Rural market access and nutritional outcomes in farm households

Rural market access and nutritional outcomes in

farm householdsWilliam A. Masters

http://sites.tufts.edu/willmasters

Amelia F. Darrouzet-Nardihttp://sites.tufts.edu/ameliadarrouzetnardi

Friedman School of Nutrition, Tufts University

Joint A4NH/ISPC workshop on agriculture-nutrition linkages22-23 September 2014

http://sites.tufts.edu/willmasters

http://sites.tufts.edu/ameliadarrouzetnardi/

http://sites.tufts.edu/ameliadarrouzetnardi/

How do rural markets influence nutrition outcomes in farm households?

1. Could improve or worsen nutrition, by various direct channelsa) Household income, wealth and purchasing powerb) Time allocation especially for women and childrenc) Relative cost of access to safe and nutritious foods

2. Could alter the ag-nutrition relationship, by effect modifiersa) Separates decision-making between farm and householdb) Alters resilience and consumption smoothing

3. Here we add to the long and old literature on #1and introduce new work on #2b

4. Will present empirical results then implicationsand start with a hypothesized causal model

Market access and farm household nutrition motivation | the global study | the DRC study

A causal model of farm household decision-making


Qty. of nutritious foods (kg/yr)

Qty. of farm household’s labor time (hrs/yr)

Qty. of farm household’s other goods (kg/yr)

Other employment(allows sale of labor to buy food)

Rural markets give households additional options, allowing them to overcome diminishing returns in working their own land

Qty. of nutritious foods(kg/yr)

Once farmers are actively trading, production decisions are “separable”

from consumption choices,linked only through purchasing power

Rural food markets(allows sale of other goods to buy food)

In self-sufficiency, production =consumption

Consumption

Production

Consumption

Production

That same separability applies whether households are buying or selling,

and allows consumption smoothing over time

Empirical identification of causal effects is difficult

• Market access tends to be closely correlated with productivity and purchasing power, but relationship may not be causal– Markets may arise and grow where people can use them– People who can use them may move towards markets

• Here, we report on two ways to identify potentially causal links between rural market access and farm household nutrition– Globally, do subnational administrative regions with an earlier history of

urbanization have healthier maternal and child heights and weights? – Within DRC, do farm households located closer to rural towns have

more resilience against seasonal shocks to child heights and weights?

• Both studies construct natural experiments, using time lags and spatial variation in risk exposure to identify effects


The global study:Does past urbanization help rural farmers today?


02

04

06

08

0n

umbe

r of

re

gion

s

-2 0 2 4 6log(years before 2000 that the average cell in a region reached 25% urbanization)

(with normal distribution superimposed, N=756 region-year observations, of regions that had reached 25% urbanization by 2000)

Log duration of urbanization around rural farm households

Note: Data shown are for 756 subnational regions in 53 countries with DHS surveys, using urbanization data from Motamed, Florax & Masters (2014)

(Mean year is 1988, earliest quartile is 1970)

Markets take time to develop, and farmers’ regions vary widely

in how long they’ve had access to towns and cities

Regions with earlier urbanizationhave taller children now


-1.6

-1.5

5-1

.5-1

.45

-1.4

-1.3

5

Chi

ld m

ean

HA

Z

6 6.5 7 7.5 8 8.5 9Log of World Bank GNI at PPP prices

(N=1171 observations from 143 DHS surveys in 57 countries with 520 subnational regions;dashed line shows subnational regions that reached 25% urbanization before 1995

Mean HAZ for rural farm children at each level of national income, by timing of urbanization

Regions with earlier urbanizationhave heavier children now


-.6

-.4

-.2

0

Chi

ld m

ean

WH

Z

6 6.5 7 7.5 8 8.5 9Log of World Bank GNI at PPP prices

(N=1171 observations from 143 DHS surveys in 57 countries with 520 subnational regions;dashed line shows subnational regions that reached 25% urbanization before 1995

Mean WHZ for rural farm children at each level of national income, by timing of urbanization

For rural farm children, being in a region with more established towns and cities is associated with

a very large weight advantage, anda small significant height advantage

The DRC study:Does proximity to town confer resilience

against seasonal shocks?

• At each farm location, the timing of a child’s birth exposes them differently to agroclimatic risk factors for malnutrition and disease

• The DRC is distinctive in that households vary widely in distance to towns and also in exposure to seasonal risks– We ask whether birth during and after wet seasons is harmful,

• For more remote households with less access to markets and services,• In regions with more seasonal variation in rainfall

– Birth timing in “placebo” regions without seasons should have no effect


Birth timing creates a natural experiment• The “treatment” is having a distinct wet season (if there

is one) occur during late pregnancy and early infancy– This is a particularly sensitive time for child development– Wet seasons are a hungry period with poorer diets– Wet seasons facilitate water- and vector-borne disease

• Market access may be protective– Households can trade to smooth consumption– Households can access health and other services

• We expect no effect of birth timing, and no protection from market access, in regions with uniform rainfall


Analytical design: Spatial difference-in-difference

Household location and child birth timing

Region has a distinct wet season? (= farther from the equator)

Yes No (“placebo”)

Child was born in or after wet season? (=Jan.-Jun. if lat.<0, Jul.-Dec. otherwise)

Yes(at risk) No Yes No

Household is closer to town? (=closer to major town)

Yes(protected?) No Yes No Yes No Yes No

Hypothesized effect of birth timing: Neg. None

Note: To test our hypothesis that market access protects against seasonality, the identifying assumptions are that birth timing occurs randomly between seasons (tested), and that seasonal risk factors would have been similar in the absence of towns (untestable).


