IFPRI Discussion Paper 02219 

December 2023 

COVID-19 and Livelihoods in Rural Guatemala  

Lessons from a Long Term Assessment and the Path to Recovery  

Maria Lucia Berrospi 

 

Francisco Ceballos 

 

Manuel A. Hernandez  

 

Cynthia Paz  

 

Markets, Trade, and Institutions Unit 



INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE 

The International Food Policy Research Institute (IFPRI), a CGIAR Research Center established in 1975, 

provides research-based policy solutions to sustainably reduce poverty and end hunger and malnutrition. 

IFPRI’s strategic research aims to foster a climate-resilient and sustainable food supply; promote healthy 

diets and nutrition for all; build inclusive and efficient markets, trade systems, and food industries; 

transform agricultural and rural economies; and strengthen institutions and governance. Gender is 

integrated in all the Institute’s work. Partnerships, communications, capacity strengthening, and data and 

knowledge management are essential components to translate IFPRI’s research from action to impact. The 

Institute’s regional and country programs play a critical role in responding to demand for food policy 

research and in delivering holistic support for country-led development. IFPRI collaborates with partners 

around the world.  

 

AUTHORS 

Maria Lucia Berrospi (m.l.berrospi@cgiar.org) is a research assistant in the Markets, Trade, and 

Institutions Unit at the International Food Policy Research Institute (IFPRI).  

Francisco Ceballos (f.ceballos@cgiar.org) is a research fellow in the Markets, Trade, and Institutions Unit 

at IFPRI. 

Manuel A. Hernandez (m.a.hernandez@cgiar.org) is a senior research fellow in the Markets, Trade and 

Institutions Unit at IFPRI. 

Cynthia Paz (c.paz@cgiar.org) is a research analyst in the Markets, Trade, and Institutions Unit at IFPRI.  

 

 

 

 

 

 

Notices  

 

1 IFPRI Discussion Papers contain preliminary material and research results and are circulated in order to stimulate discussion and 

critical comment. They have not been subject to a formal external review via IFPRI’s Publications Review Committee. Any opinions 

stated herein are those of the author(s) and are not necessarily representative of or endorsed by IFPRI.  

 
2 The boundaries and names shown and the designations used on the map(s) herein do not imply official endorsement or acceptance 

by the International Food Policy Research Institute (IFPRI) or its partners and contributors. 

 
3 Copyright remains with the authors. The authors are free to proceed, without further IFPRI permission, to publish this paper, or any 

revised version of it, in outlets such as journals, books, and other publications. 

mailto:m.l.berrospi@cgiar.org
mailto:f.ceballos@cgiar.org
mailto:m.a.hernandez@cgiar.org
mailto:c.paz@cgiar.org


 

ABSTRACT 

 

The COVID-19 pandemic has had profound effects on livelihoods and food security across rural 

populations worldwide. This study offers a long-term assessment of the impacts of the pandemic and 

the path to recovery among smallholder agricultural households in the Western Highlands of 

Guatemala. We rely on a unique longitudinal survey of 1,262 households collected over four survey 

rounds between 2019 and 2022. The results show substantial recoveries in incomes, food security, and 

dietary diversity in the region by mid-2022 compared to 2020, but at levels still worse than pre-

pandemic ones. There is also a sustained increase in the intention to emigrate. The households that were 

initially more affected in terms of food security and nutrition but recovered faster include those located 

in one (San Marcos) of the three departments and families living above the poverty line, while 

smallholders affected by the ETA and IOTA tropical storms, non-coffee producers, and indigenous 

populations have taken longer to recover. In addition, we provide quantitative estimates for a subsample 

of households interviewed during a fifth survey round at the end of 2022, showing an average decline 

of about 16 percent in total household income three years after the start of the pandemic, mainly driven 

by a decrease in agricultural income, combined with a 26 percent increase in expenditures and an 

important surge in indebtedness. Overall, the study offers valuable lessons regarding the recovery of 

vulnerable households following a major global crisis and in a context of additional shocks, remarking 

the importance of continue monitoring the situation of vulnerable households, especially those exposed 

to recurrent (weather) shocks that also have a more exhausted portfolio of coping mechanisms & 

express a higher willingness to emigrate.   

 

 

Keywords: COVID-19, income, food security, dietary diversity, migration, expenditures, debt, long-

term effects, recovery, external shocks, agricultural households, rural Guatemala 

JEL codes:  I30, O15, Q12, R20 



ACKNOWLEDGMENTS 

 

We thank Francisco Olivet and the team of O&M Estudios y Proyectos for their valuable collaboration in 

collecting the data over the multiple survey rounds, and Carlos Rivera for his technological support. This 

publication was made possible through support provided by Feed the Future through the U.S. Agency for 

International Development. The contents and conclusions of the study are the responsibility of the authors 

and do not necessarily reflect the views of the U.S. Agency for International Development or the 

International Food Policy Research Institute. 

 



1. Introduction 

COVID-19 took the world by surprise in late 2019, quickly turning into a major health shock that killed 

thousands. By 2020, it had morphed into a major global economic and social shock, putting livelihoods and 

food security at risk for millions. Vulnerable smallholders in developing countries, in particular, faced 

multiple challenges because of agricultural supply chain disruptions combined with food and agricultural 

inputs growing increasingly out of reach due to issues with availability and price hikes, with limited public 

aid and little recourse in terms of other income-generating activities. Despite the disruptiveness of the 

COVID-19 environment, beyond the initial surge in academic and public interest on the consequences of 

the pandemic, scant efforts have been geared towards analyzing its longer-term effects on livelihoods and 

food security and the path to recovery for affected households.  

This paper aims to partially fill this gap, providing a picture of the recovery process and detailed 

estimates of changes between 2019 and 2022 in rural households’ incomes, food availability and prices, 

food insecurity experiences, dietary diversity, and attitudes towards migration for a sample of agricultural 

households in the Western Highlands of Guatemala. To do this, we rely on an extensive panel of 1,262 

smallholders located across 87 communities in the departments of Huehuetenango, Quiché, and San 

Marcos, which were followed over four survey rounds from 2019 to 2022. An additional survey round 

collected at the end of 2022 among a subsample of households further permits to assess concrete 

quantitative changes on incomes, expenditures, indebtedness levels, and production decisions.  

With the onset of the pandemic, the government in Guatemala imposed several restrictions, curfews, 

and lockdowns, which combined were one of the highest restrictive environments in Central America (Hale 

et al., 2021). The Government Stringency Index, which measured the strictiness of government policy 

responses to COVID-19 on a scale 0-100 was over 96 between April and July 2020, dropped to 87 in August 

2020, and remained over 50 through the beginning of 2022. Most rural communities additionally closed or 

limited their access until mid-2021, resulting in further local resutrctions to movement and value chains’ 

disruptions.    



8 
 

The COVID-19 environment and the subsequent global and local market fluctuations were not the only 

systemic shocks affecting Guatemalan households during the period of analysis. The tropical storms 

ETA/IOTA and Julia —that hit the country in, respectively, November 2020 and October 2022— and the 

recent conflict between Russia and Ukraine that started in early 2022, which triggered supply and trade 

disruptions and important increases in international prices for energy, fertilizers, and agricultural 

commodities, also represented major events affecting households in the region.1 Following the same sample 

of households over the course of three years allows us to track the degree to which they were initially 

affected, observe any reactions, and analyze the speed and extent of their path to recovery, while 

considering their exposure to subsequent shocks. 

We build upon two previous studies that analyze the initial effects of the COVID-19 environment on 

food security and nutrition outcomes (Ceballos, Hernandez, & Paz, 2021) and the persistence of these 

effects (Ceballos, Hernandez, & Paz, 2022) using the same sample of households interviewed over the 

previous survey rounds. These studies found important qualitative decreases in incomes between 2019 and 

2020, followed by a partial recovery during 2021. To face these income shocks, households initially relied 

on savings, selling assets (mainly livestock), and public aid. By 2021, however, households were reporting 

a much higher reliance on costlier coping mechanisms such as loans from both formal and informal 

institutions. In addition, the above studies found a substantial increase in food insecurity episodes and lower 

dietary diversity after the start of the pandemic —with a decrease in animal source foods (ASF) diversity 

and an increase in fruits and vegetables (F&V) diversity—, which only partially recovered into 2021. 

Finally, a clear jump in the intention to emigrate was observed in 2021 that was not present during 2020, 

 
1 ETA and IOTA caused significant flooding, landslides, and mudflows that affected several rural areas across 16 of 

the 22 departments in Guatemala (IFRC, 2021), while Julia resulted in floodings across over two thirds of the  

municipalities in the country (United Nations, 2022). For the continuing repercussions of the Russia-Ukraine conflict, 

see IFPRI’s special series of blog posts, briefs, and events available at https://www.ifpri.org/spotlight/food-prices-war-

ukraine (accessed on September 2023). 

https://www.ifpri.org/spotlight/food-prices-war-ukraine
https://www.ifpri.org/spotlight/food-prices-war-ukraine


9 
 

arguably due to the persistence of economic hardship accompanied by the lifting of restrictions to 

movement. 

This study expands on the above in several dimensions. First, it provides a complete picture of the 

evolution of a group of key food security and nutrition indicators over the course of three years, a notably 

longer time period. Second, it studies the path of recovery for selected indicators across different groups of 

households, considering their geographic location, poverty status, and ethnicity, and whether they were 

additionally affected by extreme climatic events and the types of crops they produce. Finally, unlike the 

previous two studies, it provides quantitative estimates for a range of socioeconomic outcomes, including 

incomes, expenditures, indebtedness levels, and production decisions. 

The results show an important recovery in incomes, food security, and dietary diversity by 2022 among 

study households, but that have not yet reached pre-pandemic levels. Around two fifths of households still 

report having lower overall incomes than in 2019, a substantial improvement though compared to the 

situation in 2020 and 2021. The 2022 prevalence of severe food insecurity episodes has dipped below 2019 

levels, with that for moderate food insecurity being only marginally higher than pre-pandemic levels and 

that for mild food insecurity decreasing, compared to 2021, but still about 15 percentage points higher than 

in 2019. Dietary diversity scores have improved relative to those obtained in 2020 and 2021 and are in some 

respects even higher than the ones reported in 2019. However, the higher willingness to emigrate observed 

in 2021 persists in 2022 and more than doubles 2019 levels. The path of recovery, especially in terms of 

food security and dietary diversity, has been easier for some groups of households, such as those located in 

the department of San Marcos and those living above the poverty line in 2019, while households affected 

by the ETA and IOTA tropical storms, non-coffee producers, and indigenous populations have taken more 

time to recover. Quantitative findings for a subsample of households followed for an additional survey 

round in late 2022 point to a 16 percent average reduction in total income since 2019, mostly explained by 

lower agricultural incomes. What is unsettling is that household expenditures have increased by 26 percent, 

with indebtedness on the rise, including a 53 percent increase in the proportion of households holding debt 

and a 59 percent increase in the average volume of debt. 



