Poverty & Equity Global Practice Working Paper 162
ELICITING ACCURATE RESPONSES TO
CONSUMPTION QUESTIONS AMONG IDPS IN
SOUTH SUDAN USING “HONESTY PRIMES”
Lennart Kaplan
Utz Johann Pape
James Walsh
August 2018
Poverty & Equity Global Practice Working Paper 162
ABSTRACT
Misreporting is a well-known challenge for researchers in social sciences. This issue is especially prevalent if
incentives for misreporting exist, for example, to claim certain benefits or hide illegal behavior. Internally
displaced persons are a population that is highly dependent on aid receipts and, thus, have strong incentives to
underreport consumption levels. To improve reporting for such vulnerable populations, this paper proposes to
integrate “honesty primes” into the consumption module of the questionnaire. Honesty primes are unconscious
stimuli that induce a certain cognition or behavior. The study assesses the effectiveness of a bundle of randomly
assigned primes within a sample of internally displaced persons in South Sudan. In line with the main hypothesis,
positive and significant effects arise for low consumption quantiles, especially consumption quantities that are
more susceptible to manipulation. Hence, honesty primes can act as a cost-effective tool to induce more
accurate reporting. Further research is needed to identify more effective primes for the respective population
of interest.
This paper is a product of the Poverty and Equity Global Practice Group. It is part of a larger effort by the World Bank to
provide open access to its research and contribute to development policy discussions around the world. The authors may be
contacted at fadoho@worldbank.org and upape@worldbank.org.
The Poverty & Equity Global Practice Working Paper Series disseminates the findings of work in progress to encourage the exchange of
ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully
polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions
expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for
Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank
or the governments they represent.
‒ Poverty & Equity Global Practice Knowledge Management & Learning Team
This paper is co-published with the World Bank Policy Research Working Papers.
Eliciting Accurate Responses to Consumption Questions
among IDPs in South Sudan Using “Honesty Primes”
Lennart Kaplan, Utz Johann Pape, James Walsh1
Keywords: Consumption Measurement, Poverty, Questionnaire Design, Behavioral Intervention
JEL: C83, D63, D90, I32
1
Authors in alphabetically order. Corresponding author: Utz Pape (upape@worldbank.org). The findings,
interpretations and conclusions expressed in this paper are entirely those of the authors, and do not necessarily
represent the views of the World Bank, its Executive Directors, or the governments of the countries they
represent. This paper benefited largely from the valuable input provided by Julie Perng and Luca Parisotto as well
as discussions with Quy‐Toan Do.
1. Introduction
Accurate data on the key economic variables affecting people who have been forcibly displaced, such
as consumption and assets, are essential to understanding their situation and developing evidence‐
based policies to support them. Data inaccuracies may lead policy makers to allocate funds to the
wrong people or to the wrong programs. Poor information can lead to flawed diagnostics or incorrect
assessments of impact. The standard way in which the World Bank and other policy organizations
develop statistics is through individuals’ responses to questions in economic surveys. Self‐reported
information is vulnerable to myriad reporting inaccuracies, especially when social scientists ask
personal or intrusive questions or when respondents have (or believe they have) an incentive to report
dishonestly.2 In situations where it has been possible to compare survey responses to revealed
economic behavior, striking disparities are sometimes found. In one investigation for example, Poterba
and Summers (1986) report that misstatements regarding employment status in the Current
Population Survey led to an underestimation of the duration of unemployment by up to 80 percent
and even greater overestimates of the frequency of labor market entries and exits. In another study,
Rosenfeld, Imai, and Shapiro (2016) look at voting behaviors in a sensitive anti‐abortion referendum
held in Mississippi in 2011. They compare actual county level vote shares against survey results from
a sample frame of individuals who voted during the election (based on public records). Surveys that
used direct questioning led to an underestimation of casting a “no” vote by more than 20 percentage
points in the majority of counties.
There are a number of mechanisms through which the validity of self‐reported information in surveys
can be compromised. Some inaccuracies result from cognitive biases – for example, acquiescence or
“yea‐saying” (Bachman and O’Malley 1984; Hurd 1999), extreme responding (Cronbach 1946;
Hamilton 1968), and question order bias (Sigelman 1981). One solution to problems such as question
order bias is to randomize the order of questions (Warner 1965). Other inaccuracies emerge from
conscious but not calculated behavior. Respondents may deliberately misreport information on
sensitive subjects not to distort statistics but to maintain their reputation or to abide by political norms
(Gilens, Sniderman, and Kuklinski 1998). A common solution to this is to enable participants to cloak
their behaviors or beliefs. List experiments, endorsement experiments, and randomized experiments
are commonly used techniques for this purpose (Rosenfeld, Imai, and Shapiro 2016).
The explanations above assume that people want to tell the truth but are prevented from doing so due
to cognitive or social barriers. Some misreporting is purposeful. Individuals may misreport in a
calculated fashion to increase earnings in a study context (Mazar, Amir, and Ariely 2008) or to shape
2
This is of particular concern, for example when asking about race (Kuklinski et al. 1997) or corruption (Gingerich
2010).
2
the results of the study if they believe that it will inform policy. It is not surprising that this problem
might arise in the context of development aid, an area rife with perverse incentives (Bräutigam and
Knack 2004; Cilliers, Dube, & Siddiqi, 2015). In situations where individuals wish to influence a
particular research outcome, a guise of anonymity will not shift their behavior.
Behavioral science has been used as a policy tool to increase honesty and discourage anti‐social
behavior (World Bank 2015). This is based on research illustrating that people make decisions on the
basis of both external and internal reward mechanisms (Mazar and Ariely 2006). Even in cases where
people have a material incentive to lie, their internal drive to protect their sense of self‐integrity may
override. One example of this is when individuals’ beliefs regarding the consequences of lying affects
their behavior. In an two‐person experiment where one participant can increase her payoff by lying
but at the expense to her counterpart, Gneezy (2005) finds that individuals’ propensity to lie is sensitive
to the costs it imposes on the other person. Contextual cues affect the salience of internal incentives
(or intrinsic motivations) and thus the honesty of responses. This psychological mechanism has been
put to practical use in policy. In multiple contexts, normative messaging has been used to increase tax
payments (Hallsworth et al. 2017; Hernandez et al. 2017) or reduce littering and environment theft
(Cialdini 2003).
In this paper, we apply the tools of behavioral science to investigate the veracity of consumption
reports by internally displaced persons (IDPs). In numerous rounds of data collection in Somalia and
South Sudan, IDPs report significantly lower levels of consumption than non‐IDP households. In
previous survey rounds 45 percent of Somali IDP households report food consumption below
subsistence levels and approximately 80 percent below recommended levels (refer Figure 5). While
the data may be accurate, there are two reasons to suspect that they are not. First, such high levels of
non‐consumption would be associated with high rates of mortality due to starvation. Although being
high, the mortality rates among IDPs suggest that this is not happening systematically across the
country at such a scale (FEWS NET, 2018).3 Second, non‐IDP households that are statistically similar on
observable characteristics report higher levels of consumption than IDP households. While IDPs and
non‐IDPs may have different opportunities to generate income, it is unlikely that IDPs choose not to
smooth their resources to balance between food and non‐food consumption in a way that endangers
their life.4
3 Although data from the USAID led Famine Early Warning Systems Network (FEWS NET) suggest high level of
malnutrition, evidence on mortality across the counties is mixed (FEWS NET, 2018).
4
The underlying survey data of this study discussed at a later stage actually indicate that IDPs have a more calorie
intensive food consumption profile (refer Figure 8).
3
If it is the case that survey respondents are reporting dishonestly, the inaccuracies it generates in the
data are highly problematic. At best, it makes the data spurious and unusable. At worst, it could lead
to misallocations of aid, from more vulnerable areas to less vulnerable areas, or from solutions
emphasizing sustainability to immediate relief where immediate relief is unnecessary. Due to the
dangerous environment in South Sudan and Somalia, it is not currently possible to use alternative data
collection methods, for example ethnographic research, to investigate this puzzle in the data. The
validity of alternative investigative methods such as food diaries is vulnerable to the same incentive to
game as surveys.
One way to investigate whether people dishonestly report is to test whether consumption rates
change in response to “honesty primes.” If these primes are effective, they would be expected to
particularly affect potentially underreporting, hence, poor households. Moreover, as vulnerable
populations would have higher incentives to underreport, priming should be stronger for IDPs than for
comparable non‐IDP populations. We find the primes induce higher reporting in lower quintiles of
reported consumption. This treatment pattern is driven by aid reliant IDPs and vanishes when
considering the comparison group of non‐IDPs. The results are especially strong for consumption
quantities (items and kilograms), which are most easily subject to intentional misreporting. This
suggests that IDPs are indeed misreporting. The paper has two main limitations. First, it can only
compare the treated group against an estimate of the “true” consumption rates. Second, the
intervention is bundled. For this reason, it is impossible to isolate the causal mechanism affecting the
observed changes in reporting. Further work is needed to identify an estimate of the true level of
consumption against which to compare the primed individuals and to isolate the causal mechanisms
by which people are changing their behavior.
The paper proceeds as follows. Section 2 provides an overview about the underlying context and the
compiled data. Section 3 provides an overview about the underlying methods, while Section 4
introduces the empirical approach, which builds the foundation for the results in Section 5, which
offers an assessment of robustness and potential channels. Finally, the findings are discussed and
summarized in Section 6.
2. Context and data
On July 9, 2011 South Sudan became the 55th African independent state after seceding peacefully
from the Republic of Sudan. Facing a history of a 50‐year lasting conflict, South Sudan slid back to
instability after its peaceful independence process. This led to an internal displacement of circa 2
million, more than 15 percent of South Sudan’s population (UN OCHA, 2017). Moreover, the conflict
contributed to a deterioration of South Sudanese economic outcomes, with poverty rates reaching 82
percent in 2016, widespread severe food shortages, culminating with famine being declared in some
4
counties in 2017 (The World Bank, 2016; Devi, 2017). This makes well‐targeted crisis response and aid
allocation highly important.
The experiment sample includes 4,145 IDP and 781 non‐IDP households interviewed in 2017 in South
Sudan across the High Frequency South Sudan Survey (HFSSS), the Crisis Recovery Survey (CRS), and
the IDP Census and Sampling Study (IDPCSS). The CRS interviewed a representative sample of IDPs in
IDP camps across South Sudan. In the same period the HFS conducted interviews across urban centers
in seven of the ten former states (Figure 1). The IDPCSS conducted a census of all households in Juba
POC1. The consumption modules in questionnaires administered to respondents in the three surveys
were built in exactly the same manner so as to ensure comparability, and the fieldwork was
implemented by the same organization. The only difference across the three surveys is the population
that was sampled.
Figure 1: HFS and CRS coverage.
Note: The HFS interviewed a representative sample of households in urban centers in the states colored in blue in the map
above. The CRS interviewed households in 4 of the largest IDP camps in South Sudan, denoted by red diamonds in the map.
Major urban areas are indicated via black dots. The IDPCSS was conducted in the Juba POC1.
