analysis of household food security in selected districts of madagascar (september 2009)

8/8/2019 Analysis of Household Food Security in Selected Districts of Madagascar (September 2009)

1/21

1

Analysis of Household Food

Security in Selected Districts of

Madagascar

UN World Food Programme

Antananarivo, Madagascar

September 2009


2/21

2

Background and Overview

In June 2009, a household survey was conducted in 33 districts in 11 regions around the country inorder to measure the impact of several different shocks on household food security and to providekey household level information into a joint FAO/WFP crop and food security assessment. The focuswas to measure the impact (if any) of the political situation as well as the drought in the south thathad affected crop production.

This report contains additional analysis of household data from the June 2009 survey and focuses onidentifying areas which were food secure as well as those that were chronically food insecure andthose that were newly food insecure due to the poor performance of the agricultural season.

The report will first provide some descriptive analysis by region and district and then will outline theanalytical approach for the food security classification analysis, followed by the description of the foodsecurity groups and lastly their location within the country, by sampled district.


3/21

3

10% 9%11%

15% 14%

11% 11%

20%18%

26%

10%

3%

7%5%

3%4%

1%

4%

8% 7%

4%3%

0%

5%

10%

15%

20%

25%

30%

Alao

traMango

ro

Anala

njirofo

Androy

Anos

y

Atsim

oAn

drefa

na

Atsim

oAt

sinanana

Atsin

anan

aBo

eni

Sava

Vakin

anka

ratra

Vato

vavy

Fitovin

any

CI memb er Recen t deat h

Age of household head

24% 19%19%

15% 20%15%

23%

19% 19% 15% 18%

31% 37%21%

26%29%

24%

31%

28% 27% 34% 27%

20%18%

25% 29%28%

33%

23%

24% 25%22%

26%

23% 20% 27% 24% 17% 26%16%

27% 26% 25% 26%

0%

10%

20%

30%

40%

50%

60%

70%

80%90%

100%

Alao

traMan

goro

Anala

njirofo

Androy

Anosy

Atsim

oAn

drefa

na

Atsim

oAtsinanana

Atsin

anan

aBo

eni

Sava

Vakin

anka

ratra

Vato

vavy

Fitovin

any

< 15 years 15-24 25-34 35-44 45-54 55 + years

Household Survey Findings

Demographics

Household size: The mean household size was 5.8 persons, ranging from 4.7 in Analanjirofo to 7.1persons in Androy region while the median was 5 persons. However, about one-quarter of thesampled households in Atsimo Atsinanana andAndroyregions had 10 or more members. In contrast,about one-third of the sample households in Analanjirofo region had 1-3 members. By district, the

median household size was 7 persons for households in Beloha, Toliary II and Tsihombe while thesmallest households were found in Fenoarivo Atsinanana and Toamasina II samples (4 persons).

Female headed households: In total, 17% of the sample households were headed by women withsome variation between the regions. Twenty percent of the sample households in Androy and

Atsinanana were headed by women compared to only 14% in AnosyandAtsimo Andrefana. By districtone-third of the households in the Ambovombe and Antsirabe II samples were headed by women,compared to only 6% in Toliary II and 7% in Betafo samples.

Age of household head: The median age of household heads was 44 years with female heads beingslightly older (45 years) than male heads (42 years). Only a few households reported a head youngerthan 15 years of age. The chart below shows the ages of household heads by region. The oldest

household heads appearto be found in Atsimo

Atsinanana while the

youngest are found inAtsinanana. By district,16% of the householdheads were younger than25 years in Beloha while40% were 55 years orolder in Vohipeno district.

Presence of chronicillness: In total, 14% ofthe sampled householdshad chronically illmember1, and thesewere most likely to befound amongst

households in Vakinankaratra region (26%) and least likely found in Analanjirofo region (9%). This issummarized in the chart below. The districts with the highest percentage of households with achronically ill member include Faratsiho (36%), Betafo (29%) and Vohmar(29%) while the districts withthe lowest percentage were Soanierana Ivongo (3%) andAmbovombe (6%).