10 
 

The paper contributes to four strands of literature. First, our study adds evidence on the longer-term 

impacts of the COVID-19 environment on agricultural livelihoods and incomes, as well as those on the 

broader agrifood value chain (e.g., Ceballos, Kannan, & Kramer, 2020, 2021; Hailu, 2020; Hammond et 

al., 2022; Hirvonen, De Brauw, & Abate, 2021; Hirvonen et al., 2021; Jeyakumar, Dunna & Aneesh, 2022; 

Kumar, Padhee, & Kumar, 2020; Mahajan & Tomar, 2020; Mildendorf et al., 2022; Nchanji et al., 2021; 

Nolte, Sipangule, & Wendt, 2022; Rawal et al., 2020; Van Hoyweghen et al., 2021; and Wang et al., 2022). 

Second, the study relates to the literature on the food security and nutrition effects of COVID-19 (e.g., 

Balana et al., 2023; Bundervoet, Dávalos, & Garcia, 2022; Ceballos, Kannan, & Kramer, 2020; Ceballos, 

Hernandez, & Paz, 2021, 2022; Dasgupta & Robinson, 2022; Diaz-Bonilla et al., 2021; Hammond et al., 

2022; Hirvonen, De Brauw, & Abate, 2021; Jeyakumar et al., 2022; Laborde et al., 2020; Mekanna, 

Panchal, & Li, 2022; Rudin-Rush et al., 2022). Third, we relate to a small but growing strand of the literature 

on the impacts of compound shocks (Kruczkiewicz et al., 2021; Mahul & Signer, 2020) on businesses 

revenues (Helgeson et al., 2022), agriculture (Mishra, Bruno, & Zilberman, 2021) and food insecurity 

(Randell et al., 2021). Lastly, the paper adds to a relatively scarce literature on the quantitative effects of 

the COVID-19 environment on incomes, expenditures, asset holdings, and indebtedness (Amare et al., 

2021; Balana et al., 2023; Hoehn-Velasco et al., 2022; Tefera, Tadesse, & Asmare, 2022). 

The closest studies to ours are Amare et al. (2021), Hirvonen, De Brauw, & Abate (2021), Dasgupta & 

Robinson (2022), Rudin-Rush et al. (2022), Tefera, Tadesse, & Asmare (2022), and Han et al. (2023) who 

examine the impact of the COVID-19 pandemic on food security and nutritional intake across different 

countries using household panel data. These studies, however, either rely on solely two survey rounds 

(before and after the pandemic) or on additional high-frequency survey rounds following the pandemic but 

only until mid-2021. We extend these studies by relying on annual survey data over a larger time span (until 

2022), that permits to better assess recovery paths, and by considering a broader set of outcomes, such as 

qualitative and quantitative impacts on incomes from a variety of sources, in addition to quantitative 

impacts on expenditures, indebtedness, and production decisions. 



11 
 

The remainder of the paper is organized as follows. Section 2 describes the data used in the analysis, 

including a description of the study sample and the key outcome variables of interest. Section 3 explains 

the methodological approach, while Section 4 presents and discusses the results. Section 5 concludes. 

 

2. Data 

The sample frame used for the study corresponds to a sample of households interviewed at the end of 2019 

as part of an impact evaluation of a value chain development program for coffee and horticultural producers 

in the Western Highlands of Guatemala.2 The original sample consists of 2,142 small commercial coffee 

and horticultural farmer households who owned cell phones and were located across 87 municipalities in 

the departments of Huehuetenango, Quiche, and San Marcos.3  

The main study data comprises a total of four survey rounds —a baseline survey administered in-

person in November-December 2019 by a local team of female enumerators and three follow-up survey 

rounds conducted over the phone by the same team of enumerators during May-June 2020, 2021, and 2022.  

The data collected during the baseline survey included demographic characteristics, agricultural 

production, asset ownership, household income and expenditures, participation in social programs, food 

security, dietary diversity, perceptions, attitudes towards migration, and anthropometry measurement for 

children under five years old.  

 
2 The program is Feed the Future Guatemala Value Chains Project, funded by the United States Agency for 

International Development (USAID) and executed by two private implementors: Agropecuaria Popoyán and 

Federación de Cooperativas Agrícolas de Productores de Café de Guatemala (FEDECOCAGUA). The main objective 

of the program was to increase agricultural incomes and resilience of small farmers and their families in the Western 

Highlands and improve their nutrition status.  

3 This sample includes project beneficiaries and non-beneficiaries (control group) for the impact evaluation, who are 

part of the broader population of commercial smallholder farmers in the region. Beneficiary households interviewed 

in 2019 are a representative sample of the total population of program beneficiaries. Non-beneficiary households 

belong to neighboring communities with similar characteristics to those within the program’s reach. The working 

sample for this study has roughly an equal share of beneficiaries and non-beneficiaries. 



12 
 

The follow-up surveys were aimed at identifying the potential short-, medium-, and long-term 

economic and nutritional effects of the COVID-19 pandemic on rural households, while accounting for the 

implications of other shocks faced between 2020 and 2022, including extreme weather events and the 

Russia-Ukraine conflict and resulting global price crisis. The first follow-up survey collected qualitative 

changes in income sources (including remittances) after the lockdown, coping mechanisms to deal with 

income reductions, food security indicators, dietary diversity, perceptions, and migration intentions. 

Complete findings from this initial follow-up survey are available in Ceballos, Hernandez, & Paz (2021). 

The second follow-up survey included the same modules as the first follow-up survey together with 

additional modules inquiring about changes in agricultural activity, permanent or temporary loss of jobs or 

businesses, and exposure to ETA and IOTA tropical storms. The full findings from this second follow-up 

study are available in Ceballos, Hernandez, & Paz (2022). The third follow-up included new modules on 

availability and prices of agricultural inputs and prices of food staples considering the Russia-Ukraine 

conflict.  

Participation across all follow-up survey rounds was encouraged by providing households with a 

small airtime top-up (equivalent to 1.33 US dollars) in compensation for their time.4 Local community 

leaders were also contacted when a household was not answering repeated calls or had an inactive phone 

number. Despite these efforts, 880 households were dropped from the study sample over the course of the 

three follow-up surveys, equivalent to 40 percent of the original sample frame. The attrition rate from the 

first to the second survey wave was close to 15 percent (318 out of 2,142 households), 10 percent in the 

third survey wave (212 households), and 16 percent in the fourth survey wave (350 households).5 The final 

 
4 Households were given a small in-kind gift (calendar and pen) for the baseline in-person survey. 

5 For the 2021 and 2022 survey rounds, about half of the attrition came from households refusing to conduct another 

interview, and the rest mostly from not being able to reach the household through either phone calls or the community 

leaders’ support. In the case of the 2020 round we did not collect detailed data on attrition, but most of the excluded 

households were due to not being able to reach them by phone after several attempts. In all follow-up survey rounds, 

we attempted to contact all unreached households across different days and different times of the day. 



13 
 

working sample is a balanced panel of 1,262 households interviewed during all four survey rounds. Figure 

1 maps the communities where the study households are located across the three departments. 

Since attrition between baseline and our working sample is relatively high, we test whether 

households in these two samples are comparable in terms of observable characteristics (at baseline). As 

shown in column (4) in Appendix Table A.1, most characteristics are generally balanced across the two 

samples, with only four exceptions: a higher proportion of male-headed households, a higher household 

size, higher livestock ownership, and a higher share of households producing coffee as one of their main 

crops in the final working sample. These imbalances, however, are economically very small such that we 

do not consider any adjustments for attrition bias in the analyses but still perform some robustness checks 

as discussed in the next section. 

In addition, a subset of households in our working sample participated in a fifth survey wave of in-

person data collection, conducted during November-December 2022 as part of the endline survey for the 

impact evaluation of the value chains project. The questionnaire for this survey round was much more 

comprehensive and comparable to the baseline survey, allowing for quantitative comparisons across several 

key socioeconomic outcomes between 2019 and 2022, including households’ incomes, expenditures, 

savings and debts, crop production decisions, and livestock ownership, that we discuss in a separate 

subsection of the Results.6 The 777 households that were additionally interviewed in the endline survey 

generally share similar characteristics to those in our full working sample and in the original sample frame 

(see columns (5) and (6) in Appendix Table A.1). 

Panel A of Table 1 presents descriptive statistics of an array of household characteristics among 

the full working sample. Households are mainly located in the department of Huehuetenango (58 percent), 

followed by Quiche (32 percent), and the remaining in San Marcos (10 percent). Most of the households 

 
6 Not all households were interviewed in the fifth survey round because some could not be located in person, others 

refused to be interviewed, and there were some difficulties accessing certain locations due to weather conditions at 

the time of the fieldwork. 



14 
 

grow coffee as one of their main crops (82 percent). Household heads are primarily male (85 percent), 

middle-aged (average age of 48 years old), and their main language is not Spanish (71 percent). Dwellings 

typically have finished floors (54 percent), unfinished ceiling (86 percent), and are connected to electricity 

(84 percent) and to a water system (85 percent) but lack connection to a drainage network (69 percent). The 

average agricultural land size is 0.9 hectares, and the daily per capita household expenditure is 11.3 

quetzales (equivalent to 1.5 US dollars). These socioeconomic characteristics are roughly consistent with 

the average household characteristics in the three departments comprised in this study (according to the 

2018 Population and Housing Census) but are not necessarily representative of the broader rural population 

in the region (that includes subsistence farmers).7 

Panel B of Table 1 shows that several households in our study sample are continuously exposed to 

negative weather events, which are recurrent in the Western Highlands. More than four out of every ten 

households report having been severely affected (mainly crop losses) by the ETA/IOTA tropical storms in 

late 2020, while six out of every ten households report having had some degree of crop losses due to a 

negative weather event over the 12 months prior to the 2022 survey round.8 In terms of the level of exposure 

to COVID-19, 9 percent of households report that someone in their household was infected with COVID-

19 as of 2022, and 35 percent of households report that someone in their community was infected.9  

 
7 https://www.censopoblacion.gt/ (accessed on September 2023). See Ceballos, Hernandez, & Paz (2021, 2022) for a 

deeper discussion of the comparability of the study sample with the average rural household in the region. 

8 Half of the households that reported being affected by a negative weather event in 2022 were also affected by ETA 

and IOTA. In addition, among the subsample of households interviewed in a fifth survey round in late 2022, 71 percent 

indicated being affected by the Julia tropical storm in October 2022. 

9 The 2022 survey round also inquired about vaccinations and we find that seven out of every ten households report 

at least one member having been vaccinated (with at least one dose) as of 2022; among households with children 

between 5 and 12 years old, only four out of every ten households report at least one child receiving at least one dose. 

The vast majority of households that reported no member receiving any dose indicated that they do not want to get 

the vaccine, as opposed to an unavailability issue.    

https://www.censopoblacion.gt/


15 
 

Turning to the outcome variables, the study evaluates the evolution of several indicators based on 

information collected during each survey round, but we pay special attention to households’ income, food 

insecurity, dietary diversity, and migration intentions. For the income indicators, we rely on households’ 

self-reported qualitative changes between 2019 and 2022 in incomes from three major sources: agricultural, 

non-agricultural, and remittances.10 In terms of food insecurity, we construct three dichotomic indicators 

related to food insecurity experiences (based on the Food Insecurity Experience Scale – FIES-, Ballard, 

Kepple, & Cafiero (2013)).11 For dietary diversity, we consider three scores: (i) the household dietary 

diversity score (HDDS), a measure widely used in the nutrition literature as a proxy for diet quality, which 

focuses on the food groups consumed by a reference household member (the interviewed individual in our 

case) during the full day before the interview, (ii) the diversity in consumption of animal source foods (ASF 

score), and (iii) the diversity in consumption of fruits and vegetables (F&V score).12 Lastly, we focus on an 

indicator of intention to emigrate, based on a survey question that asks if any household member has the 

intention to migrate outside of the country, to either live or work, over the following three years. 

