HOUSEHOLD ALLOCATION AND EFFICIENCY OF TIME IN PAPUA NEW GUINEA Introduction 1 2 Photo: Thomas Perry/World Bank. © 2018 International Bank for Reconstruction RIGHTS AND PERMISSIONS This work is available under the Creative Commons Attribution 3.0 IGO license and Development/The World Bank (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons At¬tribution license, you are free to copy, distribute, 1818 H Street NW, Washington, DC 20433 transmit, and adapt this work, including for commercial purposes, under the Telephone: 202-473-1000; www.worldbank.org following conditions: Attribution – Please cite the work as follows: World Bank. 2018. Household Allocation and Efficiency of Time in Papua New Guinea. Washington, DC: World Bank. License: Creative Commons Attribution CC BY 3.0 IGO Translations – If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations – If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content – The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual SOME RIGHTS RESERVED component or part contained in the work will not infringe on the rights of those This work is a product of Consultants working for and the staff at The third parties. The risk of claims resulting from such infringement rests solely World Bank. The findings, interpretations, and conclusions expressed in with you. If you wish to reuse a component of the work, it is your responsibility this work do not necessarily reflect the views of The World Bank, its Board to determine whether permission is needed for that reuse and to obtain of Executive Directors, or the governments they represent. The World permission from the copyright owner. Examples of components can include, Bank does not guarantee the accuracy of the data included in this work. but are not limited to tables, figures, or images. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World All queries on rights and licenses should be addressed to World Bank Bank concerning the le¬gal status of any territory or the endorsement Publications, The World Bank Group. 1818 H Street NW, Washington, DC or acceptance of such boundaries. Nothing herein shall constitute or be 20433, USA; e-mail: pubrights@worldbank.org considered to be a limitation upon or waiver of the privileges and immunities Cover photo: Conor Ashleigh of The World Bank, all of which are specifically reserved. Report design: Heidi Romano 1 2 Photo: Thomas Perry/World Bank. TABLE OF CONTENTS Abstract 9 Acknowledgments 11 Abbreviations 13 Glossary 15 Executive Summary 17 I. Objective 17 II. Main Findings 18 A. Understanding Labor Dynamics in the Agricultural Sector in PNG 18 B. Impact on Production Efficiency 24 C. Impact on Household Income and Welfare 24 III. Main Recommendations 26 1. Introduction 29 1.1. Background and Context 29 1.1.1. Women in Agriculture 29 1.1.2. Women’s Empowerment 32 1.1.3. Allocation of Time within the Household 32 1.2. Study Rationale 33 1.3. Methodological Approach 33 Table of contents 3 2. Methodology and Data 35 2.1. Data 35 2.1.1. Time-Use: Framework 35 2.1.2. Time-Use: Data Collection 35 2.1.3. Women’s Empowerment Module 36 2.2. Empirical Strategy 36 3. Results 39 3.1. Gender Differences in Education, Employment and Income 39 Cocoa-growing areas 40 Coffee-growing areas 42 3.2. Women’s Empowerment and Intra-Household Decision Making 43 Cocoa-growing areas 43 Coffee-growing areas 44 Cocoa-growing areas 44 Coffee-growing areas 44 Cocoa-growing areas 49 Coffee-growing areas 49 3.3. Activities, Labor and Time-Use 54 Cocoa-growing areas 55 Coffee-growing areas 55 4 4. Assessing Time-Use Gender Discrimination 67 Cocoa-growing areas Regression Estimates 69 Coffee-growing areas Regression Estimates 70 Oaxaca-Blinder Decomposition 71 5. Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 75 Cocoa-growing areas 80 Coffee-growing areas 85 Cocoa-growing areas 86 Coffee-growing areas 87 6. Conclusion 89 References 95 Appendix 99 Appendix A Sampling Frame and Data Quality Assurance 99 Data Entry and Quality Assurance 100 Data Cleaning and Checks 100 Appendix B1 Index Weights (Cocoa) 106 Appendix B2 Index Weights (Coffee) 112 Appendix C Complementary Tables (Cocoa) 118 Appendix D Complementary Tables (Coffee) 138 Appendix E Literature Review 159 Table of contents 5 FIGURES Figure 1-A Men and women do not share the same household activities 19 Figure 2-A Men work longer hours in profitable activities 20 Figure 3-A Men and women perceive household decision making differently 22 Figure 4-A Family problems are more common in coffee-growing areas 23 Figure 1-C The welfare scale in cocoa-growing areas is positively affected by various variables 25 Figure 2-A Typical examples of activities on which a person may spend time during the course of a day 35 Figure 3-A The ‘average day’ for men and women (hours per activity) in cocoa-growing areas 56 Figure 3-B The ‘average day’ for men and women (hours per activity) in coffee-growing areas 57 Figure 3-C Time spent on daily activities by age, gender and area (in hours) – cocoa-growing areas 62 Figure 3-D Time spent on daily activities by Asset wealth quintile, gender and area (in hours) – cocoa-growing areas 63 Figure 3-E Time spent on daily activities by age, gender and area (in hours) – coffee-growing areas 64 Figure 3-F Time spent on daily activities by Asset wealth quintile, gender and area (in hours) – coffee-growing areas 65 Figure 5-A Correlates of cocoa income per tree 81 Figure 5-B Correlates of cocoa quality of pruning 82 Figure 5-C Correlates of coffee income per tree 83 Figure 5-D Correlates of coffee quality of pruning 84 BOXES Box 1-A Main Findings of The Fruit of Her Labor Report Relating to Household Allocation and Efficiency of Time in PNG 30 Box 1-B Main findings of ACIAR report 31 Box 2-A Composite Indicators 37 Box 3-A Gender differences in education, employment and income: Findings in cocoa-growing areas 40 Box 3-B Income differences between female and male-headed households in cocoa-growing areas 41 6 Box 3-C Gender Differences in Coffee-Growing Areas 42 Box 3-D Women’s empowerment and intra-household decision making in cocoa-growing areas 43 Box 3-E Activities, labor and time-use in cocoa-growing areas 55 Box 4-A Cocoa-growing areas: regression estimates 68 Box 4-B Coffee-growing areas: regression estimates 70 Box 5-A PPAP design and variables 79 Box 5-B Estimation results for household cocoa production variables 80 Box 5-C Cocoa-growing areas: main findings 86 Box 6-A Main findings 90 TABLES Table 1-A Role of partner in cocoa-growing areas 45 Table 1-B Role of partner in coffee-growing areas 47 Table 2-A Problems and decision making within the family (from the selected woman side) in cocoa-growing areas 50 Table 2-B Problems and decision making within the family (from the selected woman side) in coffee-growing areas 52 Table 3-A Daily activities: Frequencies and number of hours in cocoa-growing areas 58 Table 3-B Daily activities: Frequencies and number of hours in coffee-growing areas 60 Table 4-A Decomposition analysis of the time-use gender gap 72 Table 5-A Dependent variables and explanatory variables introduced in the regression model 76 Table of contents 7 8 Photo: Conor Ashleigh/World Bank ABSTRACT In Papua New Guinea, the only study to provide estimates of men’s and women’s time-use dates from the late 1990s and relies on data collected in 1992-93. An analysis of tasks undertaken by men and women, including domestic work, is useful to obtain a more complete picture of the labor uses of men and women. Towards this end The Household Allocation and Efficiency of Time in Papua New Guinea Report provides new insights on the gender division of labor in the agricultural sector in Papua New Guinea. This report looks at how gender-differentiated domestic work burdens impact the ability of women to allocate their labor to the cultivation, harvesting and processing of coffee and cocoa. The report identifies gender-disaggregated trends in time allocation and links these patterns to household welfare outcomes. The note also outlines recommendations to improve outcomes for women in Papua New Guinea within these two sectors. Abstract 9 10 Photo: Thomas Perry/World Bank. ACKNOWLEDGMENTS This report was prepared by Damien Echevin The team would like to thank the workshop participants (Consultant, World Bank) under the overall leadership at the March-April 2017 workshops in Kokopo. Sincere and guidance of Stephane Forman (Senior Agricultural thanks go to Johnson Riven (Gender Specialist, PPAP), Specialist, World Bank), Brenna Moore (Economist, Moniek Kindred (Nutrition Specialist, PPAP), and World Bank), Anuja Utz (Senior Operations Officer, Ali Winoto Subandoro (Nutrition Specialist, World World Bank), Kofi Nouve (Program Leader, World Bank), Bank) for their valuable recommendations. The team Gayatri Acharya (Lead Rural Development Economist, is grateful to Jethro Apinas (Project Coordinator, World Bank), and Alan Tobalbal Oliver (Agricultural PPAP), John Moxon and Jesse T. Anjen (Cocoa Board, Specialist, World Bank). PPAP), and to Potaisa Hombunaka, Bill Humphrey, Maureen Kahento and Richard Alepa (Coffee Industry The team would like to acknowledge a number of Corporation, PPAP) for their useful insights and advice. people who have provided inputs and guidance during the course of the report preparation. These include Thomas Perry (Communications Officer, World Bank) Tracey Newbury, Director of the Gender Equality and and Hamish Wyatt (Consultant, World Bank) provided Disability Inclusiveness Section of the Australian key communication advice. Gitanjali Ponnambalam Department of Foreign Affairs and Trade (DFAT). In (Country Program Assistant) provided vital addition, Sonya Sultan (Senior Social Development administrative support to the team. The report was Specialist, World Bank) and Melissa Williams (Senior edited by Rachel Kurzyp (Consultant) and designed Rural Development Specialist, World Bank) served as by Heidi Romano (Consultant). peer reviewers for the final draft of the report. This work greatly benefited from the support of various colleagues in Papua New Guinea, including Linda Ningo and Leo Darius who led the data collection team in cocoa-growing areas, and Gordon Wallangas and Verolina Mais who led the data collection team in coffee-growing areas. The team would also like to thank Harold Coulombe (Consultant) who led the impact evaluation survey for the Productive Partnership in Agriculture Project (PPAP) in Papua New Guinea. Acknowledgments 11 12 Photo: Thomas Perry/World Bank. ABBREVIATIONS ACIAR Australian Centre for International Agricultural Research ARB Autonomous Region of Bougainville CPB Cocoa Pod Borer DFAT Department of Foreign Affairs and Trade ENB East New Britain Province PGK Papua New Guinea Kina PNG Papua New Guinea PPAP Productive Partnership in Agriculture Project Abbreviations 13 14 Photo: Thomas Perry/World Bank. GLOSSARY Agreement index Composite indicator using agreement variables (partners agree on various topics) as primary indicators. Asset wealth index Composite indicator using household assets and housing characteristics as primary indicators. Composite index Linear combination of categorical data obtained from a multiple correspondence analysis or a factor analysis. Built this way, the composite indicator can be considered as the best regressed latent variable on K primary indicators, since no other explanatory variable is more informative. Family problem index Composite indicator using family problem variables (problems in family in the last two years) as primary indicators. Female decision index Composite indicator using female decision variables (decisions on various issues concerning the family) as primary indicators. Male decision index Composite indicator using male decision variables (decisions on various issues concerning the family) as primary indicators. Permission index Composite indicator using permission variables (ask permission to her/his partner to go to various places) as primary indicators. Glossary 15 16 Photo: Thomas Perry/World Bank. EXECUTIVE SUMMARY I. OBJECTIVE In the absence of detailed sex-disaggregated data on labor use, it is not possible to attribute to men or to The objective of this time-use and gender study is to women their respective contributions to value-addition better understand labor dynamics in the agricultural in these supply chains, nor to determine the specific sector in Papua New Guinea (PNG). distribution of income or other benefits from chain Specifically, the report assesses the impact of gender- activities between men and women. differentiated domestic work burdens on the ability of The Productive Partnership in Agriculture Project women to allocate their labor to the time-critical tasks (PPAP) intermediate impact evaluation survey of cultivation, harvesting and processing of agricultural provided an opportunity to conduct further research products – in particular for coffee and cocoa. The on labor allocation in the coffee and cocoa sectors report identifies gender-disaggregated trends in time to understand intra-household decision making and allocation and links these patterns to household its effects on women’s ability to participate in the welfare outcomes. It tests how different variables agricultural sector. PPAP is a $100 million project, (education, age, women’s empowerment, etc.) influence financed by The World Bank, International Fund allocation of labor to agriculture (vs. other tasks) within for Agricultural Development, European Union, households and if this influences household income Government of PNG and private stakeholders, generation and welfare. supporting smallholder cocoa and coffee development This report is the first known study of its kind. The only across the country. The PPAP evaluation survey has other study to provide estimates of men’s and women’s a sample size of around 1,480 households in three time-use in the coffee sector dates from the late 1990s areas – East New Britain Province (500 households), and relies on data collected in 1992–93 (Overfield the Autonomous Region of Bougainville (300 1998). The recent World Bank Group Fruit of Her Labor households), and the Highlands region composed report that focused on the coffee, cocoa, and fresh of Western Highlands, Jiwaka and Simbu Provinces produce sectors in PNG concluded that data on the (680 households). While the PPAP baseline survey did allocation of men’s and women’s labor to the range of provide information on the share of women receiving tasks along the supply chain, notably with respect to income from coffee, it did not address the underlying production and post-harvest processing, are virtually dynamics of intra-household decision making. non-existent. Executive Summary 17 The survey of PPAP cocoa farmers was conducted II. MAIN FINDINGS between April and August 2017 and between April and December 2017 for coffee farmers. The survey includes a rich set of modules on socio-demographic A. UNDERSTANDING LABOR DYNAMICS characteristics, occupation and work conditions, IN THE AGRICULTURAL SECTOR IN PNG household agriculture production, project participation, housing characteristics, water collection and sanitation, assets and equipment owned, participation 1. Men and Women Do Not Share the Same in associations and groups, and income and life- Activities or Tasks Within the Household satisfaction. Furthermore, two modules were added: a Men’s work is geared more towards cocoa or coffee time-use module and a women’s empowerment module. production whereas women are more focused on other agricultural activities (Figure 1-A). Women are more likely to run their own business alongside other farming activities (i.e. alternative crops), than working with cocoa or coffee. Women are generally involved in the lower-value stages of the cocoa value chain (e.g. harvesting and sale of wet beans), whereas men capture more of the value at later stages (e.g. drying and sale of dry beans). Although women understand that their activities in cocoa production are key to the quality of the end product, they are in fact more likely to engage in other agricultural production activities which give them a more direct gain. Men work longer hours in profitable activities, especially in cocoa and coffee activities, whereas women are particularly busy with domestic activities. The average number of hours spent in cocoa production by men is almost triple that of women in the cocoa-growing areas and double in the coffee- growing areas. Adding up all hours worked (including domestic work), women work on average 2.7 hours more per day than men in the cocoa-growing areas and 1.7 hours more per day in the coffee-growing areas (Figure 2-A). 18 FIGURE 1-A Understanding Labor Dynamics in the Agricultural Sector in PNG In the cocoa sector: 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 0 0 0 Men Women Self-employed 47% Self-employed 32% Self-employed in other agricultural activities 16% Self-employed in other agricultural activities 38% Other 37% Other 30% In the coffee sector: 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 0 0 0 Men Women Self-employed 40% Self-employed 14% Self-employed in other agricultural activities 16% Self-employed in other agricultural activities 41% Executive Summary 19 FIGURE 2-A MEN WORK LONGER HOURS IN PROFITABLE ACTIVITIES Cocoa Field Work Cocoa Processing 1.4 hours for Men 0.2 hours for Men 0.5 hours for Women 0.1 hours for Women Coffee Field Work Coffee Processing 1.3 hours for Men 0.2 hours for Men 0.6 hours for Women 0.2 hours for Women WOMEN ARE FREQUENTLY BUSY WITH DOMESTIC ACTIVITIES Daily Household Work in Cocoa-Growing areas Daily Household Work in Coffee-Growing areas 1.4 hours for Men 0.8 hours for Men 5.5 hours for Women 3.1 hours for Women 20 Education is an important factor to explain gender 2. Discrimination against Women’s Work differences in time-use, specifically in cocoa-growing Intra-household decision making ignores the areas. In these areas, years of schooling has a positive needs and capabilities of women which leads to and significant effect on total hours of work and formal discrimination and inefficiencies. While certain working time of women, and a negative effect on time characteristics, such as education and age, explain allocated by women to domestic work. Among other gender differences in hours worked, they don’t tell determinants are: the whole picture. Even in the presence of more • The number of children, which has a positive and educated women, the average time-use gaps significant effect on hours dedicated to domestic between women and men remain about the same. work, both for women and men; This highlights that characteristics between the genders only accounts for a small part of the time- • Literacy in Pidgin and having access to the internet, use gap. which has a negative and significant effect on total hours of work for women; and 3. Non-Cooperative Behavior • Age influences negatively total hours of work within the Household for men and domestic time for women. Intra-household decision making processes influence Unlike cocoa-growing areas, women’s empowerment the allocation of time and household efficiency. and decision-making variables have statistically When it comes to purchasing decisions, partners significant effects on time-use in coffee-growing areas. often make decisions together. However, we found that women are more likely than men to make decisions alone. Analysing purchasing decision behaviors shows that in cocoa-growing areas owning a phone or having access to the internet significantly correlates with higher bargaining power of women. In contrast, women’s bargaining power is lower when household asset wealth increases in couples, and within larger households. The patriarchy still has a strong influence in a matrilineal society because household decisions are family based and may not always recognise the power of women in matrilineage. Men and women perceive household decision making differently. These differences are symptomatic of limited rationality in decision making and non- cooperative behaviors. Whatever the destination, it appears that having to ask permission from the other partner to go to various places remains relatively frequent in cocoa-growing areas while it is less so in coffee-growing areas. However, women have to ask for permission more in both cocoa and coffee- growing areas (Figure 3-A). Executive Summary 21 FIGURE 3-A Men and Women Perceive Household Decision Making Differently Cocoa-Growing Areas 52% - 65% 63% - 79% of men have to ask for permission of women have to ask for permission Coffee-Growing Areas 26% - 37% 43% - 56% of men have to ask for permission of women have to ask for permission In cocoa-growing areas, partners frequently agree on most topics, and reported the main family problems in the last two years as being the following: 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 0 0 0 Men Women Lack of Money 42% Lack of Money 40% Illness of a Household Member 33% Illness of a Household Member 33% 22 Alcoholism and addiction of a household member and domestic violence concern women more than men. In coffee-growing areas, family problems are more common than in cocoa-growing areas in particular (Figure 4-A): FIGURE 4-A 20 40 20 40 60 80 60 100 80 100 20 40 20 40 60 80 60 100 80 0 0 0 0 Family Problems are More Common in Coffee-Growing Areas 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 0 0 0 Men Women Bad Relationships between Parents and Children 22% Bad Relationships between Parents and Children 28% Lack of money 79% Lack of money 76% Domestic Violence 40% Domestic Violence 40% Executive Summary 23 B. IMPACT ON PRODUCTION EFFICIENCY C. IMPACT ON HOUSEHOLD INCOME AND WELFARE Household efficiency and allocation of labor within the household remain closely linked; they both depend on Cocoa production and income represent only one-fifth various determinants: economic factors, bargaining of total household income in 2016, whereas coffee power and skills of household members, as well as represents two thirds. Male-headed households earn a non-economic factors. The impact these factors have higher income compared to female-headed households. on agricultural income is important as it determines the Food sufficiency is also higher among male-headed efficiency of production. households than among female-headed households In-depth analysis of household production and welfare (97 percent against 75 percent in cocoa-growing areas). requires an econometric approach that makes the In cocoa-growing areas, household income per most of the available data. Given the richness of the capita is positively affected by hours of formal work, data, omitted variable bias can be typically avoided asset wealth, living in Bougainville, the male decision when adding new regressors to the model. Endogeneity index, female selling cocoa, and female managing bias is also taken into account by using the dynamic household accounts. But it is negatively affected by nature of the data, for instance by introducing lagged the percentage of income accruing to alternative explanatory variables in regression models. However, crops production. Women who are more in control it is not possible to fully use the 2012–2017 panel of the sale of cocoa and the management of money because the time-use variables are only observable that comes from sales can increase their income and once in 2017. The instrumental variable method is household welfare. The man can still have important also difficult to implement because of the lack of valid decision making power within the household, in so far instruments. Instead, our approach lies in the reasoned as it allows an improved yield of agricultural production use of available data to measure the impact of time (farm income, and more specifically, cocoa income per allocation and other variables on household production tree). Other variables that have a significant negative and welfare. effect on per capita total household income include: The results show that higher bargaining power of • Female head; women and the availability of labor have significant positive impacts on cocoa production yield measured • Household size; by cocoa income per tree. Participation in agricultural • Permission index; groups and participation in PPAP have significant positive impacts on the number of trees in cocoa • Male involved in planning; production. The quality of pruning is also positively • Female involved in receiving payments (all things affected by involvement in PPAP and/or agricultural equal this variable is also negatively correlated group participation, as well as by other variables such with agricultural production and income); and as hours of formal work, household asset wealth, number of 13-17-year-old household members, and • Partners being “afraid to disagree.” the family problem variable. In coffee-growing areas, household income per capita is In coffee-growing areas, income per tree is positively affected by the same determinants as in cocoa-growing affected by the number of 18-59-year-old members and areas, except for women’s empowerment and decision- the women decision index and negatively affected by making variables which are not as significant. However, women’s empowerment indices such as the agreement a woman’s ability to make decisions still has a positive index and the family problem index. Regression models and statistically significant effect on income per capita, include women’s empowerment indices as explanatory while men making decisions has a negative impact variables. These indices are linear combinations on household well-being as measured by income per of categorical variables obtained from multiple capita. correspondence analysis. Using these indices, the The report further investigated the effect of different findings indicate that women’s empowerment generally variables on household welfare using the welfare scale improves household production and efficiency. which indicates the position of the household from 1 the poorest to 10 the richest, today, five years ago, and in five years (Figure 1-C). 