Digital financial services and livelihood diversification in rural Ghana

Atta-Ankomah, Richmond; Adjei-Mantey, Kwame; Amankwah, Akuffo Article Digital financial services and livelihood diversification in rural Ghana Cogent Economics & Finance Provided in Cooperation with: Taylor & Francis Group Suggested Citation: Atta-Ankomah, Richmond; Adjei-Mantey, Kwame; Amankwah, Akuffo (2024) : Digital financial services and livelihood diversification in rural Ghana, Cogent Economics & Finance, ISSN 2332-2039, Taylor & Francis, Abingdon, Vol. 12, Iss. 1, pp. 1-20, https://doi.org/10.1080/23322039.2024.2330434 This Version is available at: https://hdl.handle.net/10419/321457 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Cogent Economics & Finance ISSN: 2332-2039 (Online) Journal homepage: www.tandfonline.com/journals/oaef20 Digital financial services and livelihood diversification in rural Ghana Richmond Atta-Ankomah, Kwame Adjei-Mantey & Akuffo Amankwah To cite this article: Richmond Atta-Ankomah, Kwame Adjei-Mantey & Akuffo Amankwah (2024) Digital financial services and livelihood diversification in rural Ghana, Cogent Economics & Finance, 12:1, 2330434, DOI: 10.1080/23322039.2024.2330434 To link to this article: https://doi.org/10.1080/23322039.2024.2330434 © 2024 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group View supplementary material Published online: 05 Apr 2024. Submit your article to this journal Article views: 2410 View related articles View Crossmark data Citing articles: 6 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=oaef20 DEVELOPMENT ECONOMICS | RESEARCH ARTICLE Digital financial services and livelihood diversification in rural Ghana Richmond Atta-Ankomah a , Kwame Adjei-Mantey b and Akuffo Amankwah c a Institute of Statistical, Social and Economic Research, University of Ghana, Legon, Ghana; b Department of Applied Economics, University of Environment and Sustainable Development, Somanya, Ghana; c Living Standards Measurement Study, Development Data Group, Washington, DC, USA ABSTRACT The emergence of digital financial systems, especially mobile money, has significantly changed the financial services space in sub-Saharan Africa in ways that promote financial inclusion among the poor, including those in rural communities. Focusing on mobile money, which is the main avenue for digital financial services for rural households in Ghana, this study investigates whether digital finance affects rural households’livelihood diversification, using a nationally representative data on Ghana. Employing several econometric methods, including instrumental variable techniques to address potential endogeneity bias, the study finds mobile money to be positively associated with households’choice of non-farm business as a livelihood option as well as their engagement in livestock production. However, we find a negative relationship between mobile money and diversification in crop production including the extent to which households engaged in the crop sector are able to diversify crop production. The implication is that when rural households have access to mobile money services, they produce fewer number of different crops while exploring non-farm activities and livestock production as additional livelihood options. The results underscore the need to address constraints that limit rural households’access to digital financial tools, especially mobile money, in Ghana and in similar developing country contexts. IMPACT STATEMENT This study employs a nationally representative micro-data from rural Ghana to investigate the effect of digital finance on livelihood diversification using several econometric methods including instrumental variables to address potential endogeneity bias. The study represents a significant attempt at addressing an important gap in the existing literature which is about whether digital finance brings any value addition to livelihood diversification efforts by rural households. The key finding is that when rural agricultural households have access to digital financial services through mobile money, they produce fewer different crops, while exploring non-farm activities and livestock production as additional livelihood options. The finding implies that access to digital finance may influence rural households’ability to reallocate productive resources to areas that yield higher returns to their livelihoods and/or mitigate the risk of income loss. Expanding upon this finding with an exploration of the key correlates of diversification and access to digital finance, the study shows that addressing constraints to access to digital finance or mobile money (particularly, the limited mobile phone connectivity, electricity access or supply challenges, and low education or digital illiteracy in rural areas) is important for deepening livelihood diversification in rural Ghana. ARTICLE HISTORY Received 7 November 2023 Revised 6 February 2024 Accepted 10 March 2024 KEYWORDS Digital finance; mobile money; livelihood diversification; crops; livestock; non-farm enterprises REVIEWING EDITOR Goodness Aye, University of Agriculture, Makurdi Benue State, NIGERIA SUBJECTS Economics and Development; Economics; Development Studies; Finance; Rural Development JEL CLASSIFICATIONS G50; I31; Q00; O39 1. Introduction The advent of digital finance (DF), such as mobile money, has significantly changed the financial services space in sub-Saharan Africa (SSA), especially for the poor and the historically unbanked, allowing CONTACT Richmond Atta-Ankomah [email protected] Institute of Statistical, Social and Economic Research, University of Ghana, Legon, Ghana Supplemental data for this article can be accessed online at https://doi.org/10.1080/23322039.2024.2330434. ß2024 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. COGENT ECONOMICS & FINANCE 2024, VOL. 12, NO. 1, 2330434 https://doi.org/10.1080/23322039.2024.2330434 countries to use innovative technologies to leapfrog the financial development stages that developed countries went through (Aron, 2018; Bara, 2013). Mobile money has become an important innovative tool for facilitating financial inclusion and enhancing financial market participation (Ahmad et al., 2020; Amoah et al., 2020; Bongomin et al., 2018;N’dri & Kakinaka, 2020; Ouma et al., 2017) especially in SSA. According to the 2021 Global Findex report, SSA is leading the mobile money revolution, with the share of adults owning mobile money accounts increasing from 12% in 2014 to 33% in 2021, compared with the global share of 2% and 10% in 2014 and 2021, respectively (Asli et al., 2022). In addition, differences in financial account ownership in SSA across wealth quintiles have been halved over the past decade, driven mainly by mobile money (Asli et al., 2022). Aside from facilitating financial inclusion, there is growing empirical evidence linking digital finance to various livelihood and welfare outcomes (Atta-Ankomah, 2022; Lee et al., 2021; Ouma et al., 2017; Peprah et al., 2020). For instance, in Bangladesh, Lee et al. (2021) found that mobile money increases household consumption, reduces extreme poverty and borrowing, and shore-up savings. Similarly, Peprah et al. (2020) found a positive relationship between mobile money adoption and household outcomes, including farm output, consumption expenditure, and asset value, using data from three districts in Ghana. Ouma et al. (2017) observed that access to digital finance enhances savings behavior and the extent of saving among households in Kenya, Malawi, Uganda, and Zambia. Recent studies using a nationally representative data on Ghana also observed a