 
10 During the survey, households were asked to provide their answers to these questions by selecting one out of five 

categories: decreased a lot, decreased a little, remains the same, increased a little, and increased a lot. In some of the 

analyses below and for simplicity, we lump together the increased a little and increased a lot categories. In other 

analyses, we construct a variable capturing whether a given household reported an unambiguous income decrease, 

defined as the household reporting a decrease in one or more of the three income sources but no increases. 

11 The indicators are based on whether a household experienced mild, moderate, or severe food insecurity episodes 

during the months prior to being interviewed. Mild food insecurity is defined as the household indicating that in at 

least one instance household members had eaten only a few kinds of foods because of a lack of money or other 

resources; moderate food insecurity as having eaten less than they thought they should; and severe food insecurity as 

having not eaten despite feeling hungry. 

12 The HDDS is based on twelve food groups: (1) cereals and grains; (2) roots and tubers; (3) vegetables; (4) fruits; 

(5) meats, poultry, and offal; (6) eggs; (7) fish and seafood; (8) legumes and nuts; (9) milk and milk products; (10) oil 

and fats; (11) sugar; and (12) miscellaneous foods. For additional information on the indicator see FAO (2010). The 

ASF score is based on seven food groups: (1) beef and pork meat; (2) poultry; (3) organ meat; (4) sausages; (5) fish 

and seafood; (6) eggs; and (7) dairy products. The F&V score is based on five food groups: (1) fruits rich in vitamin 

A; (2) other fruits; (3) vegetables rich in vitamin A; (4) green leafy vegetables; and (5) other vegetables. 



16 
 

In the case of the subsample of households interviewed in the additional survey round, we pay 

special attention to quantitative changes between 2019 and 2022 in household incomes, expenditures, and 

total amount of debt. Household incomes include total monthly income and monthly income from each of 

the three major sources. Expenditures include monthly total expenditures, obtained from aggregating all 

expenditure categories reported by the household, and monthly expenditures on six major categories (food, 

housing, consumer goods and durables, education, health, and others). The debt measures includes the total 

amount of debt and the amount of debt from three different sources (formal financial institutions, friends 

and relatives, and informal lenders).  

 

3. Methodology  

The empirical analysis is divided into three sections. First, we perform before-after comparisons for several 

economic, food security, and dietary diversity outcome variables (Section 4.1). We pay special attention to 

comparisons between 2019 (pre-pandemic) and 2022, given that the short-term changes (2019 versus 2020) 

and medium-term changes (2019 versus 2021) for several of the outcomes analyzed are extensively 

discussed in Ceballos, Hernandez, & Paz (2021, 2022). For income variations, the comparisons are based 

on qualitative changes in different income sources relative to 2019, self-reported by households during each 

follow-up survey. A similar approach is followed for the availability and prices of different food groups, 

where we rely on self-reported qualitative variations during each follow-up survey. For food insecurity 

experiences, dietary diversity measures (including consumption of animal source foods and fruits and 

vegetables), and intention to emigrate, the comparisons are based on dichotomic or quantitative indicators 

collected during each survey round. 

 In a second group of analyses, we formally assess the evolution of key indicators across different 

groups of households (Section 4.2) using a multivariate regression framework. The focus in this section is 

on the quantitative measures available and include experiencing a (moderate or severe) food insecurity 

episode, dietary diversity scores, and households’ intention to emigrate. This set of analyses allows us to 

determine whether certain groups of households were more or less affected by the pandemic and whether 



17 
 

they took longer or shorter to recover. We separately consider differences in terms of household location 

(located in the department of San Marcos versus Huehuetenango and Quiche), poverty status, whether a 

household reported being affected by the tropical storms ETA and IOTA, whether the household produces 

coffee (among the main four crops produced), and whether the household considers itself to be indigenous.13 

Based on the longitudinal nature of our data and the fact that household characteristics were only 

collected at baseline, we follow a panel data approach using household random effects (i.e., assuming that 

the unobserved, time-invariant household characteristics are independent of the explanatory variables), 

while controlling for fixed effects at the community level. In particular, we estimate the following 

generalized least squares (GLS) random effects model, 

 

𝑌𝑖𝑗𝑡 = 𝛼0 + 𝛼1𝑡1 + 𝛼2𝑡2 + 𝛼3𝑡3 + ∑ 𝛼4𝑠𝑡𝑠 ∗ 𝐺𝑖
3
𝑠=1 + 𝑋𝑖𝑗𝛽 + 𝜃𝑖 +  𝑐𝑗 + 𝑢𝑖𝑗𝑡             (1) 

 

where  𝑌𝑖𝑗𝑡 represents the outcome variable of interest (i.e., food insecurity experience, dietary diversity 

score, or intention to emigrate) reported by household i located in community j at period t; 𝑡1, 𝑡2, and 𝑡3 

are dummy variables for the 2020, 2021, and 2022 survey rounds, respectively; 𝐺𝑖 is a dichotomic variable 

identifying whether the household belongs to a specific group (i.e., located in San Marcos, poor, affected 

by the tropical storms, coffee is one of their main crops produced, and indigenous); 𝑋𝑖𝑗 is a vector of 

household characteristics measured at baseline (including the indicator variable for the household group 

type); 𝜃𝑖 is a household-specific element such that 𝐸[𝜃𝑖|𝑋𝑖𝑗] = 0; 𝑐𝑗 represents community indicators that 

control for the possible influence of factors common to all households in a given community (e.g., 

 
13 A household is considered poor if their per capita daily expenditure is less than 1.9 US dollars (at 2011 national 

prices), which is the international poverty line defined by The World Bank prior to 2022. While several communities 

were hit by the tropical storms in 2020, a household is considered as affected by these events for the analysis if they 

indicated being materially affected (e.g., reported crop losses, livestock and/or backyard animal losses, dwelling 

damages, or lack of access to basic services, among others). A household is regarded as indigenous if the survey 

respondent identifies herself/himself as indigenous.  



18 
 

socioeconomic, cultural, and accessibility conditions); and 𝑢𝑖𝑗𝑡 is an error term. We consider an extensive 

set of household variables that include age, gender, education, and language of the household head, 

household size, participation in social programs, dwelling characteristics and access to public services, per 

capita household expenditure, asset ownership, agricultural variables such as landholding size, having an 

irrigation system, indicator variables for main crops produced, and if at least one member of the household 

is a temporal internal migrant (for work purposes). The standard errors are clustered at the community level 

to account for likely within-community correlation in the modeled outcomes. 

Parameter 𝛼0 captures the unconditional mean of each modeled outcome variable, while parameters 

𝛼1, 𝛼2, and 𝛼3 capture the unconditional average change in the outcome variable in 2020, 2021, and 2022 

with respect to 2019. The parameters of interest are 𝛼4𝑠, 𝑠 = 1,2,3, that permit to depict the evolution over 

time of the outcome variable distinguishing by the different groups of households considered, after 

controlling for multiple households’ characteristics and location fixed effects.14 Note that alternatively 

using household fixed effects (instead of community-level dummies) to estimate Equation (1), would result 

in omitting the vector of household characteristics 𝑋𝑖𝑗 (all time-invariant) but would not affect the 

estimation of 𝛼4𝑠, 𝑠 = 1,2,3, as these parameters are not identified from variations over time within the 

same household. 

 Finally, in Section 4.3 we take advantage of the additional survey round conducted at the end of 

2022 for a related study, which comprised a subset of our working sample. As with the late-2019 baseline 

survey, this survey was fielded in person and collected detailed quantitative information on several relevant 

socioeconomic outcomes (described in Section 2). We implement before-after comparisons of these 

outcomes between 2019 and 2022, that complement the analyses using the full working sample. In addition, 

we replicate the model in Equation (1) using as outcomes total income, aggregate expenditures, and total 

 
14 It is worth noting that the individual parameters for the group indicators can be identified except for household 

location that is measured at the department level and is absorbed in this case by the set of community-level dummies 

included in the regressions. 



19 
 

debt, to assess potential heterogeneous variations between different groups of households. We estimate the 

following GLS random effects model,    

 

𝐿𝑛(𝑄𝑖𝑗𝑡) = 𝛾0 + 𝛾1𝑡 + 𝛾2𝑡 ∗ 𝐺𝑖 + 𝑋𝑖𝑗𝛿 + 𝜃𝑖 + 𝑐𝑗 + 𝑢𝑖𝑗𝑡                 (2) 

 

where 𝑄𝑖𝑗𝑡 is the corresponding outcome variable of interest (i.e., monthly total income, monthly total 

expenditures, and total debt amount in 2019 Quetzales); 𝑡 is a dummy variable for the 2022 survey round; 

and the remaining variables are the same as those  described for Equation (1). The parameter of interest in 

Equation (2) is 𝛾2, which measures the differential variation in the modeled outcome in 2022 (relative to 

2019) across the different household groups (𝐺𝑖) considered.15 

 

3.1 Potential econometric issues 

Considering that the basis of our empirical approach relies on before-after comparisons, it is important to 

discuss likely sources of endogeneity that could affect our results. These include potential seasonality 

effects, time-varying unobservable factors correlated with the analyzed outcomes, and measurement error.   

 Given that the baseline survey was implemented at the end of 2019 while all follow-up surveys 

were implemented mid-year, the estimations could be affected by seasonal trends in food availability and 

diversity or food access (income and prices). However, as extensively discussed in Ceballos, Hernandez, 

& Paz (2021, 2022), seasonal effects among our studied sample (if any) would operate opposite to the 

observed effects in the wake of the pandemic, which provides additional support to our results. 

 Similarly, despite we account in the regression analyses for a wide set of (time-invariant) household 

characteristics and location dummies to control for possible confounding effects on the modeled outcomes, 

 
15 We focus in this case on differentiated changes between household groups as we can only compare values across 

two survey waves (years), while in Equation (1) we can assess the evolution of outcomes for different household 

groups over four survey waves.  



20 
 

we cannot completely rule out time-varying omitted variables such as changes in demographic patterns or 

migration flows influencing our results, although these were to some extent limited until the first half of 

2021 (where the vast majority of communities in the sample reported closing their borders for prolonged 

periods of time since the COVID-19 outbreak).16 Still, since around 40 percent of the households in our 

working sample report a change in household size or composition between 2019 and 2022, we conducted 

the analyses excluding this subset of households and confirmed the robustness of our main results.17 

In a related vein, while we analyze the potential influence of the tropical storms that affected part 

of our study sample on key outcomes, we recognize that the 2022 measures could also be influenced by the 

Russia-Ukraine conflict and its impact on agricultural input and food staple prices.18 This could be more 

accute for the quantitative outcomes collected during the additional survey round at the end of 2022, in that 

the comparisons between 2019 and late-2022 reported in Section 4.3 may not only be capturing long-lasting 

pandemic-related effects but also those of the conflict.  