24 FIGURE 1-C The Welfare Scale in Cocoa-Growing Areas is Positively Affected by Various Variables In cocoa-growing areas, it appears that the welfare scale today is positively affected by: Female hours of Asset wealth Living in The number The male Male involved domestic work Bougainville of members 60+ decision in planning and index decision making about cocoa production Welfare scale five years ago is positively and significantly correlated with: Asset wealth Number of Permission Agreement Family Women Females primarily index members 60+ index index problem index decision index involved in selling cocoa Welfare scale in five years in the future is negatively correlated with: Female hours of domestic work Welfare scale in five years in the future is positively correlated with: Asset wealth Living in Participation Family problem Females primarily Bougainville in PPAP index involved in selling cocoa Executive Summary 25 In coffee-growing areas, “bargaining power” or women III. MAIN RECOMMENDATIONS empowerment variables have a positive effect on the overall optimism of the household (measured From the results, it is possible to draw as self-perceived wealth in five years). The share of recommendations to improve household production alternative crop income in total income is significantly and welfare. First, household awareness and training and positively correlated with self-perceived wealth on gender dynamics and increased responsibilities today and five years ago, but has a negative effect on of women could improve welfare outcomes for self-perceived wealth in five years. Among women’s all household members. Indeed, it appears that empowerment and decision-making variables, the empowering women can improve household welfare permission index is positively correlated with wealth outcomes. For instance, the results show that scale today and five years ago, whereas the agreement household welfare outcomes are higher when women index is negatively correlated with the wealth scale have more control over cocoa sales and the resulting today, five years ago, and in five years. The family income, and empowered women are also more likely to problem index is negatively correlated with the wealth have an equal relationship with their male partner with scale five years ago. Both women decision index whom they are not afraid to disagree over household and female involved in planning and decision making decision-making. about coffee production have positive effects on self- Second, women in PNG carry a substantial burden perceived wealth today, five years ago, and of domestic work which leaves them little time to in five years. substantively engage in more value-added agricultural activities. Without a parallel effort to reduce the domestic burden, projects that seek to directly engage women in higher value agricultural activities may thus only result in a greater overall workload for women. The domestic workload may be reduced by technological interventions to reduce labour inputs, or by a more equal sharing of domestic tasks between household members through awareness-raising. 26 Photo: Thomas Perry/World Bank. Executive Summary 27 28 Photo: Thomas Perry/World Bank. 1. INTRODUCTION 1.1. BACKGROUND AND CONTEXT This report made a series of recommendations focused on improving outcomes for women in the agricultural Papua New Guinea is an agriculture-based economy: sector, which were being used to inform ongoing World the majority of the population lives in rural areas, and Bank and International Finance Corporation operations the agricultural sector accounts for approximately in the coffee, cocoa and horticulture sectors. The one-third of Gross Domestic Product. The sector report also identified the need to undertake a time-use is dominated by smallholder farming systems, with survey of the agricultural sector in PNG. almost all farmers growing subsistence food crops and an increasing number producing surpluses for sale in local markets. Smallholder farmers also engage in 1.1.1. WOMEN IN AGRICULTURE cash crop production, most notably coffee and cocoa, with around 50 percent of the labor force engaged The World Bank undertook a Country Gender in the production, processing, and sale of these two Assessment in 2012, which identified key gender commodities. issues in PNG (World Bank 2012). It specifically addressed issues related to the challenge of The Pacific region and PNG in particular, registers providing gender-inclusive access to employment and some of the worst gender indicators in the world in economic resources. The main findings relating to relation to political representation, gender-based economic opportunity remained too general to fully violence and access to economic opportunities. understand the challenges facing the agricultural Given the importance of the agricultural sector for sector, particularly cocoa, coffee, and horticultural employment and income generation, a specific study agribusinesses. on women in agriculture in PNG was commissioned by The World Bank Group in 2014. The Fruit of Her Labor: As mentioned, in 2014, The World Bank undertook Promoting Gender-Equitable Agribusiness in Papua New a more specific study on women in agriculture Guinea report focused on women’s engagement in the activities in PNG in order to achieve greater impact coffee, cocoa, and horticulture value chains (World for women from its current activities in agribusiness, Bank Group 2014). It found that labor allocation issues and to provide clear recommendations on additional fundamentally affect the performance of the coffee, interventions aimed at improving outcomes for women. cocoa, and fresh produce agribusiness supply chains in The report focused on the supply chains for coffee, PNG. Of particular importance are gender differences cocoa, and horticultural products (fresh produce) and in labor allocation and in rewards to labor, and the ways paid particular attention to the roles and constraints in which social, cultural and economic factors intersect faced by smallholders (Box 1-A). in determining labor use. Introduction 29 BOX 1-A Main Findings of The Fruit of Her Labor Report Relating to Household Allocation and Efficiency of Time in PNG + Women provide substantial labor in both + Farmers experience labor shortages. Households coffee and cocoa cultivation. Analysis of the do not have enough labor to do all the things supply chains indicates that the specific tasks they need to do the things at the right time women undertake have a substantial bearing and in the right way. However, the division of on the quality of the final product: women labor is unequal: women work more than men, are critical to improving the quality of coffee, especially when domestic work is included cocoa, and fresh produce in PNG. Women see (Overfield 1998). Cocoa Pod Borer exacerbates their main roles are in the weeding, picking, the labor constraint by requiring an even milling, and drying, land clearing, and selling more labor-intensive approach to cocoa block but not as much as men (Murray and Prior management and cultivation. 2014). These perceptions also bear out in the + Women are mostly confined to, and can only disproportionate burden of domestic work that benefit from, short supply chains. Lack of falls on women. mobility means that women are largely excluded + The ability of women in PNG to contribute to from key downstream activities along the improving the quality of coffee, cocoa, and fresh supply chains, where cocoa and coffee is sold to produce is affected by low economic incentives exporters (done by men, who, according to many for women either to allocate sufficient labor women interviewed, then pocket the cash). to these tasks or to do them well. There is a + Several key services are either absent or substantial gap between the work done by insufficient. This includes the limited reach, women in the coffee and cocoa sectors and the and focus of extension services, weak and benefit they obtain since women do much (if not inconsistent input supply, lack of new varieties most) of the work, but have much less access to, that are not readily available to farmers, and or control of, the resulting income. limited access to financial services. There are + Women’s access to the knowledge and skills also important gender-specific barriers to required to carry out these tasks is extremely accessing finance, as women tend not to own limited, as gaps in education, literacy, skills, and the land, fixed assets, or other resources that are participation in extension and training activities needed to meet collateral requirements. persist. + There are also important gender-specific dynamics at work in PNG society that affect men’s and women’s capacity to exercise economic agency differentially. PNG society is largely patriarchal, and even in matrilineal regions, men are seen as household heads and primary decision-makers. As a result, women have less access to, and control of, the resources needed to function economically, notably land and capital (financial services). 30 Another recent report by Australian Centre for International Agricultural Research (2017) titled Improving Livelihoods of Smallholder Families through Increased Productivity of Coffee-based Farming Systems in the Highlands of PNG provides interesting findings on income diversification and gender issues in coffee production areas (Box 1-B). BOX 1-B Main Findings of the ACIAR Report + There are strong economic incentives for + Households that work cooperatively and women to commit labor to vegetable and fruit harmoniously as a family, tend to have higher production because they are confident that their production (Curry and Koczberski 2004; Curry labor efforts in food production for markets will et al. 2007). Harmonious relationships among be rewarded through controlling the income family members help ensure their ongoing they earn. commitment to and participation in export crop production. + Uncertainty overpayment of women’s labor in coffee is one of the key drivers of women’s + Women and young males complained about emphasis on vegetable and fruit production in what they perceived as the unfair distribution areas with high market accessibility. of coffee income by the male head of the household. They felt that they, or their family as + When remuneration of women’s labor is a whole, were not benefiting from the income uncertain, they often withdraw all or part of the earned from coffee production. labor from export crop production and redirect it to activities where they have greater control over the income generated from their labor. Source: World Bank Group, (2014) Introduction 31 1.1.2. WOMEN’S EMPOWERMENT 1.1.3. ALLOCATION OF TIME WITHIN THE HOUSEHOLD Women’s empowerment and intra-household allocation of time may have an important impact on the Time-use studies measure two things: the quantity of performance of small agricultural units, especially in time spent on particular activities, and the quality of coffee, cocoa, and fresh produce agribusiness supply time spent on activities and the people concerned by chains in PNG. these activities. For the purpose of this study, some Women’s time allocation to different tasks can be important measurement issues can be pointed out: determined by both economic constraints (Becker • Do men’s and women’s participation in particular 1965; Gronau 1977) and cultural determinants or social agricultural activities result in less time available status (Khandker 1988; Eswaran et al. 2013). These for adequate care and feeding of young children, issues are keys to understanding the performance of or other activities which can determine household agribusiness supply chains. This is the case for five well-being? More generally, to what extent does principal reasons: women’s time-use contribute to household • Smallholders do not view their activity as production and well-being? a business, as they might value differently • What explains the differences observed between various activities during their daytime; women and men in terms of time-use? Does this • A lot of labor is allocated for social purposes, rely on different economic opportunities for women due to socio-cultural biases to allocate time; and men, or, otherwise, various legal, social or cultural determinants? • Farmers experience labor shortages, and this can be explained by gender differences in time-use; All members of the household must participate • Farming systems are highly diversified, as it is for in the time-use survey. To be able to make useful time-use within households; and comparisons, we need to know about the quantity and quality of time spent by all the women and men and • Women can only benefit from short supply chains, boys and girls within the household including domestic in part due to time (and socio-cultural) constraints. work so that a complete picture of labor use of men and women can be obtained. Identifying whether individuals are “time poor” also requires an understanding of work intensity that combines information on a full account of time spent in a given period as well as the drudgery and physical or mental effort associated with various tasks. 32 1.2. STUDY RATIONALE 1.3. METHODOLOGICAL APPROACH The objective of this time-use study was to better The PPAP Intermediate Impact Evaluation Survey understand labor dynamics in the agricultural sector provided a unique and valuable opportunity to address in PNG. Specifically, what is the impact of gender- the information gap identified in The Fruit of Her Labour differentiated domestic work burdens on the ability of report through the inclusion of a gender/time-use women to allocate their labor to the time-critical tasks module. Our time-use module has generated additional of cultivation, harvesting and processing of agricultural data on the allocation of time by men and women in the products – in particular coffee and cocoa? Our analysis coffee and cocoa sectors, including on: will help to improve understanding of: • Agricultural and non-agricultural • The balance between economic and social/other economic activities; activities for both men and women; • Domestic work related to household tasks (cooking, • Gender differences in labor use and availability elder and child care, fuel/water provisioning, in the coffee and cocoa sectors; household building/maintenance work); • Gender differences in the nature and extent of • Social and cultural activities, including church and labor constraint/shortages (whether seasonal community commitments, time spent accessing or task-specific) in these sectors; and and using social and other services (education, health); and • The implications of these different uses of time by men and women, and differences in the • Leisure/social activities. availability of time by men and women for sector strategies and expansion of economic activity in these sectors. The report aims to identify gender-disaggregated trends in time allocation and links these patterns to household welfare outcomes. The report shows how different variables (education, age, women’s empowerment, involvement in PPAP, etc.) influence allocation of labor to agriculture (vs. other tasks) within households and whether this influences household income generation. Introduction 33 34 Photo: Thomas Perry/World Bank. 2. METHODOLOGY AND DATA 2.1. DATA The broader evaluation survey includes a rich set of modules on socio-demographic characteristics, occupation and work conditions, household 2.1.1. TIME-USE: FRAMEWORK agriculture production, project participation, housing characteristics, water collection and sanitation, assets Time-use surveys are designed to account for the and equipment owned, participation to associations nature, duration, and location of all activities carried and groups, and income and life-satisfaction. out by the population during a reference period. The focus of these surveys is to understand human behavior and the lifestyle of people, especially for 2.1.2. TIME-USE: DATA COLLECTION the portion of their life for which no information is available from traditional data sources (Aguiar et al. The respondents’ daily activities were recorded through 2012; Charmes 2015; Hurst 2015; Seymour et al. 2017; face-to-face interviews, rather than asking them to fill UN 2004). A time-use survey gives a complete picture in a diary. This methodology has been used because of of the society by providing detailed information about the high level of illiteracy in PNG. The report considered how people spend their days (all 24 hours) on different a 24-hours diary with fixed 1-hour time slots. In economic and non-economic activities. Time-use each slot, respondents were asked to report if they surveys measure total time resources in terms of: performed more than one activity. A secondary activity could thus be reported. • Market activities; Time-use data was collected from all individuals in the • Productive domestic activities; and household aged 15 and over using the 24-hour recall • Leisure activities (producing satisfaction method. This method is considered more accurate as rather than goods). compared to others (for example a weekly recall period) because it is more detailed and easier for respondents This time-use study seeks to link time-use patterns to recall what they did the day before (Juster and to household welfare outcomes. This is done with Stafford 1991). It allowed the time-use data to be reference to the results of the other modules of the collected on one occasion in line with the design of the PPAP Intermediate Impact Evaluation Survey. overall PPAP Intermediate Impact Evaluation survey, which only included one visit per household (Figure2-A). FIGURE 2-A: Typical Examples Of Activities On Which A Person May Spend Time During The Course Of A Day Sleeping Eating Unpaid domestic Working in primary Unpaid care services services (for example, production (growing (care for children food preparation, of crops, animal and adult, teaching cleaning the dwelling, husbandry, and fishing) children, etc.) and shopping). and doing unpaid ‘economic’ work (such as fetching water or collecting firewood, or working unpaid in the family business) Methodology and Data 35 2.1.3. WOMEN’S EMPOWERMENT MODULE The variables belonging to the first category are the most likely to be exogenous. Although households may A random sample of women was selected to answer choose their level of education, geographic location questions in a women’s empowerment specific or the number of children, these variables remain module.1 Partners of these selected women were also relatively invariant and they are loosely instrumental for interviewed in a separate module to allow consideration policy purpose. of gender issues and intra-household bargaining. The The second set of categories are included in the women’s empowerment module includes a number of analysis as control variables. Opportunity costs as questions concerning the relationship between the measured by the salaries of outsiders can also be used partners: as control variables since they are distant proxies and • Whether she/he needs permission from her/his probably measured with error. partner to do special activities or purchases; Participation in groups and PPAP, in the third category, • Whether partners agree on various topics; are possibly endogenous due to self-selection. However, participation to PPAP should not be • Whether they are afraid to disagree; considered as strictly endogenous when it concerns the • Existence of family problems; and village and not just one given household. Indeed, using village participation to PPAP and regression controls is • Who takes important decisions in the family. a better approach in our context. Women’s empowerment might have beneficial effects Bargaining power variables in category four are the key on the household well-being. In particular, female variables of our analysis. However, their endogeneity is empowerment is particularly beneficial for children’s very likely and can’t be used to infer their causal impact health, nutrition, and education and can favor poverty with confidence. Despite this, the correlation between reduction and yield a higher level of development. time allocation and bargaining power or decision- making variables remains essential to fully understand intra-household behaviors. 