Domestic Price Components and Child Nutritional Status in Niger

Sources of Domestic Food Price Volatility
and Child Malnutrition:
evidence from Niger and Malawi
GIOVANNI ANDREA CORNIA, LAURA DEOTTI and MARIA SASSI
University of Florence, Save the Children and University of Pavia
Much attention was paid to impact of world food prices of 2007-8/201011. This obscured domestic factors’ s impact on child nutrition in SSA, i.e.
• the long term food supply impact of agricultural policies;
• huge and persistent seasonal variation in food production;
• impact of (still recurrent!) famines.
After testing impact of changes in world food prices on domestic prices
….. this paper tests influence of 3 above variables on child malnutrition
How large the transmission of world food prices on
domestic food prices in SSA?
• Large literature ….but conclusions vary on basis of specific factors such as
•
•
•
•
dependence on imported food,
magnitude of transport costs and trade margins,
world tradability of domestic staple food, and integration with global mkt
exchange rate variations,
• Thus in parts of SSA
• is difficult to argue that the world price is main driver of domestic food prices
• while domestic prices often exhibit large rises in parallell with declines of world prices
Niger and Malawi
• their structural features (small, land short, bordering bigger countries,
landlocked, low input subsist. agriculture, single staple) make them
representative of several SSA countries affected by major problems of:
- Chronic;
- Seasonal; and
- Acute;
food insecurity and child malnutrition.
Theoretical framework to asses child malnutrition
(based on Sen’s entitlement approach)
A household’s ‘control over food’ (and thus child malnutrition) depends on its
entitlement, i.e. a set of commodity bundles it commands in society
ENTITLEMENTS:
- Production-based entitlements (food produced for self-consumption)
- Exchange-based entitlements (food acquired through mkt exchange)
- wage–based entitlements
- Inheritance and transfer-based entitlements
These entitlements change (at different speeds) over the l.term, seasons & during
famines
Food entitlement bundles of a household
Food
available
Pj
Pw Th A h
Qh  Qfsch  Qjh 
 Qwh 


Pf
Pf Pf
Pf
Food produced
for selfconsumption
Good sold
Wage
labour
Transfers
received
Income
form assets
sold
Factors affecting food entitlements & child malnutrition
- Long-term food price and security (which depends on gains in
agricultural productivity – and therefore on agricultural policies);
- easonal variations in food prices (depend on storage capacity & credit
availability);
- Price changes during food acute crises/famine (depend also on food
security policies)
- Changes in value of the entitlements over these three time dimensions
Methodology of empirical analysis
• we decompose food price vector into three price sub-vectors reflecting the:
• Trend price component (reflecting – inter alia - long term gains in agricultural productivity)
• Seasonal price component (reflecting the efficiency of credit mkts and inter-temporal smoothing)
• Famine price component (reflecting the gov. capacity to respond to large price shocks )
• A residual price component (which reflects various effects, such as the price of inputs)
• We correlate these food price components (sub-vectors) to the number of
children admitted to feeding centers
Methodology (cont’d)
• Nominal food price data were deflated with the CPI
• The dynamics of trend and seasonal component suggested that the 4 price
components interact according to a multiplicative model:
FP = TPC x SPC x FPC x RPC
• Monthly Seasonal Adjust. Method allows separating seasonal (SPC) effects
from the l.term component (TPC) of the aggregate price trend.
• the long term trend (TPC) was calculated by mean of the Hodrick-Prescot
Filter and then subtracted from the aggregate price trend so as to obtain the
random (RPC) and famine (FPC) price components.
Price trend component – monthly data
Malawi: Maize (January 2003 –
December 2009)
Niger: Millet (January 2006 –
December 2010)
80
260
70
240
60
220
180
20
160
10
140
Maize Price
TPC
Note: TPC = Trend price component
Millet price
TPC
10/10
07/10
04/10
01/10
10/09
07/09
04/09
01/09
10/08
07/08
04/08
01/08
10/07
07/07
04/07
01/07
10/06
120
07/06
09/09
04/09
11/08
06/08
01/08
08/07
03/07
10/06
05/06
12/05
07/05
02/05
09/04
04/04
11/03
06/03
0
04/06
30
200
01/06
CFA
40
01/03
MWK
50
Seasonal and famine food price components and number of
children admitted per feeding centre – monthly data
CAF/NC
SPC
FPC
0.9
0.85
CAF/NC
SPC
FPC
Note: CAF/NC = child admissions by feeding centre; SPC = Seasonal price component; FPC = Famine price component
07/10
01/10
07/09
0.8
01/09
04/08
11/08
06/09
12/05
07/06
02/07
09/07
08/03
03/04
10/04
05/05
0
0.95
07/08
0.2
1
01/08
0.4
1.05
07/07
0.6
1.1
01/07
0.8
1.15
07/06
1
90
80
70
60
50
40
30
20
10
0
01/06
1.2
01/03
index number
1.4
1.2
index number
50
45
40
35
30
25
20
15
10
5
0
number
1.6
Niger: Millet (January 2006 –
December 2010)
number
Malawi: Maize (January 2003 –
December 2009)
Methodology (cont’d)
• We regress the 4 price components on the n. of child admissions
• We control for ‘hunger season dummy’ and number of feeding centers
• Do not include – due to lack of data – the value of entitlements. Yet –during
seasonal fluctuations/famines- they correlate(-) with food prices
• We ignore world prices as in the 2000s their trend did not significantly affect
domestic prices
OLS log-log regression of n.child admissions to feeding centres on price components – monthly data
(2003-2009 – Malawi 2006-2010 Niger))
Variable
Malawi
Niger
Log-log OLS estimates
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Constant term
a
5.914***
1.139***
5.710***
1.062**
-0.613
1.056**
Trend price component
b
- 0.769***
- 0.572***
-0.702***
12.05***
10.624***
11.84***
Seasonal price component
g
1.676***
1.908***
-0.189
1.926***
1.727***
1.089
Famine price component
d
0.333
0.219*
0.331*
-2.881***
-2.012**
-2.699***
Residual price component
r
0.239
0.744
0.381
-2.567
-3.341
-2.183
Number of feeding centre
l
Hunger season dummy
s
Adj R2
F- statistics
Durbin Watson
1.205***
0.427***
0.782***
0.317***
0.30
0.86
0.39
0.84
0.86
0.85
9.91***
107.93***
11.73***
72.84***
73.13***
66.6***
1.30
0.85
1.44
1.04
1.27
1.21
Results
• Despite data & methodological limitations, regressions analysis
suggests that:
• the moderate l. term food price rise in Malawi (due to rising land yields) helped
reducing child malnutrit. Opposite was true in Niger (no support to agriculture)
• Seasonal price rises are a major cause of child malnutrition in both countries. Such
effect correlates closely with hunger seasons dummy (dominated by price effect)
• The 2006-2010 famines in Niger where well responded to, it helped contain child
malnutrition. This was not the case in Malawi
• In both, the rise in feeding centers contributed to the increase in admissions
Implications for policy research
- Suitable policies aimed at intensifying agricultural production and rising
land yield to contain domestic prices & reduce child malnutrition;
- Role of long-term policies, particularly in the agricultural sector;
- Role of credit/storage policies for addressing seasonal food price
fluctuations;
- Rethinking food security policies during famines;