The conditions in camps do not allow for standard household surveys, hence an alternative survey
approach based on the Rapid Consumption Methodology was applied (Pape & Mistiaen, 2015). Here,
only 30 / 25 food and non‐food items are administered to all households. Additional 20 food and non‐
food items vary between households. More specifically, households are pre‐assigned to one of four
sub‐modules for food and non‐food consumption (each containing 20 items). Neither the enumerators
5
nor the respondents see the structure of the sub‐modules, but the assigned items are asked in a
categorically meaningful way (like cereals, fruits, etc).5
The data are used to construct four outcome measures. The surveys collect information on quantities
in terms of (i) number of consumption items and (ii) kilograms. The quantities can be used to construct
measures of (iii) monetary and (iv) caloric food consumption scaling the quantities with data on
average prices and energy levels.6 Though we are mainly interested in evaluating the impact of the
honesty primes on the total consumption value ‐ both in terms of money and food intake ‐ these
variables are difficult for respondents to falsify because these are second‐order values that are
calculated as a function of other variables, including consumption quantities and calories or prices that
are in turn deflated. All of this adds noise to the answer provided by the respondent, and they depend
in part on variables over which the respondent has no control. The consumption quantity in kilograms
is a more direct measure of the quantity consumed as expressed by the respondent, and may lead to
more accurate estimation of the impact of the honesty primes. Finally, counting the number of items
may lead to an even more accurate measure, since the variable does not undergo any cleaning at all
and is taken at face value. Furthermore, omitting an item is likely to be the easiest and quickest way
for respondents to reduce the true value of the household’s consumption.7
Poverty among IDP households is high, and 9 in 10 IDP households across South Sudan live under $1.90
USD PPP (2011) per capita per day in 2017. IDP households in the sample interviewed for the
experiment consume on average 333 SSP (2017) per capita per. IDP households reported on average
6.63 core consumption items. These figures represent about 20 percent of core items asked to the
households. Figure 2 visualizes that 39 percent of households report consumption below the
recommended daily intake of 2,100 kcal (R) and 16 percent below the subsistence level of 1,200 kcal
(S) (Ravallion & Bidani, 1994).
5
Due to the survey method applied CRS surveys contain the core consumption module and one additional
consumption module. The share in imputed consumption is on average 99.9 percent. IDP surveys contain due
to the previously outlined time constraints only core consumption items. However, by design these items
capture the lion’s share of consumption (on average approx. 94 percent of total consumption in more
comprehensive CRS surveys).
For a description of the caloric intake measure, please consult the appendix.
6
Note that the number of consumption items is not reported per capita as it does not increase proportionally
7
with household size.
6
Figure 2: Density plot of value of core food consumption.
Note: Estimates presented in the figure above are not weighted and are representative only of IDP and non‐IDP households
surveyed in the study sample. 8
3. Approach and randomization
The honesty primes introduce exogenous variation into the consumption module to try and tease out
whether consumption might be underreported in IDP households. A prime is an environmental cue
that unconsciously induces a subsequent cognition or behavior. For example, in studies with prisoners
and bankers, participants who engage in activities that prime their identity behave more dishonestly
in behavioral experiments than participants who have not participated in priming activities (Cohn,
Fehr, and Marechal 2014; Cohn, Maréchal, and Noll 2010).9 Honesty primes been found to elicit more
honest and accurate responses during questionnaires (Rasinski et al. 2005; Vinski and Watter 2012).
To investigate whether consumption might be underreported in IDP households, we introduce
exogenous variation into the consumption module. Households are randomly exposed to a bundle of
“honesty primes.” These include an emphasis on the importance of accurate answers at the beginning
of the survey, a short fictional scenario which will require passing judgment on the behavior of one of
the characters, and additional questions to tell when was the last time their household had a meal,
forcing the respondents to explicitly report that they have not eaten in the last week.
8
We do not use weights throughout the study as the research hypothesis relates not to the average treatment
effect, but particularly the pirmes’ effectiveness at the tails of the distribution.
9
Questionnaires confirmed that participants associate their identity with dishonesty.
7
•Appeal to Honesty
The importance of accurate answers and the purpose of the survey will be reiterated.
•"Thank you for taking the time to speak to us. We really appreciate the time you are
1
giving to participate in the survey. We encourage you to provide honest information. By
participating in the survey and by providing accurate information, you are playing an
important role in helping us understand the situation in South Sudan."
•Moral Prime to encourage honesty
At the beginning of the survey module concerning food consumption, the respondent
will be given a short fictional scenario which will require passing judgement.
2
•"John asks his good friend Deng if he has some money that he can lend him to help him
pay for medicine for his sick son. Deng has money but was planning to buy cigarettes
with it. He lies and tells John that he has none. Is it okay for Deng to lie to John?"
•Investigative Probing:
At the start of the survey module concerning food consumption, the respondent will be
asked to tell when was the last time their household had a meal. This question will then
also be asked for each of four major food categories: ‘Bread and Cereals’, ‘Meat’,
3
‘Fruits’, ‘Pulses and vegetables’.
•E.g., "When was the last time that any of the household members had Bread and
Cereals?"
Households are randomly exposed to behavioral treatments, in the form of a moral prime and
investigative probing, to try and elicit more truthful answers from respondents. This way they do not
constrain the choice frame, but rather alter the anchoring towards more truthful reporting (The World
Bank, 2015).
The bundle of primes addresses different behavioral processes. (1) Appeals to honesty are a standard
tool in surveys to increase data accuracy by relying on social approval (Talwar, Arruda, & Yachison, 2015).
(2) Moral primes induce unconscious cognitions, which are intended to affect subsequent behavior.
When facing incentives to lie, respondents would answer more truthfully to sustain self‐consistency.
(3) Investigative probing puts a higher salience on the question. By asking for broader categories first,
subsequent sub categories are put under more scrutiny. Self‐consistency is reinforced by relating to a
longer recall period of seven days.10 While the appeal to honesty and the moral prime target
intentional misreporting, investigative probing is addressing classical measurement error.
The sample was randomly selected into each treatment arms in two groups of approximately 50
percent, with 2,467 households in the control group and 2,459 in the treatment group. The
randomization process was built into the CAPI (Computer Assisted Personal Interview) questionnaires
10
The methodological appendix provides an overview of the relevant questions in the food consumption module.
8
administered in the surveys. As our research hypothesis suggests stronger effects of the “honesty
primes” for more vulnerable populations, we focus on IDPs for the main analysis. The availability of
the HFS sample provides a comparison group of non‐IDP households for the experiment, which will be
used for robustness checks. The treatment and control groups are relatively balanced. There is a higher
share of male headed households in the treatment group, which have also more members, though in
practical terms these differences are relatively small. As gender of the household head and household
size are correlated with poverty, these variables are included in the regression models and interacted
with the treatment to control for potential impacts.
Table 1: Balance across treatment and control arms (IDP sample).
Control Treatment Difference, p‐value
Household size 4.835 5.098 0.003**
(0.060) (0.064)
Gender of household head 0.492 0.448 0.005**
(0.011) (0.011)
Literacy of household head 0.507 0.529 0.155
(0.011) (0.011)
Household head completed some primary school 0.540 0.563 0.133
(0.011) (0.011)
Is the household head employed 0.328 0.319 0.555
(0.010) (0.010)
Share of children in household 0.364 0.373 0.309
(0.006) (0.006)
Share of elderly in household 0.011 0.010 0.582
(0.002) (0.001)
First Component of Asset Principal Component Analysis ‐0.126 ‐0.194 0.162
(0.037) (0.032)
2079 2066
N
0.502 0.498
Proportion
Standard errors in parentheses; *p<0.05, **p<0.01
4. Empirical Strategy
To assess the effect of our prime on reporting behavior, we can formulate following simple regression
equation.
∗ , (3)
9
where is the log of the outcome variable. Across different models we estimate the effect for (i) the
number of consumption items consumed [referred to in the regression equation as Cons. Num.], (ii)
consumption quantity per adult equivalent (in kilograms) [Cons. Quant.], (iii) monetary consumption
value per adult equivalent [Cons. Val.] and (iv) daily caloric intake per adult equivalent [Cons. Cal.].
Our main treatment variable is a dummy variable which takes the value of 1 if the household was
assigned to the treatment group. indicates a set of camp fixed effects, are month fixed effects,
and is the idiosyncratic error term. denotes a vector of control variables generally associated with
consumption, including household size, the gender of the household head, and the proportion of
children (under 18) in the household. Moreover, we add an asset index based on the first component
of a principal component analysis (Filmer & Pritchett, 2001; McKenzie, 2005).11 The model will be
estimated with and without controls to check the impact they may have. As the treatment might
interact with the unbalanced covariates, it makes sense to add to the regression ∗ , the interaction
of the unbalanced controls with the treatment variable (Lin & Green, 2016; Baranov, Bhalotra, Biroli,
& Maselko, 2017).
It is expected that the respondents who will be affected by the honesty primes are respondents that
would otherwise misreport and, hence, are more likely to be at the extremes of the distribution.12
Therefore, we complement our analysis with a quantile regression approach. The idea of the quantile
regression framework, which was introduced by Koenker & Bassett Jr (1978), is to take the entire
distribution of the dependent variable into account by estimating several regressions, which put more
weight to the quantile of interest. The underlying minimization problem can be stated as follows:
min | | 1 | |, (4)
: :
where θi is the quantile of interest and the weighted sums of derivations |y τ| of the outcome per
quantile. Minimizing the latter, differential effects conditional on the quantile of the dependent
variable are obtained. Further, it has the advantage of being less prone to outliers and non‐normality
of the error term. For our purpose, quantile regressions offer the advantage that they are more flexible
than simple interactions with poverty lines, which would be endogenous to consumption levels.
11
As assets (bikes, fans, rickshaws, etc.) can be more easily surveyed by enumerators, those are likely to capture
parts of the household wealth.
12
Although these hypotheses were not pre‐registered, they are based on theoretical considerations about the
mechanisms of the underlying behavioral primes.
10
5 Results
There is a slight indication that the treatment may have worked, based on consumption distributions
across treatment arms. The consumption distribution shown in Figure 3 shows a slight difference in
caloric consumption between IDP households in the treatment group and the control group, though
this is apparent only at lower levels of consumption, i.e. below the subsistence level of 1,200 kcal. The
median of calorie consumption is well above the recommended daily intake. However, still a
substantial part of the distribution of 16 percent reports below the subsistence level and 40 percent
below the recommended daily intake.13 Hence, the prime would also be relevant in the adult
equivalent setting to achieve more precise reporting, which is analyzed in our regression framework
subsequently. Taking into account the finding that consumption levels are lower than to be expected,
the most relevant treatment effects can be expected at the left tail of the distribution.
Figure 3: Caloric consumption p.c. (adult equivalents)
Note: The underlying data is based on per adult equivalents. Caloric consumption levels are labeled in the following graph as S subsistence
equivalent (1200 kcal p.c.), R recommended daily intake (2100 kcal p.c.) and M the median (2340 kcal p.c.).