Darker cells (100m2) have better market access.

Market access is measured by travel cost weighted distance to the nearest major town


Seasons depend on rainfall and temperature

equator

At the equator, average monthly rainfall fluctuates from 100 to 200 mm, and average monthly temperature fluctuates from 24 to 26 degrees Celsius.


“Winter” is a drier period, farther from the equator

equator

Away from the equator, there is a drier, colder winter, here May through August.

Latitude -6


In the other hemisphere, winter is 6 months later

equator

Here in the Northern Hemisphere, the drier season occurs from November through February.

Latitude +4


All ChildrenN=2806

Jan.-JuneNo SeasonsN=650

Jan.-JuneSeasonsN=903

July-Dec. No SeasonsN=563

July-Dec. SeasonsN=690

Child status HAZ -1.47 (1.86) -1.51 (2.02) -1.51 (1.75) -1.61 (1.92) -1.26 (1.80) WAZ -1.20 (1.38) -1.09 (1.42) -1.34 (1.31) -1.17 (1.46) -1.13 (1.34) WHZ -0.38 (1.33) -0.22 (1.39) -0.53 (1.21) -0.24 (1.41) -0.45 (1.31) Age (mos.) 29.16 (16.53) 28.81 (16.82) 28.53 (15.80) 29.70 (17.10) 29.88 (16.69) Sex (% boys) 49.4% 47.8% 49.8% 47.9% 51.4%Household Wealth quint. 2.9 (1.42) 2.61 (1.25) 3.02 (1.48) 2.67 (1.26) 3.19 (1.51) Dist. to town (km) 64.8 (52.1) 70.1 (47.9) 59.6 (51.9) 71.6 (45.3) 60.8 (59.8)Environment Conflicts 31.28 (66.9) 66.53 (97.82) 9.29 (19.47) 48.57 (84.24) 12.74 (22.99) Lat. (abs val.) 4.31 (2.64) 1.99 (1.16) 6.14 (2.01) 1.98 (1.17) 5.99 (2.02)

Note: Mean (standard deviation). Jan-June births are actually Jul.-Dec. births if the child was born in the Northern Hemisphere (N=418). Conflicts are total number of incidents between 2001 and 2007 in the respondent’s grid-cell of residence.

We split the data into groups by risk exposure..

Note: Age controls suppressed; Jan-June births are actually Jul.-Dec. births if the child was born in the Northern Hemisphere (N=418); robust pval in parentheses; errors clustered by region (N=11); *** p<0.01, ** p<0.05, * p<0.1

(1) (2) (3) (4)

VARIABLES Units/typeHAZSeasons

HAZNo Seasons

WHZSeasons

WHZNo Seasons

Conflicts Cumulative days 0.005** 0.000 0.002 -0.000Wealth quintile Categorical 0.152** 0.165** 0.102*** -0.010

Remote Binary -0.434** 0.338 0.010 0.353Born Jan.-June Binary -0.363*** 0.141 -0.0534 0.297*Born Jan.-June*Remote Interaction -0.407** 0.407* -0.144** 0.075Child is male Binary -0.192* -0.192** -0.097 -0.195**Constant Constant -0.149 0.043 0.520* 0.490

Observations N 1,593 1,213 1,593 1,213R-squared R2 0.154 0.155 0.082 0.035Number of regions N 10 7 10 7

We see a strong and significant “treatment effect” of household remoteness in areas with seasons.

Note: Data shown are coefficient estimates and 95% confidence intervals for average treatment effects in our preferred specification (Table 5), for our two dependent variables of interest followed by seven placebo variables for which no effect is expected, due to the absence of any plausible mechanism of action.

Among our robustness checks, we do placebo tests for desirable outcomes that could not be caused by birth timing

HAZW

HZ

moth

er's

educa

tion (y

rs)

moth

er cu

rrently

work

ing? (

binar

y)

moth

er's

weight (

kg)

fath

er's

educa

tion (y

rs)

years

lived

in in

terview

loca

tion

size o

f house

hold (#

of p

eople)

altitu

de (m

)-1.5

-1

-0.5

0

0.5

1

1.5

Significant effects on

child heights and weights

reported on previous slide

No effects and large variances where no effect is expected

Conclusions

• From these data, – Globally, farm households in subnational regions with a

longer history of urbanization have better nutritional status

– Within DRC, farm households that are closer to towns are more protected from seasonal shocks to nutritional status

• These results add to the large and diverse literature on farmers’ use of markets– New data will permit many new tests to refine results– But the importance of market access has strong

implications for agriculture-nutrition actions


Implications for policies and programs

• At a given level of household and community resources, facilitating market access can– Raise levels of nutritional status– Improve resilience to shocks

• Farm households can use markets in many different ways– Specialization and trade, to overcome diminishing returns on the farm– Consumption smoothing, via separability of production & consumption– Access to public services

• Future work may be able to distinguish among uses– But various uses are naturally bundled together in related transactions– And in any case policies and programs to ease market access cannot

prescribe what households do, just allow them to do it more easily!


Rural market access and nutritional outcomes in farm households

Documents

rural market access

farm householdswilliam

effects market access

global study

difficultmarket access

agnutrition relationship

drc study note

rural farmers