positive correlation between mobile money account ownership and access to credit, health insurance and remittances (Atta-Ankomah, 2022; AttaAnkomah et al., 2024). In addition, recent advances in digital financial technology have allowed poor households with mobile money accounts to receive international remittances (Ahmad et al., 2020; Amoah et al., 2020). In a more recent study, Okyere et al. (2024a) show that adoption of digital financial services has positive effect on the resilience capacity of farm households in Northern Ghana. Other existing evidence also suggests a positive effect of digital finance on household investments and production activities. Specifically, for agriculture, which is the main means of livelihood of the rural poor (Davis et al., 2017), Miller (2019) documents four channels through which digital finance could impact the sector’s development: (1) savings to help smoothen consumption over time; (2) insurance against possible crop loss; (3) payment and transactions, including cash transfers to rural households; and (4) credit to farmers. A recent study by Okyere et al. (2024b) found that mobile money adoption increases farm performance (including productivity) and welfare of households in Northern Ghana. A. Islam et al. (2018) found that mobile money increased smalland medium-scale enterprise purchases of fixed assets in Uganda, Tanzania, and Kenya, with reduced transaction costs, increased liquidity, and creditworthiness being the main enablers. Ahmad et al. (2020) and Aron (2018) provide more detailed reviews of the extant literature on the micro-economic implications of digital finance. Digital finance has the potential to revolutionize agri-food systems and livelihood diversification of rural households (Finger, 2023), but studies in this area are lacking. In particular, the role of digital finance (mobile money) in influencing rural households’livelihood diversification, and the extent or form of diversification has remained largely unexplored in the literature especially those on Ghana, although many empirical studies have established a linkage between mobile money and household welfare (Amoah et al., 2020; Atta-Ankomah & Okyere, 2022; Jack & Suri, 2014; Lee et al., 2021; Mahama & Maharjan, 2017;N’dri & Kakinaka, 2020; Peprah et al., 2020), as noted above. Also, this evidence gap exists in spite of a growing literature on the drivers of rural livelihood diversifications (Dagunga et al., 2022; Loison, 2015,2019; Martin & Lorenzen, 2016) and the importance of diversification to household economic wellbeing including resilience to various shocks (see Akaakohol & Aye, 2014; Asfaw et al., 2019; Asmah, 2011; Dagunga et al., 2020; Mohammed et al., 2021). On the drivers of livelihood diversification, Dagunga et al. (2022), for example, found a positive linkage between savings and youth income diversification in five cocoa growing regions of Ghana. Also, access to finance has been highlighted as an important driver of diversification in several other studies (see Batool & Jamil, 2019; Khatun & Roy, 2012; Loison, 2015,2019). Using panel data, Loison (2019) delved into the regional and gender differences as drivers of livelihood diversification in two agricultural regions of rural Kenya. While these studies are insightful, none explored the link between digital finance and rural livelihood diversification. Against this background, this study examines the linkage between digital finance (mobile money) and rural household livelihood diversification by focusing on three important issues in Ghana. First, we 2 R. ATTA-ANKOMAH ET AL. examine the drivers of mobile money adoption among rural households, including the gender dimensions of these determinants. Second, we examine the effect of mobile money on different livelihood diversification options (crop production, livestock production, and non-farm business operations) while exploring whether access to digital finance is associated with a de-emphasis on more traditional (agricultural) production and emphasis on less traditional alternatives in the non-farm sector. Third, we disaggregate the sample by gender of the household head to explore whether mobile money adoption has a gendered-differential impact on livelihood diversification. This study is vital from both policy and literature perspectives. As noted above, while studies on the micro impact of mobile money have expanded recently, none of the existing studies explore the linkage between mobile money and rural livelihood diversification in Ghana. Thus, the current study will contribute uniquely to the growing literature on digital finance and rural livelihood diversification in Africa. This study’s policy relevance emanates from the gender dimension it explores, the different livelihood options considered, and the fact that rural households in Ghana engage in a portfolio of livelihood activities (Davis et al., 2017). In addition, the government of Ghana is implementing a number of livelihood improvement programs and policies to strengthen mobile money service delivery in the country, making the current study relevant in that regard. The policy implications of this study are relevant to other SSA countries with a context similar to that of Ghana. This study uses nationally representative micro-data from rural Ghana and employs econometric methods (including instrumental variables) to control for the potential endogeneity between mobile money and the livelihood diversification variables. Specifically, we used the penetration of mobile phones at the household level as an instrument. We find that when rural agricultural households have access to digital financial services through mobile money, they produce fewer different crops, while exploring non-farm activities and livestock production as additional livelihood options. In other words, they can diversify into less traditional sources of livelihood. The implication is that access to digital finance may influence rural households’ability to reallocate productive resources to areas that yield higher returns to their livelihoods and/or mitigate the risk of income loss. These effects pertain to both male-headed and female-headed households, except in the case of the de-emphasis on crop production, which mainly pertains to male-headed households. These findings, together with the key results on other important correlates of diversification and access to digital finance, suggest that constraints to access to digital finance or mobile money, such as limited mobile phone connectivity, access to electricity, and low education or digital illiteracy in rural areas, need policy attention. The rest of the paper is organized as follows. Section 2 provides a brief contextual background of mobile money and livelihood diversification in Ghana. Section 3 presents the data and empirical strategy used for the study, and Section 4 discusses the results. Finally, in Section 5, we present the conclusions and policy implications of the study. 