Regression results
In order to test for the influence of control variables, the regressions are estimated with and without
control variables. When not conditioning on control variables, the results indicate significant treatment
effects only for the number of consumption items in Column (1). This outcome measure would be
easiest to falsify as it does not undergo further cleaning, e.g., in terms of deflation or calorie scaling.
When adding further controls, coefficients turn larger and imply treatment effects of 6‐14 percent.
13
Compared to the monetary consumption levels, the calorie consumption p.a. seems rather high. This is partly
attributable to the fact that IDPs’ consumption focuses on energy intensive products, where cooking oil and
sorghum constitute 45 percent of food expenditure. If we contrast the consumption shares with non‐IDPs, we
find that although the diet of non‐IDPs is less energy intensive, it comprises a higher variety (see Figure 8).
11
The interactions of the treatment and the asset index as well as household size have negative and
significant coefficients in line with previous work. For example, larger households are on average more
prone to consumption poverty and might react differentially (Lanjouw & Ravallion, 1995). The simple
treatment indicators also turn significant for the kilogram consumption quantities in Column (4) and
the monetary consumption value in Column (6).14 Yet, our main indicator of interest, the caloric food
consumption remains unaffected. This is in line with our hypothesis that the average treatment effect
should be limited and is rather uninformative as the primes are expected to particularly affect
misreporting at the tails. For this purpose, a quantile regression analysis is taken out to provide more
nuanced estimates, subsequently.
Table 2: Results from baseline estimation, model (1).
(1) (2) (3) (4) (5) (6) (7) (8)
ln(Cons. ln(Cons. ln(Cons. ln(Cons. ln(Cons. ln(Cons. ln(Cons: ln(Cons.
VARIABLES Num.) Num.) Quant.) Quant.) Val.) Value) Cal.) Cal.)
Treatment
0.035** 0.061* 0.028 0.137*** ‐0.018 0.081* 0.019 0.001
(0.016) (0.033) (0.018) (0.042) (0.018) (0.039) (0.028) (0.067)
Observations 3,955 3,955 3,955 3,955 3,955 3,955 3,955 3,955
R‐squared 0.001 0.273 0.001 0.070 0.000 0.078 0.000 0.123
State FE NO YES NO YES NO YES NO YES
Month FE NO YES NO YES NO YES NO YES
Controls NO YES NO YES NO YES NO YES
Controls
Interacted NO YES NO YES NO YES NO YES
Note: Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1‐Column (1‐2) is measured on
the household level. Column (3‐8) refer to per capita OECD adult equivalents. A full set of coefficients for control variables
can be found in Table 15 the appendix.
To capture this heterogeneity across consumption levels, quantile regressions are applied. Results are
shown in Figure 4 and Table 3.
Table 3: Results from quantile regressions of different outcome variables.
(1) (2) (3) (4)
Outcome Variables ln(Cons.Num.) ln(Cons. Quant.) ln(Cons. Val.) ln(Cons. Cal.)
Q0.1 0.165** 0.342*** 0.079 0.235*
14
Unintuitively, with regard to the monetary consumption values in columns (5), negative coefficients are
estimated, contradicting a higher consumption quantity. In line with other studies, this could be explained by
larger households buying larger quantities and, hence, consuming more while paying lower bulk purchasing
prices (Deaton & Paxson, 1998).
12
(0.064) (0.079) (0.068) (0.127)
Q0.25 0.058** 0.201*** 0.198*** 0.140*
(0.028) (0.067) (0.053) (0.080)
Q0.5 0.018 0.136** 0.119** 0.042
(0.032) (0.056) (0.050) (0.062)
Q0.75 0.047 0.114** 0.071 0.032
(0.034) (0.050) (0.051) (0.067)
Q0.9 ‐0.016 0.049 ‐0.015 0.013
(0.028) (0.050) (0.054) (0.064)
Observations 3,955 3,955 3,955 3,955
Month FE YES YES YES YES
State FE YES YES YES YES
Controls YES YES YES YES
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1‐Column (1) is
measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
The priming significantly increases reported consumption in lower quintiles. Significant treatment
effects occur mainly for the number of consumption items and the quantities in kilograms. Monetary
and caloric consumption measures are less strongly affected (
Figure 4). The latter might also be less susceptible to deliberate misreporting as they depend in part
on variables over which the respondent has no control as the pure consumption quantities are scaled
by calorie levels or deflated.15
Conditional quantile regressions are sometimes considered as uninformative as they describe the effect on the
15
distribution rather than on the individual. Hence, we also consider unconditional quantile regressions in the
appendix. Results are robust and support an upward shift in lower quantiles of the outcome variables (Table 12).
13
Figure 4: Treatment effects across quantiles.
Note: Treatment effects and confidence intervals plotted for different quantiles.
Ultimately, we are interested in the question if the prime is sufficiently strong to shift a significant
share of the distribution to more credible consumption levels both in terms of monetary and caloric
food consumption. For this purpose, we construct four dichotomous indicators. Those are equal to one
if (i) the respondent household surpasses the caloric subsistence level of 1200 kcal or (ii) the
recommended level of caloric intake of 2100 kcal. Two further dummies are created at (iii) 66.66% and
(iv) 100% of a normalized poverty line, which is scaled by the fact that only core consumption items
were assessed consistently across all surveys. Table 4 depicts results for the three thresholds using
model (3). Although the coefficients are mostly positive, only two coefficients turn significant in
Columns (2) and (3). Therefore, the results stress the nuanced effect of the prime, which only affects
certain strata of the population.
Table 4: Results using poverty thresholds, model (2) and (3).
(2) (4) (5) (6)
VARIABLES >1200kcal >2100kcal > >Poverty Line
Treatment 0.010 0.069* 0.063* 0.029
(0.027) (0.037) (0.037) (0.036)
14
Observations 3,955 3,955 3,955 3,955
R‐squared 0.067 0.098 0.118 0.135
State FE YES YES YES YES
Month FE YES YES YES YES
Controls YES YES YES YES
Controls Interacted YES YES YES YES
5. Treatment heterogeneity and robustness
Heterogenous effects
If the primes would reduce misreporting, stronger effects are to be expected among subpopulations
that have higher incentives to misreport, e.g., aid‐reliant IDPs. In order to assess this channel more
thoroughly, (i) heterogenous effects are estimated contingent on aid reliance and (ii) the sample is
compared to a non‐IDP comparison group.
Parts of the respondents from the CRS and HFS were also interviewed with regard to their previous
support through UN agencies. This dummy indicator can be used for an assessment of heterogenous
effects.16 The model is analogous to equation (3), where we add UN assistance as a further control
variable as well as an interaction term of UN assistance with the behavioral treatment. The results are
depicted in Table 5 and indicate no clear pattern. Only for the number of consumption items a positive
significant coefficient is found. This could be treated as some weak evidence that the prime is more
effective for aid exposed IDPs, but should not be overstated due to the non‐significance for the other
three outcomes of interest.
Table 5: Channel – UN assistance.
(1) (2) (3) (4)
ln(Cons. Num.) ln(Cons. Quant.) ln(Cons. Val) ln(Cons. Cal.)
0.100 0.195** 0.171* 0.105
Treatment
(0.066) (0.080) (0.081) (0.087)
0.104** ‐0.059 0.016 0.011
Treatment*UN Assistance
(0.051) (0.060) (0.061) (0.064)
Observations 2,204 2,204 2,204 2,204
R‐squared 0.38 0.086 0.098 0.108
State FE YES YES YES YES
Month FE YES YES YES YES
Controls YES YES YES YES
16
The results can only be interpreted as an explorative analysis as UN assistance was not balanced across
treatment and control groups, where treatment households have a higher probability of being previously
exposed to aid.
15
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1.
Column (1) is measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
The non‐IDP subsample offers an interesting opportunity to assess the robustness of the results.
Constraining the sample only on non‐IDPs, we can estimate our results analogous to Table 3. The
pattern of positive and significant treatment coefficients in the lower quantiles vanishes, except for
Column (1). This could be interpreted as evidence that the honesty primes applied are more efficient
for the vulnerable IDP population, which has higher incentives to indicate need than the non‐IDPs. This
would be in line with previous studies (e.g., Cilliers, Dube, & Siddiqi, 2015) suggesting a high degree of
social desirability bias in the setting of foreign assistance. Specifically, the populations exposed to
development aid, in our setting the IDPs, would be more likely to provide socially desirable answers to
signal their “worthiness” for assistance. This corresponds to Table 5, providing some weak evidence
that the primes are more effective for respondents relying on UN aid. It would be of particular interest
to examine those heterogenous effects based on more fine‐grained data on neediness and degree of
aid reliance of recipients. For this purpose, however, a “true” benchmark would be needed. As
administrative data are non‐existent or of poor quality, an alternative for future research might be to
build on measures from qualitative work as suggested by Blattman, Jamison, Koroknay‐Palicz,
Rodrigues, & Sheridan (2016). Moreover, one should be careful not to draw too strong conclusions
from these results as the number of observations is limited in this comparatively small sub‐sample.
Table 6: Quantile Regressions – reduced sample (only non‐IDPs).
(1) (2) (3) (4)
Outcome Variables ln(Cons. Num.) ln(Cons. Quant.) ln(Cons. Val.) ln(Cons. Cal.)
Q0.1 ‐0.027 ‐0.069 ‐0.026 0.032
(0.079) (0.102) (0.110) (0.113)
Q0.25 0.148** ‐0.052 0.012 ‐0.057
(0.073) (0.095) (0.107) (0.122)
Q0.5 0.067 ‐0.041 ‐0.032 0.044
(0.067) (0.081) (0.100) (0.100)
Q0.75 ‐0.071 ‐0.072 ‐0.015 ‐0.052
(0.054) (0.080) (0.092) (0.080)
Q0.9 ‐0.041 0.157 0.074 0.119
(0.047) (0.105) (0.144) (0.127)
Observations 780 780 780 770
Month FE YES YES YES YES
16
State FE YES YES YES YES
Controls YES YES YES YES
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1.
Column (1) is measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
Robustness
In line with hardly credible low consumption levels, misreporting could be considered to be more
prevalent at the tails of the distribution, hence, among the extreme values. On the one hand, it makes
sense to consider those outliers. On the other hand, one would like to avoid basing the inference
mainly on those extreme values. Ideally, one would know how to distinguish the intentionally
misreported outliers and the ones that are caused by errors in reporting or data entry. The log
normalization in the main analysis is chosen as a compromise of keeping most data possible, but
making estimates less susceptible to outliers. This suggests two natural robustness checks: (i) in a more
liberal setting, the outcomes in levels can be used and (ii) in a more conservative setting, the outliers
at the 5th and 95th percentiles are discarded. Regression results using the levels are depicted in Table
7.17
Table 7: Quantile Regressions – outcomes in levels.
(1) (2) (3) (4)
Outcome Variables Cons. Num. Cons. Quant. Cons. Val. Cons. Cal.