2.2. EMPIRICAL STRATEGY Box 2-A provides information on the composite The report assessed the determinants of the allocation indicators. of time within the household in a regression framework using a wide range of household-level and individual- level variables available in the survey. These variables can be split into four broad categories: • Socio-demographic and endowments variables (age, years of schooling, literacy, household size and composition, education, training, and information), as well as geographic location variables; • Opportunity cost variables such as hours wage of an outsider; • Participation to associations and groups, and participation to PPAP; and • Bargaining power and decision-making variables such as: agreement index, family index, decision index (refer to Box 2-A and presentation in 1. A woman was selected in each household at random in order to Appendix B), male/female involved in planning and answer a specific module concerning women’s empowerment. decision making about cocoa/coffee production, In this perspective, she must meet certain criteria: be 15 years female primarily involved in selling coffee, female of age or over, and have a partner (or have had in the past a primarily involved in receiving payments for cocoa/ partner)–whatever her age. coffee, female manage account, afraid to disagree, and found at risk. 36 BOX 2-A Composite Indicators + Asset wealth index, permission index, + Appendices B-1 and B-2 present index weight, agreement index, family problem index and mean and partial inertia of explanatory decision index are measured and presented variables. As shown in Tables B1-1 and B2-1 for in Appendices B-1 (cocoa-growing areas) and the asset- wealth index, variables considered B-2 (coffee-growing areas). Each index is a in the analysis are asset variables, housing composite indicator that is a linear combination characteristics, agricultural materials owned of categorical variables obtained from a multiple by the household, sells of animals and bank correspondence analysis (see Asselin 2009): account ownership. Almost all variables have positive weights and partial inertia is K F d Indexi=∑k=1 1k ki indicated for each of them as being less than 0.1 (higher partial inertia for television and one block variable). Tables B1-2 and B2-2 present + Where Indexi is the value of the composite index the permission index for both women and for the ith observation (household or individual), men which measures the extent to which an dki is the value of the kth dummy variable (with individual has to ask permission to her/his k=1…K) describing the variables considered partner to go somewhere. Agreement index in the analysis (for instance, asset variables, weights are presented in Tables B1-3 and B2-3. housing characteristics and equipment variables in the case of the asset wealth index), and F1k is This index gets higher when partners agree the first component of the analysis. Built this on more topics. Family problem index is also a way the composite index can be described as the linear combination of main problems in family best regressed latent variable on the K primary with mostly positive weights of explanatory indicators, since no other explained variable is variables (Tables B1-4 and B2-4). more informative. Interestingly enough, domestic violence appears the most correlated variable. Finally, decision indices for both women (women decide) and men (men decide) are presented in Tables B1-5 and B2-5. Methodology and Data 37 38 Photo: PPAP, PNG 3. RESULTS 3.1. GENDER DIFFERENCES IN EDUCATION, EMPLOYMENT AND INCOME The PPAP survey shows us the educational characteristics of the members of the households surveyed. For individuals who attend the education system, the level in which they are enrolled can be observed. For individuals who are not enrolled in the education system, the report explores the reasons for non-participation and the level attained before stopping. Results 39 COCOA-GROWING AREAS Concerning cocoa-growing areas, the main findings are as follows (Box 3-A; also see Appendix C for complementary Tables C-1 to C-6). BOX 3-A Gender Differences in Education, Employment and Income: Findings in Cocoa-Growing Areas + There is no gender gap in school attendance + Women work more often in part-time roles (around 70 percent for both women and men) than men (67.5 percent vs. 62.1 percent). which is relatively high among younger aged + No clear gap between men and women in individuals (81 percent among 6-13-year-olds terms of occupations can be reported. and 88 percent among 14-18-year-olds). + More men are self-employed in the cocoa sector + There is not a gender gap in literacy and younger than women (47.3 percent vs. 32.0 percent). women (10-24 year-olds) are likely to be literate. However, it is important to note that this Conversely, literacy is higher among men aged statistic is not representative of the whole 40-years-old and over than among women of country as part of the sample of households the same age. has been selected because they were + More men than women are attending participating to the PPAP. university (12.2 percent vs. 3.9 percent). + Women are more self-employed in other + Years of schooling completed are significantly agriculture activities compared to men higher among men than among women (37.8 percent vs. 15.9 percent). (7.9 vs. 7.4). + More men are employed in the public + The labor force participation rate is higher and private sectors than women among men (79 percent) than among women (18.3 percent vs. 12.8 percent). (75 percent) when considering 10-69-year- + Unpaid family workers represent only 8.6 old people; however, the gender gap is not percent of women and 6.4 percent of men. significant among 25-69-year-olds (both Of those who do not work, only 5.6 percent women and men participation rates are around of women and 2.5 percent of men declare 95 percent). they have to care for children. 40 Overall, if we compare households headed by a woman with those headed by a man, one can make the following observations on household income (Box 3-B). BOX 3-B Income Differences between Female and Male-Headed Households in Cocoa-Growing Areas + The greatest gender income difference is + Family structure does not differ much between that men are more likely to hold a salaried job male-headed households and female-headed (4.3 percent of women-head households vs. households, except that single parenthood 11.4 percent of men-head households). Woman among female-headed households logically and man-headed households earn roughly the decreases average size of household: on same earnings from cocoa (which is main source average, there are 4.1 members in male-headed of income for 39.1 percent of female-headed households and 3.2 members in female-headed households vs. 47.9 percent of male-headed households. households) and other agriculture products + The gender difference in wealth is significant (45.7 percent vs. 33 percent). between male-headed households and female- + Gender gaps are higher when considering headed households when they self-assess their income by source. Incomes earned by wealth in five years: women appear to be less households whose head is a man are much optimistic than men (average wealth in five higher than those earned by households headed years is evaluated at 5.8 among female-headed by a woman: in particular, incomes from cocoa households, whereas it is 6.6 among male- dry bean, coconuts, off-farm, nonfarm, hunting, headed households). and fishing. These gender differences have widened since 2011, especially for off-farm income and total income. + Food sufficiency is higher among male-headed households than among female-headed households (96.5 percent vs. 75.0 percent). Once again, gender difference is much greater in 2016 than in 2011; in 2011, self-sufficiency was very high (90 percent of households were self-sufficient). Results 41 COFFEE-GROWING AREAS In coffee-growing areas (Box 3-C), similar patterns as in cocoa-growing areas are found with some noticeable differences (see Appendix D for complementary Tables D-1 to D-6). BOX 3-C Gender Differences in Coffee-Growing Areas + Men are generally more educated and attend + Unlike cocoa-growing areas, a large proportion school more often than women, especially of income comes from coffee activity (64 percent among 19-24-year-olds. Years completed is 5.0 as compared to only 19 percent for cocoa in for men and 3.5 for women. The literacy rate is 2016). It is also interesting to note that, in 2016, also higher among men than among women. coffee represents the main source of income for 96 percent of female-headed households + The employment rate is higher among women compared to 83 percent of male-headed (95 percent) than among men (92 percent). households. + Women are more often declared as a farmer + Income per capita is higher in female-headed (78 percent) than men (68 percent), while men households than in male-headed households are more often declared as clerical workers or although this is not statistically significant in as professional workers. 2016. Furthermore, unlike cocoa-growing areas, + More men are self-employed in the coffee sector we find no gender difference in self-assessed than women by a significant amount (40 percent wealth between male-headed households and vs. 14 percent). While 19 percent of women (vs. female-headed households. 11 percent for men) are employed as unpaid family workers. 42 3.2. WOMEN’S EMPOWERMENT AND INTRA-HOUSEHOLD DECISION MAKING COCOA-GROWING AREAS Main findings for cocoa are highlighted in Box 3-D. See Appendix C for Complementary Table BOX 3-D Women’s Empowerment and Intra-Household Decision Making in Cocoa-Growing Areas + Household asset wealth has a negative + The effect of these variables is reversed when and statistically significant effect on the the decision is made by both man and woman; occurrence that a woman makes purchasing indeed, less bargaining power should force a decisions alone, the same as marriage and woman to get along with her partner for household size; a possible interpretation purchasing decisions. of this result is that bargaining power of + Living in ARB (where matrilineality is women in the household seems to decrease widespread) does not have a statistically according to these variables. significant effect on the bargaining power of + It is significantly more common for older women. As such, patriarchy does not seem to women to make buying decisions alone; have a lesser influence in matrilineal society bargaining power being reinforced, and, because household decisions are family based interestingly, it is the same when holding and may not always recognise the power of her own phone or having internet access. women in matrilineage. The impact of technology on the bargaining power of women appears to be significant. Results 43 COFFEE-GROWING AREAS A significant difference exists between women and men concerning being afraid of disagreement with Fewer variables have a significant effect on decision partner and angriness with children (18.3 percent of making in coffee-growing areas. Age has a positive women and 9.1 percent of men). 26.7 percent of women effect on the occurrence that woman makes purchasing were found at risk due to their partner’s temperament, decisions alone, whereas the impact of PPAP is whereas only 9.3 percent of men declare they were negative (i.e. more decisions are made jointly between (statistically significant difference between both). husbands and wives in PPAP areas). Living in Simbu Finally, ownership of a phone is significantly higher for has a negative and statistically significant effect on men (60.5 percent declare they have their own phone) joint decisions. than for women (38.8 percent). Only 5.1 percent of Tables 1-A and 1-B below describes the role of a partner women and 6.7 percent of men have access to the from the perspective of both selected woman and her internet, without a statistically significant difference. partner. COFFEE-GROWING AREAS COCOA-GROWING AREAS In coffee-growing areas, 88.5 percent of the selected In cocoa-growing areas, 91.3 percent of the selected women currently have a partner, 11.5 percent declared women currently have a partner, 8.7 percent declared they had one in the past. Compared to cocoa-growing they had one in the past. areas, both women and men are less likely to ask permission to go somewhere. About half of the women Having to ask permission to go in various places ask for permission, compared to less than one-third remains relatively frequent in cocoa-growing areas, among men. Furthermore, four out of five people, with rates of around 63 percent to 79 percent among women and men, generally agree with their partner. selected woman, whereas 52 percent to 65 percent of This is a lower proportion than in cocoa-growing areas. men ask permission to their woman partner. Whatever Nevertheless, in coffee-growing areas, the gender gap the destination, statistically significant differences exist appears to be not statistically significant. between women and men. Compared to cocoa-growing areas, both men Partners seem to frequently agree on most topics, and women are less likely to consult their partners: with rates ranging from around 70 percent to over to buy clothes (67 percent among women vs. 90 percent for both female and male. Significant 47 percent among men), or for children purchases differences are observed between female (selected (76 percent among women vs. 59 percent among men). woman) and male (partner of selected woman) for Nevertheless, they are much more likely of being afraid agreement on family (90.7 percent for woman and to disagree: 57 percent of women and 31 percent of 94.5 percent for male), money (88.7 percent for men are afraid to disagree with their partner because woman and 93.5 percent for male), work (86.2 percent they will be angry with them; 35 percent of women for woman and 92.6 percent for male), relationship and 23 percent of men are afraid of disagreement between parents and children (87.8 percent for with partner and angriness with children. What is woman and 92.8 percent for male), and agreement on more, many feel at risk with a partner (55 percent education of children (90.9 percent for woman and among women vs. 27 percent among men). Hence, 95.5 percent for male). Over 80 percent of women the relationship between men and women appears and men consult their partner for buying clothes (no much more confrontational in coffee-growing areas. significant difference between women and men) and over 85 percent for children’s purchases (significant Finally, in coffee-growing areas, access to the mobile gender difference: 85.8 percent among women and phone and the internet is much less common than in 90.0 percent among men). the cocoa-growing areas. In cocoa-growing areas, about one-third of women are afraid to disagree with their partner because they will be angry with them, while only one-fifth of men are (a statistically significant difference). 44 TABLE 1-A: Role of Partner in Cocoa-Growing Areas FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Has a partner: Currently 91.3 551 100.0 418 0.000 In the past 8.7 551 0.0 418 0.000 Asks Permission from Partner to Go To: The market 78.8 551 58.6 418 0.000 The health center 79.3 551 65.3 418 0.000 The community center, neighbourhood park 77.1 551 61.5 418 0.000 A place of worship 63.0 551 52.2 418 0.001 Visit relatives in the neighbourhood 73.3 551 61.7 418 0.000 Visit friends in the neighbourhood 69.3 551 60.0 418 0.003 Partners Agree On: Religion 91.7 551 93.8 418 0.203 Politics 76.0 551 81.1 418 0.056 Family 90.7 551 94.5 418 0.024 Friends 75.1 551 77.0 418 0.492 Money 88.7 551 93.5 418 0.008 House work 69.7 551 72.0 418 0.431 Work 86.2 551 92.6 418 0.001 Moral rules 78.6 551 79.4 418 0.750 Relationship between parents and children 87.8 551 92.8 418 0.008 Education of children 90.9 551 95.5 418 0.004 Results 45 FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Consult partner to buy clothes 83.8 551 81.8 418 0.409 Consult partner for children-based purchases 85.8 551 90.0 418 0.050 Afraid to disagree with partner, angry with you 34.8 551 22.2 418 0.000 Afraid to disagree with partner, angry with 18.3 551 9.1 418 0.000 your children Found at risk with partner 26.7 551 9.3 418 0.000 Has her/his own phone 38.8 551 60.5 418 0.000 Partner pays for the phone services 8.9 214 3.2 253 0.011 Access to the internet 5.1 551 6.7 418 0.294 Internet Access: At work 14.3 28 32.1 28 0.112 At home 53.6 28 60.7 28 0.595 In a relative’s house 0.0 28 0.0 28 - In a friend’s house 0.0 28 0.0 28 - In an Internet cafe 0.0 28 0.0 28 - With cellphone 64.3 28 53.6 28 0.421 Other 0.0 28 0.0 28 - Sources: PPAP Survey, 2017. 46 TABLE 1-B Role of Partner in Coffee-Growing Areas FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Has a partner Currently 88.5 392 100.0 244 0.000 In the past 11.5 392 0.0 244 0.000 Ask Permission to Partner to Go To: The market 51.8 392 29.9 244 0.000 The health center 56.4 392 36.9 244 0.000 The community center, neighbourhood park 49.0 392 29.1 244 0.000 A place of worship 42.6 392 34.0 244 0.029 Visit relatives in the neighbourhood 53.6 392 27.9 244 0.000 Visit friends in the neighbourhood 54.3 392 26.2 244 0.000 Partners agree on: Religion 89.0 392 90.2 244 0.648 Politics 64.8 392 71.3 244 0.084 Family 90.3 392 89.8 244 0.822 Friends 70.7 392 68.0 244 0.486 Money 88.3 392 91.0 244 0.268 House work 75.8 392 68.9 244 0.060 Work 80.9 392 83.6 244 0.377 Moral rules 79.6 392 80.7 244 0.724 Relationship between parents and children 83.2 392 84.0 244 0.777 Education of children 90.6 392 92.6 244 0.356 Results 47 FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Consult partner to buy closes 66.6 392 47.1 244 0.000 Consult partner for children purchases 76.0 392 58.6 244 0.000 Afraid to disagree with partner, angry with you 57.1 392 30.7 244 0.000 Afraid to disagree with partner, angry with 34.9 392 23.4 244 0.001 your children Found at risk with partner 54.8 392 27.0 244 0.000 Has her/his own phone 13.8 392 33.6 244 0.000 Partner pays for the phone services 5.6 54 1.2 82 0.199 Access to the internet 1.8 392 2.5 244 0.574 Internet Access: At work 14.3 7 33.3 6 0.454 At home 14.3 7 50.0 6 0.178 In a relative's house 0.0 7 0.0 6 - In a friend's house 0.0 7 0.0 6 - In an Internet café 0.0 7 0.0 6 - With cellphone 85.7 7 83.3 6 0.914 Other 0.0 7 0.0 6 - Sources: PPAP Survey, 2017. 48 Tables 2-A and 2-B display the problems and decision COFFEE-GROWING AREAS making within the family in both cocoa and coffee- growing areas. In coffee-growing areas, family problems are more common than in cocoa-growing areas, in particular: bad relationship between parents and children COCOA-GROWING AREAS (28 percent according to women, 22 percent according to men), lack of money (resp. 76 percent In general, the main problems in the last two years and 79 percent), and domestic violence (40 percent were lack of money (for 39.7 percent of women and both for women and men). 42.1 percent of men) and illness of a member (around 33 percent among both men and women). The absence Moreover, even if the perceptions are different of father is high for women (24.5 percent), but not for between men and women, the latter seem less able men (4.1 percent), whereas the absence of mother to make decisions than in the cocoa-growing areas (in is significantly higher for men (10.5 percent) than 78 percent of cases woman decides, against 85 percent for women (2.7 percent). Alcoholism of a member, in cocoa-growing areas). However, according to women, domestic violence, and addiction of a member are a greater proportion of them make decisions alone significantly more of a problem according to women (12 percent), whereas only a few men do (1 percent). (respectively 22.0 percent, 10.5 percent, and Violence levels are higher in the Highlands in 3.8 percent) than according to men (respectively general, so the differences may reflect cultural 10.5 percent, 6.7 percent and 1.7 percent). differences between different geographical areas Decision making within the household is also perceived and ethnic groups. differently according to women and men. In most cases, partners take decisions together. However, it is interesting to note that according to women, they are more likely than men to make decisions alone (8 percent of women and 0 percent of men). Whereas perceptions of both women and men concerning women decisions are rarely statistically significantly different (except for spending on food and toiletries), they are significantly different when we consider women’s partner decisions (except taking the kids to play or having how many children). Results 49 TABLE 2-A Problems and Decision Making within the Family (from the Selected Woman Side) in Cocoa-Growing Areas FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Problem in Family in the Last Two Years Bad relationship between parents and children 6.9 551 5.0 418 0.218 Lack of money 39.7 551 42.1 418 0.460 Alcoholism of a member 22.0 551 10.5 418 0.000 Illness of a member 33.4 551 33.0 418 0.901 Lack of work of a member 18.3 551 20.6 418 0.384 Absence of the father 24.5 551 4.1 418 0.000 Absence of the mother 2.7 551 10.5 418 0.000 Lack of time 14.0 551 14.4 418 0.867 Addiction of a member 3.8 551 1.7 418 0.038 Domestic violence 10.5 551 6.7 418 0.033 Imprisonment of a member 1.6 551 0.7 418 0.178 Infidelity 1.5 551 1.2 418 0.729 Interference from other families in your relationship 12.5 551 11.0 418 0.466 Woman's Partner Decides Buy durable household goods 89.3 551 98.6 418 0.000 How much to spend on food and toiletries 82.2 551 89.2 418 0.002 Arrange/decorate the house 68.1 551 74.2 418 0.037 Send the children to school 77.5 551 84.0 418 0.010 Take children to medical checks 73.9 551 79.7 418 0.033 Take the children to the doctor when sick 76.0 551 84.0 418 0.002 If you must work outside the home or not 73.7 551 83.7 418 0.000 Having how many children 76.6 551 80.9 418 0.106 Take the kids to play 41.6 551 44.5 418 0.361 50 FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Woman Decides Buy durable household goods 95.3 551 94.5 418 0.585 How much to spend on food and toiletries 96.7 551 98.8 418 0.025 Arrange/decorate the house 93.6 551 95.9 418 0.108 Send the children to school 86.6 551 88.0 418 0.495 Take children to medical checks 90.0 551 88.8 418 0.529 Take the children to the doctor when sick 90.9 551 91.9 418 0.604 If you must work outside the home or not 79.9 551 76.6 418 0.220 Having how many children 79.1 551 81.1 418 0.446 Take the kids to play 51.0 551 47.8 418 0.331 Sources: PPAP Survey, 2017. Results 51 TABLE 2-B Problems and Decision Making within the Family (from the Selected Woman Side) in Coffee-Growing Areas FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Problem in Family in the Last Two Years Bad relationship between parents and children 28.1 392 22.1 244 0.090 Lack of money 75.8 392 78.7 244 0.391 Alcoholism of a member 16.1 392 8.6 244 0.004 Illness of a member 30.9 392 38.1 244 0.063 Lack of work of a member 17.9 392 18.4 244 0.853 Absence of the father 19.1 392 5.3 244 0.000 Absence of the mother 1.5 392 6.1 244 0.005 Lack of time 16.3 392 24.2 244 0.018 Addiction of a member 4.1 392 7.0 244 0.132 Domestic violence 39.8 392 39.8 244 0.992 Imprisonment of a member 0.8 392 1.6 244 0.345 Infidelity 2.3 392 0.4 244 0.029 Interference from other families in your relationship 21.2 392 20.1 244 0.741 Woman's Partner Decides Buy durable household goods 79.6 392 93.0 244 0.000 How much to spend on food and toiletries 69.6 392 77.5 244 0.028 Arrange/decorate the house 54.6 392 57.0 244 0.558 Send the children to school 81.6 392 90.2 244 0.002 Take children to medical checks 62.2 392 68.4 244 0.108 Take the children to the doctor when sick 70.2 392 76.6 244 0.069 If you must work outside the home or not 48.2 392 61.1 244 0.001 Having how many children 84.7 392 93.0 244 0.001 Take the kids to play 20.9 392 28.7 244 0.029 52 FEMALE MALE DIFF (SELECTED WOMAN) (PARTNER) Mean N Mean N p-value Woman Decides Buy durable household goods 88.8 392 83.2 244 0.053 How much to spend on food and toiletries 91.6 392 86.5 244 0.050 Arrange/decorate the house 88.3 392 84.8 244 0.224 Send the children to school 91.6 392 91.4 244 0.934 Take children to medical checks 82.1 392 82.8 244 0.835 Take the children to the doctor when sick 92.6 392 96.3 244 0.039 If you must work outside the home or not 55.9 392 59.0 244 0.435 Having how many children 80.1 392 84.0 244 0.207 Take the kids to play 29.6 392 37.3 244 0.046 Sources: PPAP Survey, 2017. Results 53 3.3. ACTIVITIES, LABOR AND TIME-USE Trends in activities (frequencies and working days) are described in Appendices C and D. The focus here is on hours of work. Tables 3-A and 3-B present daily activity frequencies, number of hours for individuals participating in a given activity, and number of hours all individuals combined (participating or not) using data from the time-use module. Activities are grouped into seven categories: • Personal care (Sleeping and resting, Eating, and Personal care); • Formal work (Work as an employee); • Primary production (Cocoa field work, Cocoa processing, Other farming, Animal rearing, and Fishing); • Non-primary production (Own business work: non-agriculture or livestock); • Domestic services and care (Shopping/getting services, Sewing, weaving, other textile care, Cooking, Other domestic work: washing, cleaning, Care for children, and Care for adults/elderly); • Learning activities (School or homework); and • Other non-productive / leisure activities (Commuting/Travelling, Watching TV, Listening Radio, Reading, Sitting with family, Sports, Social visits, Practicing hobbies, Ceremony, Others, and Election). Box 3-E provides the main findings for cocoa-growing areas. 