2. Mobile money and livelihood outcomes in Ghana Ghana is an interesting country for this study for several reasons. First, mobile phone subscriptions have increased from 13 in 2005 to approximately 123 per 100 people by 2021 (World Bank Microdata Catalog, 2023). Similarly, the number of individuals with mobile money accounts soared from about 350,000 in 2012, three years after the launch of mobile money in Ghana to approximately 17.1 million in 2020 (Dabalen & Mensah, 2023; Mattern, 2018). In fact, Ghana is one of the countries in SSA that has seen a surge in mobile money usage in recent years, and the share of adults aged 15þyears with mobile money accounts increases by 47 percentage points between 2014 and 2021 (13% to 60%), while the share of accounts at financial institutions increased from 35% to 39% over the same period (Asli et al., 2022; Mattern, 2018). In addition, the joint ownership of bank accounts and mobile money accounts has seen a substantial increase from 23% in 2014 to about 30% in 2021, driven mainly by mobile money (Asli et al., 2022). This surge in mobile money usage in the country could be explained by the regulatory policies of the Bank of Ghana and the financial sector interoperability system introduced jointly by the government of Ghana, central and commercial banks, and the Ghana Chamber of Telecommunications (Oxford Business Group, 2018). In addition, the government of Ghana, in collaboration with other development partners, COGENT ECONOMICS & FINANCE 3 has introduced a number of policies, including the National Financial Inclusion and Development Strategy, the Digital Financial Services Policy, and the Cash-Lite Roadmap, in an effort to improve mobile money services delivery and expand the geographic reach (Ministry of Finance, n.d.). In 2022, the government of Ghana, through an act of parliament, introduced an electronic levy (ELevy) to generate revenue, adopting models that appear to have worked in other countries in the region. This policy was not popular among Ghanaians (Amoah et al., 2023), probably because of its less pro-poor design. Currently, there are three main telecommunication companies in Ghana (MTN Ghana, Airtel/Tigo, and Vodafone Ghana) providing mobile money services, with MTN Ghana having the largest share. Agriculture tends to be the major source of income, and hence, the livelihood for the rural poor in SSA and Ghana in particular. Davis et al. (2017) show that about 44% of households in Ghana have farming as their main source of income. In addition, diversification into non-farm income-generating activities is common among the rural poor in the country. More than 23.5% of rural households in Ghana have diverse income portfolios (Davis et al., 2017). Despite serving as the main means of livelihood for the rural poor and contributing substantially to the country’s GDP, agriculture in Ghana is dominated by smallholder farmers, and is beset with a myriad of challenges, including low adoption of sustainable agricultural practices, low level of commercialization, and climate change (Asravor et al., 2022). Resorting to and diversifying into alternative sources of livelihood, such as non-farm businesses and livestock production, is very common in rural Ghana (Mahama & Maharjan, 2017; Mahama & Nkegbe, 2021). 3. Methodology 3.1. Data This study uses secondary micro data from the seventh and most recent round of the Ghana Living Standards Surveys (GLSS 7), a nationally representative survey conducted by the Ghana Statistical Service (GSS) in 2016/2017. 1 The survey was designed to cover 15,000 households, selected through a two-stage probability sampling approach. In the first stage, 1000 enumeration areas (EAs) were randomly selected from all regions of Ghana based on the probability proportional to size (PPS) of each region. Following the household listing of the selected EAs, the second stage involved random selection of 15 households from each EA. Of the 15,000 selected households, 14,009 were successfully interviewed. Of the households interviewed, 7991 (which accounts for about 57%) were located in rural areas. This study focuses on the rural subsample to understand the potential effect of access to digital finance on different forms of livelihood diversification in the rural economy. GLSS 7 is a multi-topic household survey that provides comprehensive data on household socioeconomic characteristics and engagement in many forms of economic activities. We rely on information on households’ownership of mobile money accounts, which are the dominant means of digital financial services available to rural households in Ghana, as well as information on their involvement in crop production, livestock production, and non-farm business activities. 3.2. Regression model, measurement of variables and estimation techniques To investigate the relationship between access to digital finance and livelihood diversification, we first explored the key correlates of each of these two variables in separate regression models. However, in the model for access to digital finance, measures of livelihood diversification are included as independent variables; similarly, access to digital finance enters the models for livelihood diversification as an independent variable. We specify two main models as follows: Momoi¼dDiviþXbþvi(1) Divi¼aMomoiþXhþei(2) Equation (1) is the model for access to digital finance, which is measured by the ownership of mobile money accounts by the household, while Equation (2) models the correlates of livelihood diversification. Xin both equations represents a vector of control variables or other correlates of access to digital 4 R. ATTA-ANKOMAH ET AL. finance and livelihood diversification. Momo is a dichotomous measure that takes a value of 1 if at least one member of the household has a mobile money account, but zero otherwise. Div captures household livelihood diversification. Section 3.2.1 explains how livelihood diversification and other variables are measured. 3.2.1. Measures of livelihood diversification and other variables The measures of diversification used in the study capture diversification within a given livelihood source in the rural economy, of which three sources are considered –crop production, livestock production and non-farm enterprises. The diversification within crop production, which we refer to as crop diversification, is measured in two ways. First, we measure crop diversification as the total number of different types of crops cultivated by households, which means that for rural households that are not engaged in crop production, the score for crop diversification will be zero. Second, we used the entropy measure of diversification to create an index of crop diversification, which provides a composite measure of diversification across and within six different groups of crops (i.e. cash crops, cereals, fruits, legumes, roots, and vegetables). The entropy measure of diversification has mainly been applied to measure the distribution of firms’activities across different products or market segments (see Palepu, 1985; Raghunathan, 1995; Theil, 1967). We follow this approach to generate a diversification index for households’engagement in different types of crop production. The index is formally derived as follows: CDi¼X n i¼1 ½Piln 1 Pi  =ln nðÞ (3) Where CD i is the diversification score for household i. P i is the share of a particular crop group, say legumes, in the total number of crops grown by the household. In other words, the number of different legume crops grown by a household is expressed as a proportion of the total number of crops grown by the household. nis the number of crop groups, and there are six crop groups considered in this study, as mentioned earlier. According to Raghunathan (1995), the numerator is referred to as the entropy value, whereas the denominator provides the maximum entropy value. The values of CD range between zero and one, where zero corresponds to no diversification (that is, the household grows only a single crop from only one of the six groups), and one means that the distribution is even across the six different groups of crops (that is, the number of crops grown within each of the six groups are all equal). By definition, the entropy score exists only for