Q0.1 0.544** 0.741*** 9.440 229.126*
(0.254) (0.173) (13.305) (136.013)
Q0.25 0.298* 0.675** 60.585*** 179.447
(0.156) (0.224) (16.261) (157.584)
Q0.5 0.151 0.638* 49.404** 197.589
(0.194) (0.280) (20.477) (192.043)
Q0.75 0.341 0.700** 19.499 281.317
(0.246) (0.339) (30.762) (286.250)
Q0.9 ‐0.077 0.609 ‐30.117 ‐279.159
(0.333) (0.540) (41.581) (284.569)
Observations 3,955 3,955 3,955 3,955
Month FE YES YES YES YES
State FE YES YES YES YES
As scaling of the outcome variables is different – e.g., the outliers with regard to consumption quantity in
17
kilograms might not correspond to the consumption quantity in calories – the outliers for one measure do not
always correspond to outliers among the other measure. In order to guarantee that we still base the inference
on the same observations, outliers from all corresponding variables are dropped, which explains that the
resulting sample is smaller than 90% of the full sample.
17
Controls YES YES YES YES
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1.
Column (1) is measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
Table 8 depicts the results without outliers and indicates a slightly less nuanced pattern. In line with
our hypothesis of stronger misreporting tendencies on the extremes, Column (1) indicates significant
treatment effects at the 10th and 25th percentiles. Although significant treatment effects among higher
quintiles can be found in Columns (2) and (3), the coefficients for the 25th percentile are quantitatively
larger. Finally, with regard to caloric consumption in Column (4) statistical significance vanishes, but
the largest coefficient is to be found in the 10th percentile. Hence, although the pattern gets weakened
when excluding outliers, the prime still significantly affects the reported consumption quantities with
stronger effects in the lower quantiles.
Table 8: Quantile Regressions – without outliers.
(1) (2) (3) (4)
Outcome Variables ln(Cons. Num.) ln(Cons. Quant.) ln(Cons. Val.) ln(Cons. Cal.)
Q0.1 0.124** 0.106 0.085 0.058
(0.049) (0.067) (0.064) (0.091)
Q0.25 0.045* 0.139** 0.162*** 0.042
(0.027) (0.055) (0.044) (0.077)
Q0.5 0.000 0.065 0.119** 0.037
(0.032) (0.050) (0.046) (0.059)
Q0.75 0.028 0.077* 0.086* 0.049
(0.032) (0.043) (0.048) (0.063)
Q0.9 ‐0.027 0.064 0.027 0.039
(0.023) (0.039) (0.049) (0.051)
Observations 3,711 3,605 3,576 3,500
Month FE YES YES YES YES
State FE YES YES YES YES
Controls YES YES YES YES
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1.
Column (1) is measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
Regression techniques, which are based on assumptions for large samples drawn from finite
populations, are often not suitable in the context of randomized experiments (Heß, 2017). The
uncertainty is in this case not coming from the sampled units observed, but from the fact that we can
only observe one of the potential outcomes, which is due to the treatment applied to the different
18
units (Athey & Imbens, 2017). One approach would be to take the randomization explicitly into account
and follow R.A. Fisher’s idea of statistical inference via permutation tests of treatment allocation
(Young, 2016). The idea is to assume uncertainty about the treatment allocation and compare the
actual treatment allocation to re‐randomizations. The results of this exercise are depicted in Table 9,
underscoring the robustness of the main results.
Table 9: Results from baseline estimation, model (2), with random‐inference based p‐values.
(1) (2) (3) (4) (5) (6) (7) (8)
VARIABLES ln(Cons. ln(Cons. ln(Cons. ln(Cons. ln(Cons. ln(Cons. Val.) ln(Cons. ln(Cons.
Num.) Num.) Quant.) Quant.) Val.) Cal.) cal.)
Treatment 0.0348** 0.0614** 0.0281 0.1374*** ‐0.0178 0.0812** 0.0189 0.0007
(0.0200) (0.0560) (0.1300) (0.0020) (0.2820) (0.0340) (0.4980) (0.9940)
Observations 3,995 3,995 3,995 3,995 3,995 3,995 3,995 3,995
R‐squared 0.0012 0.2744 0.0003 0.0805 0.0006 0.0725 0.0001 0.1232
State FE NO YES NO YES NO YES NO YES
Month FE NO YES NO YES NO YES NO YES
Controls NO YES NO YES NO YES NO YES
Controls YES YES YES YES YES YES YES YES
Interacted
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1
6. Discussion and Conclusion
The conflict in South Sudan displaced circa two million persons, constituting more than 15 percent of
the country’s population (UN OCHA, 2017). Moreover, the majority of population is living in extreme
poverty (The World Bank, 2016). Humanitarian crises like this ask for well targeted policy responses,
which address the population strata with the highest need first. This, however, is no arbitrary task as
aid allocation mechanisms might set adverse incentives to underreport. Even given the extreme
context, surveyed consumption levels indicate an unusually high share below subsistence levels.
For this purpose, this study assesses the effectiveness of a bundle of “honesty primes.” In line with our
hypothesis we find significant treatment effects, which cluster in lower (potentially underreported)
consumption quintiles. Moreover, effects are stronger for the number of consumption items than for
monetary consumption quantities, where the former are more susceptible to deliberate misreporting.
Furthermore, the significant treatment effects are driven mainly by the vulnerable IDP subpopulation,
which is more likely to be in need of foreign aid. Primes can, hence, help to improve data accuracy and
inform policy to develop durable solutions. However, results should be taken with a grain of salt as it
is not possible to compare the reported consumption outcomes to more objective consumption data.
Although the mortality rates among IDPs suggest that starvation is not happening systematically across
19
the country, the precarious situation calls for further scrutiny.18 Before adjusting poverty estimates a
thorough comparison with more “objective” data from administrative, anthropometric or
observational sources is needed. While this type of data was not available in IDP camps due to the
fragile context, future research could validate this finding in other settings.
Moreover, unbundling the primes in different treatment arms could help to shed light on the
underlying causal mechanisms. The underlying design of one treatment and control arm does not allow
for further disentangling the results. However, if classical measurement error would be affected, only
treatment effects of the primes should be uniform. In contrast, heterogenous effects across quantiles
suggest that the targeting of intentional misreporting via the appeal to honesty and moral prime would
be the driver of our results. In order to design more effective primes, disentangling the pathways and
trying different combinations could be a beneficial way forward. Our research can be considered as an
early step to employ priming for better targeted policy responses in challenging contexts, which might
not only be applicable in South Sudan, but also in other contexts facing humanitarian crises.
18
FEWS NET (2018).
20
References
Athey, S., & Imbens, G. W. (2017). The state of applied econometrics: Causality and policy evaluation.
Journal of Economic Perspectives, 31(2) , 3‐32.
Bachman, J. G., & O'Malley, P. M. (1984). Yea‐Saying, Nay‐Saying and Going to Extremes: Black‐White
Differences in Response Styles. Public Opinion Quarterly 48(2) , pp. 491‐509.
Baranov, V., Bhalotra, S., Biroli, P., & Maselko, J. (2017). Maternal Depression, Women’s
Empowerment, and Parental Investment: Evidence from a Large Randomized Control Trial. Bonn:
Institute for the Study of Labor (IZA).
Blattman, C., Jamison, J., Koroknay‐Palicz, T., Rodrigues, K., & Sheridan, M. (2016). Measuring the
measurement error: A method to qualitatively validate. Journal of Development Economics, 120 , 99‐
112.
Bräutigam, D. A., & Knack, S. (2004). Foreign Aid, Institutions, and Governance in Sub‐Saharan Africa .
Economic Development and Cultural Change , 255‐285.
Cilliers, J., Dube, O., & Siddiqi, B. (2015). The white‐man effect: How foreigner presence affects
behavior in experiments. Journal of Economic Behavior and Organization (118) , pp. 397‐414.
Cohn, A., Maréchal, M. A., & Noll, T. (2010). Bad Boys: How Criminal Identity Salience Affects Rule
Violation. Review of Economic Studies 82(4) , 1289–1308.
Cronbach, L. J. (1946). Response Sets and Validity. Educational and psychological measurement (6) ,
pp. 672‐83.
Deaton, A., & Paxson, C. (1998). Economies of Scale, Household Size, and the Demand for Food. Journal
of Political Economy, 106(5) , 897‐930.
Deaton, A., & Zaidi, S. (2002). Living Standards Measurement Study Working Paper No. 135.
Washington DC: The World Bank.
Devi, S. (2017, May 20). Famine in South Sudan. The Lancet, 389(10083) , 1967–1970.
Euler, M., Krishna, V., Schwarze, S., Siregar, H., & Qaim, M. (2017). Oil Palm Adoption, Household
Welfare, and Nutrition Among Smallholder Farmers in Indonesia. World Development , 219‐235.
FEWS NET. (2018). Famine Early Warning Systems Network. Retrieved March 22, 2018, from South
Sudan Food Security Outlook ‐ October 2017 to May 2018:
http://www.fews.net/sites/default/files/documents/reports/SOUTH%20SUDAN%20Food%20Security
%20Outlook_102017_0.pdf
21
Filmer, D., & Pritchett, L. H. (2001). Estimating wealth effects without expenditure data—or tears: an
application to educational enrollments in states of India. Demography , 115‐132.
Firpo, S., Fortin, N. M., & Lemieux, T. (2009). Unconditional quantile regressions. Econometrica (77:3)
, 953‐973.
Gilens, M., Sniderman, P. M., & Kuklinski, J. H. (1998). Affirmative Action and the Politics of
Realignment. British Journal of Political Science 28(1) , pp. 159‐83.
Gneezy, U. (2005). Deception: The Role of Consequences. American Economic Review 95(1) , pp. 384‐
94.
Hamilton, D. L. (1968). Personality Attributes Associated with Extreme Response Style. Psychological
Bulleting 69(3) , pp. 192‐203.
Haughton, J., & Khandker, S. R. (2009). Handbook on Poverty & Inequality. Washington DC: World Bank
Publications.
Heß, S. (2017). Randomization inference with Stata: A guide and software. Stata Journal , 630‐651.
Hurd, M. D. (1999). Anchoring and Acquiescence Bias in Measuring Assets in Household Surveys.
Journal of Risk and Uncertainty 19(1/3) , pp. 111‐36.
Koenker, R., & Bassett Jr, G. (1978). Regression quantiles. Econometrica, 46(1) , 33‐50.
Kuklinski, J. H. (1997). Racial Prejudice and Attitudes Toward Affirmative Action. American Journal of
Political Science 41(2) , pp. 402‐19.
Lanjouw, P., & Ravallion, M. (1995). Poverty and household size. The Economic Journal , 1415‐1434.
Lin, W., & Green, D. P. (2016). Standard Operating Procedures: A Safety Net for Pre‐Analysis Plans. PS:
Political Science & Politics , 495‐500.
Mazar, N., & Ariely, D. (2006). Dishonesty in Everyday Life and Its Policy Implications. Journal of Public
Policy & Marketing 25(1) , pp. 117‐26.