54 COCOA-GROWING AREAS COFFEE-GROWING AREAS Half of all women and men are involved in primary production. A quarter of women are involved in coffee field work and a larger proportion in farming work. BOX 3-E: Domestic services and care activities are Activities, Labor and Time-Use also a priority for women (78 percent for women vs. in Cocoa-Growing Areas 28 percent for men). Women also spend less time on non-productive and leisure activities compared to men in both cocoa and coffee-growing areas (Figures 3-A + Primary production concerns about and 3-B). 44 percent of men and 37 percent of women Age and gender have some effect on the total time in cocoa-growing areas, with a statistically spent on each activity (Figures 3-C for cocoa and 3-E significant difference between who spends for coffee). Regarding primary production and formal their time doing what activities. The largest work, we detect significant differences in the hours per day spent on work by gender. In cocoa-growing areas, gap is for domestic services and care females spend fewer hours working for formal work or activities with women doing considerably primary production than males at all ages. In addition, more work in this area. we find a much flatter age profile for females than for males, reflecting the lower labor force participation + Women tend to do more multi-tasking. of females, even at young ages. However, older adults This is why the total number of hours per gradually reduce their time spent doing domestic work day added up to 26.9 hours for women and before reaching 65 years old. In coffee-growing areas, 25.4 hours for men. while age-profiles of time-use are very similar to those in cocoa-growing areas, women generally declare they + Women work on average 2.8 hours more work fewer hours for domestic services and care. in domestic activities than men. When There is a minor relationship between household asset considering only agricultural or non- wealth and total time spent on each activity by gender agricultural production activities and formal and areas (Figures 3-D for cocoa and 3-F for coffee2) In work, it is men who work more than women, cocoa-growing areas, we find that profiles are relatively flat for both males and females. Time spent on primary about 1.4 hours on average per day. production is slightly lower among richer men although it might not be statistically significant. In coffee-growing areas, time spent for domestic services and care is increasing among women. This latter observation might be due to social status determinants - women work less outside when they get richer (Eswaran et al. 2013). 2. Asset wealth index is obtained from multiple component analysis (first component of the analysis is used) which is presented in Table B1-1 for cocoa and Table B2-1 for coffee (see Box 1 and Appendix B). Results 55 FIGURE 3-A The ‘Average Day’ for Men and Women (Hours per Activity) in Cocoa-Growing Areas Men Women 0 5 10 15 20 25 30 Women Men Personal care 13.3 13.2 Formal work 0.5 0.8 Primary production 2.1 2.8 Coffee field work 0.5 1.4 Coffee processing 0.1 0.2 Other farming 1.5 1.0 Non-primary production 0.2 0.5 Domestic services and care 5.5 1.4 Learning activities 1.0 1.1 Other non-productive / leisure activities 4.3 5.4 Sources: PPAP Survey, 2017. Note: total hours can be more than 24 due to secondary activity. 56 FIGURE 3-B The ‘Average Day’ for Men and Women (Hours per Activity) in Coffee-Growing Areas Men Women 0 5 10 15 20 25 Women Men Personal care 10.0 9.6 Formal work 0.1 0.3 Primary production 2.6 2.7 Coffee field work 0.6 1.3 Coffee processing 0.2 0.2 Other farming 1.6 1.0 Non-primary production 0.4 0.5 Domestic services and care 3.1 0.8 Learning activities 0.6 0.8 Other non-productive / leisure activities 1.8 2.9 Sources: PPAP Survey, 2017. Note: total hours can be less than 24 due to misreported activity. Results 57 TABLE 3-A Daily Activities: Frequencies and Number of Hours in Cocoa-Growing Areas FEMALE MALE DIFF Mean N Mean N p-value Personal Care Frequency (%) 99.6 1077 99.9 1154 0.164 Number of hours 13.4 1073 13.3 1153 0.394 Frequency x Number of hours 13.3 1077 13.2 1154 0.561 Formal Work Frequency (%) 5.6 1077 9.3 1154 0.001 Number of hours 8.2 60 8.7 107 0.181 Frequency x Number of hours 0.5 1077 0.8 1154 0.000 Primary Production Frequency (%) 36.9 1077 43.5 1154 0.001 Number of hours 5.6 397 6.4 502 0.000 Frequency x Number of hours 2.1 1077 2.8 1154 0.000 Cocoa Field Work Frequency (%) 8.4 1077 25.7 1154 0.000 Number of hours 5.7 91 5.6 297 0.888 Frequency x Number of hours 0.5 1077 1.4 1154 0.000 Cocoa Processing Frequency (%) 1.4 1077 3.7 1154 0.000 Number of hours 4.7 15 4.8 43 0.881 Frequency x Number of hours 0.1 1077 0.2 1154 0.001 Other Farming Frequency (%) 28.9 1077 17.4 1154 0.000 Number of hours 5.1 311 5.6 201 0.043 Frequency x Number of hours 1.5 1077 1.0 1154 0.000 58 FEMALE MALE DIFF Mean N Mean N p-value Non-Primary Production Frequency (%) 3.7 1077 8.6 1154 0.000 Number of hours 5.3 40 6.2 99 0.060 Frequency x Number of hours 0.2 1077 0.5 1154 0.000 Domestic Services and Care Frequency (%) 78.3 1077 34.0 1154 0.000 Number of hours 7.1 843 4.2 392 0.000 Frequency x Number of hours 5.5 1077 1.4 1154 0.000 Learning Activities Frequency (%) 13.8 1077 15.0 1154 0.437 Number of hours 7.4 149 7.4 173 0.800 Frequency x Number of hours 1.0 1077 1.1 1154 0.414 Other Non-Productive / Leisure Activities Frequency (%) 77.1 1077 81.5 1154 0.011 Number of hours 5.5 830 6.7 940 0.000 Frequency x Number of hours 4.3 1077 5.4 1154 0.000 Sources: PPAP Survey, 2017. Results 59 TABLE 3-B Daily Activities: Frequencies and Number of Hours in Coffee-Growing Areas FEMALE MALE DIFF Mean N Mean N p-value Personal Care Frequency (%) 99.3 738 99.3 917 0.954 Number of hours 10.1 733 9.7 911 0.066 Frequency x Number of hours 10.0 738 9.6 917 0.073 Formal Work Frequency (%) 1.1 738 4.5 917 0.000 Number of hours 5.0 8 6.7 41 0.223 Frequency x Number of hours 0.1 738 0.3 917 0.000 Primary Production Frequency (%) 52.4 738 51.4 917 0.663 Number of hours 4.9 387 5.3 471 0.088 Frequency x Number of hours 2.6 738 2.7 917 0.367 Cocoa Field Work Frequency (%) 13.3 738 33.0 917 0.000 Number of hours 4.2 98 3.9 303 0.264 Frequency x Number of hours 0.6 738 1.3 917 0.000 Cocoa Processing Frequency (%) 6.2 738 7.7 917 0.229 Number of hours 2.8 46 2.7 71 0.770 Frequency x Number of hours 0.2 738 0.2 917 0.399 Other Farming Frequency (%) 45.3 738 31.0 917 0.000 Number of hours 3.5 334 3.2 284 0.101 Frequency x Number of hours 1.6 738 1.0 917 0.000 60 FEMALE MALE DIFF Mean N Mean N p-value Non-Primary Production Frequency (%) 8.4 738 9.8 917 0.319 Number of hours 4.3 62 5.4 90 0.113 Frequency x Number of hours 0.4 738 0.5 917 0.074 Domestic Services and Care Frequency (%) 77.8 738 28.2 917 0.000 Number of hours 3.9 574 2.8 259 0.000 Frequency x Number of hours 3.1 738 0.8 917 0.000 Learning Activities Frequency (%) 10.6 738 12.3 917 0.264 Number of hours 5.9 78 6.4 113 0.315 Frequency x Number of hours 0.6 738 0.8 917 0.140 Other Non-Productive / Leisure Activities Frequency (%) 52.2 738 61.8 917 0.000 Number of hours 3.4 385 4.6 567 0.000 Frequency x Number of hours 1.8 738 2.9 917 0.000 Sources: PPAP Survey, 2017. Results 61 FIGURE 3-C Time Spent on Daily Activities by Age, Gender and Area (in Hours)—Cocoa-Growing Areas 9 8 7 6 5 4 3 2 1 0 30-34 40-44 50-54 60-64 30-34 40-44 50-54 60-64 35-39 45-49 55-59 35-39 45-49 55-59 20-24 25-29 20-24 25-29 15-19 15-19 Men Women Formal work Primary production Domestic services and care Sources: PPAP Survey, 2017. 62 FIGURE 3-D Time Spent on Daily Activities by Asset Wealth Quintiles, Gender and Area (in Hours)—Cocoa-Growing Areas 16 14 12 10 8 6 4 2 0 Poorest 2 3 4 Richest Poorest 2 3 4 Richest Men Women Personal care Formal work Primary production Non-primary production Domestic services and care Learning activities Other non-productive / leisure activities Sources: PPAP Survey, 2017. Note: Asset wealth index is obtained from multiple component analysis Results 63 FIGURE 3-E Time Spent on Daily Activities by Age, Gender and Area (in Hours)—Coffee-Growing Areas 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 30-34 40-44 50-54 60-64 30-34 40-44 50-54 60-64 35-39 45-49 55-59 35-39 45-49 55-59 20-24 25-29 20-24 25-29 15-19 15-19 Men Women Formal work Primary production Domestic services and care Sources: PPAP Survey, 2017. 64 FIGURE 3-F Time Spent on Daily Activities by Asset Wealth Quintile, Gender and Area (in Hours)—Coffee-Growing Areas 12 10 8 6 4 2 0 Poorest 2 3 4 Richest Poorest 2 3 4 Richest Men Women Personal care Formal work Primary production Non-primary production Domestic services and care Learning activities Other non-productive / leisure activities Sources: PPAP Survey, 2017. Note: Asset wealth index is obtained from multiple component analysis. Results 65 66 Photo: Thomas Perry/World Bank. 4. ASSESSING TIME-USE GENDER DISCRIMINATION Using an econometric approach makes it possible to decompose the gender gap using an Oaxaca-Blinder type decomposition. The report applies the Oaxaca- Blinder method (Oaxaca 1973) to decompose the observed differential in the daily hours spent by men and women on each of the following four aggregated activities: primary production, other work, domestic work, and other non-productive and leisure activity. Assessing Time-Use Gender Discrimination 67 COCOA-GROWING AREAS REGRESSION ESTIMATES Findings on the effects of different variables on the allocation of time within the household are presented for cocoa in Appendix C using a large set of variables (Tables C-13 and C-14). Main results are as follows (Box 4-A): BOX 4-A Cocoa-Growing Areas: Regression Estimates + Age has generally a negative and significant + Being a female head and being married have a effect on domestic work and learning hours negative and significant effect on learning time for women, although not on other time-use for women; being married also has a negative variables (either working or leisure time); for effect on domestic work and learning time for men, age has a statistically significant negative men. effect on total hours of work. + The number of under 12-years-old children + Years of schooling variable has a positive and increases domestic work for both women and significant effect on total hours of work and men; it has a negative effect on time allocated formal working time, and a negative effect to personal care for women; also, it has a on domestic working time for women; no positive effect on time allocated to non-primary significant effects can be reported for men. production for men. + Literacy in Pidgin has a negative and + Other variables such as women’s empowerment significant effect on total hours of work. and decision-making variables are not correlated with allocation of time within the + Having access to the internet has a negative household for women, except that decision index and significant effect on total hours of work decreases domestic working time (the more for women; whereas, for men, it has a positive women decide the less their domestic work); influence on learning time. also, managing a bank account for women has a positive effect on their time allocated to formal work. 68 Photo: Thomas Perry/World Bank. Assessing Time-Use Gender Discrimination 69 COFFEE-GROWING AREAS REGRESSION ESTIMATES These are presented in Box 4-B. BOX 4-B Coffee-Growing Areas: Regression Estimates + Effects of age, years of schooling and literacy for + Unlike cocoa-growing areas, many women’s women are very similar to those found in cocoa- empowerment and decision-making variables growing areas; what is more, English literacy have statistically significant effects on time- has a positive and significant effect on both total use. In particular, the family problem index hours of work and primary production hours has a negative effect on women total hours of (Tables D-17 and D-18 in Appendix D). work and primary production, while it has a positive effect on formal work and domestic + The asset wealth index and the participation work hours. When women decide, they decrease to PPAP dummy have interesting effects on their domestic work hours, while when men time-use. Among women, the asset wealth index decide, they decrease their total hours of work has a positive and significant effect on total in particular primary production hours. Women hours of work, primary production, domestic involved in planning and decision making about services and personal care. These effects are coffee production has a negative and significant not statistically significant for men. It appears effect on men total hours of work; it is also that PPAP participation has a positive impact on positively correlated with men domestic women domestic work. work hours. 70 OAXACA-BLINDER DECOMPOSITION In cocoa-growing areas, the differences in characteristics represent 5.2 percent of the observed The results of the Oaxaca-Blinder decomposition of differential in time spent on domestic work, whereas the gender gap are presented in Table 4-A. Differences in coffee-growing areas it represents 12.0 percent. A between males and females in terms of time-use can large percentage of the observed differential in time be explained by various factors, i.e. differences in skills spent on other work (i.e. formal work or non-primary and observable characteristics, or discriminations. production) is explained by differences in returns Differences in choices or gender-specific constraints (97.9 percent and 124.5 percent in cocoa and coffee- (such as family constraints or other social or growing areas, respectively). The contribution of the environmental constraints) are determinants of the return component is 85.1 percent and 96.2 percent time-use gender gap, and may not be observable; they of the observed gap in time spent on non-productive are therefore considered as part of the “unexplained activities and leisure in cocoa and coffee-growing component” of the gender gap. When decomposing areas, respectively. An interpretation of these results is the time-use gender gap we obtain two components: that intra-household decision making ignores the needs gender gap differences due to differences in observed and capacities of women (which leads to discrimination characteristics (characteristics components), and inefficiencies) to the extent that differences in and gender gap differences due to differences in characteristics account for only a small part of the unobserved characteristics (returns components). gender gaps. This conclusion is consistent with our We analyzed the contribution of the characteristic and results concerning the important contribution of the return components to the observed gender differential unexplained (or return) component in the observed time in four daily activities: primary production, other work, allocation gender gaps. domestic work, and non-productive and leisure time. These returns can be interpreted as the different weights assigned to these characteristics in the individual’s decision-making process on the time spent on each activity (Table 4-A). Assessing Time-Use Gender Discrimination 71 TABLE 4-A Decomposition Analysis of the Time-Use Gender Gap COCOA Primary Other Domestic Non-Productive Production Work Work and Leisure D 0.986 0.954 -5.122 2.053 C -0.340 0.020 -0.268 0.305 R 1.326 0.933 -4.854 1.748 Contribution (in %): C/D -34.5% 2.1% 5.2% 14.9% R/D 134.5% 97.9% 94.8% 85.1% Total 100.0% 100.0% 100.0% 100.0% COFFEE Primary Other Domestic Non-Productive Production Work Work and Leisure D 0.728 0.397 -2.645 1.818 C 0.068 -0.097 -0.318 0.070 R 0.660 0.495 -2.327 1.749 Contribution (in %): C/D 9.4% -24.5% 12.0% 3.8% R/D 90.6% 124.5% 88.0% 96.2% Total 100.0% 100.0% 100.0% 100.0% Sources: PPAP Survey, 2017. Note: D (Observed difference), C (Characteristics), R (Returns). 72 Photo: Thomas Perry/World Bank. Assessing Time-Use Gender Discrimination 73 74 Photo: PPAP, PNG 5. ASSESSING THE IMPACT OF TIME ALLOCATION AND OTHER VARIABLES ON HOUSEHOLD PRODUCTION AND WELFARE A linear regression model can be a useful method to analyze correlations and a powerful tool for causality analysis with non-experimental data.3 Table 5-A presents the dependent variables and explanatory variables introduced in the regression model. In order to give some structure to the analysis, hypothesis are provided concerning the expected effect of the explanatory variables (as seen below). It merits attention that the survey sample is made up of PPAP households and non-participating households (Box 5-A). 3. For applying the regression methods with experimental data, see Imbens and Rubin (2015). Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 75 TABLE 5-A Dependent variables and explanatory variables introduced in the regression model VARIABLE VARIABLE NAME OF MEASUREMENT REFERENCE INTERPRETATION TYPE CATEGORY VARIABLE UNIT YEAR IN OF CAUSALITY SURVEY Dependent Household Yield of cocoa Kg/ha 2016 variable production Yield of coffee Kg/ha 2016 Income per tree Kina 2016 Number of trees Tree 2016 Quality of Very well pruned 2016 pruning (binary 0 or 1) Household Income per Kina 2016 well-being capita (log) Welfare scale Scale from 1 the 2016 today poorest to 10 the richest Welfare scale Scale from 1 the 2016 five years ago poorest to 10 the richest Welfare scale Scale from 1 the 2016 in five years poorest to 10 the richest Explanatory Economic Share of total Percentage 2011 A proxy for the shares of variable income earned income earned by women from alternative which is not observable; crops women’s empowerment positive effect expected Asset wealth Composite 2017 Positive wealth effect index index (0 to 1) Bargaining Age Years 2017 Age is a proxy for power experience and information; it can increase bargaining power Hours of Number of 2017 Women’s empowerment domestic work hours per day negative effect 76 VARIABLE VARIABLE NAME OF MEASUREMENT REFERENCE INTERPRETATION TYPE CATEGORY VARIABLE UNIT YEAR IN OF CAUSALITY SURVEY Hours wage Kina 2017 Increasing opportunity of outsiders cost can increase bargaining power Permission Composite index 2017 This index is negatively index (0 to 1) correlated with bargaining power Cooperation Marriage or Binary variable 2017 Marriage can imply common law (0 or 1) more cooperation and efficiency Agreement index Composite index 2017 More cooperation (0 to 1) and efficiency Family problem Composite index 2017 Less cooperation index (0 to 1) and efficiency Decision Decision index Composite index 2017 Women’s empowerment making (0 to 1) positive effect Involved in Binary variable 2017 Women’s empowerment planning and (0 or 1) (household positive effect decision making level variable) of cocoa/coffee production Female primarily Binary variable 2017 Women’s empowerment involved in selling (0 or 1) (household positive effect cocoa/coffee level variable) Female primarily Binary variable 2017 Women’s empowerment involved in (0 or 1) (household positive effect receiving payment level variable) Female manage Binary variable 2017 Women’s empowerment account (0 or 1) (household positive effect level variable) Cognitive Years of Years; Binary 2012 Better knowledge and skills schooling; literacy variable (0 or 1) information increase efficiency Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 77 VARIABLE VARIABLE NAME OF MEASUREMENT REFERENCE INTERPRETATION TYPE CATEGORY VARIABLE UNIT YEAR IN OF CAUSALITY SURVEY Age Years 2017 Experience and information increase efficiency Training Binary variable 2017 Knowledge positive effect (0 or 1) Access to Binary variable 2017 Information positive information (0 or 1) effect Phone, internet Binary variable 2017 Technology positive effect (0 or 1) Other non- Live in ARB Binary variable 2017 Matrilineality effect economic (0 or 1) (women’s empowerment variables negative effect expected) Threat of Binary variable 2017 Low cooperation violence (0 or 1) negative effect Participation Categorical 2017 Network positive effect to community variable organization Participation Binary variable 2016 Multiple positive effects to PPAP (0 or 1) of the project (see Box 2) Female head Binary variable 2017 Agency effect (0 or 1) Living in Binary variable 2017 Control variable (Province) (0 or 1) Number of family Continuous 2017 Control variable members variable 78 BOX 5-A PPAP Design and Variables + The survey sample is made up of PPAP + One of the key objectives of PPAP is that women households and non-participating households. It contribute more to increases in household is therefore important to know what elements in income through involvement in improved the design of the PPAP can reform or reinforce farming practices, processing and marketing. the status quo for women in agriculture. For that purpose, access to training and information have been provided through the + The Development Objective of the PPAP is project on production, prices, pest and disease to improve the livelihoods of smallholder management, and other agricultural livelihoods cocoa and coffee producers supported by related information. The project also serves as the project. This would be achieved through a vehicle for improving literacy among project strengthening industry coordination and supported cocoa and coffee farmers and helps institutions, facilitating linkages between to strengthen farmers’ ability to operate farm smallholder farmers and agribusiness for the businesses efficiently. provision of technologies and services, and through the provision of critical market access infrastructure. Key outcomes of the PPAP are that: • Smallholder farmers adopt efficient, market responsive and sustainable production practices leading to an increase in their income; • Demand-driven productive partnerships are scaled-up with public support; and • Key infrastructure bottlenecks in the targeted value chains are addressed. Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 79 COCOA-GROWING AREAS The main findings are as follows (Box 5-B). BOX 5-B Estimation Results for Household Cocoa Production Variables + Yield of cocoa (kg/ha dry bean equivalent) + The number of trees in cocoa production is has only a few significant determinants; also impacted positively and significantly by surprisingly, the family problem index has a household asset wealth and living in ARB. positive impact on yield and having a male Participation in the agricultural group and primarily involved in the planning and decision participation in PPAP also have significant making about cocoa production and a female positive impacts. Concerning decision making managing account also have positive effects within the household, male (partner) decision on yield of cocoa. This might be viewed as and having a female involved in planning have contradictory when trying to interpret the both negative and significant effects on the results in terms of incentives and bargaining number of trees, whereas male involved power within the household. in planning has a positive and significant impact (Figure 5-A). + When considering other production dependent variables, different variables appear to be + The quality of pruning is impacted positively significant: the asset wealth index has a positive and significantly by hours of formal work, and significant effect on cocoa income per tree, asset wealth index, number of 13-17-years- the same as the number of family members aged old household members, participation 13-69-years-old, living in ARB, having a female to agricultural association or group and involved in planning and a female managing the participation to PPAP, and also family problem account. Hence, bargaining power of women index. However, this latter variable may be an and the availability of labor seem to have indication that cocoa income induced by higher significantly positive impacts on this indicator quality can raise family problems, thus causality of cocoa production yield. still needs to be ascertained (Figure 5-B). 