Monitoring household coping strategies during complex crises final

Monitoring Household Coping Strategies during Complex Crises
United Nations Development Programme and United Nations Children’s Fund1

I. Introduction and Relevance for UNDP and UNICEF
Households cope with unanticipated shocks in a variety of ways by drawing upon
individual, community, market and public resources. Although such shocks may be of
relatively short duration, an extensive body of literature has shown that, left to
themselves, vulnerable households and individuals may cope in ways that adversely
affect human development in the long term. Such effects can become more likely during
global crises, when many are affected at the same time (aggregate shocks), and some
forms of community, market and public mechanisms may falter. The current economic
crisis is one such episode, as are recurrent episodes of high food and energy prices.

In our study, we first brought together some evidence for impacts on human development
occasioned by periods of crises. We then examined the actions – coping mechanisms-
that households take to minimize the harm from shocks to their wellbeing, by drawing on
on-going surveys that document the impact of the current crisis. Apart from their intrinsic
interest, such mechanisms are also of importance to policymakers as they link the better
observed, macro indicators of a crisis, such as declines in GDP growth, increases in
unemployment and falls in export revenues, to longer term negative impacts, such as
extended periods of poverty, poorer health, stunted children and decreases in educational
outcomes. While many of these longer term consequences may be irreversible,
understanding the coping mechanisms that lead to such outcomes may help in identifying
observable characteristics that can be monitored for timely assessment of the impact of a
crisis at the household level, as well as in devising policy interventions that could
forestall negative long term outcomes. In addition, understanding how households cope
also helps determine why different individuals within the same household may be
affected differently.

1
This paper summarizes a research project supported by UN Global Pulse’s “Rapid Impact and
Vulnerability Assessment Fund” (RIVAF) between 2010 and 2011. Global Pulse is an innovation initiative
of the Executive Office of the UN Secretary-General, which functions as an innovation lab, bringing
together expertise from inside and outside the United Nations to harness today's new world of digital data
and real-time analytics for global development. RIVAF supports real-time data collection and analysis to
help develop a better understanding of how vulnerable populations cope with impacts of global crises. For
more information visit www.unglobalpulse.org.

II. Key Findings
Our review of existing work shows extensive evidence of the harmful impacts of
aggregate shocks on human development, and is reported in greater detail in the main
paper. Going beyond this review, our study proceeds to examine evidence from two
countries during the global economic crisis of 2009. We use data collected by our
partners in the Poverty and Economic Policy (PEP) network as part of their Community-
based Monitoring Survey (CBMS) work. These surveys were conducted in a number of
countries including Indonesia, Kenya, Lao PDR, the Philippines, Tanzania, and Zambia.
Our work uses the data collected in the Philippines and in Kenya and broadly validates
the study of housing level coping behaviour as a sensitive indicator of the impact of the
crisis with important implications for both monitoring and policy design.

Coping Mechanisms in the Philippines
Macroeconomic evidence shows that the economic crisis impacted the Philippines
through a sharp decline in its exports; and a fall in remittances from overseas workers,
particularly those working in the USA. Based on these key channels, an initial scoping
study identified 13 barangays (wards/villages) most likely to feel the effects of the crisis,
as well as indicators to be monitored at the household and community levels. Consistent
with the CBMS methodology, all households in the selected sites were included in the
survey, thus covering 4,954 households with 21,454 members2. The survey was
conducted in April 2009, with a reference period from November 2008 to April 2009.
When asked directly about any impacts felt from the crisis, a large number of households
reported that they were affected by the crisis, with 31.03 per cent reporting no impact or a
mild impact and 65.74 per cent reporting a moderate to severe impact 3.

The survey asked questions about a wide array of coping strategies related to augmenting
income (for example through sales of assets, borrowing, or drawing upon savings);
changing food consumption (for example through consuming staples only, reducing
portion sizes or consuming one’s own harvest); and altering the use of education (for
example through transferring from a private to a public school, spending less on books
and consumables or withdrawing from school) and health (for example through shifting
from private providers to government clinics, using generic medication in place of
branded options or using alternative medicines). Household characteristics such as
incomes, location, composition, levels of education and other features were also
recorded.