Finally, self-reports on qualitative income changes, food insecure experiences, and consumption 

patterns could be subject to measurement error due to a likely negative perception bias because of the 

pandemic, especially during the onset of COVID-19, while migration intentions could be misreported due 

to the sensitivity of the question as there is a high emigration rate (mostly irregular) from rural areas in 

 
16 See Ceballos, Hernandez, & Paz (2021, 2022) for additional discussion on the closure of communities. 

17 As an additional robustness exercise, we confirmed that the observed variations in food insecurity experiences, 

dietary diversity scores, and intentions to emigrate persisted when using the full unbalanced panel of households. This 

is in line with the fact that the households that remained in the sample generally share similar observable characteristics 

than those that were lost along the three follow-up survey rounds. 

18 As noted earlier, in the third follow-up survey implemented in mid-2022 we included questions around the potential 

impacts of the Russia-Ukraine conflict and find that the conflict seems to have contributed to a significant increase in 

the price of agricultural inputs, especially chemical fertilizers, as well as an increase in food prices. When asked about 

any variations in their farming production decisions, only half of the respondents report having implemented any 

change (mainly using less fertilizer or cheaper ones). Additional details are available upon request.  



21 
 

Guatemala.19 Yet, provided these sources of errors are not correlated with other covariates, the parameters 

in the corresponding regression analyses would be estimated with less precision but remain unbiased. 

 

4. Results 

This section presents the empirical results. We first present before-after comparisons on several outcome 

variables of interest related to incomes, food availability and prices, food security, dietary diversity, and 

migration intentions. Second, we examine the evolution over time of some of these outcomes across 

different groups of households to characterize those profiles that resulted most affected by the pandemic or 

took longer to recover. Third, we assess quantitative changes in incomes, expenditures, indebtedness levels, 

and production decisions using a subsample of households interviewed both in 2019 and late 2022, and 

assess heterogeneous variations across different household groups. 

 

4.1. Before and after comparisons  

Figure 2 presents self-reported qualitative changes in household income by source for the years 2020, 2021, 

and 2022, relative to 2019. Overall, we observe a consistent recovery in incomes by 2022, with a 

substantially lower share of households facing lower incomes relative to 2019 than that in previous survey 

rounds (despite the resulting global crisis from the Russia-Ukraine conflict). While in 2020, 61, 63, and 93 

percent of households reported experiencing lower incomes from agricultural, non-agricultural, and 

remittances sources, respectively, compared to the pre-pandemic period, only 41, 26, and 46 percent of 

households report lower incomes in 2022.20, Still, on aggregate, 41.2 percent (520 households) report an 

 
19 According to statistics from the United Nations, net emigration was close to 856,000 people between 2002 and 

2021 in Guatemala, equivalent to 5 percent of their current population 

(https://population.un.org/wpp/Download/Standard/MostUsed/, accessed on September 2023).  

20 The decrease in remittances is though not aligned with national reports that indicate that family remittances 

accounted for about 18% of the gross domestic product in Guatemala in 2021 (Maldonado & Harris, 2022).   

https://population.un.org/wpp/Download/Standard/MostUsed/


22 
 

unambiguous income decrease (i.e., a decrease in one or more of their three income sources but no 

increases) in 2022.21  

Figure 3 reports the type of coping mechanisms that households indicating an unambiguous 

reduction in income have been using to deal with these income losses across survey rounds. The use of 

savings stands as an important coping mechanism (used by about a third of reporting households), which 

has remained relatively stable over the years. Selling assets and/or livestock and borrowing from formal or 

informal sources are also relevant mechanisms that nonetheless show a small reduction in 2022 compared 

to 2021, potentially indicating the approach to a point of exhaustion after 3 years of pandemic. This view 

is somewhat supported by the substantial increase in the percentage of households reporting doing nothing 

to cope with income losses (from 20-25 to 33 percent), in addition to the notable drying up of aid flows 

from public and private organizations (after reaching up to 23 percent of households in previous years). In 

this sense, the data indicate a slight shift towards costlier coping mechanisms over the years, an issue to 

which we return below when analyzing quantitative changes in indebtedness among a subset of households. 

Next, we describe the evolution of several food security and dietary diversity indicators from 2019 

to 2022. Beyond the important reductions in households’ income generating capacity during the COVID-

19 pandemic, numerous studies have pointed to substantial negative effects on food security and nutrition 

(e.g., Amare et al., 2021; Balana et al., 2023; Hammond et al., 2022; Tefera, Tadesse, & Asmare, 2022). 

Such effects relate both to a reduction in accessibility to food, stemming from reduced incomes and 

increased prices, and a decline in food availability, due to the initial halt of food value chains and 

transportation networks. At the same time, said difficulties affected households in different ways, increasing 

 
21 In terms of geographical variations, in Appendix Figure A.1 we observe that income declines were worst in the 

department of San Marcos, particularly in agricultural incomes and remittances, where more than seven out of ten 

households indicate that their income decreased (either a lot or a little) compared to 2019. Given that San Marcos is a 

department relatively more developed and connected to markets than the other two departments, the COVID-19 

environment (as well as the global crisis triggered by the conflict in Eastern Europe) could have had a larger and more 

long-lasting effect on economic activities. It follows that the overall income recovery described above is mainly driven 

from households located in Huehuetenango and Quiche. 



23 
 

food insecurity among vulnerable households and affecting diets for most households, not only due to 

changes in food availability and accessibility but also due to potential substitution effects from changes in 

relative prices between food groups. 

Figure 4 reports the evolution of food availability (panel A) and prices (panel B) in local markets, 

as reported by households, across five key food groups in Guatemalan diets: dairy products, meats, fruits 

and vegetables (F&V), roots and tubers, and grains and cereals. By 2022 (similarly to 2021), eight out of 

every ten households indicated that food availability remained the same or had increased compared to 2019 

across the five food groups. This confirms the sustained post-pandemic recovery in the Guatemalan food 

system, where in 2020 more than nine out of every ten households indicated an overall decrease in food 

availability (not reported). Regarding food prices, however, most households continue to report higher 

prices relative to 2019 across all food groups, arguably fueled by the global food and fertilizer price 

increases around the pandemic and the conflict in Eastern Europe.22 Although not reported in the interest 

of space, households also report higher prices for major staple foods in 2022 compared to 2021 (with oil, 

beans, and poultry exhibiting the highest increases).  

Panel A of Table 2 presents, in turn, the prevalence of mild, moderate, and severe food insecurity 

episodes across survey rounds and the test results for whether the post-pandemic values are statistically 

different than the pre-pandemic ones. All three observed dimensions of food insecurity (mild, moderate, 

and severe) increased substantially following the pandemic but gradually improved over time, with 2022 

levels comparable to those in 2019. In particular, the 2022 share of households reporting not eating at some 

point despite feeling hungry (i.e., severe food insecurity) dropped to 7 percent, down from 23 percent in 

the previous year and even lower than the 11 percent observed in 2019. In the case of the prevalence of 

moderate and mild food insecurity, these are lower than those observed in 2021 and 2020 but remain slightly 

higher than that in 2019. The share of households reporting eating less than they thought they should (i.e., 

 
22 The average monthly price inflation in the country was 1.8 percent in mid-2020, 5.2 percent in mid-2021, and 4.7 

percent in mid-2022 (Central Bank of Guatemala, https://banguat.gob.gt/es/page/inflacion, accessed on September 

2023). 

https://banguat.gob.gt/es/page/inflacion


24 
 

moderate food insecurity) decreased to 39 percent in 2022, compared to 58 percent in 2021 and 86 percent 

in 2020, but it is still higher than the 35 percent observed in 2019, although the difference is not statistically 

significant at conventional levels. The share of households reporting only eating a few kinds of foods 

because of a lack of resources (i.e., mild food insecurity) decreased by about 10 percentage points between 

2021 and 2022, reaching 73 percent, which is still above the 58 percent observed in 2019. 

Panel B of the table shows average dietary diversity scores over the same period. In this case too, 

dietary diversity in 2022 has improved relative to 2020 and 2021, and even when compared to 2019 for 

some scores. For instance, the average household dietary diversity score (HDDS) was 6.9 out of twelve 

possible food groups in 2022, above the 6.8 score observed in 2019 and after reaching a low of 6.5 in 2020.23 

In the case of animal source foods (ASF), households are consuming on average 1.05 out of seven possible 

animal source food groups (mainly eggs), a small recovery compared to 2020 and 2021 findings but still 

below the 1.25 score observed in 2019. The average household consumption score for fruits and vegetables 

(F&V) is 2.4 out of five possible food groups in 2022, higher than the pre-pandemic period (2.1 in 2019) 

but slightly lower than 2020 and 2021 levels.24 The switch away from ASF to F&V diversity observed since 

2020 thus persists.25 Regarding women 15-49 years old and children 6-23 months old in the household, we 

find that the average dietary diversity score for women (WDDS) continued to remain around 4.5 out of nine 

 
23 While the average HDDS in 2022 is not statistically different to the 2019 score, when comparing the distribution of 

scores across years we find a higher concentration of households around the average in 2022 (i.e., consuming between 

5 and 9 food groups), as opposed to 2019 where the consumption of different food groups was more dispersed 

(including a larger fraction of households consuming 10 and 11 food groups as well as 3 and 4 groups). See Appendix 

Figure A.2. 

24 In Appendix Figure A.3, we present the share of households that consumed each of the seven ASF groups (Panel A) 

and five F&V groups (Panel B) over the four survey rounds. We observe a recovery in the consumption across all 

seven ASF groups in 2022, but the share of households consuming each food group is still lower than in 2019. For 

fruits and vegetables, we continue to find increases in consumption among all F&V groups since the onset of COVID, 

except for the group of vegetables rich in vitamin A. 

25 We cannot conclude, however, whether this substitution has resulted in a higher or lower micronutrient intake as we 

did not collect quantities consumed during the surveys. 



25 
 

possible food groups during all survey rounds, in contrast to a sustained increase in the dietary diversity for 

children (DDS), which reached 3.9 out of seven possible food groups in 2022 (compared to 3.4 in 2019), 

arguably due to a prioritization of young children in the household over the past years.26 

Finally, Panel C of the table presents the share of households reporting whether any household 

member intends to emigrate in the subsequent years. Interestingly, the important increase in the fraction of 

households reporting a member intending to emigrate observed in 2021 (9.8 percent, versus 4 percent in 

2019), remains in 2022 (8.5 percent). The higher intentions to leave the country are rather revealing 

considering that this question is normally prone to underreporting, and they could reflect recurrent 

economic and financial difficulties faced by households despite the apparent improvement in several food 

security and nutrition indicators in 2022. 