80 FIGURE 5-A Correlates of Cocoa Income Per Tree Share of alternative crop income in total income Hours of domestic work Female*hours of domestic work Hours of formal work Female*hours of formal work Asset-wealth index Living in ARB Female head Married or common law Number of members 12 years and less Number of members 13-17 years Number of members 18-59 years Number of members 60 years and more Participation to agriculture association or group Participation to non-agric association or group Participation to PPAP Permission index Agreement index Family problem index Decision index (woman decide) Decision index (partner decide) Female involved in planning and decision Male involved in planning and decision Female primarily involved in selling Female primarily involved in receiving Female manage account Afraid to disagree Found at risk -4 -2 0 2 4 Source: PPAP Survey, 2017 Note: OLS regression coefficient estimates and 95%confidence intervals. Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 81 FIGURE 5-B Correlates of Cocoa Quality of Pruning Share of alternative crop income in total income Hours of domestic work Female*hours of domestic work Hours of formal work Female*hours of formal work Asset-wealth index Living in ARB Female head Married or common law Number of members 12 years and less Number of members 13-17 years Number of members 18-59 years Number of members 60 years and more Participation to agriculture association or group Participation to non-agric association or group Participation to PPAP Permission index Agreement index Family problem index Decision index (woman decide) Decision index (partner decide) Female involved in planning and decision Male involved in planning and decision Female primarily involved in selling Female primarily involved in receiving Female manage account Afraid to disagree Found at risk -20 0 20 40 Source: PPAP Survey, 2017 Note: OLS regression coefficient estimates and 95%confidence intervals. 82 FIGURE 5-C Correlates of Coffee Income Per Tree Share of alternative crop income in total income Hours of domestic work Female*hours of domestic work Hours of formal work Female*hours of formal work Asset-wealth index Living in Western Highlands Living in Jiwaka Living in Simbu Female head Married or common law Number of members 12 years and less Number of members 13-17 years Number of members 18-59 years Number of members 60 years and more Participation to agriculture association or group Participation to non-agric association or group Participation to PPAP Permission index Agreement index Family problem index Decision index (woman decide) Decision index (partner decide) Female involved in planning and decision Male involved in planning and decision Female primarily involved in selling Female primarily involved in receiving Female manage account Afraid to disagree Found at risk -15 -10 -5 0 5 Source: PPAP Survey, 2017 Note: OLS regression coefficient estimates and 95%confidence intervals. Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 83 FIGURE 5-D Correlates of Coffee Quality of Pruning Share of alternative crop income in total income Hours of domestic work Female*hours of domestic work Hours of formal work Female*hours of formal work Asset-wealth index Living in Western Highlands Living in Jiwaka Living in Simbu Female head Married or common law Number of members 12 years and less Number of members 13-17 years Number of members 18-59 years Number of members 60 years and more Participation to agriculture association or group Participation to non-agric association or group Participation to PPAP Permission index Agreement index Family problem index Decision index (woman decide) Decision index (partner decide) Female involved in planning and decision Male involved in planning and decision Female primarily involved in selling Female primarily involved in receiving Female manage account Afraid to disagree Found at risk -40 -20 0 20 40 Source: PPAP Survey, 2017 Note: OLS regression coefficient estimates and 95%confidence intervals. 84 COFFEE-GROWING AREAS Figures 5-C and 5-D on the previous pages present regression results for household coffee production variables. Income per tree is increasing with the number of 18-59-year-old members and women decision index. Agreement index and family problem index both have a negative and significant effect on income per tree. The very good quality of pruning is positively correlated with household size. Female involved in planning and decision also has a negative effect on pruning, while female involved in receiving payment for coffee has a positive effect. Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 85 COCOA-GROWING AREAS Box 5-C provided the main findings for cocoa-growing areas BOX 5-C Cocoa-growing areas: main findings + Household income per capita has many + The welfare scale today is positively and statistically significant determinants. The significantly impacted by hours of domestic income share of alternative crop has a negative work (although negatively by female hours of and significant effect, while the effect of the domestic work), asset wealth, living in ARB, number of hours of formal work appears to be number of members 60+, and male decision positive. Other positive effects concern: asset index. wealth, living in ARB, and male decision index, + Welfare scale five years ago is positively and female selling cocoa, and female managing significantly correlated with asset wealth index, account. When a woman is more in control number of members 60+, permission index, of the sale of cocoa and the management of agreement index, family problem index, female money that comes from it, there is an increase decision index, female primarily involvement in in income and household welfare. The man can selling cocoa. It is negatively and significantly have an important power of decision making correlated with female involvement in planning within the household, in so far as it improves and female primarily involved in receiving the yield of agricultural production (farm payments. income, and more specifically cocoa income per tree). Other variables have a significant negative + Welfare scale in five years is positively and effect on per capita total household income: significantly impacted by hours of domestic female head, household size, permission index, work (although negatively by female hours of male involved in planning, female involved in domestic work), asset wealth, living in ARB, receiving payments, and afraid to disagree. participation to PPAP, family problem index, female primarily involved in selling cocoa. It is negatively and significantly impacted by female primarily involved in receiving payments. 86 COFFEE-GROWING AREAS In coffee-growing areas, household income per capita has fairly the same determinants as in cocoa-growing areas, except for women’s empowerment and decision- making variables which are not as significant. Among those latter variables, agreement index and women decision index have a positive and significant effect on income per capita, while men decision index (men decide) has a negative and significant impact. Concerning wealth scale, the share of alternative crop income in total income has a positive and significant effect today and five years ago, but a negative effect in five years. Female hours of formal work also have a negative effect on wealth scale in five years. Among women’s empowerment and decision-making variables, permission index has a positive effect on wealth scale today and five years ago, whereas agreement index has a negative effect on wealth scale today, five years ago, and in five years. Family problem index has a negative effect on wealth scale five years ago. Both women decision index and female involved in the planning and decision making about coffee production has a positive effect on wealth scale today, five years ago, and in five years. Males involved in the planning and decision making about coffee production has a negative effect on the wealth scale five years ago. Assessing the Impact of Time Allocation and Other Variables on Household Production and Welfare 87 88 Photo: Thomas Perry/World Bank. 6. CONCLUSION The objective of this time-use and gender study was to better understand labor dynamics in the agricultural sector in PNG. This report uses data from two separate modules of the PPAP follow-up survey on time- use and women’s empowerment in order to better understand intra-household decision making and the ability of women to allocate their labor to the time- critical tasks of agricultural production, and whether these determinants, among other factors, influence household production and welfare. Main findings are presented in Box 6-A. Conclusion 89 BOX 6-A Main Findings + Men’s work is more geared towards cocoa or + There is an important gap between the coffee production than women who are more proportion of household income earned from employed in other agricultural activities. In cocoa and income earned from coffee. In 2016, the ARB and ENB regions, 47 percent of men cocoa represented 19 percent of total household and 32 percent of women are self-employed in income in the cocoa sector, and it was the main the cocoa sector, and 16 percent of men and 38 source of income for 47 percent of households, percent of women are self-employed in other whereas coffee represented 64 percent of agriculture activities. In the Highlands region, income in the coffee sector, and it was the 40 percent of men are self-employed in the main source of income for up to 83 percent of coffee sector, against only 14 percent of women, households. while 16 percent of men and 41 percent of + Despite the contribution of certain women are self-employed in other agriculture characteristics, such as education and age, to activities. explain gender differences in hours worked, the + Men work longer hours in profitable activities, unexplained part of the gender gap remains especially cocoa and coffee activities, whereas the most important. An interpretation is that women are particularly busy with domestic intra-household decision making ignores the activities. Men are more responsible for needs and capacities of women (which leads to profitable activities such as cocoa or coffee, discrimination and inefficiencies) to the extent while women have a more diversified schedule, that differences in characteristics account for especially with long hours of domestic work. In only a small part of the gender gaps. cocoa and coffee activities, men are also more + Beyond the explanation of time-use gap involved in the tasks that require skills and add between men and women, it is important to more value to production. take into account intra-household decision + Income per capita and food sufficiency rate is making processes. This can provide a better generally higher among households headed by understanding of the factors which influence a man, which can be attributed to the fact that the allocation of time and, more generally, men are more involved in profitable activities household efficiency. such as cocoa or coffee production. It is also a + More can be learned on non-cooperative consequence of higher education among men behaviors from the analysis of the women’s than among women. empowerment module. It appears having to ask permission to go to various places remains relatively common in cocoa-growing areas, while it is less so in coffee-growing areas. 90 + Higher bargaining power of women and the + In coffee-growing areas, household income availability of labor seem to have significant per capita has fairly the same determinants positive impacts on cocoa and coffee production as in cocoa-growing areas, except for women’s yield. While women’s empowerment indices, empowerment and decision-making variables such as agreement index and family problem which are not as significant. Among those latter index, have a negative effect on yield in coffee variables, agreement index and women decision areas. index have positive and significant effect on income per capita, while men decision index + In cocoa-growing areas, household income per (men decide) has a negative and significant capita is determined negatively by the income impact. share of alternative crops, while positively by hours of formal work, asset wealth, living + Both women decision index and female involved in Bougainville, male decision index, female in planning and decision making about coffee selling cocoa, and female managing household production have positive effects on wealth scale accounts. today, five years ago, and in five years. The fact that woman is more in control of the sale of cocoa and the management of money that comes from it can indeed increase income declared and also actual household welfare. Conclusion 91 From these findings it is possible to draw some principal recommendations: Focus on women’s empowerment to improve household welfare outcomes: The results show that household welfare outcomes are higher when women have more control over the sale of cocoa and the resulting income. They also show that more control and bargaining power of women significantly correlates with better access to a mobile phone and/or the internet, and empowered women are also more likely to have an equal relationship with their male partner, with whom they are not afraid to disagree over household decision making. Given the entrenched nature of intra-household gender dynamics and attitudes in PNG, it is likely that household awareness- raising and training on gender dynamics, and greater responsibilities for women could improve welfare outcomes for all household members. Reduce the domestic work burden for women before they can engage in more value-added agricultural activities: Women in PNG carry a substantial burden of domestic work, and are generally primarily responsible for cooking, washing, cleaning, and caring for other household members. This leaves them little time to substantively engage in more value-added agricultural activities. Without a parallel effort to reduce the domestic burden, initiatives that seek to directly engage women in higher value agricultural activities may only result in a greater overall workload for women, as they will be expected to continue their usual tasks as well as take on additional ones. The domestic workload may be reduced by technological interventions to reduce labour inputs, or by a more equal sharing of domestic tasks between household members. 92 Photo: Stephane Forman/World Bank. Conclusion 93 94 Photo: Stephane Forman/World Bank. REFERENCES ACIAR (2017). Improving Livelihoods of Smallholder Charmes, J. (2015). Time Use Across the World: Families through Increased Productivity of Coffee- Findings of a World Compilation of Time Use based Farming Systems in the Highlands of PNG. Surveys. UNDP Human Development Report Canberra, Australia. Office Background Paper. Aguiar, M., Hurst, E., and Karabarbounis, L. (2012). Chen, Z., and Wooley, F. (2001). A Cournot-Nash Recent Development in the Economics of Time Use. Model of Family Decision Making. Economic Journal, Annual Review of Economics, 4: 373-397. 111(474): 722-748. Aguiar, M. A., Hurst, E., and Karabarbounis L. (2013). Chiappori, P.-A. (1988). Rational Household Labor Time Use During The Great Recession. American Supply. Econometrica, 56: 63-89. Economic Review, 103(5): 1664−1696. Chiappori, P.-A. (1992). “Collective Labor Supply and Asselin, L-M. (2009). Analysis of Multidimensional Welfare. Journal of Political Economy, 100: 437-467. Poverty: Theory and Case Studies. Springer. Chiappori, P-A., Donni, O., and Komunjer, I. (2012). Bardasi, E., and Wodon, Q. (2005). Measuring Time Learning from a Piece of Pie. Review of Economic Poverty and Analyzing its Determinants: Concepts and Studies, 79(1): 162-195. Application to Guinea. In Blackden, C. M., and Wodon, Q. (ed.): Gender, Time Use, and Poverty in Sub-Saharan Chiappori, P-A., and Donni, O. (2006). Les modèles Africa. World Bank Working Paper No 73. non unitaires de comportement du ménage: un survol de la littérature. L’Actualité Economique, Société Becker, D. (1965). A Theory of Allocation of Canadienne de Science Economique, vol. 82(1): 9-52. Time. Economic Journal, 75(299): 493-517. Curry, G. N., and Koczberski, G. (2004). Mobilising Browning, M., and Chiappori, P.-A. (1998). Smallholder Labour in Oil Palm Production: Results “Efficient Intra-Household Allocations: of the Mobile Card Trial, Hoskins, West New Britain, A General Characterization and Empirical Tests. Papua New Guinea. Department of Social Sciences, Econometrica, 66: 1241-1278. Curtin University of Technology. Carter, M. R., and Katz, E. G. (1997). Separate Spheres Curry, G.N., Koczberski, G., Omuru, E. and Nailina, and the Conjugal Contract: Understanding the R.S. (2007). Farming or Foraging? Household Labour Impact of Gender-Biased Development. In Haddad, and Livelihood Strategies amongst Smallholder L., Hoddinott, and Alderman, H. (ed): Intrahousehold Cocoa Growers in Papua New Guinea. Perth, Black Resource Allocation in Developing Economics: Swan Press. Models, Methods, and Policy. IFPRI, The John Hopkins University Press, Baltimore and London. References 95 Duflo, E., and Udry, C. (2004). Intrahousehold Resource Murray-Prior, Roy, 2014. IFC Agribusiness PNG: Monpi Allocation in Cote d’Ivoire: Social Norms, Separate Coffee Exports and Monpi Cocoa Exports Baseline Accounts and Consumption Choices. NBER Working Study, Draft Final Report, Prepared for International Paper No. 10498. Finance Corporation, AgriBiz RD&E Services, (mimeo). Eswaran, M., Ramaswami, B., Wadhwa, W. (2013). Oaxaca, R. (1973). Male-Female Wage Differentials Status, Caste, and the Time Allocation of Women in Urban Labor Markets. International Economic in Rural India. Economic Development and Cultural Review, 14: 693–709. Change, 61 (2): 311-333. Overfield, D. (1998). An Investigation of the Gronau, R. (1977). Leisure, Home Production, and Household Economy: Coffee Production and Work – the Theory of the Allocation of Time Revisited. Gender Relations in Papua New Guinea. Journal of Journal of Political Economy, 85 (6): 1099-1123. Development Studies 38(5): 52-70. Haddad, L., and Kanbur, R. (1994). Are Better Off Pollak, R. A. (2012). Allocating Time: Individuals’ Households More Unequal or Less Unequal? Oxford Technologies, Household Technology, Perfect Economic Papers, 46(3): 445-58. Substitutes, and Specialization. Annals of Economics and Statistics, 105/106: 75-97. Hurst, E. (2015). Measuring Time Use in Household Surveys. Journal of Economic and Social Measurement, Pollak, R. A. (2013). Allocating Household Time: 40: 151–170. When Does Efficiency Imply Specialization? NBER Working Paper 19178. Imbens, G. W., and Rubin, D. B. (2015). Causal Inference for Statistics, Social, and Biomedical Sciences: An Quinsumbing, A. R., and Maluccio, J. A. (2000). Introduction. Cambridge University Press, New York. Intrahousehold Allocation and Gender Relations: New Empirical Evidence from Four Developing Countries. Juster, F. T., and Stafford, F. P. (1991). The Allocation IFPRI, FCND Discussion Papers No. 84. of Time: Empirical Findings, Behavioral Models, and Problems of Measurement. Journal of Economic Seymour, G, Malapit, H., and Quisumbing, A. (2017). Literature, 29 (2): 471-522. Measuring Time Use in Development Settings. Policy Research working paper No. WPS 8147, World Bank, Khandker, S. R. (1988). Determinants of Women’s Time Washington D.C. Allocation in Rural Bangladesh. Economic Development and Cultural Change, 37: 111-126. Udry, C. (1996). Gender, Agricultural Production, and the Theory of the Household. Journal of Political Lundberg, S. J., and Pollak, R. A. (1993). Separate Economy, 104(5): 1010-1046. Spheres Bargaining and the Marriage Market. Journal of Political Economy, 101: 988-1010. UN. 2004. Guide for Producing Statistics on Time Use. New York. Lundberg, S. J., and Pollak, R.A. (1994). Non- cooperative bargaining models of marriage. American World Bank (2011). World Development Report 2012: Economic Review Papers and Proceedings, 84: 132-137. Gender Equality and Development. Washington DC. Lundberg, S., and Pollak, R. A. (2001). World Bank (2012). Papua New Guinea - Country Efficiency in Marriage. Review of Economics Gender Assessment for the Period 2011-2012. of the Household, 1(3): 153-167. Washington DC. Lundberg, Pollak and Wales, 1997. Do Husbands and World Bank Group, (2014). The Fruit of Her Labor: Wives Pool Their Resources? Evidence from the United Promoting Gender-Equitable Agribusiness in Papua Kingdom Child Benefit. Journal of Human Resources, New Guinea. Washington DC. 32(3): 463-480. 