2
It was also possible to keep track of some individual households from a previous survey in 2006 (and
earlier for a few barangays) and as a result, panel data is available on 2702 households, with one round of
pre-crisis and one round of post-crisis data, which will be further analyzed.
3
Within the 2702 households for which information from 2006 is also available, and which reported a
moderate or severe impact from the GFC, there is a small, but perceptible shift in the real per capita
income distribution functions towards the left, with the effect being especially marked below the poverty
line of 18,000 pesos. The self reported impact is thus borne out by changes in the poverty status of
households, and we take the self reported impact to be a reliable indicator for inferring coping behaviour
occasioned by the crisis.

Table 1 shows the frequency of some of the coping behaviours of interest, both in the
aggregate and also by (per capita) income quintiles. This table shows that in all quintiles,
households responded to the crisis by adopting some combination of coping mechanisms.

Table 1: Coping strategies (percentages), by income quintile

Income Quintile
Coping Strategy Total Lowest 2 3 4 Highest
At least one food related
85.99 81.85 85.4 85.76 90.3 86.67
strategy
At least one education related
25.05 33.17 28.3 24.43 24.44 14.67
strategy
At least one health related
60.4 56.85 57.91 64.29 64.44 58.37
strategy
Borrowed money 37.34 38.15 41.74 38.26 39.72 28.63
Used savings 13.84 16.44 14.6 12.95 12.69 12.56
Pawned Assets 4.08 2.06 3.92 4.72 5.59 4.05
Sold Assets 2.56 3.62 3.59 2.21 1.93 1.45
Looked for additional work 5.62 5.73 8.72 5.29 5.17 3.12

Income-related Coping: Borrowing money is one of the two most prevalent coping
strategies, with 37.34 per cent of households reporting borrowing to meet various
expenses. Across different income groups, differences were observed in how assets
were used to tide over the crisis: the poor are more likely to sell assets while the rich are
relatively more likely to pawn. The data also shows (Table 2) that community borrowing
(from a friend, relative, neighbor, a cooperative or an NGO) is by far the most common
borrowing behaviour, with the lower quintiles relatively much more likely to borrow
from the community while the higher quintiles are more likely to borrow from a private
source, possibly at higher interest rates. We also see that private borrowing is much more
common in urban areas than rural areas.

Table 2: Households (percentage) borrowing from different sources

Income
Quintile
All (%) Lowest 2 3 4 Highest Rural (%)Urban (%)
Tried to borrow money 38.46 39.88 43.36 39.46 39.92 29.46 38.04 38.91
Borrowed money 37.34 38.15 41.74 38.26 39.72 28.63 36.87 37.85
Community borrowing 21.03 22.19 19.95 22.14 21.02 19.56 20.25 21.85
Private borrowing 11.39 7.10 10.49 13.48 12.79 13.69 5.54 15.18
Others 3.37 2.06 3.42 2.89 4.01 4.85 3.72 3.01

Food-related Coping: A large proportion of households (85.99 per cent) used at least one
food related coping strategy – with significant differences in those likely to be more
common across income groups, as shown in Figure 1. It is quite evident from this that

some strategies appear more likely to be adopted at lower levels of income than others –
for example, reducing portions, and consuming staples only.

Each one of these can be quite harmful, especially with respect to nutrition, with
potentially severe consequences for the very young, the pregnant, the lactating and those
suffering from chronic illnesses such as HIV/AIDS. Moreover, during an extended
period, the vulnerability of a household to subsequent shocks as well as the likelihood of
long-term negative outcomes can increase as a result of resorting to such methods.

Figure 1: Food-related coping mechanisms

Food-Related Coping Strategies
Percentage of Households

Consumed
staple food only
Combined
meals
Reduced
portion
Consumed own
harvest
Other

Decile (Per capita income)

Education and Health-related Coping: As far as education related coping mechanisms
are concerned, 25.05 per cent of households overall reported utilizing at least one such
measure. Similar to the food related strategies, education is more likely to be impacted in
poorer households. 60.4 per cent of all households have to use at least one health related
strategy. Table 3 shows a breakdown of households reporting different degrees of impact
from the economic crisis.

It appears that the difference in coping behavior observed across income groups is not a
result of differential impact across income groups. The impact of the crisis (at least self
reported) is independent of household income level.