 

4.2. Evolution of key outcomes by groups of households 

We now turn to analyze the evolution of three key outcome variables between different groups of 

households over the four survey waves. We are interested in assessing whether the evolution of food 

insecurity (moderate or severe), dietary diversity, and migration intentions followed a different pattern for 

different groups of households. We choose specific household groups which a priori one could expect to 

have been more (or less) affected by the pandemic or could have taken longer (or shorter) to recover. We 

 
26 Similar to the reference household member for the HDDS, the same woman 15-49 years old (in 1,017 households) 

was followed over the four survey rounds. A different child 6-23 months old was, however, selected during each survey 

round due to the narrow age range to construct the dietary indicator and the wide time span of the study, that would 

otherwise have constrained the number of observations for the analysis; the number of households with children 6-23 

months are 225 in 2019, 168 in 2020, 168 in 2021, and 143 in 2022. The nine food groups in the WDDS include: (1) 

cereals, grains, roots, and tubers; (2) dark green vegetables; (3) other fruits and vegetables rich in vitamin A; (4) other 

fruits and vegetables; (5) organ meats; (6) meat, poultry, and fish; (7) eggs; (8) legumes, nuts, and seeds; and (9) milk 

and milk products. The seven food groups the children DDS include: (1) cereals, grains, roots, and tubers; (2) legumes 

and nuts; (3) dairy products (milk, yogurt, cheese); (4) flesh foods (meat, fish, poultry, and liver/organ meats); (5) 

eggs; (6) vitamin A rich fruits and vegetables; and (7) other fruits and vegetables. 



26 
 

accordingly divide households into two groups based on their loation (department), poverty status, and 

ethnicity, in addition to whether they were affected by ETA and IOTA and whether they are coffee 

producers.  

Figure 5 shows the results of this exercise, based on the regression framework described in 

Equation (1) that accounts for a wide set of household controls and location fixed effects. Every data point 

in the plotted figures corresponds to the estimated outcome variable for each survey year and household 

group, all else being equal.27 See Appendix Table A.2 for the full regression results.  

In Panel A of Figure 5 we distinguish by household location. While households in the department 

of San Marcos were initially most affected by the pandemic compared to those in Huehuetenango and 

Quiche, in terms of an increase in the prevalence of food insecurity experiences and a reduction in dietary 

diversity, they also show a faster recovery. As mentioned above, San Marcos is overall more developed and 

connected to markets, which are two factors that could explain the observed pattern. Yet, households in San 

Marcos exhibit a considerably higher intention to emigrate over the past two years, and the gap with the 

other two departments has further accentuated during 2022, in line with the larger share of households 

reporting income decreases compared to the other two departments. 

  Panel B of the figure reveals that non-poor households showed a larger initial increase in the 

prevalence of moderate and severe food insecurity but recovered faster. A somewhat similar pattern is 

observed for dietary diversity, where non-poor households experienced a larger initial decrease but 

eventually recovered. In the case of intention to emigrate, both groups of households roughly followed a 

similar pattern until 2022, when the proportion of non-poor households intending to emigrate continued 

growing (although the change is not statistically significant compared to 2019) while the one for poor 

households decreased. 

 
27 In this framework, the 2019 reported outcome variables are equivalent to the sample averages for each household 

group in that year. 



27 
 

As noted earlier, households in the study region are regularly exposed to weather shocks, and in 

our working sample 43 percent of the households reported being significantly affected by ETA and IOTA 

tropical storms in late 2020. In Panel C, we accordingly show the evolution of our key outcomes for the 

group of households that reported being affected and not being affected by these storms. The tropical storms 

that occurred between the 2020 and 2021 survey rounds appeared to have delayed the recovery in food 

insecurity experiences. Similarly, households affected by ETA and IOTA show a larger decrease and a 

slower recovery in their dietary diversity. The intention to emigrate also remains higher (through 2022) 

among households affected by the tropical storms. This evidence suggests that extreme weather events not 

only affect households negatively when they occur, but also reduce their capacity to recover from other 

shocks, remarking the importance of promoting climate mitigation and adaptation strategies across 

vulnerable regions. 

Lastly, Panels D and E of the figure depict the evolution of the three outcomes distinguishing 

between coffee producers and non-producers and between indigenous and non-indigenous households, 

respectively. We find that coffee producers were less affected by the pandemic (in line with the increase in 

international coffee prices over the past years) but reveal a higher intention to emigrate in recent years, 

which could be masking other more recent difficulties faced by them.28 Non-coffee producers, in contrast, 

have taken longer to recover in terms of food security and nutrition but do not show higher emigration 

intentions. In terms of ethnicity, indigenous households were affected more strongly by the pandemic and 

exhibit a slower recovery in food insecurity and dietary diversity than non-indigenous households. We 

additionally explored heterogeneous variations considering other household characteristics (e.g., gender, 

age, education of household head, or whether a household produces fruits and vegetables or animal source 

foods), but we did not find major differences in this regard.29 

 

 
28 For the evolution of international coffee prices, see http://www.ico.org/new_historical.asp (accessed on September 

2023). 

29 Results are available upon request. 

http://www.ico.org/new_historical.asp


28 
 

4.3. Quantitative estimates 

This section presents quantitative estimates for changes in incomes, expenditures, indebtedness, crop 

production decisions, and livestock ownership between 2019 and 2022. These estimates are based on the 

subsample of 777 households in our working sample who additionally participated in the fifth survey wave 

conducted during the end of 2022. As discussed above, these comparisons should be interpreted keeping in 

mind that they may include both the long-lasting effects of the pandemic and the ongoing impacts of the 

conflict in Eastern Europe.  

Table 3 presents average values for a range of quantitative outcomes reported for 2019 and 2022 

and indicates whether the difference between years is statistically significant at conventional levels. 

Considering the non-negligible inflation in Guatemala after the start of COVID-19, all 2022 monetary 

values are deflated to 2019 levels using the national consumer price index (CPI), thus allowing for a direct 

comparison across the two years.30 It is worth noting that the indebtedness and crop production information 

for 2019 was asked through recall during the 2022 survey, so the resulting indicators could be subject to 

anchoring bias and should thus be interpreted with caution (see , e.g., Godlonton, Hernandez, & Murphy, 

2018).31 

Panel A of the table focuses on quantitative differences in household income and expenditures. 

Overall, we observe a moderate decline of almost 16 percent in the average total household income between 

the pre-pandemic period and late 2022. When distinguishing by source, the income derived from 

agricultural activities shows a similar decline to total income, while the one derived from non-agricultural 

activities exhibits a small but not statistically significant increase of 4 percent. Income from remittances 

and from transfers, donations, and other sources also show decreases of 25 and 72 percent, respectively, 

but these represent a much smaller fraction of the total income for the average household in our sample. 

Total monthly expenditures, in contrast, show a substantial increase of almost 26 percent between 2019 and 

 
30 Monetary outcomes for 2022 were divided by 1.18, reflecting the total change in the CPI between the date of both 

surveys, and are trimmed beyond the 1st and 99th percentile to account for potential outliers. 

31 All other quantitative measures were collected during both the 2019 and 2022 surveys. 



29 
 

2022. The expenditure categories that show the largest, statistically-significant increases include food 

expenses, health expenses, and other expenses (such as transportation and utilities), while education 

expenses is the only category that shows a decrease. Housing expenses (such as rentals and dwelling 

improvements) more than doubled across the two years but the change is not significant. 

Panel B compares savings and debt indicators. In line with an increase in expenditures and a 

decrease in income, we observe increases both in the extensive and intensive debt margins. The proportion 

of households reporting having debt from any source increased by more than 10 percentage points between 

2019 and 2022, equivalent to a 53 percent increase from the 20 percent of households that reported holding 

debt in 2019. Most of this increase is driven by an increase in the percentage of households reporting 

holding debt from formal financial institutions, with only a slight but not significant increase in the fraction 

of households holding debt from family or relatives, and no seeming increase in the proportion of 

households holding debt from informal lenders. In terms of indebtedness levels, the overall debt amount 

increased by more than 59 percent, again mostly driven by debt from formal financial institutions and only 

marginal increases in the other sources of debt elicited. Regarding savings, on aggregate, the proportion of 

households reporting holding savings increased from 18 percent in 2019 to 24 percent in 2022 (equivalent 

to a 30 percent increase). Yet, among households reporting having savings in 2019 and/or 2022 (212 

households), 24 percent of them indicate that their amount of savings decreased.32 

Panels C and D analyze changes in crop production decisions and livestock ownership, 

respectively, with the aim of assessing whether the pandemic pushed households to diversify or specialize 

their agricultural production, or shift it towards certain products which could have potentially fared better 

during the study period. We generally find few differences in crop portfolios between the two years, both 

in terms of the number of crops produced (including total and by different categories) and the proportion 

 
32 Indicators around savings should be interpreted with caution, since due to the high rate of extortions in Guatemala 

households are usually reticent to disclose whether they hold savings or whether their savings have increased or 

decreased. These concerns should not apply to debt indicators, which should suffer from considerably less reporting 

bias. 



30 
 

of households reporting producing a given category of crops or coffee and cardamom, the two most 

important cash crops cultivated in the Western Highlands. The only exceptions are a slight decrease in the 

fraction of households producing subsistence crops (maize and beans) and the number of subsistence crops 

produced, and a slight increase of 1.4 percentage points in the fraction of households producing cardamom. 

While statistically significant, such differences do not seem to be economically important. Noticeably, we 

observe a very large increase in the proportion of households reporting owning hens (with no differences 

in ownership of bulls and cows or pigs). This pattern could be linked to a potential diversification of income 

generation activities that households may have been pursuing after the pandemic, and could be related to 

the fact that poultry farming was encouraged by the value chains program in which almost half of our study 

sample participated.33 

Finally, similar to Section 4.2, in Table 4 we assess differentiated changes in total income, 

expenditures, and debt levels between different groups of households, based on the regression framework 

described in Equation (2). In line with the findings above, we see that income decreased to a larger extent 

for households located in San Marcos and non-poor households. The changes in household expenditures, 

however, show varying patterns across groups. While expenditures increased for households in 

Huehuetenango and Quiche (who also exhibit a lower prevalence of income declines), expenditures in San 

Marcos did not change significantly. In addition, non-poor households show a decline in expenditures, in 

contrast to poor households, who report an increase. Coffee producers also exhibit a larger increase in 

expenditures than non-coffee producers, while we do not find major differences in expenditure changes 

among households affected or not by the ETA and IOTA tropical storms and by ethnicity.34 Finally, we 

generally do not find major variations in the increase in debt levels across groups, except for location where 

 
33 Most of the increase in the share of households owning hens in our sample is driven by the beneficiaries of the value 

chains program. 

34 The indicator variable capturing whether a household was affected by the ETA and IOTA tropical storms also 

includes whether the household was affected by the Julia tropical storm in October 2022. 



31 
 

households in San Marcos did not increase their debt level in contrast to households in Huehuetenango and 

Quiche. 

 

5. Conclusions 

In the year following the COVID-19 outbreak, the academic literature flourished with articles analyzing 

the short-term effects of the COVID environment on livelihoods, food security, and nutrition across the 

developing world. Shortly after, however, the stream of studies came to a halt, with scant recent efforts to 

assess extended effects of the COVID environment and of the path to recovery. This study intends to fill 

this gap by providing a long-term assessment of the effects of the pandemic and the recovery process in a 

context of subsequent shocks, across a sample of rural households in the Western Highlands of Guatemala. 