Recent death of a household member: In total, only 5% of the sampled households indicated that ahousehold member haddied in recent monthsranging from a high of 8%in Boeniregion to a low of1% in Atsinanana region(see above). By district,14% of the households inVohmar reported the

death of a member,followed by 11% in

Marovoay and 10% inSoanierana Ivongo. Only1% if sampled householdsin Antsirabe II, Farafangana,

Mahanoro, and Manakaradistricts reported the recent death of a member.

1 Chronic illness refers to illness for three months or more.


4/21

4

Livelihoods

Key livelihood activities the households engaged in during the past six months prior to the surveywere investigated in order to understand the households livelihood strategies in providing for itsfood and income needs.

The five most important livelihood sources for the sample households were Handicraft/artisan (45%),sales of agricultural products (41%), daily wage labour (32%), animal sales (15%) and agricultural

labour (4%). By region, the top four livelihood activities and the percentage of households engaging inthem are in the table below. For most regions, the most common activities are Handicrafts/skilledlabour or Sales of agricultural products.

Main livelihood Second Third Fourth

Alaotra Mangoro Craft/skilled = 55% Wage labour = 38% Agric WL = 29% Fishing = 18%

Analanjirofo Agric sales = 35% Craft/skilled = 34% Wage labour = 32% Fishing = 9%

Androy Craft/skilled = 76% Agric sales = 41% Wage labour = 37% Animal sales = 15%

Anosy Agric sales = 46% Craft/skilled = 44% Wage labour = 40% Fishing = 13%

Atsimo Andrefana Craft/skilled = 51% Agric sales = 44% Animal sales = 24% Wage labour = 21%

Atsimo Atsinanana Craft/skilled = 67% Wage labour = 47% Fishing = 13% -

Atsinanana Wage labour = 42% Craft/skilled = 37% Agric sales = 29% Fishing = 15%

Boeni Agric sales = 69% Craft/skilled = 44% Animal sales = 43% Wage labour = 38%

Sava Agric sales = 44% Craft/skilled = 38% Wage labour = 22% Agric WL = 9%

Vakinankaratra Agric sales = 87% Animal sales = 48% Craft/skilled = 29% Wage labour = 10%

Vatovavy Fitovinany Wage labour = 47% Craft/skilled = 36% Agric sales = 22% Salary = 4%

The households were asked to estimate the amount of income from the various sources and fromthis information the per capita monthly income was calculated. By main income source, the activitieswith the highest per capita monthly income were Private sector salary (39,500 Ariary/month), Publicsector salary (33,900 Ariary/month), and Pension (33,300 Ariary/month). The activity with the lowestincome was Wage labour which paid only 11,500 Ariary per month.

By region, the mean and median per capita monthly incomes are presented in the chart below. Theseare only estimates from reported information and are by no means exact but rather are useful in

comparing the regions.

27,000

19,000

7,900

12,600

23,800

11,400

22,200

24,400

16,200

19,500

6,600

15,300

8,300

5,000

7,000

12,500

8,300

15,20014,300

7,500

9,600

3,600

-

5,000

10,000

15,000

20,000

25,000

30,000

Alao

traMango

ro

Anala

njirofo

Andr

oy

Anosy

Atsim

oAn

drefana

Atsim

oAt

sinanana

Atsin

anan

aBo

eni

Sava

Vakin

anka

ratra

Vatovavy

Fitovin

any

Per capita monthly income Mean Per capita monthly income Median


5/21

5

8% 11%

67%

47% 51%61%

21% 17%12% 11%

62%

64%

85%

32%

49% 43%38%

76%

62% 77% 77%

36%29%

4% 1% 4% 6% 1% 3%

22%12% 13%

3%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Alao

traMangor

o

Anala

njirofo

Androy

Anosy

Atsim

oAnd

refana

Atsim

oAt

sinanan

a

Atsin

anana

Boeni

Sava

Vakin

ankaratr

a

Vatovavy

Fitovinan

y

Asset wealth poor Asset wealth medium Asset wealth rich

Households in the Alaotra Mangoro region have the highest mean and median per capita monthlyincomes. Households in Atsinanana have the second highest median per capita monthly incomesfollowed by households in Boeni. The lowest incomes were found amongst households in VatovavyFitovinanyandAndroyregions.