households that grow at least one crop; hence, CD actually measures the extent of crop diversification among households that are into crop production. For diversification within livestock production, we use the total number of different types of livestock raised by the household as a measure. Additionally, we examine livestock diversification using a measure that explores the extent of diversification among households that reported raising livestock. However, we used principal component analysis (PCA) to generate this measure instead of the entropy method for two reasons: (1) the number of different types of livestock that the households were asked to report on was much smaller (24 different livestock) compared to the 42 different crops; (2) for 10 of these different livestock, the data showed that the number of households raising each of the 10 constituted less than a percentage of all rural households, leaving only 14 for classification into subgroups. As a result, we performed PCA on the remaining 14 different livestock types to generate a score (based on the first principal component) that measures the extent of livestock diversification among households that raise livestock. Each of the 14 livestock types entered our PCA as a binary variable, where one means the household raised a particular livestock, and zero if otherwise. With regards to diversification involving non-farm activities, we used only a dichotomous measure which takes a value of 1 if the household owns at least a non-farm enterprise and zero if otherwise. We rely only this dichotomous measure because while the survey asked households to report on up to four non-farm enterprises, the data showed that only six percent of the households had two or more nonfarm enterprises. Those with no non-farm enterprises and those with only one non-farm enterprise constituted 67% and 27% of the sample, respectively. COGENT ECONOMICS & FINANCE 5 Table 1 provides further details regarding the variable labels and descriptions of livelihood diversification variables. Similarly, Table 1 provides information on how all the other variables including the control variables are defined as well as with their summary statistics. 3.2.2. Estimation techniques and strategy for addressing potential endogeneity Table 2 provides the details of the specific estimation techniques applied to Equations (1) and (2). Each estimation technique was selected based on its suitability for the nature of the dependent variable being modeled. For dichotomous variables, logit regression method was applied while Poison regression method was applied to count dependent variables and the Ordinary Least Squares method was applied to models with continuous dependent variables. Thus, the models for access to digital finance and Table 1. Descriptive statistics on variables used. Description Mean SD Min Max momo equals 1 if at least a household (HH) member has a mobile money account, otherwise zero 0.17 0.38 0 1 crop_total is the number of different types of crops HH cultivated (there were total of 42 different types of crops the HH was required to report on). 2.51 2.28 0 13 total_diverse is an entropy measure of crop diversification across 6 different groups of crops - it captures the extent of diversification across the individual crop types as well as the 6 different groups of crops 0.38 0.23 0 0.92 n_livestock is the number of different types of livestock raised by the HH (there were total of 24 different types of livestock the HH was required to report on). 1.16 1.78 0 11 livestock_div is a PCA score based on the first principal component using ownership of 14 different types of livestock. −0.41 1.37 −2.26 6.73 own_enterp2 takes a value of 1 if the HH owns a non-farm enterprise; otherwise zero 0.33 0.47 0 1 female equals 1 if the gender of the HH head is female, otherwise zero. 0.28 0.45 0 1 age is the HH head’s age in completed years 47.41 16.34 15 99 age_squared is the squared of age 2515 1707 225 9801 married_con equals 1 if HH head is in marital or consensual relationship, otherwise zero. 0.69 0.46 0 1 Religion of HH head at three levels, measured as dummies: no_religion 0.07 0.25 0 1 christians 0.63 0.48 0 1 other_reli 0.30 0.46 0 1 Education of HH head at four levels, measured as dummies: no_education (HH head has no education) 0.40 0.49 0 1 primary (HH head completed primary education) 0.17 0.38 0 1 JHS_JSS_Middle (HH headed completed junior high school or equivalent) 0.30 0.46 0 1 sec_voc_plus (HH head completed at least senior high school, vocational school or equivalent) 0.08 0.26 0 1 hhsize is the number of persons who make up the HH. 4.69 3.11 1 28 any_shock equals 1 if at least a HH member above five years was ill or injured in the past 12 months, otherwise zero 0.44 0.5 0 1 el_ngrid equals 1 if HH community is connected to national power grid, otherwise zero 0.81 0.39 0 1 fin_inst equals 1 for availability of formal financial institution in HH community, otherwise zero. 0.29 0.46 0 1 pri_com equals 1 for a community with primary school, otherwise zero 0.87 0.34 0 1 JHS_com equals 1 for a community with Junior High School, otherwise zero 0.73 0.44 0 1 SHS_com equals 1 for a community with Senior High School, otherwise zero 0.30 0.46 0 1 Ecological zone where HH is located at three levels, measured as dummies: rural_coastal 0.14 0.35 0 1 rural_forest 0.38 0.49 0 1 rural_savanna 0.47 0.5 0 1 m_penetration is mobile phone penetration rate at the community/EA 0.49 0.18 0.02 0.94 prop_mm is HH members owning phone as a proportion of all HH members 0.39 0.33 0 1 lnfarmsize is natural log of total farm size owned or operated by HH 1.25 1.01 0 9.21 lntotal_cost is natural log of the total agricultural input cost of the HH 4.11 2.93 0 9.05 ln_welfare is natural log of annual real per capita adult equivalent consumption expenditure 7.66 0.88 3.66 11.05 Dummies for 10 regions of Ghana: 2 western 0.10 0.30 0 1 central 0.09 0.28 0 1 greater_accra 0.02 0.13 0 1 volta 0.11 0.32 0 1 eastern 0.10 0.30 0 1 ashanti 0.09 0.28 0 1 brong_ahafo 0.09 0.29 0 1 northern 0.12 0.33 0 1 upper_east 0.14 0.34 0 1 upper_west 0.15 0.35 0 1 Note: ‘SD’stands for standard deviation, and ‘min’and ‘max’respectively, stand for minimum and maximum. 6 R. ATTA-ANKOMAH ET AL. ownership of non-farm enterprises were estimated using logit regressions, whereas those for the count measures of diversification (ie Crop_total and n_livestock) were estimated using Poisson regressions and those for continuous measures of diversification (ie Total_diverse and livestock_div) were estimated using OLS. By definition, Equations (1) and (2) indicate a potential simultaneity problem, which could be a major source of endogeneity bias, between each of the diversification variables and the mobile money variable. Without addressing this problem, Equations (1) and (2) only help us learn about the association between access to digital finance and our measures of livelihood diversification. We therefore address this potential endogeneity bias using instrumental variable models. Given our interest in the effect of access to digital finance (mobile money) on livelihood diversification, we sought for an instrument for mobile money, and not for livelihood diversification. Following recent studies (such as Aker & Mbiti, 2010; Atta-Ankomah, 2024; Bukari & Koomson, 2020), we considered two candidates as potential instruments for mobile money: the proportion of households in a community/enumeration area that owns mobile phones (labelled in Table 1 as m_penetration) and the proportion of household members that own mobile phone (labelled in Table 1 as prop_mm). While these two variables are measures of mobile phone penetration, m_penetration provides a measure at the level of the community/enumeration area level, whereas prop_mm measures mobile phone penetration at the household level. In our preliminary analysis, both variables passed the under-identification test; however, only prop_mm passed the weak identification test. Hence, the results reported in this study used prop_mm as the instrument for mobile money. Our instrumental variable models were estimated using the limited-information maximum likelihood (LIML) estimator. Our choice of LIML was informed by the argument that the LIML estimator tends to produce less biased estimates, is more robust to weak instruments, and produces confidence intervals with better coverage rates than two-stage least squares (Poi, 2006; Stock et al., 2002; Stock & Yogo, 2005). In addition, the LIML estimator performs better than both two-stage least squares and generalized methods of moments in finite samples but is equivalent to a two-stage least squares method in large samples (Cameron & Trivedi, 2010). 