We exploit a rich array of panel data collected over four survey rounds between 2019 and 2022, that allows 

us to depict the evolution of a range of economic, food security, and nutrition outcomes, including income 

variations from various sources, the prevalence of food insecurity experiences, dietary diversity, and 

attitudes towards migration. A subsample of the study households was additionally interviewed during a 

fifth survey round at the end of 2022, further allowing us to examine quantitative changes in incomes, 

expenditures, indebtedness levels, and production decisions. Importantly, the analysis is enriched by 

considering the exposure of households to additional shocks, such as the tropical storms that hit the country 

during the study period and the implications of the recent Russia-Ukraine conflict. 

The results reveal an important recovery in incomes, food security, and dietary diversity among 

households in the study region, although the levels have overall not yet reached pre-pandemic ones. In 

particular, around one third of households still report lower incomes in 2022 compared to 2019. While food 

availability levels have returned to normal, prices have continued their upward trend across all food groups, 

in line with global food price inflation. The prevalence of severe food insecurity is lower than in 2019, but 

that of moderate and, in particular, mild food insecurity still remains higher than pre-pandemic levels. 

Dietary diversity scores have generally recovered, although the switch away from animal source foods 

(ASF) to a more diverse consumption of fruits and vegetables (F&V) persists since the onset of the 



32 
 

pandemic. The increase in migration intentions observed in 2021 continues into 2022, more than doubling 

2019 levels.  

The path of recovery, in turn, has not been equal for all, particularly in terms of food insecurity and 

dietary diversity. Households located in the department of San Marcos and who were above the poverty 

line in 2019 were initially more affected by the pandemic but recovered faster, while those affected by the 

ETA and IOTA tropical storms, non-coffee producers, and indigenous populations have taken longer to 

recover than their peer groups. 

Quantitative estimates for a subset of households interviewed in an additional survey round indicate a 

16 percent reduction in average total household income between 2019 and 2022, mostly driven by a decline 

in agricultural income. More worrisome, the reduction in income has been accompanied by an increase in 

household expenditures (26 percent) and an increase in indebtedness from formal institutions in both the 

extensive and intensive margin. The proportion of households reporting holding debt has increased by 10 

percentage points (from a baseline of 20 percent), while the average level of debt held has overall increased 

by 59 percent. 

All in all, we observe an important degree of recovery three years after the start of the pandemic, 

although this was attenuated to an extent by the additional global and local shocks faced by households. 

Nevertheless, the recovery trend should not serve to put policymakers’ minds at rest. While food security 

and dietary diversity have been increasing and approaching pre-pandemic levels, incomes are still lagging, 

combined with higher expenditures and indebtedness. In a context of a relative drying up of public and 

private aid, our findings show that more focus should be put into programs targeting vulnerable groups. 

The increasing debt burden on study households similarly demands a closer look at the terms and conditions 

of loans from formal institutions (which could be including some instances of predatory lending) and points 

to the importance of expanding more flexible financial mechanisms. Finally, the evidence shows a slower 

recovery path among households affected by the ETA and IOTA tropical storms, which remarks the need 

of promoting climate mitigation and adaptation initiatives, especially in a setting of recurrent extreme 

weather events such as the Western Highlands in Guatemala. In the wake of future climate or other systemic 



33 
 

shocks, special attention should be paid to affected households, since these tend to be the most vulnerable 

households in the region and after several years of the COVID-19 pandemic and subsequent shocks face a 

much more exhausted portfolio of coping mechanisms.   



34 
 

References 

Amare, M., Abay, K.A., Tiberti, L., & Chamberlin, J. (2021). COVID-19 and food security: Panel data 

evidence from Nigeria. Food Policy, 101, 102099. 

 

Ballard, T.J., Kepple, A.W., & Cafiero, C. (2013). The Food Insecurity Experience Scale: Development 

of global standard for monitoring hunger worldwide. FAO Technical Paper Version 1.1, October. 

 

Balana, B.B., Ogunniyi, A., Oyeyemi, M., Fasoranti, A., Edeh, H. & Andam, K. (2023). COVID-19, food 

insecurity and dietary diversity of households: Survey evidence from Nigeria. Food Security, 15, 

219–241 (2023).  

 

Bundervoet, T., Dávalos, M.E., & Garcia, N. (2022). The short-term impacts of COVID-19 on 

households in developing countries: An overview based on a harmonized dataset of high-frequency 

surveys. World Development, 153, 105844. 

 

Ceballos, F., Hernandez, M.A., & Paz, C. (2021). Short-term impacts of COVID-19 on food security and 

nutrition in rural Guatemala: Phone-based farm household survey evidence. Agricultural Economics, 

52(3): 477-494. 

Ceballos, F., Hernandez, M.A., & Paz, C. (2022). COVID-19 and extreme weather: Impacts on food 

security and migration attitudes in rural Guatemala. IFPRI Discussion Paper No. 2126. 

Ceballos, F., Kannan, S., & Kramer, B. (2020). Impacts of a national lockdown on smallholder farmers’ 

income and food security: Empirical evidence from two states in India. World Development, 136, 

105069. 

Ceballos, F., Kannan, S., & Kramer, B. (2021). Crop prices, farm incomes, and food security during the 

COVID‐19 pandemic in India: Phone‐based producer survey evidence from Haryana State. 

Agricultural Economics, 52(3), 525-542. 

Diaz-Bonilla, E., Flores, L., Paz, F., Piñeiro, V., & Zandstra, T. (2021). Guatemala: The impact of 

COVID-19 and policy implications: Second report (Vol. 21). International Food Policy Research 

Institute, Washington, DC. 

Dasgupta, S., & Robinson, E.J. (2022). Impact of COVID-19 on food insecurity using multiple waves of 

high frequency household surveys. Scientific Reports, 12(1), 1-15. 

FAO. (2010). Guidelines for measuring household and individual dietary diversity. Food and Agriculture 

Organization, Rome. http://www.fao.org/3/a-i1983e.pdf. 

http://www.fao.org/3/a-i1983e.pdf


35 
 

Godlonton, S., Hernandez, M.A., & Murphy, M. (2018). Anchoring bias in recall data: Evidence from 

Central America. American Journal of Agricultural Economics, 100(2), 479-501. 

Hailu, G., (2020). Economic thoughts on COVID‐19 for Canadian food processors. Canadian Journal of 

Agricultural Economics, 68(2), 163-169. 

Hale, T., Angrist, N., Goldszmidt, R., Kira, B., Petherick, A., Phillips, T., Webster, S., Cameron-Blake, 

E., Hallas, L., Majumdar, S., & Tatlow, H. (2021). A global panel database of pandemic policies 

(Oxford COVID-19 Government Response Tracker). Nature Human Behavior 5, 529-538.  

Hammond, J., Siegal, K., Milner, D., Elimu, E., Vail, T., Cathala, P., Gatera, A., Karim, A., Lee, J.-E., 

Douxchamps, S., Tu, M.T., Ouma, E., Lukuyu, B., Lutakome, P., Leitner, S., Wanyama, I., Thi, T.P., 

Phuc, P.T. H., Herrero, M., & van Wijk, M. (2022). Perceived effects of COVID-19 restrictions on 

smallholder farmers: Evidence from seven lower- and middle-income countries. Agricultural 

Systems, 198, 103367.  

Han, X., Guo, Y., Xue, P., Wang, X., & Zhu, W. (2022). Impacts of COVID-19 on Nutritional Intake in 

Rural China: Panel Data Evidence. Nutrients, 14(13), 2704. 

Helgeson, J., Aminpour, P., Fung, J., Roa Henriquez, A., Zycherman, A., Butry, D., Nierenberg, C., & 

Zhang, Y. (2022). Natural hazards compound COVID-19 impacts on small businesses 

disproportionately for historically underrepresented group operators. Natural Hazards Review, 23(1), 

1-15. 

Hirvonen, K., De Brauw, A., & Abate, G.T., (2021). Food consumption and food security during the 

COVID‐19 pandemic in Addis Ababa. American Journal of Agricultural Economics 103(3), 772-789. 

Hirvonen, K., Minten, B., Mohammed, B., & Tamru, S. (2021). Food prices and marketing margins 

during the COVID‐19 pandemic: Evidence from vegetable value chains in Ethiopia. Agricultural 

Economics, 52(3), 407-421. 

Hoehn-Velasco, L., Silverio-Murillo, A., Balmori de la Miyar, J.R., & Penglase, J. (2022). The impact of 

the COVID-19 recession on Mexican households: evidence from employment and time use for men, 

women, and children. Review of Economics of the Household, 20, 763-797. 

IFRC, (2021). 6-months Operation Updated, Central America: Hurricanes Eta & Iota. Situation Report, 

International Federation of Red Cross and Red Crescent Societies. 

https://reliefweb.int/report/guatemala/central-america-hurricanes-eta-iota-6-months-operation-update-

mdr43007  

Jeyakumar, A., Dunna, D., & Aneesh, M. (2022). Loss of Livelihood, Wages, and Employment During 

the COVID-19 Pandemic in Selected Districts of Chhattisgarh in India, and Its Impact on Food 

Insecurity and Hunger. Frontiers in Public Health, 10, 810772.  

https://reliefweb.int/report/guatemala/central-america-hurricanes-eta-iota-6-months-operation-update-mdr43007
https://reliefweb.int/report/guatemala/central-america-hurricanes-eta-iota-6-months-operation-update-mdr43007


36 
 

Kruczkiewicz, A., Klopp, J., Fisher, J., Mason, S., McClain, S., Sheekh, N.M., Moss, R., Parks, R.M., & 

Braneon, C. (2021). Opinion: Compound risks and complex emergencies require new approaches to 

preparedness. Proceedings of the National Academy of Sciences, 118(19). 

Kumar, A., Padhee, A.K., & Kumar, S. (2020). How Indian agriculture should change after COVID-19. 

Food Security, 12(4), 837-840. 

Laborde, D., Martin, W., Swinnen, J., & Vos, R. (2020). COVID-19 risks to global food security. 

Science, 369(6503), 500-502. 

Mahajan, K., & Tomar, S. (2021). COVID‐19 and supply chain disruption: Evidence from food markets 

in India. American Journal of Agricultural Economics, 103(1), 35-52. 

Mahul, O., & Signer, B. (2020). The perfect storm: How to prepare against climate risk and disaster 

shocks in the time of COVID-19. One Earth, 2(6), 500-502. 

Maldonado, R., & Harris, J. (2022). Las Remesas a Latinoamérica y el Caribe en 2021: El Esfuerzo de los 

Migrantes Durante la Crisis del COVID-19. Nota Técnica N° IDB-TN-2493, Inter-American 

Development Bank (IDB), Washington, DC. 

Mekanna, A.N., Panchal, S.K., & Li, L. (2022). Beyond lockdowns: a systematic review of the impacts of 

COVID-19 lockdowns on dietary pattern, physical activity, body weight, and food security. Nutrition 

Reviews, 81(7):790-803. 

Middendorf, B.J., Traoré, H., Middendorf, G., Jha, P.K., Yonli, D., Palé, S., & Vara Prasad, P.V. (2022). 