Mazar, N., Amir, O., & Ariely, D. (2008). The Dishonesty of Honest People: A Theory of Self‐Concept
Maintenance. Journal of Marketing Research 45(6) , pp. 633‐44.
McKenzie, D. J. (2005). Measuring inequality with asset indicators. Population Economics, 18(2) , 229–
260.
Pape, U., & Mistiaen, J. (2015). Measuring Household Consumption and Poverty in 60 Minutes: The
Mogadishu. Washington DC: World Bank / Proceedings of ABCA Conference 2015.
22
Rasinski, K. A., Visser, P. S., Zagatsky, M., & Rickett, E. M. (2005). Using implicit goal priming to improve
the quality of self‐report data. Journal of Experimental Social Psychology, 41(3) , 321‐327.
Ravallion, M., & Bidani, B. (1994). How robust is a poverty profile? The world bank economic review,
8(1) , 75‐102.
Rosenfeld, B., Imai, K., & Shapiro, J. N. (2016). An Empirical Validation Study of Popular Survey
Methodologies for Sensitive Questions. American Journal of Political Science 60(3) , pp. 783‐802.
Siegelman, L. (1981). Question‐Order Effects on Presidential Popularity. Public Opinion Qarterly , pp.
199‐207.
Stecklov, G., Weinreb, A., & Carletto, C. (2017). Can incentives improve survey data quality in
developing countries? Results from a field experiment in India. Series A ‐ Statistics in Society , 1‐24.
Talwar, V., Arruda, C., & Yachison, S. (2015). The effects of punishment and appeals for honesty on
children’s truth‐telling behavior. Journal of Experimental Child Psychology , 209‐217.
The World Bank. (2016, October 20). The World Bank in South Sudan. Retrieved February 24, 2018,
from South Sudan Overview: http://www.worldbank.org/en/country/southsudan/overview
The World Bank. (2015). World Development Report: Mind, Society, and Behavior. Washington DC: The
World Bank.
UN OCHA. (2017, December). United Nations Office for the Coordination of Humanitarian Affairs.
Retrieved February 24, 2018, from South Sudan: http://www.unocha.org/south‐sudan
United Nations. (2017). UN News. Retrieved March 15, 2018, from Famine declared in region of South
Sudan – UN: https://news.un.org/en/story/2017/02/551812‐famine‐declared‐region‐south‐sudan‐un
Van Den Broeck, G., & Maertens, M. (2017). Moving Up or Moving Out? Insights into Rural
Development and Poverty Reduction in Senegal. World Development , 95‐109.
Vinski, M., & Watter, S. (2012). Priming Honesty Reduces Subjective Bias in Self‐Report Measures of
Mind Wandering. Consciousness & Cognition 21(1) , 451‐55.
Voors, M. J., Nillesen, E. E., Verwimp, P., Bulte, E. H., Lensink, R., & Soest, D. P. (2012). Violent Conflict
and Behavior: A Field Experiment in Burundi. American Economic Review, 102(2) , 941‐964.
Warner, S. L. (1965). Randomized Response: A Survey Technique for Eliminating Evasive Answer Bias.
Journal of the American Statistical Association 60(309) , pp. 63‐69.
23
White, H. (1980). A Heteroskedasticity‐Consistent Covariance Matrix Estimator and a Direct Test for
Heteroskedasticity. Econometrica , 817‐38.
Young, A. (2016). Channeling fisher: Randomization tests and the statistical insignificance of seemingly
significant experimental results. London: London School of Economics, Working Paper.
24
Appendix:
Technical appendix
Construction of the caloric food intake measure
While monetary poverty lines are a key metric, when identifying the poor, caloric food poverty
headcounts are of equal relevance in our context. We create a food intake approximation by
multiplying the quantities of food items from the core consumption survey with average caloric values
of these products. The caloric intake c of household i is estimated as follows: caloric intake
∑ item ∗ calories ∗ quantity .
_
Forty‐three percent of household members are children, who naturally have lower consumption levels
than adults. We can account for this by using adult equivalents (AE) and rely on OECD scale, which
scales consumption of additional adults per household by factor 0.7 and of children by factor 0.5
(Haughton & Khandker, 2009).
1 0.7 1 0.5 (1)
1 0.7 1 0.5
Analytical Appendix
Caloric food poverty in Somalia
Using the same approach, we derive caloric food intake measures, which motivated the notion that
misreporting might be prevalent.
Figure 5: Calorie consumption ‐ IDPs Somalia
Source: Authors' calculations using HFS Somalia Wave 1.
25
Balance across survey strata
Table 10: Treatment distribution by survey strata.
Treatment for honesty primes
State/Camp Control Treatment Total
No. No. No.
JubaPOC 223 263 486
Wau 294 284 578
CRS
Bor 292 257 549
Bentiu 294 297 591
IDPCSS Juba POC1 – IDPCSS 976 965 1,941
Warrap 60 60 120
Northern Bahr el Ghazal 50 61 111
Western Bahr el Ghazal 62 58 120
HFS ‐ Wave 4 Lakes 50 54 104
Western Equatoria 54 50 104
Central Equatoria 38 40 78
Eastern Equatoria 74 70 144
Total 2467 2459 4926
Reaction to the honesty primes
An overwhelming majority of respondents answered in a positive manner to the moral prime. Less
than 10 percent of respondents answered that it is okay for the character in the fictional scenario to
lie to his friend. 19
[Moral prime]: I will give you a little scenario and would like to know what you think: John asks his good friend
Deng if he has some money that he can lend him to help him pay for medicine for his sick son. Deng has money
but was planning to buy cigarettes with it. He lies and tells John that he has none. Is it okay for Deng to lie to
John?
Percent N
Yes, it is okay for Deng to lie to John. 8.8 217
No, it is not okay for Deng to lie to John. 91.2 2,240
Total 100 2,457
This might be interpreted in two ways. First, it might point to a low fraction of respondents, who would
be willing to lie, which would reduce the potential of finding significant treatment effects. Second, it
could indicate that the prime would increase the propensity to report truthfully. However, as studies
suggest a high social desirability bias in the aid allocation setting (Stecklov, Weinreb, & Carletto, 2017;
19 The respondents, who find a lie inappropriate, have a higher share of male and unemployed household heads.
Moreover, IDPs have a significantly lower probability to find a lie acceptable. For a more detailed description of
characteristics between respondents, who affirmed and rejected the lie, please see the appendix.
26
Cilliers, Dube, & Siddiqi, 2015), implications should not be drawn too early and will be discussed in
subsequent sections.
Appropriateness of lying:
It is puzzling that IDPs have on average a lower probability to report that they would find a lie
appropriate when compared to non‐IDPs (see Table 11). This is in line with more pro‐social
preferences of conflict affected populations found by Voors, Nillesen, Verwimp, Bulte, Lensink, &
Soest (2012). However, this might be misleading, as the analysis of channels indicates that the
significant treatment effects are attributable to the IDP subsample, which seem to be more likely to
misreport.
Table 11: Distribution of respondents, who would find a lie (in‐)appropriate.
Yes, it is No, it is Overall (1) vs. (2),
okay for not okay p‐value
Deng to for Deng
lie to to lie to
John. John.
Household size 5.041 5.123 5.119 0.696
(0.228) (0.061) (0.059)
Gender of household head 0.327 0.456 0.445 0.000***
(0.032) (0.011) (0.010)
Literacy of household head 0.544 0.532 0.533 0.734
(0.034) (0.011) (0.010)
Household head completed some primary school 0.565 0.568 0.568 0.919
(0.034) (0.010) (0.010)
Is the household head employed 0.184 0.279 0.270 0.003***
(0.026) (0.009) (0.009)
Share of children in household 0.315 0.356 0.353 0.042*
(0.019) (0.006) (0.006)
Share of elderly in household 0.014 0.015 0.015 0.890
(0.007) (0.002) (0.002)
Level of Education of Household Head 2.060 1.967 1.975 0.205
(0.075) (0.022) (0.021)
non‐IDP Population 0.212 0.155 0.160 0.029*
(0.028) (0.008) (0.007)
N 217 2238 2455
Proportion 0.088 0.912 1.000
Robustness of results using an unconditional quantile regression
Conditional quantile regressions are sometimes critiqued on the ground as they would consider the
treatment effect conditional on the distribution and not on the individual ranking. Therefore, we also
replicate the main regressions within an unconditional quantile regression framework (Firpo, Fortin, &
Lemieux, 2009). Table 12depict the results of unconditional quantile regressions. The results indicate
27
a comparable pattern to Table 3. Especially, the quantities of consumption items and kilograms
experience positive treatment effects in lower quantiles. Although higher quantiles are affected as well
in Column (2), the largest effects can be found in the 10% quantile, which would be consistent with
the hypothesis of more honest answers among potentially underreporting households.
Table 12: Results from unconditional quantile regressions of different outcome variables.
(1) (2) (3) (4)
Outcome Variables ln(Cons.Num.) ln(Cons. Quant.) ln(Cons. Val.) ln(Cons. Cal.)
Q0.1 0.105** 0.259*** 0.076 0.134
(0.046) (0.090) (0.079) (0.145)
Q0.25 0.078** 0.210*** 0.169*** 0.075
(0.032) (0.067) (0.062) (0.077)
Q0.5 0.004 0.104** 0.118** 0.071
(0.035) (0.053) (0.056) (0.063)
Q0.75 ‐0.012 0.132** 0.067 0.025
(0.040) (0.066) (0.059) (0.089)
Q0.9 0.024 0.075 ‐0.003 0.062
(0.044) (0.087) (0.077) (0.119)
Observations 3,955 3,955 3,955 3,955
Month FE YES YES YES YES
State FE YES YES YES YES
Controls YES YES YES YES
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1‐Column (1) is
measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
28
Figure 6: Treatment effects across quantiles.
Note: Treatment effects and confidence intervals plotted for different quantiles.
Robustness of results in extended IDP and non‐IDP subsample
Table 13 reports the corresponding results of a quantile regression for an extended sample of IDPs and
Non‐IDPs. Results correspond to the previously found larger coefficients in the lower quintiles.
Coefficients are of similar size and the pattern remains qualitatively similar. However, statistical
significance is reduced in columns (3) and (4) with regard to the indicators that are measured with
more noise (e.g., monetary consumption values and caloric consumption).
Table 13: Quantile Regressions – extended sample IDPs and Non‐IDPs
(1) (2) (3) (4)
Outcome Variables ln(Cons. Num.) ln(Cons. Quant.) ln(Cons. Val.) ln(Cons. Cal.)
Q0.1 0.136*** 0.254*** 0.072 0.153
(0.049) (0.058) (0.067) (0.094)
Q0.25 0.085*** 0.123** 0.085 0.044
(0.031) (0.049) (0.052) (0.064)
Q0.5 0.024 0.088* 0.092** 0.037
(0.029) (0.049) (0.043) (0.053)
Q0.75 0.018 0.094** 0.052 0.028
29
(0.031) (0.042) (0.044) (0.052)
Q0.9 ‐0.019 0.058 ‐0.026 0.035
(0.024) (0.050) (0.048) (0.050)
Observations 4,735 4,735 4,735 4,735
Month FE YES YES YES YES
State FE YES YES YES YES
Controls YES YES YES YES
Interacted Controls YES YES YES YES
Robust standard errors in parentheses (White, 1980): *** p<0.01, ** p<0.05, * p<0.1.