96 Photo: Thomas Perry/World Bank. Introduction 97 98 Photo: Thomas Perry/World Bank. Appendix A SAMPLING FRAME AND DATA QUALITY ASSURANCE For social surveys of large populations, sample size For data collection, interviewers were hired and requirements are generally determined as a proportion supervised in two separate data collection teams (i.e. of the square root of the population. The formula to cocoa and coffee teams). They have participated in the determine sample size is: n=√N*1.5, where n is the same training meetings and they have carried out the sample number and N is the population number. For field test of the questionnaires in both areas separately. the Baseline Survey, this number was 1,200 households The pilot survey enabled data collection teams to test in total, considering that smallholders have very small the different modules of the evaluation questionnaire production blocks. A total of 800 households were (including the time-use and gender module) to improve enumerated across four target provinces for coffee the formulation of questions, and to test the social and 400 households across two target provinces for acceptability of certain issues. By doing so, this activity cocoa. The follow-up survey has extended the samples highlighted the significant aspects and issues of the to approximately 1,100 households for coffee and 800 survey which needed to be addressed before the data households for cocoa. collection phase. In particular, some questions could be excluded if they were considered irrelevant given The data collection activities were completely the objective of the evaluation during the pretest separated from the PPAP implementation itself. and field test, and if these questions were likely to Furthermore, questionnaire and instructions were not increase distrust and thus reduce the response rate. directly provided by the implementation agency. These Other questions have been reformulated to reflect the conditions of independence between management country context. Each reformulation has been done of project implementation and management of data during the pretest and throughout the days of the field collection are prerequisite for good quality data. test that have allowed testing a modified questionnaire One issue was that Karak University, which provided each day and, the last day, to get a quasi-final some of the staff for data collection, was at the same questionnaire. time lead partner in the ENB province. In order to limit the data collection bias, it was decided that the Bougainville teams would take charge of data collection in the communities with Karak University as a lead partner. By doing this, the potential desirability bias has been significantly mitigated. Appendix A 99 DATA ENTRY AND QUALITY ASSURANCE DATA CLEANING AND CHECKS The use of tablets has eased the process and allowed The data cleaning procedure has consisted of various various checks when entering the data during the steps for all modules: inconsistency checks, range interviews. This decision should have increased the checks, skip checks and other miscellaneous checks. In level of accuracy of the data. Tablets have been tested particular, various checks have been done: in the field with the teams of interviewers. • The partner of the head must be of the opposing The Computer-Assisted Personal Interviewing is sex (note that this error is mostly correctable with also programmed in a way to automatically skip the names); unnecessary sections. In fact, the skipped questions • Head’s children/grandchildren younger than the were not shown to the interviewers, therefore head + head’s parents and grandparents older than eliminating a risk of an error, and reducing the burden the head (no error found of this type); on the interviewer to understand the skips themselves. In addition, multiple choice questions were used • Checking if the last completed grade is coherent allowing only the members of an appropriate age with the age (some errors might be due to a or type to answer. This strongly limits the possible different interpretation of schools grades); mistakes from interviewers and avoids the necessity of • No lead partner for control village households verifying the coherence of such questions. (This is, however, a declaration error and thus LP The only type of errors that can be detected at the variable can be replaced when we know what are end of the Computer-Assisted Personal Interviewing the LPs in the community); and is coherence problems that could not be programed • Field observation was done on one or more fields (manual entry errors are still possible but there is for current producers. no way to detect most of those absent asking the interviewee once more). A few miscellaneous checks of The response rate to the questionnaire was relatively this type have been performed on the database. high as nearly 92.6 percent of sampled households Overall, the verification process was a three-step have answered the questionnaire in cocoa-growing process: areas (741 households answered the questionnaire out of 800 initially sampled). However, note that new • A quick verification by the supervisor upon households can be replaced, and among the 400 receiving the filled questionnaires from tablets; households initially sampled from the baseline, 315 • A verification program running after each day of have the same head, thus 26 were replaced or no entry which outputs a file with a list of errors for longer have the same head of households at the time verification; and of the interview in 2017, and 59 were not interviewed (Table A-1). • Data cleaning which is far less thorough and complex than after manual data collection and data In coffee-growing areas, the response rate to the entry of all questionnaires. questionnaire is relatively low at 67.1 percent (Table A-2). Note that the Jimy district in Jiwaka province was not sampled, which represents 81 baseline households, hence the response rate is, in fact, a bit higher (around 75 percent). Other households initially sampled in the baseline were not found at the time due to the pre-election period in May—June 2017, which has had a much greater impact on households’ response rate in the Highlands region. Furthermore, some new controls were added to the planned sample but are not considered here. Among the 529 households who answered the questionnaire (in particular the income section), 509 have the same head. 100 TABLE A-1 Sample Size and Response Rates in Cocoa-Growing Areas Baseline households New households Planned/Sampled 400 400 Answered the questionnaire 341 (85.3%) 400 (100.0%) Answered with the same head as in the baseline 315 (78.8%) - Sources: PPAP Cocoa Survey, 2012 (baseline), 2017. TABLE A-2 Sample Size and Response Rates in Coffee-Growing Areas Baseline households New households Planned/Sampled 788 151 Answered the questionnaire 529 (67.1%) 151 (100.0%) Answered with the same head as in the baseline 509 (64.4%) - Sources: PPAP Coffee Survey, 2012 (baseline), 2017. Table A-3 presents results concerning partnerships Table A-3 also indicates that, at the end of interviews, and participation in cocoa and coffee-growing areas. a high percentage of households would be happy to We observe that only 46.6 percent of households participate in a survey in 2019: 94.2 percent in cocoa- have heard about PPAP in PPAP cocoa-growing growing areas and 89.8 percent in coffee-growing areas (of those, 59.1 percent declare having a lead areas, with no significant difference between PPAP and partner under PPAP), and a proportion as low as 18.9 non-PPAP. This indicates that the questionnaire was percent in PPAP coffee-growing areas (of those, 12.3 relatively well received by the interviewed households. percent declare having a lead partner under PPAP). These individual statements may not reflect the reality of project participation, thus, for PPAP impact assessment, a more “objective” variable can be used based on information gathered prior to data collection on community participation to the PPAP. The PPAP variable used in this report, therefore, concerns the community and not the household specifically. Appendix A 101 TABLE A-3 Partnerships and Participation to Survey (% of households) COCOA-GROWING AREAS PPAP Non-PPAP DIFF Mean N Mean N p-value Partnerships Heard of PPAP 46.6 414 29.1 327 0.000 Has a Lead Partner under the PPAP 59.1 193 46.3 95 0.041 Received tools from LP 48.2 114 38.6 44 0.274 Replaced tools 10.9 55 11.8 17 0.925 Received seedlings from LP 79.8 114 50.0 44 0.000 Used seedlings 91.2 91 90.9 22 0.966 Happy to participate in survey in 2019 93.2 411 95.4 327 0.191 COFFEE-GROWING AREAS PPAP Non-PPAP DIFF Mean N Mean N p-value Partnerships Heard of PPAP 18.9 428 20.2 252 0.678 Has a Lead Partner under the PPAP 12.3 81 21.6 51 0.180 Received tools from LP 30.0 10 9.1 11 0.239 Replaced tools 0.0 3 0.0 1 - Received seedlings from LP 10 10 0.0 11 0.317 Used seedlings 100 1 - 0 - Happy to participate in survey in 2019 89.9 425 89.6 249 0.894 Sources: PPAP Cocoa & Coffee Surveys, 2017. 102 Non-response concerning key variables is relatively Therefore, it is recommended to use all the answers low. One example is income with a response rate in the time-use analysis (robustness checks could of 96.2 percent in the Cocoa Survey compared to be provided by subtracting a typical day). Also, note 89.0 percent in the Coffee Survey. Concerning the that as few as 5.3 percent of individuals answered in women’s empowerment module, 550 out of 551 of reference to a Sunday in the Cocoa Survey which can eligible sampled women answered this part of the be considered de facto as a typical day. 13.2 percent questionnaire in the Cocoa Survey, compared to 390 of individuals answered in reference to a Sunday in the out of 392 in the Coffee Survey. Among them, 91.3 Coffee Survey. Other indications concerning answers percent declared they currently have a partner in the quality are questions concerning busyness and a Cocoa Survey, compared to 88.5 percent in the Coffee comfortable amount of time during the day: in cocoa- Survey; among 503 selected woman partners, 418 growing areas, 7.2 percent consider they were not busy answered the questionnaire (response rate of selected enough (compared to 30.0 percent in coffee-growing woman partners is thus 83.1 percent) in the Cocoa areas), and 8.1 percent had no comfortable amount Survey, compared to 244 among 347 selected woman of time (compared to 23.9 percent in coffee-growing partners in the Coffee Survey. areas). Although this is very subjective, these answers are indications of a typical day as well. Finally, 41.1 Time-use sheet has been answered by 2231 individuals percent of respondents have a watch in cocoa-growing out of 2264 eligible in the Cocoa Survey, (response rate areas, compared to 23.3 percent in coffee-growing is 98.5 percent) which is very high given the length of areas, which may be a better position to assess the the questionnaire as a whole. The response rate is 86.7 time spent on various activities. percent in the Coffee Survey. As shown in Table A-4 below, a typical day represents 23.1 percent of total answers in the cocoa-growing areas and 21.7 percent in coffee-growing areas. It is difficult to determine precisely what constitutes a typical day. Appendix A 103 TABLE A-4 Time-Use Variables (% of individuals)—Cocoa Survey OVERALL MEN WOMEN PPAP NON-PPAP Time sheet answered 98.5 98.8 98.6 98.8 98.2 Monday 16.9 16.9 16.9 17.6 16.0 Tuesday 16.4 17.2 15.4 18.5 13.6 Wednesday 16.0 16.0 16.0 18.4 13.0 Thursday 15.2 14.7 15.8 13.6 17.3 Friday 18.1 17.9 18.3 15.1 21.9 Saturday 12.1 11.7 12.4 12.2 11.9 Sunday 5.3 5.5 5.2 4.6 6.3 Atypical day 23.1 22.4 23.9 24.7 21.0 Was too busy 34.9 35.9 33.8 35.7 33.8 Not busy enough 7.2 7.1 7.2 7.9 6.2 Had a comfortable amount of time 49.8 50.0 49.7 48.7 51.3 No comfortable amount of time 8.1 7.0 9.3 7.7 8.6 Has a watch 41.1 50.5 31.0 39.3 43.4 Sources: PPAP Cocoa Survey, 2017. 104 TABLE A-5 Time-Use Variables (% of individuals)—Coffee Survey OVERALL MEN WOMEN PPAP NON-PPAP Time sheet answered 86.7 90.4 84.3 86.2 88.6 Monday 12.7 12.5 13.0 14.8 5.5 Tuesday 12.7 13.2 12.1 14.0 8.2 Wednesday 20.0 19.8 20.3 19.0 23.8 Thursday 11.8 12.1 11.3 12.5 9.0 Friday 18.4 18.3 18.4 18.1 19.4 Saturday 11.2 11.4 10.9 10.9 12.0 Sunday 13.2 12.6 14.0 10.7 22.1 A typical day 21.7 23.1 19.9 21.7 21.6 Was too busy 36.8 36.7 36.8 37.3 34.8 Not busy enough 30.0 30.3 29.7 28.7 34.5 Had a comfortable amount of time 9.3 8.4 10.5 9.5 8.8 No comfortable amount of time 23.9 24.6 22.9 24.4 21.9 Has a watch 23.3 32.6 11.6 23.8 21.4 Source: PPAP Coffee Survey, 2017 Appendix A 105 Appendix B1 INDEX WEIGHTS (COCOA) TABLE B1-1 Asset Wealth Index WEIGHT MEAN INERTIA % Stove 5.8 0.066 0.047 Refrigerator 5.1 0.109 0.058 Microwave oven 11.6 0.004 0.023 Fan 8.1 0.030 0.047 Television 3.7 0.279 0.060 Cassette/CD player 3.8 0.084 0.027 VCR/DVD 4.3 0.146 0.052 Camera 6.1 0.063 0.049 Radio 0.9 0.598 0.008 Computer 5.9 0.076 0.054 Mobile Phone 1.6 0.737 0.015 Bicycle 1.0 0.231 0.008 Motorcycle 4.5 0.012 0.014 Car 5.4 0.043 0.026 Truck/bus 5.6 0.026 0.018 Boat/dinghy 4.7 0.007 0.006 Good quality walls 1.6 0.119 0.009 Good quality roof 1.8 0.838 0.018 Good quality floor 1.6 0.107 0.009 106 WEIGHT MEAN INERTIA % Cooking fuel (gas or electricity) 6.0 0.030 0.026 Electric lights 1.9 0.552 0.023 Piped water -0.3 0.032 0.002 Flush toilet 6.0 0.034 0.035 Latrine -0.1 0.780 0.015 Wheel barrow 2.2 0.393 0.028 Chainsaw 2.6 0.161 0.021 Knapsack - Good quality e.g. CP3 2.1 0.383 0.038 Knapsack - Low quality e.g. Chinese brand 0.1 0.070 0.009 Secateurs - Good quality 2.0 0.308 0.035 Secateurs - Low quality e.g. Chinese brand 0.5 0.053 0.015 Bow saw - Good quality 2.4 0.217 0.038 Bow saw - Low quality e.g. Chinese brand -0.5 0.043 0.016 Spade 1.4 0.825 0.016 Bush knife 1.9 0.974 0.006 Cocoa bags 1.4 0.093 0.011 Canvas or drying sheets 2.5 0.078 0.017 Harvest containers/buckets 0.5 0.238 0.013 Hand pulper machine 1.6 0.008 0.003 Motor generated pulper machine 0.4 0.003 0.002 Other 2.6 0.013 0.004 Other cocoa processing equipment 2.6 0.004 0.004 Pole Pruner 1.8 0.016 0.004 One block -0.2 0.637 0.038 Sells poultry 0.5 0.120 0.003 Sells pigs -0.2 0.312 0.003 Bank account 2.1 0.603 0.026 Appendix B 107 TABLE B1-2 Permission Index WEIGHT MEAN INERTIA % Ask permission to her/his partner to go to: The market 16.8 0.701 0.168 The health center 17.6 0.733 0.173 The community center, neighbourhood park 17.4 0.704 0.182 A place of worship 14.7 0.582 0.146 Visit relatives in the neighbourhood 16.3 0.683 0.167 Visit friends in the neighbourhood 15.7 0.653 0.162 Partner pays for phone -1.5 0.028 0.000 108 TABLE B1-3 Agreement Index WEIGHT MEAN INERTIA % Partners agree on: Religion 9.2 0.926 0.059 Politics 5.4 0.783 0.052 Family 10.5 0.924 0.079 Friends 6.2 0.760 0.084 Money 8.3 0.909 0.063 House work 6.0 0.707 0.091 Work 9.1 0.890 0.088 Moral rules 6.3 0.790 0.079 Relationship between parents and children 10.4 0.901 0.103 Education of children 11.1 0.931 0.086 Not consult her partner to buy clothes 5.6 0.830 0.053 Not consult her partner for children purchases 7.1 0.877 0.063 Not afraid to disagree with partner, angry with you 1.9 0.294 0.037 Not afraid to disagree with partner, angry with your children 1.6 0.144 0.035 Not found at risk with partner 1.4 0.193 0.032 Appendix B 109 TABLE B1-4 Family Problem Index WEIGHT MEAN INERTIA % Problem in family in the last two years: Bad relationship between parents and children 16.7 0.061 0.132 Lack of money -0.6 0.407 0.028 Alcoholism of a member 12.0 0.170 0.166 Illness of a member 3.6 0.334 0.034 Lack of work of a member 5.1 0.194 0.051 Absence of the father 4.1 0.158 0.029 Absence of the mother -0.6 0.061 0.016 Lack of time 8.0 0.142 0.069 Addiction of a member 16.6 0.029 0.077 Domestic violence 18.3 0.089 0.229 Imprisonment of a member 14.0 0.012 0.030 Infidelity 11.7 0.013 0.026 Interference from other families in your relationship 10.8 0.119 0.111 110 TABLE B1-5 Decision Index WEIGHT MEAN INERTIA % Woman decides: Buy durable household goods 3.8 0.950 0.059 How much to spend on food and toiletries 9.7 0.977 0.074 Arrange/decorate the house 9.7 0.947 0.054 Send the children to school 13.9 0.873 0.131 Take children to medical checks 18.1 0.896 0.217 Take the children to the doctor when sick 20.5 0.914 0.236 If you must work outside the home or not 8.4 0.787 0.080 Having how many children 7.9 0.800 0.054 Take the kids to play 7.9 0.498 0.094 Partner decides: Buy durable household goods 19.0 0.934 0.131 How much to spend on food and toiletries 12.2 0.853 0.114 Arrange/decorate the house 8.3 0.707 0.085 Send the children to school 12.2 0.803 0.136 Take children to medical checks 12.1 0.765 0.165 Take the children to the doctor when sick 13.3 0.795 0.182 If you must work outside the home or not 9.5 0.781 0.087 Having how many children 7.4 0.784 0.050 Take the kids to play 6.1 0.430 0.051 Appendix B 111 Appendix B2 INDEX WEIGHTS (COFFEE) TABLE B2-1 Asset Wealth Index WEIGHT MEAN INERTIA % Stove 9.6 0.025 0.027 Refrigerator 9.4 0.024 0.029 Microwave oven Fan 15.1 0.001 0.010 Television 6.6 0.126 0.052 Cassette/CD player 5.8 0.088 0.028 VCR/DVD 7.2 0.069 0.037 Camera 5.5 0.060 0.019 Radio 0.8 0.571 0.021 Computer 7.2 0.024 0.017 Mobile Phone 1.3 0.484 0.019 Bicycle 3.5 0.034 0.007 Motorcycle 5.6 0.009 0.016 Car 9.2 0.018 0.022 Truck/bus 4.5 0.007 0.018 Boat/dinghy 8.2 0.001 0.020 Good quality walls 2.6 0.022 0.006 Good quality roof 2.7 0.335 0.021 Good quality floor 5.9 0.057 0.025 112 WEIGHT MEAN INERTIA % Cooking fuel (gas or electricity) 8.9 0.034 0.034 Electric lights 4.1 0.219 0.033 Piped water -0.1 0.085 0.005 Flush toilet 7.4 0.004 0.017 Latrine -0.2 0.982 0.013 Wheel barrow 6.1 0.076 0.027 Chainsaw 7.3 0.010 0.008 Knapsack - Good quality e.g. CP3 3.6 0.185 0.026 Knapsack - Low quality e.g. Chinese brand 1.3 0.104 0.025 Secateurs - Good quality 5.1 0.119 0.034 Secateurs - Low quality e.g. Chinese brand 1.6 0.065 0.031 Bow saw - Good quality 3.0 0.193 0.021 Bow saw - Low quality e.g. Chinese brand 2.0 0.076 0.025 Spade 1.4 0.951 0.026 Bush knife 1.2 0.960 0.028 Cocoa bags 1.2 0.716 0.022 Canvas or drying sheets 0.7 0.591 0.026 Harvest containers/buckets 2.4 0.428 0.026 Hand pulper machine 1.5 0.350 0.022 Motor generated pulper machine 3.5 0.012 0.005 Other 0.1 0.004 0.002 Other cocoa processing equipment 5.2 0.001 0.006 One block 1.8 0.710 0.096 Sells poultry 2.2 0.126 0.009 Sells pigs 0.7 0.479 0.006 Bank account 3.9 0.246 0.029 Appendix B 113 TABLE B2-2 Permission Index WEIGHT MEAN INERTIA % Ask permission to her/his partner to go to: The market 17.0 0.435 0.179 The health center 17.2 0.491 0.189 The community center, neighbourhood park 17.3 0.415 0.183 A place of worship 16.2 0.394 0.155 Visit relatives in the neighbourhood 15.4 0.438 0.149 Visit friends in the neighbourhood 15.2 0.437 0.146 Partner pays for phone 1.8 0.006 0.001 114 TABLE B2-3 Agreement Index WEIGHT MEAN INERTIA % Partners agree on: Religion 11.0 0.896 0.090 Politics 2.8 0.674 0.019 Family 11.8 0.904 0.096 Friends 6.0 0.699 0.065 Money 7.4 0.894 0.040 House work 6.2 0.733 0.066 Work 8.9 0.822 0.090 Moral rules 9.2 0.803 0.106 Relationship between parents and children 10.0 0.838 0.109 Education of children 9.4 0.915 0.053 Not consult her partner to buy clothes 4.0 0.591 0.051 Not consult her partner for children purchases 4.4 0.694 0.053 Not afraid to disagree with partner, angry with you 3.3 0.470 0.062 Not afraid to disagree with partner, angry with your children 2.2 0.304 0.042 Not found at risk with partner 3.4 0.442 0.059 Appendix B 115 TABLE B2-4 Family Problem Index WEIGHT MEAN INERTIA % Problem in family in the last two years: Bad relationship between parents and children 11.8 0.259 0.154 Lack of money 3.1 0.770 0.068 Alcoholism of a member 12.2 0.132 0.093 Illness of a member 5.5 0.338 0.045 Lack of work of a member 8.7 0.181 0.105 Absence of the father 9.4 0.139 0.060 Absence of the mother 18.4 0.033 0.059 Lack of time 1.5 0.194 0.064 Addiction of a member 9.9 0.052 0.045 Domestic violence 10.1 0.397 0.151 Imprisonment of a member 15.6 0.011 0.038 Infidelity -2.5 0.016 0.020 Interference from other families in your relationship 9.4 0.208 0.096 116 TABLE B2-5 Decision Index WEIGHT MEAN INERTIA % Woman decides: Buy durable household goods 10.1 0.868 0.093 How much to spend on food and toiletries 13.0 0.897 0.124 Arrange/decorate the house 11.9 0.871 0.119 Send the children to school 16.0 0.916 0.155 Take children to medical checks 12.3 0.825 0.170 Take the children to the doctor when sick 15.2 0.942 0.098 If you must work outside the home or not 6.1 0.571 0.074 Having how many children 9.9 0.817 0.107 Take the kids to play 5.6 0.325 0.061 Partner decides: Buy durable household goods 13.0 0.850 0.102 How much to spend on food and toiletries 9.9 0.729 0.107 Arrange/decorate the house 8.5 0.557 0.099 Send the children to school 16.1 0.850 0.147 Take children to medical checks 12.9 0.647 0.189 Take the children to the doctor when sick 13.2 0.727 0.172 If you must work outside the home or not 7.0 0.532 0.068 Having how many children 11.8 0.880 0.064 Take the kids to play 7.6 0.238 0.052 Appendix B 117 Appendix c COMPLEMENTARY TABLES (COCOA) TABLE C-1 Literacy and School Attendance FEMALE MALE DIFF Mean N Mean N p-value Literacy (10 years old or more)* Total 90.1 1257 90.4 1321 0.778 10 to 24 years 91.7 460 87.4 492 0.028 25 to 39 years 93.7 350 92.9 297 0.691 40 to 59 years 85.5 372 91.8 404 0.005 60 or more 85.3 75 91.4 128 0.206 School attendance (6 to 24 years) Total 72.6 555 68.8 597 0.160 6 to 13 years 81.3 209 81.4 226 0.984 14 to 18 years 88.8 205 86.5 192 0.484 19 to 24 years 36.2 141 34.1 179 0.698 Sources: PPAP Survey, 2017. *Reads and write English or Pidgin. 118 TABLE C-2 School Level and Completion FEMALE MALE DIFF Mean N Mean N p-value School level (6 to 24 years) None/Kindergarten 5.3 1328 5.5 1400 0.859 Primary 46.9 1328 41.4 1400 0.004 Secondary 40.2 1328 38.4 1400 0.322 University/tertiary 3.9 1328 12.2 1400 0.000 Other 3.6 1328 2.5 1400 0.092 Years completed (among those 7.4 901 7.9 969 0.000 who stopped studying) Sources: PPAP Survey, 2017. TABLE C-3 Participation, Employment and Unemployment FEMALE MALE DIFF Mean N Mean N p-value All (10-69 years old) Participation rate 74.5 1245 78.6 1292 0.017 Employment rate 93.8 928 93.2 1015 0.625 Unemployment rate 0.4 928 0.8 1015 0.309 Adult (25-69 years old) Participation rate 94.8 785 96.5 800 0.093 Employment rate 95.4 744 95.6 772 0.876 Unemployment rate 0.3 744 0.6 772 0.273 Sources: PPAP Survey, 2017. Appendix C 119 TABLE C-4 Employment Characteristics FEMALE MALE DIFF Mean N Mean N p-value Schedule Full time (35+ hours) 32.5 870 37.9 946 0.016 Part time 67.5 870 62.1 946 0.016 Occupation Farmer 76.6 870 72.6 946 0.054 Fisherman 0.2 870 0.8 946 0.070 Hunter 0.0 870 0.0 946 - Forestry worker 0.0 870 0.2 946 0.157 Services & sales worker 4.3 870 3.9 946 0.714 Clerical worker 1.6 870 0.5 946 0.027 Technician 0.5 870 4.1 946 0.000 Professional 5.7 870 6.2 946 0.660 Manager 0.2 870 0.8 946 0.070 Student 7.1 870 8.4 946 0.329 Other 3.8 870 2.3 946 0.071 120 FEMALE MALE DIFF Mean N Mean N p-value Employment status Employee (Wage), public sector 4.5 870 4.2 946 0.791 Employee (Wage), private, Cocoa 1.0 870 3.1 946 0.002 Employee (Wage), private, other agricultural 2.9 870 4.4 946 0.074 Employee (Wage), private, non-agricultural 4.4 870 6.6 946 0.040 Self-employed, Cocoa 32.0 870 47.3 946 0.000 Self-employed, other agricultural 37.8 870 15.9 946 0.000 Self-employed, non-agricultural 3.9 870 5.6 946 0.089 Unpaid family worker 8.6 870 6.4 946 0.080 Apprentice 0.0 870 0.4 946 0.045 NGO 0.2 870 0.3 946 0.721 Coop 0.0 870 0.2 946 0.157 Student 3.9 870 5.0 946 0.272 Other 0.8 870 0.6 946 0.669 Sources: PPAP Survey, 2017. Appendix C 121 TABLE C-5 Reasons for not Working FEMALE MALE DIFF Mean N Mean N p-value Reasons for not working: Domestic work 0.9 322 0.4 285 0.365 Personal / family affairs 0.3 322 0.7 285 0.503 Pregnancy / delivery 0.9 322 0.0 285 0.082 Caring for children 5.6 322 2.5 285 0.047 Illness 3.7 322 3.9 285 0.932 Disability 4.3 322 6.7 285 0.214 Too young 0.0 322 0.4 285 0.317 Remittances 4.7 322 3.5 285 0.474 Old aged/ pension 78.9 322 80.0 285 0.734 Student 0.6 322 2.1 285 0.121 Other 0.0 322 0.0 285 - Sources: PPAP Survey, 2017. 122 TABLE C-6 Household Composition, Income and Satisfaction TOTAL FEMALE HEAD MALE HEAD GENDER GAP Mean N Mean N Mean N p-value Main source of income Earning from cocoa 47.3 710 39.1 46 47.9 664 0.245 Earning from other 33.8 710 45.7 46 33.0 664 0.098 agriculture products Earning from livestock 0.0 710 0.0 46 0.0 664 - Earning from fishing 0.8 710 0.0 46 0.9 664 0.014 Earning from non-agriculture 5.5 710 6.5 46 5.4 664 0.771 business Salaries/wages/commissions 11.0 710 4.3 46 11.4 664 0.031 Earning from rents 0.6 710 0.0 46 0.6 664 0.045 (house/assets/properties) Remittances from abroad 0.0 710 0.0 46 0.0 664 - Domestic remittances 0.8 710 4.3 46 0.6 664 0.220 Pension 0.0 710 0.0 46 0.0 664 - Aid in nature / cash 0.0 710 0.0 46 0.0 664 - Freely from forest 0.0 710 0.0 46 0.0 664 - Other 0.1 710 0.0 46 0.2 664 0.317 Wealth scale (from 1 the poorest to 10 the richest) Today 4.4 728 4.0 46 4.4 682 0.158 Five years ago 3.7 728 3.7 46 3.7 682 0.979 In five years 6.6 727 5.8 46 6.6 681 0.015 Appendix C 123 TOTAL FEMALE HEAD MALE HEAD GENDER GAP Mean N Mean N Mean N p-value Income by source Cocoa - wet bean 245.5 731 166.4 49 251.1 682 0.248 Cocoa - dry bean 1809.9 731 438.5 49 1908.4 682 0.000 Coconuts 1453.8 731 541.7 49 1519.4 682 0.000 Off-farm employment 4272.5 731 237.8 49 4562.4 682 0.000 Non-farm income 2659.4 731 526.5 49 2812.7 682 0.000 e.g. trade store. PMV Hunting and fishing 131.0 731 26.5 49 138.5 682 0.025 Gifts. Customary 75.7 731 105.1 49 73.6 682 0.582 payments. remittances Balsa 6.2 731 0.0 49 6.6 682 0.264 Other agricultural 245.4 731 303.2 49 241.2 682 0.718 Other 30.9 731 0.0 49 33.1 682 0.114 Total income per capita 3170.7 725 945.7 45 3317.9 680 0.000 Household composition Number of members 0.8 731 0.7 49 0.8 682 0.526 12 years and less Number of members 0.5 731 0.3 49 0.5 682 0.019 13-17 years Number of members 2.5 731 1.9 49 2.5 682 0.004 18-59 years Number of members 0.3 731 0.3 49 0.3 682 0.878 60 years and more Sources: PPAP Survey, 2017. 