Tafesse, B., Getahun, F., Teklu, W., 1992. Losses from foot-and-mouth disease in a mixed farming area of eastern Ethiopia. Trop. Anim.Health Prod. 24, 144.

MoARD, 2007. Livestock Development Master Plan Study. Phase I Report –Data Collection and Analysis. Volume B – Meat Production November2007. Ministry of Agriculture and Rural Development, Government ofEthiopia.

Negassa, A., Jabbar, M., 2008. Livestock Ownership, Commercial Off-TakeRates and Their Determinants in Ethiopia. Research Report 9.ILRI (International Livestock Research Institute), Nairobi, Kenyahttp://cgspace.cgiar.org/bitstream/handle/10568/237/ThemeIIIRReport%209.pdf?sequence=1

Negassa, A., Rashid, S., Gebremedhin, B., 2011. Livestock Production andMarketing. Development Strategy and Governance Division, Inter-national Food Policy Research Institute – Ethiopia Strategy SupportProgram II, Ethiopia, ESSP II Working Paper 26. http://www.ifpri.org/sites/default/files/publications/esspwp26.pdf

OIE, 2009. Foot and Mouth Disease: Aetiology Epidemiology DiagnosisPrevention and Control References. http://www.oie.int/fileadmin/Home/eng/Animal Health in the World/docs/pdf/FOOT ANDMOUTH DISEASE FINAL.pdf

Perry, B.D., Rich, K.H., 2007. Poverty impacts of foot-and-mouth diseaseand the poverty reduction implications of its control. Vet. Rec. 160,238–241.

Perry, B.D., Randolph, T.F., Ashley, S., Chimedza, R., Forman, T., Morri-son, J., Poulton, C., Sibanda, L., Stevens, C., Tebele, N., Yngstrom, I.,2003. The Impact and Poverty Reduction Implications of Foot andMouth Disease Control in Southern Africa, with Special Reference toZimbabwe. International Livestock Research Institute (ILRI), Nairobi,

Kenya.

Rast, L., Windsor, P.A., Khounsy, S., 2010. Limiting the impact offoot and mouth disease in Northern Lao People’s DemocraticRepublic by Vaccination: a case study. Transbound. Emerg. Dis. 57,147–153.

Veterina

437–449.Young, J.R., Suon, S., Andrews, C.J., Henry, L.A., Windsor, P.A., 2013. Assess-

36 W.T. Jemberu et al. / Preventive

Roeder, P.L., Abraham, G., Mebratu, G.Y., Kitching, R.P., 1994. Foot-and-mouth disease in Ethiopia from 1988 to 1991. Trop. Anim. Health Prod.26, 163–167.

Rufael, T., Catley, A., Bogale, A., Sahle, M., Shiferaw, A., 2008. Foot andmouth disease in the Borana pastoral system in southern Ethiopiaand implications for livelihoods and international trade. Trop. Anim.Health Prod. 40, 29–38.

Rushton, J., 2009. The Economics of Animal Health and Production. CABI,Nosworthy Way, Wallingford, Oxfordshire, UK, pp. 193–200.

Scoones, I., Wolmer, W., 2006. Livestock, Disease, Trade and Markets: Pol-

icy Choices for the Livestock Sector in Africa. IDS Working Paper 269.Institute for Development Studies, Brighton.

Shankar, B., Morzaria, S., Fiorucci, A., Hak, M., 2012. Animal disease andlivestock-keeper livelihoods in Southern Cambodia. Int. Soc. Dev. Rev.34, 39–63.

ry Medicine 116 (2014) 26–36

Thomson, G.R., Bastos, A.D.S., 2005. Foot-and-mouth disease. In: Coet-zer, J.A.W., Tustin, R.C. (Eds.), Infectious Diseases of Livestock. OxfordUniversity Press, New York.

Thrusfield, M., 2005. Veterinary Epidemiology, third edition. BlackwellScience, Oxford, UK, pp. 228–242.

Vosloo, W., Bastos, A.D.S., Sangare, O., Hargreaves, S.K., Thom-son, G.R., 2002. Review of the status and control of footand mouth disease in sub-Saharan Africa. Rev. Sci. Tech. 21,

ment of financial impact of foot and mouth disease on smallholdercattle farmers in Southern Cambodia. Transbound. Emerg. Dis. 60,166–174.