Column (1) is measured on the household level. Column (2‐4) refer to per capita OECD adult equivalents.
Robustness to per capita instead of per adult equivalents
There is some uncertainty about the per adult equivalent scaling in the data. Ideally the distribution
might be estimated from more fine‐grained data on the intra‐household consumption distribution.
This is often not available, and, as Deaton & Zaidi (2002) summarize, “no satisfactory” scaling method
is identified so far. Therefore, the OECD scaling methodology is still frequently used (e.g., Euler,
Krishna, Schwarze, Siregar, & Qaim, 2017; Van Den Broeck & Maertens, 2017). Yet, potential concerns
with regard to the main results’ stability with regard to the scaling by adult equivalents remains.
Therefore, our outcome measures can be also constructed using agnostic per capita scales. In line with
the low consumption levels, the median of per capita calorie intake (1,589 kcal. per day) is well below
the recommended daily intake of 2,100 kcal (Ravallion & Bidani, 1994). Almost one‐third of
respondents (30.1 percent) report a calorie intake below the daily subsistence level of 1,200 kcal per
day. In contrast, several respondents report overly high consumption levels, which surpass
conventional consumption levels by far (> 4,000 kcal. per day). This supports previous evidence that
misreporting is prevalent. As with the number of consumption items, the graph indicates that there is
a slight shift in reported consumption among the treated regarding very low consumption levels.
Figure 7: Calory Consumption p.c.
30
Analogous to the estimation in Table 3, Table 14 depicts the results of a quantile regression using
agnostic per capita scales. The estimates indicate that the treatment effects remain stable and
respondents would report statistically significantly higher quantities in Column (1) and Column (2) if
treated. Hence, scaling does not explain our results, but is a factor to take into account, when
interpreting the outcomes.
Table 14: Results from quantile regressions of different outcome variables (pc scales).
(1) (2) (3)
Outcome Variables ln(Cons. Quant. p.c..) ln(Cons. Val. p.c.) ln(Cons. Cal. p.c.)
Q0.1 0.358*** 0.040 0.207
(0.087) (0.068) (0.135)
Q0.25 0.161*** 0.160** 0.076
(0.059) (0.053) (0.081)
Q0.5 0.124*** 0.079 0.073
(0.057) (0.054) (0.066)
Q0.75 0.050 0.055 0.021
(0.049) (0.054) (0.071)
Q0.9 0.057 ‐0.003 0.027
(0.063) (0.051) (0.081)
Observations 3,955 3,955 3,955
Month FE YES YES YES
State FE YES YES YES
Controls YES YES YES
Interacted Controls YES YES YES
Table 15: Results – full set of (interacted) controls
(1) (2) (3) (4)
Log of Log of Log of Log of
VARIABLES cons_num cons_q_kg_pc_ae cons_value_pc_ae cons_cal_pc_ae
Treatment = 1, Treatment 0.061* 0.137*** 0.081** 0.001
(0.033) (0.042) (0.039) (0.067)
Household size 0.033*** ‐0.024*** ‐0.044*** ‐0.105***
(0.005) (0.005) (0.005) (0.008)
Gender of household head 0.009 0.043* 0.022 ‐0.047
31
(0.018) (0.026) (0.025) (0.039)
Share of children in household 0.106** 0.243*** 0.190*** 0.027
(0.046) (0.053) (0.054) (0.085)
1st component of asset PCA 0.026*** 0.013 0.022*** 0.039***
(0.005) (0.008) (0.008) (0.012)
1.treat#c.hhsize ‐0.008 ‐0.015** ‐0.009 0.006
(0.006) (0.007) (0.007) (0.011)
1.hhh_gender#0b.treat ‐0.020 ‐0.007 ‐0.008 ‐0.059
(0.028) (0.036) (0.036) (0.054)
1.treat#c.pchild ‐0.001 ‐0.066 ‐0.107 ‐0.016
(0.059) (0.074) (0.073) (0.116)
1.treat#c.Asset_PCA_1 ‐0.017** ‐0.003 ‐0.007 ‐0.003
(0.007) (0.011) (0.010) (0.016)
Observations 3,955 3,955 3,955 3,955
R‐squared 0.274 0.073 0.080 0.123
State FE YES YES YES YES
Month FE YES YES YES YES
Controls YES YES YES YES
Robust standard errors in
parentheses
*** p<0.01, ** p<0.05, * p<0.1
Household size and purchasing prices per kilo price
In order to find out if the data bear out the pattern that larger households pay lower prices, e.g., due
to bulk purchasing, we regress the log of the reported price on household size, state, month and
consumption good specific fixed effects.
ln price ∝ β hhsize γ δ θ ε (5)
ln price ∝ β hhsize γ δ θ ε
The results are depicted in Table 16 and indicate a negative average correlation. This supports the
choice of interacting unbalanced controls with the treatment indicator.
Table 16: Correlation of household size and purchasing prices per kilo.
(1)
VARIABLES
lnprice
Household size ‐0.003**
32
(0.001)
Observations 24,409
R‐squared 0.548
State FE YES
Month FE YES
Item FE Yes
Robust standard errors in parentheses (White, 1980):
*** p<0.01, ** p<0.05, * p<0.1
Consumption shares of IDP and non‐IDP populations
Figure 8 describes the consumption shares of IDPs and non‐IDPs. While the figure shows that the diet
of IDPs is slightly less diverse than the diet of non‐IDPs, it is also revealed that large shares of IDP
budgets are spent on goods that offer a high caloric intake per SSP spent, e.g., sorghum and cooking
oil. This also leads to a counterintuitive pattern found in the data, where IDPs consume less than non‐
IDPs in terms of monetary value, but more in terms of caloric food intake.
Figure 8: Consumption Shares (SSP values)
Note: The figure lists the consumption shares of items, which constitute at least 1% of household consumption.
33
Poverty & Equity Global Practice Working Papers
(Since July 2014)
The Poverty & Equity Global Practice Working Paper Series disseminates the findings of work in progress
to encourage the exchange of ideas about development issues. An objective of the series is to get the
findings out quickly, even if the presentations are less than fully polished. The papers carry the names of
the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in
this paper are entirely those of the authors. They do not necessarily represent the views of the
International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or
those of the Executive Directors of the World Bank or the governments they represent.
This series is co‐published with the World Bank Policy Research Working Papers (DECOS). It is part of a
larger effort by the World Bank to provide open access to its research and contribute to development
policy discussions around the world.
For the latest paper, visit our GP’s intranet at http://POVERTY.
1 Estimating poverty in the absence of consumption data: the case of Liberia
Dabalen, A. L., Graham, E., Himelein, K., Mungai, R., September 2014
2 Female labor participation in the Arab world: some evidence from panel data in Morocco
Barry, A. G., Guennouni, J., Verme, P., September 2014
3 Should income inequality be reduced and who should benefit? redistributive preferences in Europe
and Central Asia
Cojocaru, A., Diagne, M. F., November 2014
4 Rent imputation for welfare measurement: a review of methodologies and empirical findings
Balcazar Salazar, C. F., Ceriani, L., Olivieri, S., Ranzani, M., November 2014
5 Can agricultural households farm their way out of poverty?
Oseni, G., McGee, K., Dabalen, A., November 2014
6 Durable goods and poverty measurement
Amendola, N., Vecchi, G., November 2014
7 Inequality stagnation in Latin America in the aftermath of the global financial crisis
Cord, L., Barriga Cabanillas, O., Lucchetti, L., Rodriguez‐Castelan, C., Sousa, L. D., Valderrama, D.
December 2014
8 Born with a silver spoon: inequality in educational achievement across the world
Balcazar Salazar, C. F., Narayan, A., Tiwari, S., January 2015
Updated on August 2018 by POV GP KL Team | 1
9 Long‐run effects of democracy on income inequality: evidence from repeated cross‐sections
Balcazar Salazar,C. F., January 2015
10 Living on the edge: vulnerability to poverty and public transfers in Mexico
Ortiz‐Juarez, E., Rodriguez‐Castelan, C., De La Fuente, A., January 2015
11 Moldova: a story of upward economic mobility
Davalos, M. E., Meyer, M., January 2015
12 Broken gears: the value added of higher education on teachers' academic achievement
Balcazar Salazar, C. F., Nopo, H., January 2015
13 Can we measure resilience? a proposed method and evidence from countries in the Sahel
Alfani, F., Dabalen, A. L., Fisker, P., Molini, V., January 2015
14 Vulnerability to malnutrition in the West African Sahel
Alfani, F., Dabalen, A. L., Fisker, P., Molini, V., January 2015
15 Economic mobility in Europe and Central Asia: exploring patterns and uncovering puzzles
Cancho, C., Davalos, M. E., Demarchi, G., Meyer, M., Sanchez Paramo, C., January 2015
16 Managing risk with insurance and savings: experimental evidence for male and female farm
managers in the Sahel
Delavallade, C., Dizon, F., Hill, R., Petraud, J. P., el., January 2015
17 Gone with the storm: rainfall shocks and household well‐being in Guatemala
Baez, J. E., Lucchetti, L., Genoni, M. E., Salazar, M., January 2015
18 Handling the weather: insurance, savings, and credit in West Africa
De Nicola, F., February 2015
19 The distributional impact of fiscal policy in South Africa
Inchauste Comboni, M. G., Lustig, N., Maboshe, M., Purfield, C., Woolard, I., March 2015
20 Interviewer effects in subjective survey questions: evidence from Timor‐Leste
Himelein, K., March 2015
21 No condition is permanent: middle class in Nigeria in the last decade
Corral Rodas, P. A., Molini, V., Oseni, G. O., March 2015
22 An evaluation of the 2014 subsidy reforms in Morocco and a simulation of further reforms
Verme, P., El Massnaoui, K., March 2015
Updated on August 2018 by POV GP KL Team | 2
23 The quest for subsidy reforms in Libya
Araar, A., Choueiri, N., Verme, P., March 2015
24 The (non‐) effect of violence on education: evidence from the "war on drugs" in Mexico
Márquez‐Padilla, F., Pérez‐Arce, F., Rodriguez Castelan, C., April 2015
25 “Missing girls” in the south Caucasus countries: trends, possible causes, and policy options