124 TABLE C-7 Who Makes Purchasing Decision, Including Durable Goods? WOMAN ALONE BOTH WOMAN & MAN Estimate P-value Estimate P-value Hours of domestic work 0.001 0.506 -0.001 0.651 Hours of formal work -0.003 0.622 0.005 0.373 Share of alternative crop income in total income -0.001 0.554 0.001 0.540 Hours wage by outsider (female) 0.000 0.954 0.000 0.796 Hours wage by outsider (male) 0.001 0.444 -0.001 0.599 Asset wealth index -0.002 0.017 0.002 0.009 Living in ARB 0.010 0.650 -0.029 0.201 Married or common law -0.545 0.000 0.619 0.000 Number of members 12 years and less -0.005 0.539 0.007 0.379 Number of members 13-17 years -0.034 0.012 0.025 0.065 Number of members 18-59 years -0.028 0.005 0.023 0.023 Number of members 60 years and more -0.085 0.000 0.074 0.001 Participation to agriculture association or group -0.011 0.640 -0.004 0.850 Participation to non-agric association or group -0.016 0.441 0.012 0.576 Participation to PPAP 0.012 0.555 -0.010 0.628 Age 0.004 0.000 -0.003 0.004 Years of schooling 0.004 0.330 -0.005 0.202 Training on cocoa 0.038 0.322 -0.015 0.695 Information on cocoa -0.014 0.619 -0.011 0.689 Has a phone 0.050 0.033 -0.055 0.021 Has access to internet 0.137 0.005 -0.117 0.018 Intercept 0.498 0.000 0.413 0.000 N 525 525 R-square 0.3668 0.4024 Sources: PPAP Survey, 2017. Appendix C 125 TABLE C-8 Trends in Activities (frequencies) Performed by Household Members 2011 2016 % Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Clearing Land 11.4 15.5 0.017 20.1 26.0 0.001 0.491 Lining 5.6 9.2 0.007 10.3 20.6 0.000 0.001 Shade establishment 5.2 8.1 0.021 6.7 14.2 0.000 0.011 Nursery operations 2.3 4.3 0.028 4.6 8.6 0.000 0.150 Holing and planting 5.8 8.5 0.037 14.3 20.4 0.000 0.108 Weeding (establishment 4.9 7.9 0.017 14.9 18.5 0.019 0.734 phase) Fertilizing/mulching 0.0 0.4 0.083 0.5 1.2 0.050 0.393 (establishment phase) Pest and disease management 1.2 3.6 0.002 1.8 5.1 0.000 0.416 (establishment phase) Weeding (production phase) 21.2 28.8 0.001 29.2 36.0 0.000 0.792 Pruning 7.3 20.8 0.000 6.7 32.8 0.000 0.000 Fertilizing/mulching 0.1 0.7 0.074 0.8 1.1 0.483 0.562 (production phase) Spraying agro-chemicals 2.3 6.9 0.000 1.8 13.2 0.000 0.000 Pest and disease management 2.7 6.8 0.000 3.3 11.0 0.000 0.014 (production phase) Soil and water conservation 0.3 0.7 0.215 0.2 1.4 0.001 0.152 Harvesting (including burying 17.9 19.7 0.365 26.8 28.5 0.366 0.962 husks for CPB control) Fermentary operations 6.6 10.8 0.003 11.6 20.3 0.000 0.033 Selling cocoa 16.2 24.3 0.000 19.3 28.3 0.000 0.749 Sources: PPAP Survey, 2012-2017. 126 TABLE C-9 Trends in Activities (days of work) Performed by Household Members 2011 2016 Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Clearing Land 4.9 4.8 0.849 10.0 11.1 0.569 0.549 Lining 3.3 3.6 0.403 7.4 4.9 0.281 0.226 Shade establishment 3.3 3.6 0.633 7.3 5.5 0.444 0.391 Nursery operations 4.6 4.3 0.666 24.3 26.9 0.750 0.719 Holing and planting 3.7 4.3 0.343 4.2 5.5 0.037 0.383 Weeding (establishment 3.1 4.1 0.176 10.9 12.6 0.544 0.800 phase) Fertilizing/mulching 0.0 2.0 - 2.6 3.9 0.334 - (establishment phase) Pest and disease management 2.0 2.5 0.190 21.8 11.6 0.489 0.468 (establishment phase) Weeding (production phase) 4.9 5.1 0.799 17.7 16.1 0.480 0.455 Pruning 3.8 4.6 0.062 9.2 8.7 0.723 0.379 Fertilizing/mulching 2.0 2.0 1.000 8.2 6.2 0.405 0.424 (production phase) Spraying agro-chemicals 3.1 3.3 0.757 22.3 15.1 0.631 0.622 Pest and disease management (production 2.5 3.3 0.105 9.6 16.0 0.021 0.046 phase) Soil and water conservation 1.5 1.0 0.317 4.0 9.6 0.011 0.007 Harvesting (including burying 2.4 2.8 0.313 13.3 13.1 0.897 0.751 husks for CPB control) Fermentary operations 4.4 6.4 0.112 19.9 22.6 0.363 0.847 Selling cocoa 1.2 1.7 0.042 7.3 6.1 0.070 0.021 Sources: PPAP Survey, 2012-2017. Appendix C 127 TABLE C-10 Trends in Activities Performed by Household Members (Frequencies x Days of Work) 2011 2016 % Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Clearing Land 0.6 0.9 0.084 2.0 2.9 0.077 0.220 Lining 0.2 0.4 0.001 0.8 1.0 0.392 0.857 Shade establishment 0.2 0.3 0.022 0.5 0.8 0.171 0.501 Nursery operations 0.1 0.2 0.216 1.1 2.3 0.047 0.060 Holing and planting 0.2 0.4 0.018 0.6 1.1 0.000 0.023 Weeding 0.1 0.3 0.014 1.6 2.3 0.160 0.282 (establishment phase) Fertilizing/mulching 0.0 0.0 0.317 0.0 0.0 0.047 0.067 (establishment phase) Pest and disease management 0.0 0.1 0.003 0.4 0.6 0.513 0.660 (establishment phase) Weeding (production phase) 1.2 1.6 0.018 5.2 5.8 0.426 0.846 Pruning 0.3 1.0 0.000 0.6 2.9 0.000 0.000 Fertilizing/mulching 0.0 0.0 0.279 0.1 0.1 0.965 0.884 (production phase) Spraying agro-chemicals 0.1 0.2 0.002 0.4 2.0 0.000 0.001 Pest and disease management (production 0.1 0.2 0.000 0.3 1.8 0.000 0.000 phase) Soil and water conservation 0.0 0.0 0.664 0.0 0.1 0.003 0.004 Harvesting (including burying 0.5 0.6 0.137 3.6 3.7 0.767 0.956 husks for CPB control) Fermentary operations 0.3 0.8 0.006 2.3 4.6 0.000 0.002 Selling cocoa 0.2 0.4 0.000 1.4 1.7 0.134 0.668 Sources: PPAP Survey, 2012-2017. 128 TABLE C-11 Trends in Activities Performed by Outsiders 2011 2016 Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Hours of paid labour 39.1 58.2 0.420 4.3 39.4 0.000 0.534 by outsiders Average hourly wage 7.1 14.0 0.437 21.3 17.5 0.456 0.295 for outsiders (PGK) Sources: PPAP Survey, 2012-2017. TABLE C-12 Activity-Based Decision Making 2011 2016 % Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Primarily involved - - - 6.2 12.1 0.000 - in selling livestock Primarily involved 8.6 26.2 0.000 8.9 20.5 0.000 0.003 in selling cocoa Primarily involved in 12.4 22.7 0.000 9.3 20.1 0.000 0.807 receiving payments for cocoa Involved in planning and decision making about 22.8 7.6 0.000 26.4 38.2 0.000 0.000 cocoa production Operate the account 9.1 14.8 0.000 15.1 25.3 0.000 0.024 Sources: PPAP Survey, 2012-2017. Appendix C 129 130 TABLE C-13 OLS Regressions on Time-Use (Women) TOTAL HOURS PRIMARY FORMAL NON-PRIMARY DOMESTIC PERSONAL LEARNING OF WORK PRODUCTION WORK PRODUCTION WORK CARE Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Age 0.027 0.100 0.010 0.487 0.008 0.341 0.008 0.160 -0.151 0.000 -0.009 0.000 -0.007 0.640 Years of schooling 0.181 0.014 -0.025 0.711 0.235 0.000 -0.029 0.260 -0.311 0.045 -0.004 0.650 0.073 0.301 Literacy in English 0.889 0.089 0.800 0.097 -0.187 0.501 0.276 0.133 2.606 0.019 0.029 0.671 -0.565 0.262 Literacy in Pidgin -1.880 0.006 -1.179 0.061 -0.589 0.104 -0.112 0.639 -0.372 0.796 -0.044 0.625 0.219 0.738 Training on cocoa -0.543 0.346 0.006 0.990 -0.467 0.128 -0.082 0.685 0.517 0.671 0.004 0.963 0.035 0.949 Information on cocoa 0.590 0.164 0.275 0.482 0.179 0.430 0.136 0.362 -1.060 0.238 -0.020 0.725 -0.418 0.308 Has a phone 0.478 0.181 -0.248 0.452 0.370 0.052 0.355 0.005 -0.537 0.478 0.029 0.538 -0.370 0.283 Has access to internet -1.729 0.022 -0.694 0.319 -0.408 0.311 -0.627 0.019 -0.176 0.912 -0.111 0.267 2.317 0.002 Female head -1.248 0.134 -1.264 0.100 0.249 0.574 -0.233 0.427 2.736 0.121 -0.235 0.033 -0.088 0.913 Married or common law 0.000 1.000 0.254 0.714 -0.045 0.909 -0.208 0.431 2.464 0.121 -0.498 0.000 -0.134 0.853 Number of members -0.142 0.266 -0.027 0.820 -0.069 0.310 -0.046 0.304 1.222 0.000 0.004 0.814 -0.250 0.043 12 years and less Number of members 13-17 years 0.204 0.324 0.057 0.763 0.077 0.483 0.069 0.342 0.582 0.183 0.052 0.057 -0.426 0.033 Number of members 18-59 years -0.151 0.316 -0.030 0.831 -0.065 0.416 -0.056 0.290 0.235 0.460 -0.015 0.438 0.228 0.117 Number of members 0.067 0.844 0.071 0.822 0.013 0.945 -0.017 0.891 0.322 0.657 0.178 0.000 -0.011 0.974 60 years and more Hours wage by outsider (female) 0.025 0.232 0.025 0.195 0.002 0.834 -0.002 0.749 0.069 0.126 0.000 0.933 0.014 0.483 Hours wage by outsider (male) -0.019 0.249 -0.011 0.462 -0.005 0.558 -0.003 0.647 -0.053 0.136 -0.001 0.593 -0.042 0.010 Participation to agriculture 0.375 0.282 0.318 0.324 0.088 0.635 -0.031 0.803 0.013 0.986 -0.002 0.961 -0.028 0.933 association or group Asset wealth index -0.008 0.575 -0.023 0.073 0.006 0.396 0.009 0.067 0.019 0.526 0.000 0.815 0.001 0.914 Living in ARB 0.448 0.233 0.003 0.993 0.311 0.121 0.134 0.311 -0.051 0.948 -0.023 0.647 0.567 0.118 Permission index -0.007 0.228 -0.004 0.500 -0.004 0.237 0.000 0.888 0.002 0.882 0.000 0.670 0.013 0.017 Agreement index 0.009 0.363 0.001 0.882 0.006 0.242 0.002 0.669 -0.013 0.562 0.002 0.114 0.011 0.294 Family problem index -0.009 0.380 -0.007 0.450 -0.003 0.603 0.001 0.784 0.010 0.647 0.002 0.236 -0.002 0.834 Decision index (woman decide) 0.006 0.591 0.004 0.657 -0.003 0.659 0.004 0.305 -0.045 0.046 -0.001 0.378 0.007 0.477 Decision index (partner decide) -0.002 0.831 -0.007 0.383 0.006 0.186 -0.001 0.749 0.017 0.361 0.002 0.036 0.001 0.938 Female Involved in planning and decision making 0.273 0.450 0.208 0.532 -0.114 0.554 0.178 0.161 -0.608 0.426 0.062 0.196 0.152 0.663 about cocoa production (household level variable) Male Involved in planning and decision making -0.285 0.501 -0.150 0.700 0.011 0.960 -0.146 0.329 1.206 0.179 -0.076 0.177 0.045 0.912 about cocoa production (household level variable) Female Primarily involved in selling cocoa (household -0.688 0.379 -0.290 0.688 -0.514 0.218 0.117 0.673 0.364 0.826 0.044 0.674 0.736 0.330 level variable) Female Primarily involved in receiving payments for cocoa 0.732 0.347 0.665 0.355 0.420 0.312 -0.353 0.198 -0.123 0.941 0.017 0.869 -0.198 0.792 (household level variable) Female manage account 0.677 0.063 0.194 0.564 0.467 0.016 0.017 0.897 -0.804 0.297 -0.029 0.543 -0.124 0.723 (household level variable) Afraid to disagree 0.022 0.952 0.065 0.845 0.017 0.930 -0.060 0.638 0.307 0.688 0.023 0.632 -0.374 0.282 Found at risk -0.001 0.732 0.000 0.925 -0.001 0.633 0.000 0.998 -0.002 0.817 0.000 0.408 -0.001 0.831 Intercept 1.157 0.405 2.798 0.029 -1.398 0.059 -0.243 0.619 12.433 0.000 0.643 0.001 10.217 0.000 N 540 540 540 540 540 540 540 R-square 0.0913 0.0565 0.1674 0.0679 0.1637 0.1608 0.1058 Sources: PPAP Survey, 2017. Appendix C 131 132 TABLE C-14 OLS Regressions on Time-Use (Men) TOTAL HOURS PRIMARY FORMAL NON-PRIMARY DOMESTIC PERSONAL LEARNING OF WORK PRODUCTION WORK PRODUCTION WORK CARE Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Age -0.075 0.004 -0.029 0.197 -0.048 0.010 0.001 0.894 -0.008 0.729 -0.003 0.645 0.019 0.383 Years of schooling -0.062 0.485 -0.141 0.067 0.000 0.996 0.080 0.030 0.076 0.360 0.012 0.566 -0.144 0.055 Literacy in English -0.121 0.877 0.184 0.789 -0.049 0.931 -0.256 0.433 0.525 0.478 0.027 0.880 0.440 0.510 Literacy in Pidgin -1.682 0.116 -1.717 0.065 -0.041 0.957 0.076 0.863 0.662 0.510 0.040 0.872 -0.269 0.767 Training on cocoa 0.431 0.393 0.487 0.269 -0.066 0.856 0.010 0.960 -0.092 0.847 -0.020 0.866 -0.001 0.999 Information on cocoa 0.425 0.425 0.479 0.302 -0.053 0.891 -0.002 0.993 -0.143 0.775 -0.235 0.057 -0.037 0.935 Has a phone 0.076 0.875 -0.619 0.142 0.684 0.049 0.011 0.956 -0.550 0.227 -0.102 0.363 -0.065 0.875 Has access to internet -0.549 0.565 0.156 0.851 -0.099 0.885 -0.606 0.126 -0.493 0.584 0.834 0.000 0.568 0.483 Female head -3.370 0.479 -0.520 0.900 -5.289 0.122 2.438 0.217 -0.954 0.831 -0.315 0.775 5.494 0.174 Married or common law 2.049 0.170 1.664 0.200 0.911 0.395 -0.526 0.394 -2.854 0.043 -0.826 0.017 -0.328 0.795 Number of members -0.113 0.538 -0.151 0.343 -0.176 0.182 0.215 0.005 0.428 0.014 -0.026 0.547 -0.178 0.253 12 years and less Number of members 13-17 years -0.330 0.248 -0.386 0.121 -0.019 0.926 0.074 0.529 -0.395 0.143 0.009 0.887 -0.001 0.995 Number of members 18-59 years 0.592 0.011 0.412 0.041 0.161 0.334 0.019 0.844 0.130 0.550 0.020 0.715 0.028 0.887 Number of members 0.480 0.386 0.107 0.824 0.334 0.401 0.039 0.866 -0.210 0.686 -0.056 0.662 0.295 0.530 60 years and more Hours wage by outsider (female) 0.175 0.731 -0.124 0.780 0.270 0.459 0.029 0.892 -0.096 0.841 0.041 0.725 -0.285 0.508 Hours wage by outsider (male) -0.224 0.709 0.324 0.535 -0.750 0.082 0.202 0.416 -0.441 0.435 0.081 0.557 -0.918 0.072 Participation to agriculture 0.192 0.667 0.013 0.974 0.081 0.800 0.098 0.597 -0.517 0.220 0.249 0.016 -0.353 0.351 association or group Participation to non-agric 0.018 0.386 -0.028 0.119 0.015 0.296 0.030 0.000 -0.015 0.434 0.000 0.922 0.000 0.983 association or group Participation to PPAP -0.201 0.697 0.548 0.223 -0.718 0.053 -0.031 0.883 -0.566 0.245 -0.058 0.627 1.378 0.002 Asset wealth index -0.003 0.637 -0.002 0.718 -0.002 0.693 0.001 0.760 -0.008 0.223 0.000 0.876 0.008 0.167 Living in ARB 0.021 0.253 0.006 0.724 0.010 0.481 0.006 0.425 0.009 0.618 0.006 0.171 0.018 0.260 Permission index 0.017 0.390 -0.020 0.248 0.009 0.503 0.027 0.001 -0.031 0.092 -0.005 0.317 0.001 0.959 Agreement index 0.007 0.676 -0.018 0.254 0.032 0.014 -0.006 0.393 -0.012 0.461 -0.003 0.482 0.007 0.649 Family problem index -0.006 0.714 0.019 0.213 -0.039 0.002 0.014 0.055 0.000 1.000 0.001 0.734 0.014 0.340 Decision index (woman decide) -0.125 0.821 0.398 0.406 -0.207 0.601 -0.316 0.166 -0.445 0.391 0.058 0.648 0.098 0.833 Decision index (partner decide) 0.880 0.190 1.153 0.049 -0.652 0.177 0.378 0.174 0.726 0.251 -0.095 0.540 -0.588 0.302 Female Involved in planning and decision making 0.735 0.496 -0.165 0.861 1.065 0.170 -0.165 0.712 -0.567 0.577 -0.112 0.653 1.156 0.207 about cocoa production (household level variable) Male Involved in planning and decision making about -0.503 0.643 -0.484 0.609 -0.452 0.562 0.433 0.336 0.030 0.977 -0.023 0.927 -1.286 0.163 cocoa production Female Primarily involved in selling cocoa (household level -0.582 0.257 -0.405 0.366 -0.217 0.556 0.040 0.851 -0.489 0.313 0.144 0.226 -0.167 0.701 variable) Female Primarily involved in receiving payments for cocoa -0.472 0.425 0.091 0.860 -0.296 0.485 -0.266 0.277 0.674 0.226 0.213 0.119 -0.601 0.231 (household level variable) Female manage account -0.015 0.083 -0.008 0.274 -0.003 0.586 -0.003 0.343 -0.009 0.259 0.000 0.959 0.006 0.379 (household level variable) Afraid to disagree 4.529 0.093 3.967 0.092 2.684 0.166 -2.122 0.058 5.487 0.031 0.469 0.452 11.108 0.000 Found at risk Intercept 419 419 419 419 419 419 419 0.0978 0.161 0.1198 0.1394 0.1077 0.116 0.0898 N -0.075 0.004 -0.029 0.197 -0.048 0.010 0.001 0.894 -0.008 0.729 -0.003 0.645 0.019 0.383 R-square -0.062 0.485 -0.141 0.067 0.000 0.996 0.080 0.030 0.076 0.360 0.012 0.566 -0.144 0.055 Sources: PPAP Survey, 2017. Appendix C 133 134 TABLE C-15 Household Cocoa Production Regressions YIELD OF COCOA INCOME PER TREE NUMBER OF TREES QUALITY OF PRUNING Estimate P-value Estimate P-value Estimate P-value Estimate P-value Inc. share alt. crop 0.142 0.912 -0.034 0.246 -3.367 0.434 0.065 0.703 Hrs. domestic work -2.735 0.427 -0.040 0.604 0.653 0.955 -0.025 0.961 Female * Hrs dom work 3.874 0.273 0.031 0.692 1.021 0.931 0.020 0.970 Hrs. formal work 4.774 0.384 -0.056 0.648 -30.178 0.099 1.591 0.048 Female * Hrs formal work 3.142 0.726 -0.332 0.099 11.924 0.690 -0.729 0.591 Asset wealth index -1.627 0.063 0.048 0.015 10.196 0.000 0.573 0.000 Living in ARB 11.530 0.644 2.484 0.000 292.402 0.000 -14.206 0.000 Female head 129.680 0.096 0.980 0.576 45.946 0.858 -1.355 0.910 Married or common law -61.215 0.274 -0.596 0.636 -26.988 0.885 10.375 0.194 HHSize 12- 3.461 0.700 0.040 0.845 -49.277 0.096 -0.370 0.776 HHSize 13-17 20.037 0.149 0.712 0.023 -63.683 0.167 4.324 0.028 HHSize 18-59 9.012 0.388 0.594 0.011 19.310 0.574 -3.739 0.010 HHSize 60+ -29.051 0.196 0.797 0.115 -101.153 0.171 0.340 0.917 Particip. agric group -30.589 0.180 -0.625 0.223 198.259 0.009 6.608 0.042 Particip. non-agric group 5.475 0.823 -0.632 0.252 17.924 0.826 -0.350 0.921 Participation to PPAP 17.763 0.415 0.195 0.691 232.233 0.001 7.658 0.015 Permission index 0.562 0.096 0.003 0.662 -0.970 0.387 0.000 0.996 Agreement index -1.124 0.153 0.003 0.882 1.072 0.683 -0.056 0.615 Family problem index 1.917 0.015 0.034 0.053 -2.386 0.363 0.229 0.047 Woman decide index -0.984 0.188 0.001 0.953 1.769 0.476 0.068 0.528 Man decide index -0.090 0.888 0.028 0.051 -5.137 0.016 -0.092 0.315 Female planning 17.144 0.482 2.399 0.000 -230.970 0.004 -0.189 0.958 Male planning 160.583 0.001 1.369 0.217 421.794 0.008 1.626 0.867 Female selling 10.622 0.816 -1.210 0.238 156.500 0.304 -3.383 0.580 Female receiving -0.889 0.985 -1.440 0.165 -20.218 0.896 -5.203 0.403 Female account 71.983 0.004 1.557 0.006 41.201 0.626 -1.766 0.626 Afraid to disagree 2.474 0.926 0.175 0.770 -108.502 0.217 1.880 0.606 Found at risk -0.104 0.742 -0.009 0.199 0.925 0.371 -0.017 0.693 Intercept 128.293 0.496 -4.260 0.121 234.754 0.832 8.337 0.751 N 677 677 686 583 R-square 0.073 0.157 0.106 0.144 Sources: PPAP Survey, 2017. Appendix C 135 136 TABLE C-16 Household Welfare Regressions – Cocoa-Growing Areas INCOME PER WEALTH SCALE WEALTH SCALE WEALTH SCALE CAPITA (LOG) TODAY 5 YEARS AGO IN 5 YEARS Estimate P-value Estimate P-value Estimate P-value Estimate P-value Inc. share alt. crop -0.045 0.000 -0.002 0.826 0.004 0.600 -0.005 0.537 Hrs. domestic work 0.007 0.592 0.056 0.001 0.016 0.343 0.042 0.020 Female * Hrs dom work -0.006 0.639 -0.038 0.029 -0.023 0.185 -0.037 0.047 Hrs. formal work 0.043 0.019 0.029 0.198 -0.040 0.081 0.022 0.357 Female * Hrs formal work -0.011 0.721 0.030 0.448 -0.007 0.857 0.008 0.844 Asset wealth index 0.037 0.000 0.060 0.000 0.042 0.000 0.057 0.000 Living in ARB 0.384 0.000 0.433 0.000 -0.125 0.284 0.515 0.000 Female head -1.124 0.000 0.451 0.171 0.440 0.191 0.259 0.487 Married or common law 0.029 0.888 -0.375 0.161 0.065 0.808 -0.077 0.791 HHSize 12- -0.248 0.000 -0.044 0.281 0.047 0.244 -0.040 0.367 HHSize 13-17 -0.236 0.000 0.050 0.420 0.011 0.862 -0.020 0.773 HHSize 18-59 -0.091 0.019 0.084 0.075 -0.058 0.224 0.079 0.129 HHSize 60+ -0.178 0.022 0.204 0.036 0.394 0.000 0.110 0.295 Particip. agric group 0.080 0.371 0.119 0.278 -0.196 0.077 0.077 0.517 Particip. non-agric group -0.078 0.388 0.006 0.959 -0.073 0.516 0.069 0.565 Participation to PPAP -0.138 0.083 0.115 0.235 0.025 0.797 0.227 0.036 Permission index -0.003 0.033 0.000 0.955 0.004 0.011 -0.001 0.407 Agreement index 0.001 0.669 0.001 0.776 0.007 0.037 0.006 0.137 Family problem index 0.005 0.119 0.001 0.720 0.012 0.001 0.008 0.036 Woman decide index 0.005 0.089 -0.003 0.413 0.012 0.001 -0.001 0.891 Man decide index 0.005 0.017 0.008 0.008 -0.004 0.202 0.006 0.075 Female planning 0.098 0.311 -0.167 0.146 -0.278 0.017 0.228 0.066 Male planning -0.378 0.001 0.327 0.018 -0.014 0.919 0.158 0.289 Female selling 0.665 0.001 0.474 0.056 0.665 0.006 0.666 0.022 Female receiving -0.690 0.001 -0.394 0.112 -0.500 0.039 -0.846 0.003 Female account 0.248 0.006 -0.036 0.748 0.194 0.089 -0.025 0.836 Afraid to disagree -0.196 0.036 -0.078 0.510 0.039 0.742 -0.130 0.310 Found at risk 0.000 0.720 -0.002 0.292 -0.002 0.202 -0.001 0.566 Intercept 7.825 0.000 3.721 0.000 1.383 0.136 5.795 0.000 N 928 928 928 928 R-square 0.339 0.268 0.223 0.247 Sources: PPAP Survey, 2017. Appendix C 137 Appendix D COMPLEMENTARY TABLES (COFFEE) TABLE D-1 Literacy and School Attendance FEMALE MALE DIFF Mean N Mean N p-value Literacy (10 years old or more)* Total 63.6 973 76.1 1145 0.000 10 to 24 years 88.0 325 89.9 386 0.424 25 to 39 years 58.9 380 81.8 302 0.000 40 to 59 years 45.7 230 64.5 361 0.000 60 or more 10.5 38 45.8 96 0.000 School attendance (6 to 24 years) Total 70.7 417 76.9 484 0.038 6 to 13 years 77.4 159 80.7 197 0.442 14 to 18 years 86.6 142 86.9 160 0.948 19 to 24 years 42.2 116 58.3 127 0.012 Sources: PPAP Survey, 2017. *Reads and write English or Pidgin. 138 TABLE D-2 School Level and Completion FEMALE MALE DIFF Mean N Mean N p-value School level (6 to 24 years) None/Kindergarten 30.5 1033 18.4 1216 0.000 Primary 40.8 1033 42.0 1216 0.543 Secondary 26.6 1033 33.7 1216 0.000 University/tertiary 1.9 1033 4.5 1216 0.000 Other 0.2 1033 1.3 1216 0.002 Years completed 3.5 648 5.0 673 0.000 (among those who stopped studying) Sources: PPAP Survey, 2017. TABLE D-3 Participation, Employment and Unemployment FEMALE MALE DIFF Mean N Mean N p-value All (10-69 years old) Participation rate 91.4 965 89.1 1115 0.072 Employment rate 94.7 882 92.0 993 0.022 Unemployment rate 0.7 882 1.6 993 0.056 Adult (25-69 years old) Participation rate 97.0 640 96.6 729 0.628 Employment rate 96.8 621 93.9 704 0.012 Unemployment rate 0.6 621 1.4 704 0.158 Sources: PPAP Survey, 2017. Appendix D 139 TABLE D-4 Employment Characteristics FEMALE MALE DIFF Mean N Mean N p-value Schedule Full time (35+ hours) 22.6 835 23.0 914 0.865 Part time 77.4 835 77.0 914 0.865 Occupation Farmer 78.3 835 67.9 914 0.000 Fisherman 0.0 835 0.1 914 0.317 Hunter 0.0 835 0.0 914 - Forestry worker 0.1 835 0.2 914 0.613 Services & sales worker 1.8 835 2.8 914 0.144 Clerical worker 0.4 835 2.3 914 0.000 Technician 0.1 835 1.4 914 0.001 Professional 1.3 835 4.5 914 0.000 Manager 0.0 835 0.5 914 0.025 Student 1.6 835 1.3 914 0.669 Other 16.4 835 18.7 914 0.206 140 FEMALE MALE DIFF Mean N Mean N p-value Employment status Employee (Wage), public sector 1.1 835 4.7 914 0.000 Employee (Wage), private, Coffee 4.9 835 6.3 914 0.192 Employee (Wage), private, other agricultural 5.4 835 3.9 914 0.152 Employee (Wage), private, non-agricultural 1.0 835 1.8 914 0.149 Self-employed, Coffee 13.9 835 39.6 914 0.000 Self-employed, other agricultural 41.0 835 16.1 914 0.000 Self-employed, non-agricultural 4.1 835 3.7 914 0.704 Unpaid family worker 19.0 835 11.4 914 0.000 Apprentice 0.1 835 1.1 914 0.008 NGO 0.0 835 0.1 914 0.317 Coop 0.0 835 0.0 914 - Student 1.9 835 1.8 914 0.797 Other 7.7 835 9.5 914 0.166 Sources: PPAP Survey, 2017. Appendix D 141 TABLE D-5 Reasons for not Working FEMALE MALE Mean N Mean N p-value Reasons for not working: Domestic work 2.4 84 2.3 129 0.979 Personal / family affairs 2.4 84 0.8 129 0.384 Pregnancy / delivery 2.4 84 0.0 129 0.155 Caring for children 1.2 84 3.9 129 0.197 Illness 0.0 84 1.6 129 0.156 Disability 7.1 84 10.9 129 0.347 Too young 0.0 84 0.0 129 - Remittances 13.1 84 7.0 129 0.158 Old aged/ pension 64.3 84 66.7 129 0.723 Student 7.1 84 7.0 129 0.963 Other 0.0 84 0.0 129 - Sources: PPAP Survey, 2017. 142 TABLE D-6 Household Composition, Income and Satisfaction TOTAL FEMALE HEAD MALE HEAD GENDER GAP Mean N Mean N Mean N p-value Main source of income Earning from coffee 83.1 668 96.3 27 82.5 641 0.001 Earning from other 9.0 668 3.7 27 9.2 641 0.156 agriculture products Earning from livestock 1.5 668 0.0 27 1.6 641 0.001 Earning from fishing 0.0 668 0.0 27 0.0 641 - Earning from non-agriculture 1.8 668 0.0 27 1.9 641 0.000 business Salaries/wages/commissions 4.2 668 0.0 27 4.4 641 0.000 Earning from rents (house/ 0.3 668 0.0 27 0.3 641 0.157 assets/properties) Remittances from abroad 0.0 668 0.0 27 0.0 641 - Domestic remittances 0.0 668 0.0 27 0.0 641 - Pension 0.0 668 0.0 27 0.0 641 - Aid in nature / cash 0.0 668 0.0 27 0.0 641 - Freely from forest 0.0 668 0.0 27 0.0 641 - Other 0.1 668 0.0 27 0.2 641 0.317 Wealth scale (from 1 the poorest to 10 the richest) Today 3.3 667 3.7 27 3.3 640 0.132 Five years ago 2.7 667 2.9 27 2.7 640 0.329 In five years 5.2 667 5.3 27 5.1 640 0.590 Appendix D 143 TOTAL FEMALE HEAD MALE HEAD GENDER GAP Mean N Mean N Mean N p-value Income by source Coffee cherry 694.4 670 556.8 29 700.6 641 0.721 Coffee parchment 1058.2 670 1091.8 29 1056.7 641 0.903 Coffee green bean 86.3 670 260.0 29 78.4 641 0.325 Off-farm employment 715.7 670 93.4 29 743.9 641 0.028 Non-farm income e.g. trade 166.9 670 172.4 29 166.6 641 0.974 store. PMV Hunting and fishing 70.9 670 0.0 29 74.2 641 0.001 Gifts. Customary payments. 0.0 670 0.0 29 0.0 641 - remittances Balsa 5.3 670 0.0 29 5.6 641 0.009 Other agricultural 41.9 670 51.7 29 41.4 641 0.849 Other 30.8 670 0.0 29 32.2 641 0.005 Total income per capita 884.2 670 710.8 29 892.1 641 0.332 Household composition Number of members 0.9 670 0.4 29 0.9 641 0.000 12 years and less Number of members 0.4 670 0.4 29 0.4 641 0.933 13-17 years Number of members 2.3 670 1.7 29 2.4 641 0.007 18-59 years Number of members 0.2 670 0.1 29 0.2 641 0.146 60 years and more Sources: PPAP Survey, 2017. 