4. Results and discussion This section presents and discusses the findings of the study. We first present brief descriptive results on the relationship between our diversification variables and access to digital finance, after which we present the results of the econometric analysis. Under the econometric results, the correlates of access to digital finance in rural Ghana are first presented and discussed, followed by the correlates of livelihood diversification. The section then discusses the results of the effect of access to digital finance on livelihood diversification in models that account for potential endogeneity. 4.1. Descriptive analysis The average proportion of rural households owning mobile phones across the various communities/enumeration areas was 49%, while only 17% of the households owned at least one mobile money account (see Table 1). This suggests that there were many households in rural Ghana with mobile phones that did not own or operate a mobile money account as of 2016/2017, the survey year. The chart in the bottom right of Figure 1 shows that the mobile money account ownership rate differs slightly between Table 2. Estimation strategy. Dependent variable Variable name Nature of dependent variable Relevant equation Estimation model Access to digital finance Momo Dichotomous Equation (1) Logit Crop diversification Crop_total Count Equation (2) Poisson Total_diverse Continuous Equation (2) OLS Livestock diversification n_livestock Count Equation (2) Poison livestock_div Continuous Equation (2) OLS Non-farm enterprise Own_enter2 Dichotomous Equation (1) Logit Note: Instrumental variable (IV) estimations are also performed for all dependent variables (see Section 3.2.1 for details). COGENT ECONOMICS & FINANCE 7 endogeneity. Thus, we conclude that access to digital finance is endogenous in models (1), (2), and (3). This also implies that instrumental variable models are more appropriate for determining the effect of digital finance on crop diversification and livestock diversification. In the case of the model for ownership of an NFE and the extent of livestock diversification (using the PCA score), we do not reject the null hypothesis of exogeneity. The exogeneity of digital finance in non-farm enterprises and the intensity of livestock diversification models imply that the results in Columns (4) and (5) of Table 4 hold. The tests for under-identification, with the null hypothesis that the model is under-identified, yield a pvalue of less than .01 for all the models, and hence, we reject the null hypothesis and conclude that the models are not under-identified. In other words, there is a high correlation between the instrument and endogenous variable, confirming that our instrument is relevant. Similarly, the weak identification test, with the null hypothesis that the model is weakly identified, is rejected for all the models because the F-statistic for each of the models is far greater than 10, the conventional threshold for determining weak identification. The IV LIML results in Table 5 show that access to digital finance tends to increase the number of different types of livestock reared. The results in Table 5 Column 4 indicates that the effect disappears with respect to the extent of livestock diversification measured using PCA scores, which restricts the sample to households that own at least one type of livestock. This means that the significant effect in the model for the number of different livestock (see Column 3 of Table 5) may be largely driven by the variation between households with no livestock and those with any number of livestock. In contrast to the results on diversification toward livestock, we find that access to digital finance reduces diversification between crops and within crop groups. In other words, households with access to digital finance are less likely to have diversified crop production. Thus, access to digital financial services is associated with diversification into non-crop production avenues and this may be driven by allocative efficiency that may be linked to access to digital financial services, including access to credit, spreading risk, or both. In particular, it is reasonable to believe that rural agricultural households specialize in the production of particular crops if, through access to digital finance, they are able to engage in NFEs that generate more income than when they grow only crops. Indeed, evidence suggests that financial inclusion, especially access to agricultural credit, leads farmers to intensify the cultivation of specific crops and lower diversification within crop production (Gebreselassie, 2006; Loison, 2019; Xie et al., 2019). 4.2.4. Digital finance and NFEs - sub-sample results for only rural agricultural households To reinforce our findings on the effect of digital finance on diversification through NFEs, we restrict the sample to rural agricultural households only and re-examine this relationship. The results are presented in Table 6. 4 These results serve as further robustness check. Column (1) is for households that are in at least one type of agricultural activity (ie crop production, or livestock production), while column (2) is for households that are in both types of agricultural activities. The findings show that having access to digital finance has a positive and significant effect on diversifying livelihood sources among rural Table 6. Logit regression results on effect of digital finance on ownership of NFEs (sub-sample results for only rural agricultural households). Independent variables (1) (2) Household is into either crop or livestock production Household is into both crop and livestock production momo 0.280 0.411 (0.094) (0.128) Constant −6.048 −5.730 (0.653) (0.826) Control variables and regional fixed effects Yes Yes Observations 5944 2841 Pseudo R-squared 0.102 0.0987 Number of clusters 549 432 Chi square statistics 474.9 248.0 pValue for chi square 0 0 Note: (1) Robust (clustered) standard errors are in parentheses; (2) p<.01, p<.05, p<.1; and (3) endogeneity test results from IV LIML regressions show that mobile money was not endogenous –see Table A4 in the Online Appendix. 