Das Gupta, M., April 2015
26 Measuring inequality from top to bottom
Diaz Bazan, T. V., April 2015
27 Are we confusing poverty with preferences?
Van Den Boom, B., Halsema, A., Molini, V., April 2015
28 Socioeconomic impact of the crisis in north Mali on displaced people (Available in French)
Etang Ndip, A., Hoogeveen, J. G., Lendorfer, J., June 2015
29 Data deprivation: another deprivation to end
Serajuddin, U., Uematsu, H., Wieser, C., Yoshida, N., Dabalen, A., April 2015
30 The local socioeconomic effects of gold mining: evidence from Ghana
Chuhan-Pole, P., Dabalen, A., Kotsadam, A., Sanoh, A., Tolonen, A.K., April 2015
31 Inequality of outcomes and inequality of opportunity in Tanzania
Belghith, N. B. H., Zeufack, A. G., May 2015
32 How unfair is the inequality of wage earnings in Russia? estimates from panel data
Tiwari, S., Lara Ibarra, G., Narayan, A., June 2015
33 Fertility transition in Turkey—who is most at risk of deciding against child arrival?
Greulich, A., Dasre, A., Inan, C., June 2015
34 The socioeconomic impacts of energy reform in Tunisia: a simulation approach
Cuesta Leiva, J. A., El Lahga, A., Lara Ibarra, G., June 2015
35 Energy subsidies reform in Jordan: welfare implications of different scenarios
Atamanov, A., Jellema, J. R., Serajuddin, U., June 2015
36 How costly are labor gender gaps? estimates for the Balkans and Turkey
Cuberes, D., Teignier, M., June 2015
37 Subjective well‐being across the lifespan in Europe and Central Asia
Bauer, J. M., Munoz Boudet, A. M., Levin, V., Nie, P., Sousa‐Poza, A., July 2015
Updated on August 2018 by POV GP KL Team | 3
38 Lower bounds on inequality of opportunity and measurement error
Balcazar Salazar, C. F., July 2015
39 A decade of declining earnings inequality in the Russian Federation
Posadas, J., Calvo, P. A., Lopez‐Calva, L.‐F., August 2015
40 Gender gap in pay in the Russian Federation: twenty years later, still a concern
Atencio, A., Posadas, J., August 2015
41 Job opportunities along the rural‐urban gradation and female labor force participation in India
Chatterjee, U., Rama, M. G., Murgai, R., September 2015
42 Multidimensional poverty in Ethiopia: changes in overlapping deprivations
Yigezu, B., Ambel, A. A., Mehta, P. A., September 2015
43 Are public libraries improving quality of education? when the provision of public goods is not enough
Rodriguez Lesmes, P. A., Valderrama Gonzalez, D., Trujillo, J. D., September 2015
44 Understanding poverty reduction in Sri Lanka: evidence from 2002 to 2012/13
Inchauste Comboni, M. G., Ceriani, L., Olivieri, S. D., October 2015
45 A global count of the extreme poor in 2012: data issues, methodology and initial results
Ferreira, F.H.G., Chen, S., Dabalen, A. L., Dikhanov, Y. M., Hamadeh, N., Jolliffe, D. M., Narayan, A.,
Prydz, E. B., Revenga, A. L., Sangraula, P., Serajuddin, U., Yoshida, N., October 2015
46 Exploring the sources of downward bias in measuring inequality of opportunity
Lara Ibarra, G., Martinez Cruz, A. L., October 2015
47 Women’s police stations and domestic violence: evidence from Brazil
Perova, E., Reynolds, S., November 2015
48 From demographic dividend to demographic burden? regional trends of population aging in Russia
Matytsin, M., Moorty, L. M., Richter, K., November 2015
49 Hub‐periphery development pattern and inclusive growth: case study of Guangdong province
Luo, X., Zhu, N., December 2015
50 Unpacking the MPI: a decomposition approach of changes in multidimensional poverty headcounts
Rodriguez Castelan, C., Trujillo, J. D., Pérez Pérez, J. E., Valderrama, D., December 2015
51 The poverty effects of market concentration
Rodriguez Castelan, C., December 2015
52 Can a small social pension promote labor force participation? evidence from the Colombia Mayor
program
Pfutze, T., Rodriguez Castelan, C., December 2015
Updated on August 2018 by POV GP KL Team | 4
53 Why so gloomy? perceptions of economic mobility in Europe and Central Asia
Davalos, M. E., Cancho, C. A., Sanchez, C., December 2015
54 Tenure security premium in informal housing markets: a spatial hedonic analysis
Nakamura, S., December 2015
55 Earnings premiums and penalties for self‐employment and informal employees around the world
Newhouse, D. L., Mossaad, N., Gindling, T. H., January 2016
56 How equitable is access to finance in turkey? evidence from the latest global FINDEX
Yang, J., Azevedo, J. P. W. D., Inan, O. K., January 2016
57 What are the impacts of Syrian refugees on host community welfare in Turkey? a subnational
poverty analysis
Yang, J., Azevedo, J. P. W. D., Inan, O. K., January 2016
58 Declining wages for college‐educated workers in Mexico: are younger or older cohorts hurt the
most?
Lustig, N., Campos‐Vazquez, R. M., Lopez‐Calva, L.‐F., January 2016
59 Sifting through the Data: labor markets in Haiti through a turbulent decade (2001‐2012)
Rodella, A.‐S., Scot, T., February 2016
60 Drought and retribution: evidence from a large‐scale rainfall‐indexed insurance program in Mexico
Fuchs Tarlovsky, Alan., Wolff, H., February 2016
61 Prices and welfare
Verme, P., Araar, A., February 2016
62 Losing the gains of the past: the welfare and distributional impacts of the twin crises in Iraq 2014
Olivieri, S. D., Krishnan, N., February 2016
63 Growth, urbanization, and poverty reduction in India
Ravallion, M., Murgai, R., Datt, G., February 2016
64 Why did poverty decline in India? a nonparametric decomposition exercise
Murgai, R., Balcazar Salazar, C. F., Narayan, A., Desai, S., March 2016
65 Robustness of shared prosperity estimates: how different methodological choices matter
Uematsu, H., Atamanov, A., Dewina, R., Nguyen, M. C., Azevedo, J. P. W. D., Wieser, C., Yoshida, N.,
March 2016
66 Is random forest a superior methodology for predicting poverty? an empirical assessment
Stender, N., Pave Sohnesen, T., March 2016
67 When do gender wage differences emerge? a study of Azerbaijan's labor market
Tiongson, E. H. R., Pastore, F., Sattar, S., March 2016
Updated on August 2018 by POV GP KL Team | 5
68 Second‐stage sampling for conflict areas: methods and implications
Eckman, S., Murray, S., Himelein, K., Bauer, J., March 2016
69 Measuring poverty in Latin America and the Caribbean: methodological considerations when
estimating an empirical regional poverty line
Gasparini, L. C., April 2016
70 Looking back on two decades of poverty and well‐being in India
Murgai, R., Narayan, A., April 2016
71 Is living in African cities expensive?
Yamanaka, M., Dikhanov, Y. M., Rissanen, M. O., Harati, R., Nakamura, S., Lall, S. V., Hamadeh, N., Vigil
Oliver, W., April 2016
72 Ageing and family solidarity in Europe: patterns and driving factors of intergenerational support
Albertini, M., Sinha, N., May 2016
73 Crime and persistent punishment: a long‐run perspective on the links between violence and chronic
poverty in Mexico
Rodriguez Castelan, C., Martinez‐Cruz, A. L., Lucchetti, L. R., Valderrama Gonzalez, D., Castaneda
Aguilar, R. A., Garriga, S., June 2016
74 Should I stay or should I go? internal migration and household welfare in Ghana
Molini, V., Pavelesku, D., Ranzani, M., July 2016
75 Subsidy reforms in the Middle East and North Africa Region: a review
Verme, P., July 2016
76 A comparative analysis of subsidy reforms in the Middle East and North Africa Region
Verme, P., Araar, A., July 2016
77 All that glitters is not gold: polarization amid poverty reduction in Ghana
Clementi, F., Molini, V., Schettino, F., July 2016
78 Vulnerability to Poverty in rural Malawi
Mccarthy, N., Brubaker, J., De La Fuente, A., July 2016
79 The distributional impact of taxes and transfers in Poland
Goraus Tanska, K. M., Inchauste Comboni, M. G., August 2016
80 Estimating poverty rates in target populations: an assessment of the simple poverty scorecard and
alternative approaches
Vinha, K., Rebolledo Dellepiane, M. A., Skoufias, E., Diamond, A., Gill, M., Xu, Y., August 2016
Updated on August 2018 by POV GP KL Team | 6
81 Synergies in child nutrition: interactions of food security, health and environment, and child care
Skoufias, E., August 2016
82 Understanding the dynamics of labor income inequality in Latin America
Rodriguez Castelan, C., Lustig, N., Valderrama, D., Lopez‐Calva, L.‐F., August 2016
83 Mobility and pathways to the middle class in Nepal
Tiwari, S., Balcazar Salazar, C. F., Shidiq, A. R., September 2016
84 Constructing robust poverty trends in the Islamic Republic of Iran: 2008‐14
Salehi Isfahani, D., Atamanov, A., Mostafavi, M.‐H., Vishwanath, T., September 2016
85 Who are the poor in the developing world?
Newhouse, D. L., Uematsu, H., Doan, D. T. T., Nguyen, M. C., Azevedo, J. P. W. D., Castaneda Aguilar, R.
A., October 2016
86 New estimates of extreme poverty for children
Newhouse, D. L., Suarez Becerra, P., Evans, M. C., October 2016
87 Shedding light: understanding energy efficiency and electricity reliability
Carranza, E., Meeks, R., November 2016
88 Heterogeneous returns to income diversification: evidence from Nigeria
Siwatu, G. O., Corral Rodas, P. A., Bertoni, E., Molini, V., November 2016
89 How liberal is Nepal's liberal grade promotion policy?
Sharma, D., November 2016
90 Pro-growth equity: a policy framework for the twin goals
Lopez-Calva, L. F., Rodriguez Castelan, C., November 2016
91 CPI bias and its implications for poverty reduction in Africa
Dabalen, A. L., Gaddis, I., Nguyen, N. T. V., December 2016
92 Building an ex ante simulation model for estimating the capacity impact, benefit incidence, and cost
effectiveness of child care subsidies: an application using provider‐level data from Turkey
Aran, M. A., Munoz Boudet, A., Aktakke, N., December 2016
93 Vulnerability to drought and food price shocks: evidence from Ethiopia
Porter, C., Hill, R., December 2016
94 Job quality and poverty in Latin America
Rodriguez Castelan, C., Mann, C. R., Brummund, P., December 2016
95 With a little help: shocks, agricultural income, and welfare in Uganda
Mejia‐Mantilla, C., Hill, R., January 2017
Updated on August 2018 by POV GP KL Team | 7
96 The impact of fiscal policy on inequality and poverty in Chile
Martinez Aguilar, S. N., Fuchs Tarlovsky, A., Ortiz‐Juarez, E., Del Carmen Hasbun, G. E., January 2017