144 TABLE D-7 Who Makes Purchasing Decision, Including Durable Goods? WOMAN ALONE BOTH WOMAN & MAN Estimate P-value Estimate P-value Hours of domestic work -0.005 0.226 0,003 0,475 Hours of formal work 0.002 0.951 -0,009 0,827 Share of alternative crop income in total income -0.001 0.585 0,002 0,523 Hours wage by outsider (female) 0.000 0.982 0,000 0,817 Hours wage by outsider (male) 0.000 0.965 0,000 0,841 Asset wealth index -0.001 0.418 0,001 0,334 Living in Western Highlands -0.008 0.873 0,027 0,650 Living in Jiwaka -0.030 0.464 0,043 0,341 Living in Simbu 0.150 0.002 -0,138 0,008 Married or common law -0.121 0.107 0,183 0,027 Number of members 12 years and less -0.003 0.856 0,007 0,653 Number of members 13-17 years 0.002 0.950 0,025 0,392 Number of members 18-59 years -0.037 0.049 0,057 0,006 Number of members 60 years and more -0.035 0.428 -0,002 0,968 Participation to agriculture association or group -0.058 0.318 0,087 0,171 Participation to non-agric association or group -0.017 0.618 0,037 0,337 Participation to PPAP -0.086 0.014 0,112 0,004 Age 0.006 0.002 -0,007 0,001 Years of schooling 0.002 0.660 -0,001 0,918 Training on coffee 0.088 0.679 -0,100 0,666 Information on coffee 0.052 0.245 -0,038 0,447 Has a phone 0.040 0.431 -0,009 0,869 Has access to internet 0.266 0.061 -0,261 0,094 Intercept 0.158 0.171 0,695 0,000 N 373 373 R-square 0.1458 0.1616 Sources: PPAP Survey, 2017. Appendix D 145 TABLE D-8 Trends in Activities (frequencies) Performed by Household Members 2011 2016 % Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Clearing Land 20.3 36.7 0.000 19.1 28.0 0.000 0.005 Transplanting 9.5 22.7 0.000 6.0 14.0 0.000 0.010 Shade establishment and 3.7 29.3 0.000 4.6 16.9 0.000 0.000 control Nursery operations 1.5 6.3 0.000 1.4 2.8 0.022 0.001 Weeding 37.6 57.6 0.000 50.6 52.1 0.520 0.000 Fertilizing/mulching 0.4 4.4 0.000 7.2 10.6 0.010 0.640 Fencing 1.0 8.8 0.000 1.9 5.6 0.000 0.001 Digging / cleaning drains 11.2 37.8 0.000 8.8 22.6 0.000 0.000 Pruning / renovating 6.5 44.8 0.000 10.9 39.7 0.000 0.000 Pest and disease 0.4 4.6 0.000 0.9 1.4 0.333 0.000 management Coffee picking 62.3 63.2 0.664 71.6 69.4 0.275 0.277 Wet milling (pulping and 34.3 42.9 0.000 29.7 41.0 0.000 0.354 washing) Drying (sun drying and 36.3 42.1 0.004 39.7 38.4 0.554 0.018 bagging) Farm record keeping 0.3 1.0 0.032 0.7 0.8 0.785 0.248 Sources: PPAP Survey, 2012-2017. 146 TABLE D-9 Trends in Activities (days of work) Performed by Household Members 2011 2016 Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Clearing Land 8.6 9.7 0.576 27.5 20.9 0.190 0.154 Transplanting 7.6 7.9 0.794 17.5 12.9 0.529 0.509 Shade establishment and 5.5 4.8 0.565 16.5 8.7 0.340 0.383 control Nursery operations 4.4 8.7 0.012 3.0 22.6 0.089 0.189 Weeding 12.7 10.9 0.061 11.7 10.0 0.178 0.966 Fertilizing/mulching 5.3 7.5 0.213 10.3 17.7 0.295 0.482 Fencing 8.7 7.7 0.687 12.9 6.6 0.489 0.580 Digging / cleaning drains 9.1 8.2 0.663 13.9 12.3 0.704 0.873 Pruning / renovating 7.9 8.1 0.840 7.8 9.1 0.557 0.659 Pest and disease 11.8 7.3 0.384 1.9 16.9 0.288 0.195 management Coffee picking 16.7 16.2 0.619 11.1 9.5 0.162 0.476 Wet milling (pulping and 7.7 7.2 0.498 5.8 5.6 0.809 0.703 washing) Drying (sun drying and 12.7 11.0 0.183 8.6 8.6 0.904 0.219 bagging) Farm record keeping 2.5 1.8 0.629 37.5 22.9 0.576 0.596 Sources: PPAP Survey, 2012-2017. Appendix D 147 TABLE D-10 Trends in Activities Performed by Household Members (Frequencies x Days of Work) 2011 2016 % Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Clearing Land 1.7 3.5 0.001 5.3 5.8 0.624 0.358 Transplanting 0.7 1.8 0.000 1.0 1.8 0.211 0.604 Shade establishment 0.2 1.4 0.000 0.8 1.5 0.101 0.319 and control Nursery operations 0.1 0.5 0.000 0.0 0.6 0.079 0.648 Weeding 4.8 6.3 0.002 5.9 5.2 0.302 0.008 Fertilizing/mulching 0.0 0.3 0.000 0.7 1.9 0.138 0.276 Fencing 0.1 0.7 0.000 0.2 0.4 0.519 0.045 Digging / cleaning drains 1.0 3.1 0.000 1.2 2.8 0.002 0.391 Pruning / renovating 0.5 3.6 0.000 0.9 3.6 0.000 0.539 Pest and disease 0.1 0.3 0.000 0.0 0.2 0.275 0.790 management Coffee picking 10.5 10.3 0.789 8.0 6.6 0.107 0.294 Wet milling 2.6 3.1 0.118 1.7 2.3 0.009 0.730 (pulping and washing) Drying 4.6 4.6 0.955 3.4 3.3 0.798 0.857 (sun drying and bagging) Farm record keeping 0.0 0.0 0.087 0.3 0.2 0.727 0.685 Sources: PPAP Survey, 2012-2017. 148 TABLE D-11 Trends in Activities Performed by Outsiders 2011 2016 Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Hours of paid labour by 253.3 834.5 0.420 13.3 34.3 0.308 0.437 outsiders Average hourly wage for 0.7 1.7 0.275 47.3 83.9 0.464 0.476 outsiders (PGK) Sources: PPAP Survey, 2012-2017. TABLE D-12 Activity-Based Decision Making 2011 2016 % Female Male Diff Female Male Diff Diff-in-diff p-value p-value p-value Primarily involved - - - 5.6 21.6 0.000 - in selling livestock Primarily involved 5.5 30.1 0.000 6.6 39.2 0.000 0.000 in selling coffee Primarily involved in 17.9 13.6 0.000 8.7 37.4 0.000 0.000 receiving payments for coffee Involved in planning and decision making about 27.6 37.4 0.000 35.7 50.4 0.000 0.050 coffee production Operate the account 2.0 8.4 0.000 1.6 11.3 0.000 0.004 Sources: PPAP Survey, 2012-2017. Appendix D 149 150 TABLE D-13 OLS Regressions on Time-Use (Women) TOTAL HOURS PRIMARY FORMAL NON-PRIMARY DOMESTIC PERSONAL LEARNING OF WORK PRODUCTION WORK PRODUCTION WORK CARE Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Age 0.026 0.301 0.012 0.607 0.002 0.371 0.011 0.368 -0.066 0.007 -0.016 0.001 -0.023 0.387 Years of schooling -0.052 0.560 -0.074 0.384 0.029 0.004 -0.007 0.874 -0.070 0.425 0.037 0.036 -0.044 0.642 Literacy in English 1.548 0.011 1.911 0.001 -0.127 0.057 -0.236 0.438 -0.463 0.440 -0.035 0.773 -0.299 0.646 Literacy in Pidgin -1.488 0.014 -1.607 0.005 -0.101 0.130 0.220 0.466 0.493 0.408 -0.144 0.232 0.259 0.688 Training on coffee -0.788 0.715 -0.399 0.845 1.345 0.000 -1.733 0.108 -1.840 0.386 -0.486 0.259 -1.626 0.481 Information on coffee -1.371 0.019 -1.645 0.003 0.116 0.068 0.158 0.585 0.254 0.656 -0.191 0.099 -1.129 0.069 Has a phone 1.082 0.084 0.658 0.267 0.006 0.935 0.418 0.180 -0.994 0.106 0.100 0.421 0.038 0.955 Has access to internet -2.508 0.112 -3.118 0.038 1.491 0.000 -0.881 0.263 -1.256 0.418 1.957 0.000 -1.233 0.464 Female head 1.555 0.187 1.532 0.170 0.178 0.169 -0.154 0.793 2.803 0.016 -0.025 0.916 1.727 0.170 Married or common law 1.482 0.209 0.853 0.446 0.210 0.105 0.419 0.477 0.589 0.612 -0.020 0.932 -1.001 0.427 Number of members 12 years 0.289 0.102 0.116 0.488 0.007 0.721 0.166 0.061 0.623 0.000 0.011 0.754 0.607 0.001 and less Number of members 13-17 years 0.180 0.586 0.050 0.873 0.049 0.172 0.080 0.627 -0.291 0.369 0.206 0.002 0.621 0.079 Number of members 18-59 years -0.165 0.489 0.110 0.628 -0.042 0.106 -0.232 0.052 0.084 0.721 0.173 0.000 0.362 0.156 Number of members 60 years -0.020 0.970 0.326 0.525 -0.077 0.193 -0.268 0.320 -0.392 0.461 0.548 0.000 0.288 0.618 and more Hours wage by outsider (female) 0.020 0.369 0.012 0.570 0.001 0.751 0.007 0.514 0.005 0.833 -0.006 0.211 -0.004 0.850 Hours wage by outsider (male) -0.018 0.423 -0.012 0.560 -0.001 0.716 -0.005 0.675 -0.007 0.729 0.005 0.243 0.006 0.811 Participation to agriculture -0.120 0.861 0.027 0.967 -0.130 0.087 -0.017 0.960 1.035 0.127 0.015 0.915 0.155 0.834 association or group Participation to non-agric -0.129 0.772 0.034 0.935 0.019 0.694 -0.183 0.412 -0.369 0.400 0.142 0.110 -0.948 0.047 association or group Participation to PPAP -0.768 0.073 -0.356 0.380 -0.016 0.728 -0.395 0.065 1.147 0.007 0.094 0.271 -0.157 0.731 Asset wealth index 0.039 0.010 0.031 0.029 0.002 0.210 0.005 0.465 0.058 0.000 -0.005 0.102 0.037 0.022 Living in Western Highlands -0.914 0.174 -0.908 0.155 0.080 0.280 -0.087 0.796 1.126 0.089 -0.289 0.031 0.998 0.165 Living in Jiwaka 2.064 0.000 1.658 0.002 0.022 0.716 0.384 0.169 1.635 0.003 -0.206 0.065 2.977 0.000 Living in Simbu -0.838 0.205 -0.579 0.356 0.026 0.716 -0.285 0.387 1.377 0.035 -0.167 0.205 1.113 0.116 Permission index 0.011 0.078 0.004 0.453 -0.001 0.416 0.007 0.022 0.006 0.327 -0.001 0.418 0.009 0.176 Agreement index -0.008 0.405 -0.007 0.458 -0.002 0.054 0.001 0.868 0.014 0.152 -0.005 0.008 0.001 0.961 Family problem index -0.031 0.026 -0.030 0.024 0.005 0.002 -0.006 0.377 0.031 0.025 -0.002 0.512 -0.021 0.148 Decision index (woman decide) 0.016 0.148 0.003 0.789 -0.001 0.413 0.014 0.010 -0.031 0.005 0.001 0.761 -0.024 0.046 Decision index (partner decide) -0.021 0.026 -0.003 0.756 0.000 0.640 -0.018 0.000 0.003 0.723 -0.002 0.194 -0.033 0.001 Female Involved in planning and decision making 0.158 0.757 0.112 0.818 0.044 0.430 0.002 0.993 0.605 0.230 -0.124 0.223 1.245 0.023 about coffee production (household level variable Male Involved in planning anddecision making -0.240 0.825 1.064 0.302 -0.135 0.259 -1.169 0.032 -0.329 0.758 0.363 0.094 1.619 0.164 aboutcoffee production (household level variable) Female Primarily involved in selling coffee 0.024 0.979 0.532 0.546 -0.215 0.036 -0.293 0.527 0.944 0.302 -0.217 0.241 -0.521 0.600 (household level variable) Female Primarily involved in receiving payments for coffee -0.887 0.295 -0.846 0.292 -0.085 0.362 0.043 0.918 -0.895 0.283 0.496 0.003 0.545 0.547 (household level variable) Female manage account 0.145 0.901 -0.565 0.610 0.313 0.015 0.397 0.496 -0.881 0.443 0.016 0.946 1.312 0.294 (household level variable) Afraid to disagree 0.552 0.378 0.283 0.634 0.029 0.670 0.240 0.442 -0.466 0.449 -0.129 0.302 -0.836 0.212 Found at risk 0.001 0.851 0.003 0.630 -0.001 0.364 -0.001 0.735 -0.005 0.392 0.002 0.110 -0.003 0.679 Intercept 2.123 0.262 1.149 0.522 0.076 0.716 0.899 0.341 3.078 0.099 0.449 0.234 10.840 0.000 N 383 383 383 383 383 383 383 R-square 0.2192 0.202 0.4743 0.1512 0.2594 0.3117 0.2734 Appendix D Sources: PPAP Survey, 2017. 151 152 TABLE D-14 OLS Regressions on Time-Use (Men) TOTAL HOURS PRIMARY FORMAL NON-PRIMARY DOMESTIC PERSONAL LEARNING OF WORK PRODUCTION WORK PRODUCTION WORK CARE Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Estimate P-value Age -0.042 0.219 -0.027 0.399 0.006 0.668 -0.022 0.168 0.003 0.846 -0.003 0.165 0.025 0.357 Years of schooling -0.068 0.536 -0.089 0.380 0.015 0.740 0.006 0.911 0.007 0.872 -0.009 0.240 -0.126 0.150 Literacy in English 0.222 0.809 0.163 0.847 -0.098 0.797 0.157 0.710 0.599 0.102 0.060 0.344 0.469 0.522 Literacy in Pidgin -0.720 0.435 -0.248 0.770 0.041 0.915 -0.513 0.225 -0.210 0.565 -0.017 0.782 0.409 0.576 Training on coffee 0.837 0.355 0.564 0.498 -0.150 0.689 0.423 0.309 -0.416 0.247 0.071 0.251 -1.047 0.146 Information on coffee -0.479 0.470 -1.387 0.023 0.367 0.181 0.542 0.075 0.169 0.520 -0.037 0.417 -0.525 0.318 Has a phone -0.185 0.792 -0.193 0.764 0.034 0.906 -0.026 0.935 0.028 0.918 0.043 0.368 -0.342 0.538 Has access to internet 1.646 0.401 0.446 0.805 0.679 0.403 0.522 0.561 -0.580 0.455 -0.054 0.687 -1.446 0.353 Female head 2.347 0.303 1.897 0.365 0.012 0.990 0.437 0.675 0.416 0.645 -0.070 0.654 0.506 0.780 Married or common law 5.872 0.035 5.245 0.040 0.412 0.719 0.215 0.866 0.228 0.836 -2.352 0.000 -0.970 0.659 Number of members 0.162 0.540 0.036 0.883 0.165 0.133 -0.039 0.750 0.240 0.023 -0.004 0.808 0.624 0.003 12 years and less Number of members 13-17 years 0.159 0.753 0.778 0.095 -0.218 0.299 -0.401 0.084 -0.512 0.011 0.027 0.429 0.049 0.902 Number of members 18-59 years 0.325 0.394 0.437 0.213 0.043 0.783 -0.155 0.375 -0.040 0.792 -0.001 0.976 0.516 0.090 Number of members 0.085 0.931 0.787 0.387 -0.384 0.348 -0.317 0.484 0.097 0.805 0.061 0.367 0.800 0.308 60 years and more Hours wage by outsider (female) 0.623 0.556 1.302 0.181 -0.006 0.988 -0.673 0.166 0.398 0.342 0.073 0.315 0.853 0.309 Hours wage by outsider (male) -1.177 0.087 -1.221 0.053 -0.073 0.796 0.117 0.709 0.169 0.533 -0.030 0.528 0.499 0.359 Participation to agriculture -0.763 0.262 -0.689 0.270 -0.060 0.831 -0.014 0.965 0.102 0.705 -0.029 0.535 -0.055 0.919 association or group Participation to non-agric 0.020 0.367 -0.011 0.592 0.015 0.099 0.016 0.123 0.002 0.844 -0.001 0.414 0.025 0.158 association or group Participation to PPAP 1.416 0.218 1.412 0.182 0.056 0.906 -0.051 0.922 0.772 0.091 0.064 0.413 0.770 0.399 Asset wealth index 2.889 0.001 2.098 0.009 0.166 0.644 0.625 0.117 0.622 0.071 0.008 0.896 1.184 0.086 Living in Western Highlands 0.323 0.765 0.354 0.721 0.470 0.293 -0.501 0.311 0.650 0.129 -0.035 0.640 0.327 0.702 Living in Jiwaka 0.010 0.323 0.011 0.239 0.001 0.766 -0.002 0.634 -0.002 0.524 -0.001 0.074 -0.003 0.693 Living in Simbu 0.021 0.176 0.019 0.185 -0.006 0.380 0.008 0.277 0.010 0.115 0.002 0.157 0.015 0.232 Permission index -0.025 0.246 -0.033 0.098 -0.009 0.289 0.017 0.081 -0.002 0.797 -0.003 0.058 -0.018 0.285 Agreement index 0.015 0.405 0.003 0.876 0.002 0.735 0.010 0.231 0.004 0.538 0.000 0.758 0.003 0.836 Family problem index -0.056 0.001 -0.033 0.037 -0.001 0.860 -0.022 0.006 -0.011 0.114 0.002 0.165 -0.041 0.003 Decision index (woman decide) -1.622 0.029 -1.110 0.103 -0.119 0.698 -0.393 0.246 0.577 0.050 -0.040 0.425 1.039 0.077 Decision index (partner decide) -9.576 0.033 -11.759 0.005 0.614 0.739 1.569 0.444 0.388 0.826 -0.139 0.649 -3.466 0.329 Female Involved in planning and decision making 1.187 0.444 0.001 1.000 0.311 0.627 0.875 0.219 -0.421 0.493 0.142 0.183 1.456 0.237 about coffee production (household level variable) Male Involved in planning and decision making -1.562 0.278 -1.112 0.400 -0.247 0.678 -0.204 0.757 -0.247 0.665 0.110 0.264 -0.574 0.615 about coffee production (household level variable) Female Primarily involved in selling coffee -0.086 0.959 0.831 0.582 -0.748 0.273 -0.169 0.823 0.327 0.616 -0.025 0.826 1.734 0.185 (household level variable) Female Primarily involved in receiving payments for coffee 1.692 0.180 1.612 0.164 0.234 0.653 -0.155 0.788 1.141 0.023 -0.059 0.496 1.119 0.263 (household level variable) Female manage account -0.004 0.767 0.000 0.975 -0.001 0.912 -0.004 0.543 -0.006 0.329 0.002 0.045 -0.034 0.004 (household level variable) Afraid to disagree 12.137 0.026 13.068 0.009 -1.136 0.612 0.205 0.934 -1.305 0.543 2.492 0.000 12.112 0.005 Found at risk -0.042 0.219 -0.027 0.399 0.006 0.668 -0.022 0.168 0.003 0.846 -0.003 0.165 0.025 0.357 Intercept -0.068 0.536 -0.089 0.380 0.015 0.740 0.006 0.911 0.007 0.872 -0.009 0.240 -0.126 0.150 N 243 243 243 243 243 243 243 R-square 0.2199 0.2407 0.0828 0.2007 0.2314 0.5525 0.3169 Appendix D Sources: PPAP Survey, 2017. 153 154 TABLE D-15 Household Coffee Production Regressions YIELD OF COCOA INCOME PER TREE NUMBER OF TREES QUALITY OF PRUNING Estimate P-value Estimate P-value Estimate P-value Estimate P-value Inc. share alt. crop -0.119 0.730 -0.006 0.867 -38.9 0.112 -0.267 0.169 Hrs. domestic work -0.998 0.652 0.163 0.453 -238.3 0.128 0.791 0.518 Female * Hrs dom work 0.251 0.908 -0.121 0.570 186.2 0.225 -0.823 0.492 Hrs. formal work -2.789 0.227 -0.234 0.302 96.8 0.553 3.817 0.005 Female * Hrs formal work -0.195 0.977 -0.532 0.434 1416.9 0.004 -8.112 0.034 Asset wealth index 0.057 0.771 0.009 0.631 72.6 0.000 0.203 0.058 Living in Western Highlands 7.670 0.412 -2.746 0.003 -1182.8 0.071 -9.123 0.082 Living in Jiwaka -0.820 0.910 -0.577 0.417 -1519.7 0.003 -14.343 0.000 Living in Simbu 36.719 0.000 -2.074 0.015 -161.8 0.793 -8.749 0.071 Female head 78.512 0.000 -0.302 0.857 126.4 0.916 4.040 0.667 Married or common law 37.488 0.024 1.037 0.525 1298.8 0.269 8.649 0.346 HHSize 12- 0.524 0.823 -0.064 0.781 -7.0 0.966 5.079 0.000 HHSize 13-17 -7.342 0.093 -0.023 0.957 -567.8 0.066 1.382 0.569 HHSize 18-59 4.791 0.136 0.862 0.006 -76.3 0.736 4.114 0.020 HHSize 60+ -3.209 0.663 -0.480 0.508 1956.6 0.000 -1.279 0.754 Particip. agric group -4.823 0.582 0.923 0.284 -1853.6 0.003 2.784 0.562 Particip. non-agric group -4.947 0.368 0.870 0.107 442.6 0.254 -0.754 0.806 Participation to PPAP -9.472 0.099 0.339 0.548 -1876.6 0.000 -6.770 0.035 Permission index -0.023 0.770 0.015 0.053 8.8 0.110 0.074 0.089 Agreement index -0.340 0.015 -0.062 0.000 13.8 0.164 -0.109 0.183 Family problem index -0.357 0.049 -0.041 0.021 -29.8 0.020 -0.094 0.352 Woman decide index 0.105 0.477 0.029 0.046 -61.1 0.000 -0.066 0.433 Man decide index 0.166 0.216 0.017 0.200 0.2 0.986 -0.057 0.447 Female planning 2.637 0.691 0.120 0.854 -778.3 0.096 2.440 0.510 Male planning 13.805 0.423 -8.099 0.000 764.7 0.530 5.108 0.591 Female selling -10.316 0.425 -1.162 0.361 594.2 0.516 -20.210 0.005 Female receiving -4.409 0.700 -2.060 0.067 72.7 0.928 17.479 0.006 Female account -1.087 0.944 -1.884 0.214 964.9 0.376 7.333 0.386 Afraid to disagree 7.290 0.417 1.263 0.152 -90.3 0.887 -1.911 0.704 Found at risk 0.071 0.428 0.002 0.823 -1.7 0.793 0.013 0.791 Intercept -31.208 0.310 8.789 0.002 5986.0 0.336 8.016 0.736 N 578 578 580 556 R-square 0.373 0.147 0.279 0.131 Sources: PPAP Survey, 2017. Appendix D 155 156 TABLE D-16 Household Welfare Regressions – Coffee-Growing Areas INCOME PER WEALTH SCALE WEALTH SCALE WEALTH SCALE CAPITA (LOG) TODAY 5 YEARS AGO IN 5 YEARS Estimate P-value Estimate P-value Estimate P-value Estimate P-value Inc. share alt. crop -0.035 0.000 0.018 0.014 0.037 0.000 -0.024 0.003 Hrs. domestic work -0.021 0.584 0.006 0.892 0.015 0.797 0.042 0.456 Female * Hrs dom work 0.036 0.331 -0.052 0.271 -0.064 0.266 -0.062 0.260 Hrs. formal work 0.108 0.007 0.008 0.876 -0.015 0.798 -0.007 0.899 Female * Hrs formal work 0.076 0.521 -0.118 0.419 0.091 0.610 -0.633 0.000 Asset wealth index 0.028 0.000 0.028 0.000 0.007 0.134 0.049 0.000 Living in Western Highlands -0.418 0.008 0.034 0.864 0.779 0.001 0.049 0.829 Living in Jiwaka 0.170 0.171 0.012 0.936 0.544 0.003 -0.586 0.001 Living in Simbu -0.514 0.000 -0.210 0.243 -0.484 0.027 1.440 0.000 Female head -0.032 0.914 0.327 0.364 0.469 0.285 0.880 0.034 Married or common law 0.007 0.979 0.414 0.232 0.292 0.489 0.878 0.026 HHSize 12- -0.307 0.000 -0.121 0.012 0.008 0.891 0.050 0.380 HHSize 13-17 -0.315 0.000 -0.139 0.128 -0.032 0.774 -0.051 0.630 HHSize 18-59 -0.143 0.009 -0.027 0.691 0.181 0.027 0.058 0.444 HHSize 60+ -0.426 0.001 -0.156 0.306 -0.166 0.372 0.004 0.981 Particip. agric group -0.173 0.252 -0.249 0.182 -0.050 0.826 -0.996 0.000 Particip. non-agric group 0.212 0.023 0.079 0.491 0.039 0.779 -0.067 0.609 Participation to PPAP -0.072 0.455 0.095 0.420 0.299 0.038 -0.215 0.113 Permission index -0.001 0.703 0.011 0.000 0.011 0.000 0.003 0.124 Agreement index 0.005 0.031 -0.022 0.000 -0.028 0.000 -0.029 0.000 Family problem index 0.001 0.740 -0.007 0.069 -0.012 0.007 0.002 0.667 Woman decide index 0.009 0.000 0.019 0.000 0.010 0.011 0.016 0.000 Man decide index -0.005 0.025 -0.001 0.775 0.006 0.089 0.006 0.060 Female planning 0.062 0.589 0.392 0.005 0.403 0.017 0.321 0.049 Male planning 0.430 0.144 -0.254 0.486 -0.929 0.037 0.014 0.973 Female selling 0.126 0.566 -0.437 0.095 -0.480 0.133 -0.701 0.025 Female receiving -0.095 0.617 0.003 0.989 -0.240 0.395 0.048 0.863 Female account -0.185 0.469 -0.108 0.729 -0.515 0.175 -0.141 0.680 Afraid to disagree -0.063 0.679 0.011 0.951 -0.143 0.527 0.517 0.014 Found at risk -0.002 0.150 0.003 0.084 0.004 0.094 0.002 0.268 Intercept 5.930 0.000 2.418 0.002 2.895 0.000 4.517 0.000 N 602 600 600 600 R-square 0.329 0.308 0.274 0.443 Sources: PPAP Survey, 2017. Appendix D 157 158 Photo: PPAP, PNG Appendix E LITERATURE REVIEW Economic literature sets minimum principles from Over the past twenty years, most of the research which a theory of household behavior could be drawn. effort on this topic has been dedicated to testing This is important since meaningful models of the efficiency (Chiappori and Donni 2006). In fact, this family can be used to predict future demographic assumption is justified by the main authors in this trends (or understand past ones) or to analyze the literature by the Folk THEOREM: efficiency stems effect of policies on household decision making from a repeated non-cooperative game between (including female labor supply, fertility, intra-household players with the perfect symmetry of information. inequality, etc.). The standard unitary model, which Yet, at least two other streams of the literature come treats multi-person households as a single decision to contradict this view. Some authors argue that the maker, has been abundantly rejected (Browning and conditions under which efficiency can be presumed Chiappori 1998; Lundberg, Pollak and Wales 1997). Its are specific: many decisions in life are too rare, too most convincing replacement, the collective model, engaging or too irreversible to allow cooperation (e.g., respects individual preferences and relies on the location decisions, work decisions, fertility, etc.). In sole assumption of efficiency of household decisions contrast, strategic decisions are expected to take (Chiappori 1988 and 1992). Other cooperative models place (Lundberg and Pollak 1994), and the outcome is in which the bargaining model is specified, i.e. in expected to be inefficient (Lundberg and Pollak 1993 which further axioms to efficiency are assumed, do and 2001; Chen and Wooley 2001, and Haddad and not necessarily offer additional testable restrictions Kanbur 1994). than those stemming from the efficiency assumption (Chiappori, Donni and Komunjer 2011). Appendix E 159 Another, more extreme view, is that couples are Time poverty can be understood as the fact that some fundamentally non-cooperative. Tests of efficiency individuals do not have enough time for rest and leisure in developing countries tend to support this view. after taking into account the time spent working, They mainly concern productive decisions among whether in the labor market, for domestic work, or agricultural households (Udr 1996; Duflo and Udry for other activities such as fetching water and wood 2004). For instance, Carter and Katz (1997) remind us (Bardasi and Wodon 2005). The availability of better that gender-based norms, divisions, and conflicts are data on time-use in developing countries makes the important in the determination of household resource possibility to measure time poverty for a wide range allocation. This latter view states that the household is of countries.4 Other important research has been better conceived as consisting of separate, gendered done on the impact of technological change on the spheres of decision making and activity that are value of time. In some cases, the diversity of activities related to one another by a “conjugal contract” — the may reduce the total amount of time available for one terms under which household members exchange particular activity and in others, technological change goods, incomes, and services among themselves. can increase time availability due to the decrease Quinsumbing and Maluccio (2000) found that the of domestic work. All these time-use changes have central role of women is determining household well- an impact on household satisfaction. In developing being, and that when higher relative resources are countries, many activities (such as socio-cultural, controlled by women they tend to increase the shares community or domestic activities) are not valued by spent on the education of children. the market, although they are valued by the individuals and communities. Thus, it is important to measure the Considering possible non-cooperative behaviors extent of these activities in order to better assess the within the household, important research questions impact of various exogenous changes on both time-use need to be asked about the way individuals trade their and well-being. time on the market as well as within the household. In particular: how individuals produce satisfaction with time and goods? And, how various activities aggregate to produce well-being? Whether domestic time can be attributed value or not isn’t important in order to assess the impact of time-use on well-being (Aguiar et al. 2013). Valuing domestic time allows us to estimate a complete income that is far less unequally distributed than a monetary income. Valuing domestic (or leisure) time can diminish by about 20-30 percent household inequality measured on complete income instead of monetary income. During the great recession in the US, satisfaction decreased by about 6 percent whereas income decreased by 10-12 percent; this relatively low elasticity can be explained by the fact that monetary income losses have been compensated by more domestic activities and leisure time. Also, it is important to measure the value of non-market activities to explain well-known paradoxes such as the Easterlin paradox which considers that household life- satisfaction does not generally improve when monetary 4. The linear regression model was a useful method to analyze income increases (although household satisfaction correlations and a powerful tool for causality analysis with non- might improve when complete income increases or experimental data. See http://timeuse-2009.nsms.ox.ac.uk/ relative income increases). information/studies/ for further information. 160 Photo: Thomas Perry/World Bank. Introduction 161 162 Photo front cover: Conor Ashleigh/World Bank / Back cover: Stephane Forman /World Bank.