Impacts of the COVID‐19 pandemic on vegetable production systems and livelihoods: Smallholder 

farmer experiences in Burkina Faso. Food and Energy Security, 11(1), e337.  

Mishra, A., Bruno, E., & Zilberman, D. (2021). Compound natural and human disasters: Managing 

drought and COVID-19 to sustain global agriculture and food sectors. Science of the Total 

Environment, 754, 142210. 

Nchanji, E.B., Lutomia, C.K., Chirwa, R., Templer, N., Rubyogo, J.C., & Onyango, P. (2021). Immediate 

impacts of COVID-19 pandemic on bean value chain in selected countries in sub-Saharan Africa. 

Agricultural Systems, 188 (2021), 103034. 

Nolte, K., Sipangule, K., & Wendt, N. (2022). Agricultural households in times of crisis. The COVID-19 

pandemic, livelihoods and land-use decisions. Kiel Institute for the World Economy, 261843. 

Randell, H., Jiang, C., Liang, X.Z., Murtugudde, R., & Sapkota, A. (2021). Food insecurity and 

compound environmental shocks in Nepal: Implications for a changing climate. World Development, 

145, 105511. 

Rawal, V., Kumar, M., Verma, A., & Pais, J. (2020). COVID-19 lockdown: Impact on agriculture and 

rural economy. Society for Social and Economic Research, 13(2), 34. 



37 
 

Rudin-Rush, L., Michler, J.D., Josephson, A., & Bloem, J.R. (2022). Food insecurity during the first year 

of the COVID-19 pandemic in four African countries. Food Policy, 111, 102306. 

Tefera, S.A., Tadesse, T.B., & Asmare, G.W. (2022). Prevalence of household food insecurity in Ethiopia 

during the COVID-19 pandemic: Evidence from panel data. Scientific African, 16, e01141.  

United Nations (2022). Guatemala: Ciclon Tropical Julia. Informe de Situación. United Nations Office for 

the Coordination of Humanitarian Affairs. https://reliefweb.int/report/guatemala/guatemala-ciclon-

tropical-julia-informe-de-situacion-no-01-al-11-de-octubre-2022 

Van Hoyweghen, K., Fabry, A., Feyaerts, Wade, I., & Maertens, M. (2021). Resilience of global and local 

value chains to the Covid‐19 pandemic: Survey evidence from vegetable value chains in Senegal. 

Agricultural Economics, 52(3), 423-440. 

Wang, C., He, X., Song, X., Chen, S., & Luo, D. (2022). Dynamic livelihood impacts of COVID-19 on 

different rural households in mountainous areas of China. PLoS ONE, 17(9), e0273816. 

 

https://reliefweb.int/report/guatemala/guatemala-ciclon-tropical-julia-informe-de-situacion-no-01-al-11-de-octubre-2022
https://reliefweb.int/report/guatemala/guatemala-ciclon-tropical-julia-informe-de-situacion-no-01-al-11-de-octubre-2022


Figure 1. Households in the working sample 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

Note: This figure maps the location of all visited communities in 2019, 2020, 2021, and 2022. The location of a 

community is determined by taking the average coordinates across all households interviewed in it. 

  



39 
 

Figure 2. Changes in income by source (relative to 2019) 

 

 

Note: This figure shows the percentage of interviewed households that reported changes in incomes in 2020, 2021, and 2022 

with respect to 2019, by income source. The total number of observations is based on 1,262 interviewed households in the 

case of the change in agricultural and non-agricultural incomes; and on 392, 353, and 387 households, respectively, in 2020, 

2021, and 2022 for households who report receiving remittances in the baseline survey and/or each corresponding follow-up 

survey. Answer categories in the questionnaire were symmetric, with households being able to declare that their income had 

increased either a lot or a little; however, these are lumped together in a single “Increased” category  for clarity of exposition. 

 

 

 

  

35%

24%

13%

19%

11%

36%

28%

58%

35%

18%

8%

17%

0% 20% 40% 60% 80% 100%

Jun - 2021

33%

18%

15%

13%

8%

26%

30%

55%

47%

25%

19%

12%

0% 20% 40% 60% 80% 100%

Jun - 2022

72%

50%

32%

21%

13%

29%

5%

32%

31%

4%

9%

0% 20% 40% 60% 80% 100%

Remittances

Non-Agricultural Income

Agricultural income

Jun - 2020



40 
 

Figure 3. Coping mechanisms against income losses 

 
Note: This figure shows the percentage of interviewed households that reported using different coping mechanisms 

(multiple choice answers) when reporting an unambiguous decrease in their total household income (relative to 2019) 

during each follow-up survey round. The values are based on a sample of 866, 681, and 520 households that reported 

an unambiguous decrease in total income in, respectively, 2020, 2021, and 2022. 

  

33%

1%

2%

20%

21%

36%

19%

23%

7%

33%

30%

26%

25%

21%

2%

14%

31%

34%

0% 20% 40% 60% 80% 100%

None

Aid from government/non-government

organization

Aid from a relative or friend

Borrowing money

Selling goods/assets

Using savings

2020 2021 2022



41 
 

Figure 4. Changes in food availability and prices (relative to 2019) 
 

Panel A. Percentage of households reporting changes in food availability  

 

 

 

Panel B. Percentage of households reporting increases in food prices 

 

Note: Panel A shows the percentage of interviewed households that reported changes in food availability at their local 

market in 2021 and 2022 (relative to 2019) by food group. Panel B shows the percentage of households that reported 

increases in food prices at their local market in 2020, 2021, and 2022 (relative to 2019) by food group. Answer 

categories for the food prices question were: increased, decreased, remains the same, or don’t know; however, the 

figure only shows the “Increased” category for clarity of exposition. The values in both panels are based on a sample 

of 1,262 interviewed households.

90%

89%

89%

92%

88%

98%

98%

98%

98%

97%

92%

85%

84%

72%

62%

0% 20% 40% 60% 80% 100%

Grains and cereals

Roots and tubers

Fruits and vegetables

Meats

Dairy products

2020 2021 2022

29%

24%

22%

28%

22%

54%

59%

67%

63%

69%

18%

17%

11%

9%

9%

0% 50% 100%

Grains and cereals

Roots and tuber

Fruits and vegetables

Meats

Dairy products

2022 vs 2019

29%

27%

26%

25%

22%

55%

58%

59%

61%

64%

16%

15%

15%

14%

13%

0% 20% 40% 60% 80% 100%

Grains and cereals

Roots and tuber

Fruits and vegetables

Meats

Dairy products

2021 vs 2019



42 
 

Figure 5. Food insecurity, dietary diversity, and intention to emigrate over time by household groups 

Panel A: Evolution of outcomes by location 

 

 

Panel B: Evolution of outcomes by poverty level 

 

38.0%

**
87.4%

**
61.0%

41.7%

23.0%

**
84.3%

**
39.3% 26.4%

0%

20%

40%

60%

80%

100%

2019 2020 2021 2022

Moderate or severe food insecurity

6.6
*

6.4 6.6 6.8

7.7
**
7.0

**
7.3 7.6

0

2

4

6

8

10

12

2019 2020 2021 2022

Household Dietary Diversity Score

20.2%

**
93.5%

25.8%
17.7%

37.6%

**
86.2%

**
61.4%

41.9%

0%

20%

40%

60%

80%

100%

2019 2020 2021 2022

Moderate or severe food insecurity

7.7

**
6.1

**
6.8

*
7.5

6.7

6.6

6.7 6.9

0

2

4

6

8

10

12

2019 2020 2021 2022

Household Dietary Diversity Score

3.2%
**

0.0%

**
23.4%

**
25.8%

4.0%
*

2.4%

**
8.3%

*
6.6%

0%

5%

10%

15%

20%

25%

30%

2019 2020 2021 2022

Intention to emigrate

3.6%

2.2% **
9.8% **

7.8%

6.2%

**
1.7%

10.1%
12.4%

0%

5%

10%

15%

20%

2019 2020 2021 2022

Intention to emigrate



43 
 

Panel C: Evolution of outcomes by exposure to tropical storms 

 

Panel D: Evolution of outcomes by being a coffee-producer 

 

 

 

37.9%

**
83.5%

**
69.7%

39.9%

34.4%

**
89.6%

**
49.0% 39.3%

0%

20%

40%

60%

80%

100%

2019 2020 2021 2022

Moderate or severe food insecurity

35.0%

**
86.3%

**
54.8%

36.3%

39.8%

**
89.8%

**
72.1%

54.4%

0%

20%

40%

60%

80%

100%

2019 2020 2021 2022

Moderate or severe food insecurity

4.4% *
2.4%

**
10.4%

**
9.4%

1.8% 0.9%

7.1%

4.4%
0%

5%

10%

15%

20%

2019 2020 2021 2022

Intention to emigrate

6.8
*

6.6 6.8

*
7.0

6.6
**
6.0

**
6.0 6.3

0

2

4

6

8

10

12

2019 2020 2021 2022

Household Dietary Diversity Score

3.3%

2.4%

**
11.4%

**
10.3%

4.5%

**
1.9%

*
8.6% 7.1%

0%

5%

10%

15%

20%

2019 2020 2021 2022

Intention to emigrate

6.7 **
6.3

*
6.4

6.7

6.9 6.7 6.9 7.0

0

2

4

6

8

10

12

2019 2020 2021 2022

Household Dietary Diversity Score



44 
 

Panel E: Evolution of outcomes by ethnicity 

 

 

Note: This figure shows the evolution over time of key outcome variables for different groups of households. The sample includes 1,262 households interviewed 

across all four survey rounds. In Panel A, the total number of households located in San Marcos is 124 versus 1,138 in Quiche or Huehuetenango. In Panel B, the 

total number of poor households is 1,084 versus 178 non-poor households. In Panel C, the total number of households affected by the tropical storms ETA and 

IOTA is 544 versus 718 non-affected households. In Panel D, the total number of coffee producer households is 1,036 versus 226 non-coffee producer households. 

In Panel E, the total number of indigenous households is 1,002 versus 226 non- indigenous households. See the main text for more details about the household 

groups. The 2019 values reported in all figures are equivalent to the sample average for each household group in that year. The 2020, 2021 and 2022 values are 

obtained from generalized least squared (GLS) random effects estimations of Equation (1) for each outcome variable, accounting for a host of household 

socioeconomic characteristics collected during the 2019 baseline survey and community fixed effects. Asterisks indicate whether the outcome change between the 

2019 baseline and each follow-up survey are statistically significant at the 95% (**) or 90% (*) levels.

38.8%

**
88.7% **

63.3%

41.2%

24.6%

**
80.0%

*
37.3%

33.1%

0%

20%

40%

60%

80%

100%

2019 2020 2021 2022

Moderate or severe food insecurity

6.6 **
6.3

6.5 6.7

7.7 7.3 7.4 7.8

0

2

4

6

8

10

12

2019 2020 2021 2022

Household Dietary Diversity Score

3.5%

2.3%

**
9.0%

**
8.7%

5.8%

**
1.5%

**
13.1%

7.7%

0%

5%

10%

15%

20%

2019 2020 2021 2022

Intention to emigrate



45 
 

Table 1. Household characteristics 

Note: This table shows summary statistics of different household characteristics collected during the baseline survey 

in 2019 (Panel A) and of household reports indicating exposure to weather shocks and COVID-19 (Panel B) for the 

full study sample. The indicator for a household being affected by the tropical storms ETA and IOTA is based on 

household reports from the 2021 follow-up survey. Indicators for COVID-19 exposition (infection) are based on 

household reports from the 2021 and 2022 follow-up surveys. 