97 Conditionality as targeting? participation and distributional effects of conditional cash transfers
Rodriguez Castelan, C., January 2017
98 How is the slowdown affecting households in Latin America and the Caribbean?
Reyes, G. J., Calvo‐Gonzalez, O., Sousa, L. D. C., Castaneda Aguilar, R. A., Farfan Bertran, M. G., January
2017
99 Are tobacco taxes really regressive? evidence from Chile
Fuchs Tarlovsky, A., Meneses, F. J., March 2017
100 Design of a multi‐stage stratified sample for poverty and welfare monitoring with multiple
objectives: a
Bangladesh case study
Yanez Pagans, M., Roy, D., Yoshida, N., Ahmed, F., March 2017
101 For India's rural poor, growing towns matter more than growing cities
Murgai, R., Ravallion, M., Datt, G., Gibson, J., March 2017
102 Leaving, staying, or coming back? migration decisions during the northern Mali conflict
Hoogeveen, J. G., Sansone, D., Rossi, M., March 2017
103 Arithmetics and Politics of Domestic Resource Mobilization
Bolch, K. B., Ceriani, L., Lopez‐Calva, L.‐F., April 2017
104 Can Public Works Programs Reduce Youth Crime? Evidence from Papua New Guinea’s Urban Youth
Employment Project
Oleksiy I., Darian N., David N., Sonya S., April 2017
105 Is Poverty in Africa Mostly Chronic or Transient? Evidence from Synthetic Panel Data
Dang, H.‐A. H., Dabalen, A. L., April 2017
106 To Sew or Not to Sew? Assessing the Welfare Effects of the Garment Industry in Cambodia
Mejía‐Mantilla, C., Woldemichael, M. T., May 2017
107 Perceptions of distributive justice in Latin America during a period of falling inequality
Reyes, G. J., Gasparini, L. C., May 2017
108 How do women fare in rural non‐farm economy?
Fuje, H. N., May 2017
109 Rural Non‐Farm Employment and Household Welfare: Evidence from Malawi
Adjognon, G. S., Liverpool‐Tasie, S. L., De La Fuente, A., Benfica, R. M., May 2017
Updated on August 2018 by POV GP KL Team | 8
110 Multidimensional Poverty in the Philippines, 2004‐13: Do Choices for Weighting, Identification and
Aggregation Matter?
Datt, G., June 2017
111 But … what is the poverty rate today? testing poverty nowcasting methods in Latin America and the
Caribbean
Caruso, G. D., Lucchetti, L. R., Malasquez, E., Scot, T., Castaneda, R. A., June 2017
112 Estimating the Welfare Costs of Reforming the Iraq Public Distribution System: A Mixed Demand
Approach
Krishnan, N., Olivieri, S., Ramadan, R., June 2017
113 Beyond Income Poverty: Nonmonetary Dimensions of Poverty in Uganda
Etang Ndip, A., Tsimpo, C., June 2017
114 Education and Health Services in Uganda: Quality of Inputs, User Satisfaction, and Community
Welfare Levels
Tsimpo Nkengne, C., Etang Ndip, A., Wodon, Q. T., June 2017
115 Rental Regulation and Its Consequences on Measures of Well‐Being in the Arab Republic of Egypt
Lara Ibarra, G., Mendiratta, V., Vishwanath, T., July 2017
116 The Poverty Implications of Alternative Tax Reforms: Results from a Numerical Application to
Pakistan
Feltenstein, A., Mejia‐Mantilla, C., Newhouse, D. L., Sedrakyan, G., August 2017
117 Tracing Back the Weather Origins of Human Welfare: Evidence from Mozambique?
Baez Ramirez, J. E., Caruso, G. D., Niu, C., August 2017
118 Many Faces of Deprivation: A multidimensional approach to poverty in Armenia
Martirosova, D., Inan, O. K., Meyer, M., Sinha, N., August 2017
119 Natural Disaster Damage Indices Based on Remotely Sensed Data: An Application to Indonesia
Skoufias, E., Strobl, E., Tveit, T. B., September 2017
120 The Distributional Impact of Taxes and Social Spending in Croatia
Inchauste Comboni, M. G., Rubil, I., October 2017
121 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine
Fuchs, A., Meneses, F. September 2017
122 Fiscal Incidence in Belarus: A Commitment to Equity Analysis
Bornukova, K., Shymanovich, G., Chubrik, A., October 2017
Updated on August 2018 by POV GP KL Team | 9
123 Who escaped poverty and who was left behind? a non‐parametric approach to explore welfare
dynamics using cross‐sections
Lucchetti, L. R., October 2017
124 Learning the impact of financial education when take-up is low
Lara Ibarra, G., Mckenzie, D. J., Ruiz Ortega, C., November 2017
125 Putting Your Money Where Your Mouth Is Geographic Targeting of World Bank Projects
to the Bottom 40 Percent
Öhler, H., Negre, M., Smets, L., Massari, R., Bogetić, Z., November 2017
126 The impact of fiscal policy on inequality and poverty in Zambia
De La Fuente, A., Rosales, M., Jellema, J. R., November 2017
127 The Whys of Social Exclusion: Insights from Behavioral Economics
Hoff, K., Walsh, J. S., December 2017
128 Mission and the bottom line: performance incentives in a multi-goal organization
Gine, X., Mansuri, G., Shrestha, S. A., December 2017
129 Mobile Infrastructure and Rural Business Enterprises Evidence from Sim Registration Mandate in
Niger
Annan, F., Sanoh, A., December 2017
130 Poverty from Space: Using High-Resolution Satellite Imagery for estimating Economic Well-Being
Engstrom, R., Hersh, J., Newhouse, D., December 2017
131 Winners Never Quit, Quitters Never Grow: Using Text Mining to measure Policy Volatility and its Link
with Long-Term Growth in Latin America
Calvo-Gonzalez, O., Eizmendi, A., Reyes, G., January 2018
132 The Changing Way Governments talk about Poverty and Inequality: Evidence from two Centuries of
Latin American Presidential Speeches
Calvo-Gonzalez, O., Eizmendi, A., Reyes, G., January 2018
133 Tobacco Price Elasticity and Tax Progressivity In Moldova
Fuchs, A., Meneses, F., February 2018
134 Informal Sector Heterogeneity and Income Inequality: Evidence from the Democratic Republic of
Congo
Adoho, F., Doumbia, D., February 2018
135 South Caucasus in Motion: Economic and Social Mobility in Armenia, Azerbaijan and Georgia
Tiwari, S., Cancho, C., Meyer, M., February 2018
Updated on August 2018 by POV GP KL Team | 10
136 Human Capital Outflows: Selection into Migration from the Northern Triangle
Del Carmen, G., Sousa, L., February 2018
137 Urban Transport Infrastructure and Household Welfare: Evidence from Colombia
Pfutze, T., Rodriguez-Castelan, C., Valderrama-Gonzalez, D., February 2018
138 Hit and Run? Income Shocks and School Dropouts in Latin America
Cerutti, P., Crivellaro, E., Reyes, G., Sousa, L., February 2018
139 Decentralization and Redistribution Irrigation Reform in Pakistan’s Indus Basin
Jacoby, H.G., Mansuri, G., Fatima, F., February 2018
140 Governing the Commons? Water and Power in Pakistan’s Indus Basin
Jacoby, H.G., Mansuri, G., February 2018
141 The State of Jobs in Post-Conflict Areas of Sri Lanka
Newhouse, D., Silwal, A. R., February 2018
142 “If it’s already tough, imagine for me…” A Qualitative Perspective on Youth Out of School
and Out of Work in Brazil
Machado, A.L., Muller, M., March 2018
143 The reallocation of district-level spending and natural disasters: evidence from Indonesia
Skoufias, E., Strobl, E., Tveit, T. B., March 2018
144 Gender Differences in Poverty and Household Composition through the Life-cycle A Global
Perspective
Munoz, A. M., Buitrago, P., Leroy de la Briere, B., Newhouse, D., Rubiano, E., Scott, K., Suarez-Becerra,
P., March 2018
145 Analysis of the Mismatch between Tanzania Household Budget Survey and National Panel Survey
Data in Poverty & Inequality Levels and Trends
Fuchs, A., Del Carmen, G., Kechia Mukong, A., March 2018
146 Long-Run Impacts of Increasing Tobacco Taxes: Evidence from South Africa
Hassine Belghith, N.B., Lopera, M. A., Etang Ndip, A., Karamba, W., March 2018
147 The Distributional Impact of the Fiscal System in Albania
Davalos, M., Robayo-Abril, M., Shehaj, E., Gjika, A., March 2018
148 Analysis Growth, Safety Nets and Poverty: Assessing Progress in Ethiopia from 1996 to 2011
Vargas Hill, R., Tsehaye, E., March 2018
149 The Economics of the Gender Wage Gap in Armenia
Rodriguez-Chamussy, L., Sinha, N., Atencio, A., April 2018
Updated on August 2018 by POV GP KL Team | 11
150 Do Demographics Matter for African Child Poverty?
Batana, Y., Cockburn, J., May 2018
151 Household Expenditure and Poverty Measures in 60 Minutes: A New Approach with Results from
Mogadishu
Pape, U., Mistiaen, J., May 2018
152 Inequality of Opportunity in South Caucasus
Fuchs, A., Tiwari, S., Rizal Shidiq, A., May 2018
153 Welfare Dynamics in Colombia: Results from Synthetic Panels
Balcazar, C.F., Dang, H-A., Malasquez, E., Olivieri, S., Pico, J., May 2018
154 Social Protection in Niger: What Have Shocks and Time Got to Say?
Annan, F., Sanoh, A., May 2018
155 Quantifying the impacts of capturing territory from the government in the Republic of Yemen
Tandon, S., May 2018
156 The Road to Recovery: The Role of Poverty in the Exposure, Vulnerability and Resilience to Floods in
Accra
Erman, A., Motte, E., Goyal, R., Asare, A., Takamatsu, S., Chen, X., Malgioglio, S., Skinner, A., Yoshida,
N., Hallegatte, S., June 2018
157 Small Area Estimation of Poverty under Structural Change
Lange, S., Pape, U., Pütz, P., June 2018
158 The Devil Is in the Details; Growth, Polarization, and Poverty Reduction in Africa in the Past Two
Decades
F. Clementi F., Fabiani, M., Molini, V., June 2018
159 Impact of Conflict on Adolescent Girls in South Sudan
Pape, U., Phipps, V., July 2018
160 Urbanization in Kazakhstan; Desirable Cities, Unaffordable Housing, and the Missing Rental Market
Seitz, W., July 2018
161 SInequality in Earnings and Adverse Shocks in Early Adulthood
Tien, B., Adoho, F., August 2018
162 Eliciting Accurate Responses to Consumption Questions among IDPs in South Sudan Using “Honesty
Primes”
Kaplan, L., Pale, U., Walsh, J., Auguste 2018
Updated on August 2018 by POV GP KL Team | 12
163 What Can We (Machine) Learn about Welfare Dynamics from Cross-Sectional Data?
Lucchetti, L., August 2018
164 Infrastructure, Value Chains, and Economic Upgrades
Luo, X., Xu, X., August 2018
165 The Distributional Effects of Tobacco Taxation; The Evidence of White and Clove Cigarettes in
Indonesia
Fuchs, A., Del Carmen, G., August 2018
For the latest and sortable directory,
available on the Poverty & Equity GP intranet site. http://POVERTY
WWW.WORLDBANK.ORG/POVERTY
Updated on August 2018 by POV GP KL Team | 13