Household Assets and Livestock

The survey collected data on asset ownership from each household (19 assets, both productive and

non productive). The data was then analysed considering whether the households own that particularasset or not. Then households were classified as being asset poor (0-4 different types of assets), assetmedium (5-9) or asset rich (10 or more).

Of the total sampledhouseholds 35% werefound to be asset poor57% were asset mediumand 8% were asset rich.According to the chart onthe right, households in

Androy were the mostlikely to be asset poor(67%) followed by those inVatovary Fitovinany (62%)

while those in AlaotraMangoro are the mostlikely to be asset rich(29%) and the least likelyto be asset poor (8%). By district, 78% of the households in the Ambovombe sample were asset poorcompared to only 1% in Vavatenina. By comparison, 44% of the households inAmparafaravola districtwere asset rich while there were no asset rich households in the samples from Nosy Varika, Vohipeno,

Mahanoro, Tsihombe andAmbovombe districts.

Livestock ownership was low overall, with only 40% of the sampled households owning cattle, only5% owning sheep or goats and 11% owning pigs. Sixty percent of the households owned chickensthough and 14% owned ducks. By region, nearly 60% of the sample households inAtsimo Andrefanaand Vakinankaratra owned cattle compared to only 15% in Antsinanana. Sheep and goat ownershipwas also highest in Atsimo Andrefana where 19% of households owned sheep and 20% owned goats.

Households inAndroywere the only others who owned sheep (10%) or goats (14%) in any numbers.At district level, cattle ownership was exceptionally high in Ampanihy (82%) followed by Faratsiho(71%) and Benenitra (71%) and lowest in Vatomandry(11%) andMahanoro (15%) districts.

The chart below shows the relationship between household asset wealth and livestock ownership.There was no relationship between sheep or goat ownership and asset wealth so they are notincluded.

43%

73%

3%

13%

27%

63%

81%

29%

49%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Asset poor Asset medium Asset rich

Cattle

Pigs

Chickens


6/21

6

0

1

2

3

4

5

6

7

rice maize cassava beans/peas vegetables meat oils/fats sugar

poor borderline acceptable

2% 1% 4% 1% 1% 7%16% 20%

1%10%

64%

43%31%

41%

15% 16%

21%

44%19% 53%

90%

34%

57%68%

56%

84% 83%72%

39%

60%46%

0%

20%

40%

60%

80%

100%

Vakin

anka

ratra

Vatova

vyFitovin

any

Atsim

oAt

sinan

ana

Atsin

anan

a

Anala

njirofo

Alao

traMango

roBo

eni

Atsim

oAn

drefa

na

Androy

Anosy

Sava

poor borderline acceptable

Dietary diversity and food frequency

Research has shown that dietary diversity and frequency are a good proxy measures of foodconsumption and food security at household level dietary diversitythe number of different foods orfood groups consumed over a given reference period, can act as an alternative indicator of foodsecurity under a variety of circumstances.

Food consumption data was collected and analyzed using the standard WFP methodology: the variety

and frequency of different foods and food groups consumed over a 7-day recall period was recordedto calculate a weighted food consumption score. Weights were based on the nutritional density ofthe foods. Standard cut-points or thresholds were established to enable analysis of trends and toprovide a benchmark for success. Households are then classified as having either poor, borderlineor acceptable consumption based on the analysis of the data.

Households classified as having poor food consumption were basing their diet eating only maizethree days per week, vegetables4-5 days and rice only one dayper week. This is generallyregarded as a sign of extremehousehold food insecurity.Households with borderlineconsumption are eating theequivalent of rice on a daily basis

plus vegetables 5 days a week,sugar/sugar products about threedays per week, oils/fats 1-2 daysand cassava one day per week.Only households classified ashaving acceptable consumption

were having, along with daily intake of rice and sugar, 4 days of oils/fats and vegetables, 3 days ofmeat, 2 days of cassava and only one day of beans/peas per week.

Overall, only 5% of the households were classified as having poor consumption while 33% hadborderline consumption and 62% achieved acceptable consumption. Households with acceptableconsumption were significantly less likely (p < 0.01) to have a female head than those with poor orborderline consumption. In addition, those with poor consumption were significantly more likely (p

analysis of household food security in selected districts of madagascar (september 2009)

Documents