Table 3: Households (percentage) reporting different degrees of impact

Bottom Income quintile
Total 40 Top 60 Lowest 2 3 4 Highest
Not + mild 31.03 35.63 28.91 38.56 30.77 27.64 25.51 33.71
Moderate +severe 65.74 64.63 68.31 57.68 68.02 70.16 71.36 63.28

Figures 2 and 3 summarize the relative prevalence of income, food, health and education
related coping strategies among the poor and the rich. The red bars show the ratio of the
prevalence of a coping strategy among the bottom and top quintile of per capita income.
The blue bar shows the ratio between the bottom 40 per cent and the top 60 per cent of
the population, by per capita income.

To the right of the bold vertical line are those strategies where the (unconditional)
probability is observed to be higher among the poor relative to the rich. Apart from the
food-related strategies discussed earlier, we observed that among the poor, education-
related coping is more likely and health-related coping is less likely.

The poor appear to already be at minimal levels of expenditure with respect to health-
related services. A similar analysis with respect to income-related coping strategies
shows that the poor are more likely to rely on selling assets, and looking for additional
work.

Figure 2: Income-related coping across the poor and the rich

Philippines: Adoption of coping strategies by the poor and
the rich

Borrowed money
Used savings
Pawned Assets
Sold Assets
Looked for additional work

0.0 0.5 1.0 1.5 2.0 2.5 3.0
Ratio

?irst quintile to ?ifth bottom 40 to top 60

Figure 3: Coping mechanisms across the poor and the rich

Philippines: Adoption of coping strategies by the poor and
the rich

At least one food related strategy
Consumed staple food only
Combined meals
Reduced portion
Consumed own harvest
Other
At least one education related strategy
At least one health related strategy

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
Ratio

?irst quintile to ?ifth bottom 40 to top 60

Coping Mechanisms in Kenya
The data collection for the study was carried out from July 2009 to January 2010, in Tana
River district (one of the poorest districts in Kenya with about 72 per cent of the
population living below the poverty line), and surveyed 5,882 households in six sub-
locations. One notable characteristic of the region is that many households reported being
affected by factors other than the economic crisis - 90.2 per cent by drought, 38.84 per
cent by ethnic conflict and so on. In these circumstances it becomes especially hard to
determine what part of the coping behavior, if any, can be attributed to the economic
crisis.

Notwithstanding such difficulties in attribution (present to some extent in the Philippines
as well), the coping strategies follow a broadly similar pattern as discussed above – with
differences in relative frequencies, that may be reflecting differences in room to
maneuver due to higher levels of continuing poverty.

In Kenya, selling belongings in general and selling belongings specifically to buy food
are widely reported coping strategies, reported by 20.59 and 33.40 per cent of
households4 respectively. In addition, 26.28 per cent of households reported a decline in
their monthly food expenditure. Common use of food-related coping is one of the many
similarities in the behavior of households in the Philippines and Kenya. At the same time,
there are distinct and important differences between the two countries, indicating the
relevance and importance of country-specific approaches.

As seen in Table 4, 11.75 per cent of households report borrowing as a coping strategy.
Borrowing is more common in higher income quintiles, probably reflecting the
borrowing constraints faced by lower income households. The evidence from Kenya

4
This question (Q36) had a different reference period than other questions and asked if the household had
sold any belonging buy food over the last one month.

shows that it is easier for the upper quintiles to borrow to tide over shocks. A
significantly larger proportion of the relatively well-off also seem to have savings that
they can spend in times of need. On the other hand, potentially harmful coping behavior
seems to be more common in the bottom two income quintiles of households. This was
also seen in the Philippines.

Table 4: Coping Strategies (percentages), by Income Quintile

Income
Quintile
Coping Strategy Total Lowest 2 3 4 Highest
Decline in monthly food expenditure 26.28 38.87 30.25 18.80 20.09 22.67
Decline in monthly education
expenditure 13.29 8.87 7.14 7.63 12.26 12.59
Shifted at least one child from private to
public school 4.38 0.68 1.10 1.13 1.79 2.67
Withdrew at least one child from
school 1.43 5.63 4.25 3.73 4.09 3.28
Decline in monthly health expenditure 15.93 23.70 15.04 14.64 13.70 11.64
Borrowed money 11.75 6.73 7.90 12.39 13.28 18.19
Used savings 15.60 3.50 6.46 13.08 22.72 32.07
Sold belongings 20.59 13.90 25.23 27.04 20.34 15.60
Sold belongings to buy food in the
last month 33.40 40.41 38.57 35.62 29.02 22.67
Sought additional work 3.41 2.05 4.50 3.64 3.23 3.36