14 R. ATTA-ANKOMAH ET AL. agricultural households (that is, when non-agricultural households are excluded from the analysis), and the effect is stronger for households that engage in both types of agricultural activities. It is possible that those households that are involved in both crop and livestock production have garnered the experience of diversifying even though this diversification was limited to the agricultural sector. Therefore, they find it easier to diversify into NFEs when they have access to digital finance and opportunities for inclusion in the financial system. These results show that our overall finding of the strong effect of digital finance in promoting livelihood diversification among rural households is consistent for all rural households and rural agricultural households. 4.2.5. Gender heterogeneity in the effect of digital finance on livelihood diversification Table 7 summarizes the heterogeneity in the effect of digital finance on livelihood diversification by the gender of the household head. 5 Where the endogeneity test yields an endogenous result for access to digital finance, the IV results are interpreted, whereas the OLS/logit/Poison results are interpreted for cases where the endogeneity test showed ‘not endogenous’results. The results show a significant effect on ownership of NFE for both maleand female-headed households, although the effect appears stronger among female-headed households, which are consistent with the findings of Loison (2019) and Lay et al. (2008). While the results demonstrate no significant effect of digital finance on crop diversification among female-headed households, we observe a negative relationship among male-headed households, implying that the negative effect of digital finance on crop diversification observed in the previous section is mainly driven by male-headed households. Regarding livestock production, the results show significant diversification in the number of different livestock raised among male-headed households while digital finance has no significant effect on livestock diversification among households headed by females. An implication of these findings is that increased access to digital finance among females in rural Ghana will lead to greater degree of livelihood diversification away from agriculture. Previous research has shown that FHHs in rural areas tend to be more vulnerable in terms of access to resources than other households (Adjei-Mantey et al., 2022). Thus, support to such households through improved access to mobile money and other digital financial services can help expand opportunities for economic participation by females, and lead to increased female economic empowerment and improvement in living standards. This is in line with a recent study by M. S. Islam et al. (2022) that found that participation in NFEs has significant and positive effects on rural women’s income. NFE as an important avenue for women’s economic participation is further underpinned by existing structural challenges (such as the patriarchal nature of land ownership) in Ghana that prevent women from earning maximum returns from farm work. 5. Conclusion The emergence of digital financial systems, especially mobile money, has significantly changed the financial services space in SSA in ways that provide opportunities to promote financial inclusion among the poor, especially those in rural areas. Focusing on mobile money, which is the main avenue for digital Table 7. Heterogeneous effects of momo on livelihood diversification by gender of household head. Type of variable Gender of household head IV results OLS/logit/ PoisonEndogeneity test result coefficient crop_total Male Endogenous −3.218 −0.027 Female Not endogenous −0.789 −0.000 total_diverse Male Endogenous −0.206 −0.006 Female Not endogenous −0.010 −0.006 n_livestock Male Endogenous 1.292 −0.011 Female Not endogenous 0.688−0.011 livestock_div Male Not endogenous 0.648 0.084 Female Not endogenous 0.030 0.136 own_enterp2 Male Not endogenous 0.034 0.296 Female Not endogenous 0.364 0.465 Note. (1) Robust (clustered) standard errors are in parentheses; (2) p<.01, p<.05, p<0.1; and (3) all models passed the identification tests (refer to Table A3 in the Online Appendix for detailed results). COGENT ECONOMICS & FINANCE 15 financial services for rural households in Ghana, this study investigates whether digital finance influences rural households’livelihood diversification, using nationally representative data from Ghana. This study also delves into the effect on different forms of livelihood diversification among rural households. Using several econometric methods, including instrumental variable techniques, the study found mobile money to be positively associated with households’choice of non-farm enterprises as a livelihood option, as well as engagement in livestock production. However, we find a negative relationship between mobile money and diversification in crop production, as well as the extent of crop diversification by households within the crop sector. In other words, when agricultural households have access to mobile money services, they tend to produce fewer number of different crops while exploring non-farm activities and livestock production as additional livelihood options. Generally, the results show that when rural households become more included in the financial system through mobile money, they tend to diversify into less traditional sources of livelihood. This effect pertains to both male-headed and female-headed households, except in the case of diversification in crop production, where we found the negative effect to be mainly driven by male-headed households. In addition to mobile money, other important correlates of diversification worth highlighting are the education of the household head, the experience of a health shock by the household, expenditure on agricultural inputs, and the geographic location of the household. Education was also found to be an important correlate of access to digital finance in addition to other factors such as agricultural input cost, access to the national electricity grid, and the proportion of household/community members owning a mobile phone. These findings have several important policy implications. The diversification away from crop production and the diversification into non-farm business activities and livestock production suggests that rural households may be reallocating productive resources into areas that yield higher returns to their livelihoods and/or mitigate the risk of income losses, and that access to digital finance facilitates this process. While this shift may enhance rural livelihoods, it may also free land resources for more commercialized agricultural production beyond subsistence agriculture practiced by many rural households in developing countries such as Ghana. This means that the Government of Ghana should implement policies to support the adoption and use of digital financial services by rural households as part of broader strategies to revolutionize agriculture and the rural economy rather imposing an E-levy on mobile money transactions. Such supportive policies could also lead to increases in women’s economic participation and empowerment by facilitating their engagement in NFE activities. The results also highlight the need to address constraints that limit rural households’access to digital financial tools, especially mobile money, for which ‘non-formal’education of the less educated on how to use these digital tools is crucial. To do this, we recommend a collaborative effort between government and major telcos operating mobile money services in Ghana. In addition, government should enhance investment in the provision of complementary infrastructure required for broadening access to mobile money services in rural communities, such as making electricity more accessible as well as incentivizing telcos to make mobile phones and networks more accessible in rural communities. While the data used in this