  

Characteristics  Mean 
Standard  

Min. Max. 
Deviation 

Panel A: Household characteristics at baseline (2019) 

Household head is male 0.849 0.358 0.000 1.000 

Household head age (in years) 48.034 13.428 16.000 94.000 

Household head has no education 0.369 0.483 0.000 1.000 

Household head did not complete elementary education 0.299 0.458 0.000 1.000 

Household head completed elementary education or higher 0.332 0.471 0.000 1.000 

Household head main spoken language is Spanish 0.288 0.453 0.000 1.000 

Household size 5.884 2.708 1.000 18.000 

Household is beneficiary of social school program 0.265 0.442 0.000 1.000 

Dwelling has finished walls 0.457 0.498 0.000 1.000 

Dwelling has finished ceiling 0.138 0.345 0.000 1.000 

Dwelling has finished floors 0.539 0.499 0.000 1.000 

Dwelling is connected to electricity 0.840 0.367 0.000 1.000 

Dwelling is connected to water network 0.851 0.356 0.000 1.000 

Dwelling is connected to drainage network 0.313 0.464 0.000 1.000 

Cooking fuel of household is electricity or gas 0.032 0.177 0.000 1.000 

Daily per capita expenditures (in Quetzales) 11.254 12.325 0.167 108.972 

Household owns a TV or radio 0.728 0.445 0.000 1.000 

Household owns a vehicle 0.243 0.429 0.000 1.000 

Household owns livestock 0.595 0.491 0.000 1.000 

Agricultural land size (in hectares) 0.907 1.339 0.007 21.840 

Agricultural land has irrigation system  0.147 0.354 0.000 1.000 

Coffee is among main crops produced 0.821 0.384 0.000 1.000 

Cardamom is among main crops produced 0.119 0.324 0.000 1.000 

Corn is among main crops produced 0.471 0.499 0.000 1.000 

Beans is among main crops produced 0.176 0.381 0.000 1.000 

Household had an internal migrant in the last 3 years 0.040 0.197 0.000 1.000 

Household is located in Huehuetenango 0.582 0.493 0.000 1.000 

Household is located in Quiche 0.320 0.467 0.000 1.000 

Household is located in San Marcos 0.098 0.298 0.000 1.000 

Panel B: Weather shocks and COVID-19 exposure 

Household affected by tropical storms ETA and IOTA in 2020 0.431 0.495 0.000 1.000 

Household affected by weather event over 12 months prior to 2022 survey 0.601 0.490 0.000 1.000 

Someone in household infected with COVID-19 as of 2022 0.089 0.284 0.000 1.000 

Someone in the community infected with COVID-19 as of 2022 0.353 0.478 0.000 1.000 

Number of observations    1,262 



46 
 

Table 2. Food insecurity, dietary diversity, and intention to emigrate over time 

 

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

Indicator 2019 2020 2021 2022 

 

Panel A: Prevalence of food insecurity  

Mild food insecurity 0.577 0.899*** 0.830*** 0.732***  

 (0.494) (0.301) (0.375) (0.443)  

Moderate food insecurity 0.349 0.865*** 0.579*** 0.392  

 (0.477) (0.342) (0.494) (0.488)  

Severe food insecurity 0.113 0.212*** 0.231*** 0.074**  

 (0.317) (0.409) (0.422) (0.263)  

Panel B: Dietary diversity indicators  

Household Dietary Diversity Score (HDDS), scale 0-12 6.803 6.513** 6.677 6.910  

 
(1.854) (1.602) (1.578) (1.560)  

Animal Source Food (ASF) score, scale 0-7 1.255 0.833*** 0.872*** 1.048***  

 
(1.023) (0.789) (0.849) (0.926)  

Fruits and Vegetables (F&V) score, scale 0-5 2.136 2.496*** 2.482*** 2.395**  

 (1.442) (1.340) (1.418) (1.336)  

Women Dietary Diversity Score (WDDS), scale 0-9 4.503 4.510 4.591 4.546  

 (1.430) (1.253) (1.271) (1.271)  

Children Dietary Diversity Score (DDS), scale 0-7  3.378 3.863*** 3.554 3.930***  

 (1.632) (1.233) (1.467) (1.226)  

Panel C: Intention to emigrate  

Intention to emigrate over subsequent three years 0.040 0.021** 0.098*** 0.085***  

  (0.195) (0.145) (0.298) (0.279)  

Note: This table reports the average and standard deviation (in parentheses) of different outcome variables of interest 

over the four survey rounds. Asterisks in columns (2), (3), and (4) indicate the statistical significance of tests of 

equality of means between baseline (2019) and each follow-up survey, based on estimations of Probit regressions (for 

outcome variables in Panel A and C) and OLS regressions (for outcome variables in Panel B), with standard errors 

clustered by community. The prevalence of food insecurity reported in Panel A corresponds to situations experienced 

over the months prior to each survey round and are part of the eight Food Insecurity Experience Scale (FIES) items 

proposed by Ballard et al. (2013). See the main text for the food groups comprised in each dietary diversity score 

reported in Panel B. Panel C reports whether a household member has the intention to migrate out of the country over 

the next three years. All reported indicators are based on 1,262 households surveyed in the four rounds, except for 

WDDS and children DDS. The number of surveyed households with women is 1,017 for the four rounds; the number 

of households with children is 225 in 2019, 168 in 2020,168 in 2021, and 143 in 2022. ***, **, and * denote statistical 

significance at, respectively, the 99%, 95%, and 90% levels. 

 

 

 



47 
 

Table 3. Quantitative changes in income, expenditures, indebtedness, crop production 

decisions, and livestock owernship between 2019 and 2022 

 

  (1) (2)   

Indicator 2019 2022   

 
Panel A: Monthly income and expenditures  

 

Total income (in 2019 Quetzales) 3,068.789 2,564.368 ***  

 (3,452.328) (3,203.572)   

Agricultural income (in 2019 Quetzales) 1,542.244 1,271.855 **  

 (1,998.749) (2,155.379)   

Non-agricultural income (in 2019 Quetzales) 766.168 793.489   

 (1,510.047) (1,588.172)   

Income from remittances (in 2019 Quetzales) 609.184 456.013   

 (1,882.374) (1,328.379)   

Income from transfers, donations, and other sources (in 2019 Quetzales) 151.193 43.011 ***  

 (688.670) (267.897)   

Total expenditures (in 2019 Quetzales) 1,543.316 1,937.097 ***  

 (1,201.197) (1,788.483)   

Food expenses (in 2019 Quetzales) 624.315 719.612 ***  

 (508.358) (536.826)   

Housing expenses (in 2019 Quetzales) 52.646 110.616   

 (212.923) (734.462)   

Consumer goods and durables expenses (in 2019 Quetzales) 166.487 199.880 *  

 (242.262) (459.699)   

Education expenses (in 2019 Quetzales) 48.187 30.299 *  

 (126.012) (196.245)   

Health expenses (in 2019 Quetzales) 97.290 143.937 **  

 (177.582) (504.547)   

Other expenses (in 2019 Quetzales) 554.392 732.754 ***  

  (614.193) (914.164)    

  

(Cont.) 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



48 
 

 

 (Cont.) 

 

 

  

  (1) (2)  

Indicator 2019 2022  

 
Panel B: Savings and debts  

 

Holds savings 0.181 0.237 ***  

 (0.386) (0.425)   

Holds debt 0.196 0.299 ***  

 (0.397) (0.458)   

Holds debt from formal financial institutions 0.149 0.223 ***  

 (0.357) (0.416)   

Holds debt from informal lenders 0.010 0.010   

 (0.101) (0.101)   

Holds debt from friends and/or relatives 0.048 0.079   

 (0.213) (0.269)   

Total debt (in 2019 Quetzales) 4,067.638 6,460.247 ***  

 (14,945.935) (18,396.063)   

Debt from formal financial institutions (in 2019 Quetzales) 3,294.353 5,314.831 ***  

 (14,161.076) (17,055.918)   

Debt from informal lenders (in 2019 Quetzales) 152.042 275.187   

 (1,869.903) (3,912.808)   

Debt from friends and/or relatives (in 2019 Quetzales) 621.244 870.229 **  

 (4,658.907) (6,477.964)   

Total debt (in 2019 Quetzales), conditional on reporting debt in 2019 and/or 2022 11,800.791 18,742.084 ***  

 (23,623.609) (27,444.121)   



49 
 

  (1) (2)   

Indicator 2019 2022   

 
Panel C: Crop production  

 

Number of crops produced 2.077 2.023   

 (1.232) (1.197)   

Number of subsistence crops produced 0.830 0.782 **  

 (0.836) (0.837)   

Number of other grains and tubers produced 0.022 0.026   

 (0.147) (0.159)   

Number of fruits and vegetables produced 0.306 0.287   

 (0.799) (0.722)   

Produces subsistence crops 0.551 0.517 **  

 (0.498) (0.500)   

Produces coffee 0.739 0.741   

 (0.439) (0.439)   

Produces cardamon 0.144 0.158 **  

 (0.351) (0.365)   

Produces other grains and tubers 0.022 0.026   

 (0.147) (0.159)   

Produces fruits & vegetables 0.176 0.185   

 (0.381) (0.389)    

Panel D: Livestock ownership  
 

Owns bulls/cows 0.073 0.076   

 (0.261) (0.265)   

Owns hens 0.521 0.767 ***  

 (0.500) (0.423)   

Owns pigs 0.225 0.268   

  (0.418) (0.443)    

Note: This table reports the average and standard deviation (in parentheses) of different outcome variables of interest 

in 2019 and 2022. Monetary values (i.e., incomes, expenditures, and debts) reported for year 2022 are adjusted to 

2019 prices using the national Consumer Price Index (CPI). Income and expenditures reported include both monetary 

and non-monetary values. Monthly total expenditures correspond to the aggregation of food, housing (e.g., housing 

or land rental, dwelling construction, expansion and improvements), consumer goods and durables (e.g., clothing, 

home furnishing, household assets), education (e.g., tuition and fees, books, school supplies, uniforms), health (e.g., 

medicines, health exams, hospitalizations), and other expenses (e.g., communication, transportation, utilities, 

household and personal services, miscellaneous) reported by the household. Subsistence crops include maize and 

beans. Indicators in Panel B and Panel C are based on 2019 recall data (reported in 2022) and 2022 concurrent data. 

Monetary indicators are trimmed beyond the 1st and 99th percentile to account for potential outliers. Asterisks indicate 

the statistical significance of tests of equality of means between baseline (2019) and the fifth survey wave (December 

2022), derived from estimations of Probit regressions (for dichotomic outcome variables) and OLS regressions (for 

amounts and/or numeric outcome variables), with standard errors cluste