The results from the Philippines and Kenya are consistent with the recent emerging
literature on the impact of shocks on households. For example Crompton et al (2011)
draws together evidence from a number of studies on the effects of the 2007-2008 food
price spike. They find that high food prices increased malnutrition in young children, and
poverty. They report that nearly all households surveyed reported eating less-preferred
food as well as the use of credit and savings as widespread coping mechanisms.
Preliminary evidence available from the global financial crisis reaffirms these themes
while introducing some new ones.

Conclusions
The evidence received and analyzed thus far indicates that coping mechanisms can
indeed provide a fairly sensitive indicator of how households respond to negative shocks;
with significant differences between those adopted by the poor and the rich. While there
are important country specific differences in the details, it also appears that the poor are
more likely to have only a limited number of options, which makes them more prone to
adopting strategies that are likely to result in adverse human development outcomes.
Relatively quick surveys may be able to identify the most relevant coping strategies in a

particular context, as well as suggesting concrete policy options that may help in
mitigating the harmful effects.

The longer such coping measures last, the more likely that there will be long-term
negative effects. Moreover, as the shock abates, the speed at which coping measures are
wound down becomes an important parameter in determining how quick the recovery
from the negative consequences will be. UNDP and UNICEF will remain involved in
continuing surveys in these countries to address some of these issues.

III. Key Challenges

Unavailability of Baseline Information - In the absence of a baseline, it becomes difficult
to distinguish between regular behaviour and coping behaviour. In the Philippines,
CBMS has interviewed a panel of households around the year 2006 and then in 2009.
However, the questions related to coping behavior that are part of the GFC (2009)
questionnaire are not part of the ‘core’ questionnaire that was administered in 2006 and
2009. As a result, we are unable to assess how the relative frequency (for example, of
using savings) or magnitude (example, the size of loans) of a particular household
activity is different from the norm. Moreover, if the relevant behaviour changes
seasonally, we have no way of taking this into account5.
Difficulty in Attribution - In the Philippines, a large number of respondents identified
themselves as being impacted by the crisis. This was supported by a shift in the
distribution of real per capita income, giving us confidence that the coping (or the part of
household behaviour that can be considered coping and not just usual behaviour) was in
response to the economic crisis. Households were also asked if they were affected by a
number of direct transmission channels such as the loss of a job, decline in remittances, a
decline in the frequency of remittances etc. It is puzzling that only 375 households (7.57
per cent) reported being affected by at least one of these direct channels. Either these
questions on the direct impacts are missing some channels from the crisis or households
are feeling the impact of other shocks. This is hard to determine as no information is
collected on price levels or other contemporaneous shocks. Finally, the criteria for
choosing the 13 sentinel sites are not very clear. In fact, collecting data at other sites, not
expected to show impacts from the crisis, would have provided useful control group data.
In the case of Kenya, it is difficult to determine the nature of shocks that the household is
reacting to. As seen above, the survey responders reported being subject to drought,
conflict and a number of other shocks. There was no specific question on how they were
affected by the economic crisis.

5
According to the FAO, for the Philippines, “the wet-season rice crop in the north lasts from June to
November and the dry-season crop from January to May-June. In the south it is the reverse: wet-season
crops last from October-November to March-April and dry-season crops from May-June to November”. It
is therefore plausible that at least some of the households in the sample would have resorted to coping
behaviours as a matter of routine in the agricultural lean season. However, we are unable to identify the
relative contribution of this factor to the observed behaviours.

Intensity of Harmful Coping – Food-related and education-related coping—each with its
own long-term consequences —are commonly observed. It would have been more
informative if the frequency of such behaviour had been noted. For example, the impact
on a household that reports eating less-preferred food once or twice in the last six months
feels it very differently than a household that reports doing this repeatedly in the last six
months. Similarly, the frequency of other coping, such as how often school is being
missed by children, matters.

Monitoring household coping strategies during complex crises final

More Related Content

What's hot (20)

Similar to Monitoring household coping strategies during complex crises final (20)

More from UN Global Pulse (20)

Recently uploaded (20)

Monitoring household coping strategies during complex crises final