study came from the latest GLSS (a nationally representative household survey in Ghana), it is important to acknowledge that the data was collected in 2016/2017, hence, may not capture recent developments within the mobile money space. Therefore, while this study offers significant insights into the socioeconomic impact of mobile money in the rural economy, it may be worthwhile to replicate the study when a new nationally representative data becomes available in order to learn about the impact of recent developments in the mobile money space. Also, the measures of diversification used in this study largely focused on diversification within broad livelihood activities areas and not across these areas, which is an area further studies can explore. Additionally, further studies may directly explore impact of mobile money on income diversification, which can encapsulate other sources of income such as remittances and assets. Notes 1. The authors obtained an anonymized version of the data from GSS and did not directly engage any of the participants of the survey. Hence, the research activities conducted by the authors of this study are very less likely to lead to a violation of any research ethics. 16 R. ATTA-ANKOMAH ET AL. 2. This study uses the ten regions that existed at the time of the GLSS 7 survey. The regional boundaries have since been re-demarcated to create additional six regions in the country. 3. The full regression results are displayed in Table A2 in the Online Appendix. 4. The full results are presented in Table A3 in the online appendix. 5. The full results are shown in the Online Appendix in Tables A4 and A5. Author contributions Richmond Atta-Ankomah: Conception, methodology, data management and analysis, writing draft and discussion; revising and editing the intellectual Kwame Adjei-Mantey: Literature review, writing draft and discussion, revising and editing the intellectual contents Akuffo Amankwah: Literature review, writing draft and discussion, revising and editing the intellectual contents Disclosure statement No potential conflict of interest was reported by the author(s). Funding This study received no funding. About the authors Richmond Atta-Ankomah is a development economist and Senior Research Fellow at the Institute of Statistical, Social and Economic Research, University of Ghana. His research interest is diverse but largely converges around development issues Concerning households and firms in sub Saharan Africa. Kwame Adjei-Mantey is researcher at the Department of Applied Economics, University of Environment and Sustainable Development, Somanya, Ghana. His research interests are in the areas of energy and environmental economics, development economics and behavioral economics. Akuffo Amankwah is an Economist for the Living Standards Measurement Study (LSMS), the World Bank’s flagship household survey program housed at the Development Data Group. His primary areas of research are poverty, labor, aquaculture, agriculture and rural development, and methodological studies to improve household surveys. ORCID Richmond Atta-Ankomah http://orcid.org/0000-0001-9982-9377 Data availability statement The main data underlying this study are available upon request from the Ghana Statistical Service (https://microdata.statsghana.gov.gh/index.php/catalog/97/study-description). References Adjei-Mantey, K., Kwakwa, P. A., & Adusah-Poku, F. (2022). Unraveling the effect of gender dimensions and wood fuel usage on household food security: Evidence from Ghana. Heliyon,8(11), e11268. https://doi.org/10.1016/j.heliyon.2022.e11268 Ahmad, H. A., Green, C., & Jiang, F. (2020). Mobile money, financial inclusion and development: A review with reference to African experience. Journal of Economic Surveys,34(4), 753–792. https://doi.org/10.1111/joes.12372 Ahmed, M. T., Bhandari, H., Gordoncillo, P. U., Quicoy, C. B., & Carnaje, G. P. (2018). Factors affecting extent of rural livelihood diversification in selected areas of Bangladesh. SAARC Journal of Agriculture,16(1), 7–21. https://doi.org/ 10.3329/sja.v16i1.37419 Akaakohol, M. A., & Aye, G. C. (2014). Diversification and farm household welfare in Makurdi, Benue State, Nigeria. Development Studies Research. An Open Access Journal,1(1), 168–175. https://doi.org/10.1080/21665095.2014. 919232 Aker, J. C., & Mbiti, I. M. (2010). Mobile phones and economic development in Africa. Journal of Economic Perspectives,24(3), 207–232. https://doi.org/10.1257/jep.24.3.207 COGENT ECONOMICS & FINANCE 17 Amoah, A., Korle, K., & Asiama, R. K. (2020). Mobile money as a financial inclusion instrument: What are the determinants? International Journal of Social Economics,47(10), 1283–1297. https://doi.org/10.1108/IJSE-05-2020-0271 Amoah, A., Kwablah, E., Amoah, B., & Adjei-Mantey, K. (2023). Willingness to pay for electronic transaction levy: Empirical evidence from Ghana. African Journal of Economic and Management Studies,14(4), 663–679. https://doi. org/10.1108/AJEMS-09-2022-0359 Aron, J. (2018). Mobile money and the economy: A review of the evidence. The World Bank Research Observer,33(2), 135–188. https://doi.org/10.1093/wbro/lky001 Asfaw, S., Scognamillo, A., Di Caprera, G., Sitko, N., & Ignaciuk, A. (2019). Heterogeneous impact of livelihood diversification on household welfare: Cross-country evidence from Sub-Saharan Africa. World Development,117, 278–295. https://doi.org/10.1016/j.worlddev.2019.01.017 Asli, D., Klapper, L., Singer, D., & Ansar, S. (2022). The Global Findex Database 2021: Financial inclusion,digital payments, and resilience in the age of COVID-19. World Bank. https://doi.org/10.1596/978-1-4648-1897-4 Asmah, E. E. (2011). Rural livelihood diversification and agricultural household welfare in Ghana. Journal of Development and Agricultural Economics,3(7), 325–334. Asravor, R. K., Boakye, A. N., & Essuman, J. (2022). Adoption and intensity of use of mobile money among smallholder farmers in rural Ghana. Information Development,38(2), 204–217. https://doi.org/10.1177/0266666921999089 Atta-Ankomah, R. (2022). Credit for households in Ghana: Has mobile money (momo) improved inclusive access? Scientific African,16, e01230. https://doi.org/10.1016/j.sciaf.2022.e01230 Atta-Ankomah, R. (2024). Mobile money and intra-household employment diversification in Ghana. African Journal of Science, Technology, Innovation and Development,1–17. https://doi.org/10.1080/20421338.2024.2317657 Atta-Ankomah, R., Asante-Addo, C., Okyere, C. Y., & Asante-Poku, N. A. (2024). Does digital financial technology improve access to finance among households experiencing idiosyncratic health shocks in Ghana? Unpublished Manuscript, Last modified, January 9, 2024. Atta-Ankomah, R., & Okyere, Y. C. (2022). Welfare effects of financial inclusion services in Ghana: A comparative analysis of mobile money and other financial services. Global Social Welfare,10(1), 83–92. https://doi.org/10.1007/ s40609-022-00234-x Baffour, P. T., Rahaman, W. A., & Mohammed, I. (2021). Impact of mobile money access on interna remittances, consumption expenditure and household welfare in Ghana. Journal of Economic and Administrative Sciences,37(3), 337–354. https://doi.org/10.1108/JEAS-04-2020-0045 Bara, A. (2013). Mobile money for financial inclusion: Policy and regulatory perspective in Zimbabwe. African Journal of Science, Technology, Innovation and Development,5(5), 345–354. https://doi.org/10.1080/20421338.2013.829287 Batool, S., & Jamil, F. (2019). Rural employment and income diversification in Pakistan. Pakistan Journal of Agricultural Sciences,56(02), 503–510. https://doi.org/10.21162/PAKJAS/19.7484 Bongomin, O. C. G., Ntayi, M. J., Munene, C. J., & Malinga, A. C. (2018). Mobile money and financial inclusion in SubSaharan Africa: The moderating role of social networks. Journal of African Business,19(3), 361–384. https://doi.org/ 10.1080/15228916.2017.1416214 Bukari, C., & Koomson, I. (2020). Adoption of mobile money for healthcare utilization and spending in rural Ghana. In S. A. Churchill (Ed.), Moving from the millennium to the sustainable development goals (pp. 37–60). Palgrave Macmillan. Cameron, A. C., & Trivedi, P. K. (2010). Microeconometrics using Stata (rev. ed.). Stata Press. Dabalen, A., & Mensah, J. T. (2023). Ten facts about digital technology adoption in Ghana. Africa Can End Poverty. Dagunga, G., Ayamga, M., & Danso-Abbeam, G. (2020). To what extent should farm households diversify? Implications on multidimensional poverty in Ghana. World Development Perspectives,20, 100264. https://doi.org/ 10.1016/j.wdp.2020.100264 Dagunga, G., Azumah, S. B., Zakaria, A., Boateng, A. N., Mensah, B. K., Boateng, S. E., Tsekpo, M. E., & Kankam, P. (2022). Savings and economic diversification among youth in Ghana: Implications for policy and practice. Cogent Economics & Finance,10(1), 2123887. https://doi.org/10.1080/23322039.2022.2123887 Davis, B., Di Giuseppe, S., & Zezza, A. (2017). Are African households (not) leaving agriculture? Patterns of households’income sources in rural Sub-Saharan Africa. Food Policy,67, 153–174. https://doi.org/10.1016/j.foodpol.2016. 09.018 Dirribsa, S., & Tassew, B. (2015). Analysis if livelihood diversification strategy of rural households: A case study of Ambo District, Oomiya Regional State, Ethiopia. International Journal of Current Research and Academic Review, 3(8), 406–426. Finger, R. (2023). Digital innovations for sustainable and resilient agricultural systems. European Review of Agricultural Economics,50(4), 1277–1309. https://doi.org/10.1093/erae/jbad021 Gebreselassie, S. (2006). Intensification of smallholder agriculture in Ethiopia: Options and scenarios (Future Agricultures Consortium Discussion Paper 007). Future Agricultures. Gebru, G. W., Ichoku, H. E., & Phil-Eze, P. O. (2018). Determinants of livelihood diversification strategies in Eastern Tigray Region of Ethiopia. Agriculture & Food Security,7(1), 62. https://doi.org/10.1186/s40066-018-0214-0 Grzybowski, L. (2015). The role of network effects and consumer heterogeneity in the adoption of mobile phones: Evidence from South Africa. Telecommunications Policy,39(11), 933–943. https://doi.org/10.1016/j.telpol. 2015.08.010 18 R. ATTA-ANKOMAH ET AL. Islam, A., Muzi, S., & Meza, J. L. R. (2018). Does mobile money use increase firms’investment? Evidence from enterprise surveys in Kenya, Uganda, and Tanzania. Small Business Economics,51(3), 687–708. https://doi.org/10.1007/ s11187-017-9951-x Islam, M. S., Islam, S., Fatema, K., & Khanum, R. (2022). Rural women participation in farm and off-farm activities and household income in Bangladesh. Heliyon,8(9), e10618. https://doi.org/10.1016/j.heliyon.2022.e10618 Jack, W., & Suri, T. (2014). Risk sharing and transactions costs: Evidence from Kenya’s mobile money revolution. American Economic Review,104(1), 183–223. https://doi.org/10.1257/aer.104.1.183 Kassegn, A., & Abdinasir, U. (2023). Determinants of rural households’livelihood diversification strategies: In the case of north Wollo zone, Amhara National Regional State, Ethiopia. Cogent Economics and Finance,11(1), 2185347. https://doi.org/10.1080/23322039.2023.2185347 Khatun, D., & Roy, B. C. (2012). Rural livelihood diversification in West Bengal: Determinants and constraints. Agricultural Economics Research Review,25(1), 115–124. Lay, J., Mahmoud, T. O., & M’Mukaria, G. M. (2008). Few opportunities, much desperation: The dichotomy of nonagricultural activities and inequality in western Kenya. World Development,36(12), 2713–2732. https://doi.org/10. 1016/j.worlddev.2007.12.003 Lee, J. N., Morduch, J., Ravindran, S., Shonchoy, A., & Zaman, H. (2021). Poverty and migration in the digital age: Experimental evidence on mobile banking in Bangladesh. American Economic Journal: Applied Economics,13(1), 38–71. https://doi.org/10.1257/app.20190067 Loison, S. A. (2015). Rural livelihood diversification in sub-Saharan Africa: A literature review. The Journal of Development Studies,51(9), 1125–1138. Loison, S. A. (2019). Household livelihood diversification and gender: Panel evidence from rural Kenya. Journal of Rural Studies,69, 156–172. Mahama, A. T., & Maharjan, K. L. (2017). Determinants of livelihood diversification in Ghana from the national livelihood strategies and spatial perspective. Journal of International Development and Cooperation,23(1), 75–90. Mahama, A. T., & Nkegbe, K. P. (2021). Impact of household livelihood diversification on welfare in Ghana. Scientific African,13, e00858. https://doi.org/10.1016/j.sciaf.2021.e00858 Mattern, M. (2018). How Ghana became one of Africa’s top mobile money markets. CGAP (Consultative Group to Assist the Poor) Blogpost. What You Need to Know. Martin, S. M., & Lorenzen, K. (2016). Livelihood diversification in rural Laos. World Development,83, 231–243. https:// doi.org/10.1016/j.worlddev.2016.01.018 Miller, H. (2019). What works for digital financial services in agricultural development? USAID Feed the Future Initiative. Ministry of Finance. (n.d). Toward a cash-lite Ghana –Building an inclusive digital payments ecosystem: Digital payments roadmap. https://mofep.gov.gh/sites/default/files/acts/Ghana_Cashlite_Roadmap.pdf Mohammed, K., Batung, E., Kansanga, M., Nyantakyi-Frimpong, H., & Luginaah, I. (2021). Livelihood diversification strategies and resilience to climate change in semi-arid northern Ghana. Climatic Change,164(3-4), 1–23. https:// doi.org/10.1007/s10584-021-03034-y Mothobi, O., & Grzybowski, L. (2017). Infrastructure deficiencies and adoption of mobile money in Sub-Saharan Africa. Information Economics and Policy,40,71–79. https://doi.org/10.1016/j.infoecopol.2017.05.003 Myeni, S., Makate, M., & Mahonye, N. (2020). Does mobile money promote financial inclusion in Eswatini? International Journal of Social Economics,47(6), 693–709. https://doi.org/10.1108/IJSE-05-2019-0310 N’dri, L. M., & Kakinaka, M. (2020). Financial inclusion, mobile money, and individual welfare: The case of Burkina Faso. Telecommunications Policy,44(3), 101926. https://doi.org/10.1016/j.telpol.2020.101926 Nguyen, D. L., Nguyen, T. T., & Grote, U. (2022). Shocks, household consumption, and livelihood diversification: A comparative evidence from panel data in rural Thailand and Vietnam. Economic Change and Restructuring,56(5), 3223–3255. https://doi.org/10.1007/s10644-022-09400-9 Okyere, C. Y., Atta-Ankomah, R., & Asante-Addo, C. (2024a). Does Digital Financial Inclusion Improve Food Security and Household Resilience? Evidence from Northern Ghana. Unpublished Manuscript, Last Modified, January 2, 2024. Okyere, C. Y., Atta-Ankomah, R., & Asante-Addo, C. (2024b). Mobile money adoption, farm performance and household welfare in Northern Ghana. Unpublished Manuscript, Last modified, January 10, 2024. Ouma, S. A., Odongo, T. M., & Were, M. (2017). Mobile financial services and financial inclusion: Is it a boon for savings mobilization? Review of Development Finance,7(1), 29–35. https://doi.org/10.1016/j.rdf.2017.01.001 Oxford Business Group. (2018). Ghana rolls out mobile money interoperability platform.https://oxfordbusinessgroup. com/articles-interviews/ghana-rolls-out-mobile-money-interoperability-platform. Palepu, K. (1985). Diversification strategy, profit performance and the entropy measure. Strategic Management Journal,6(3), 239–255. 10.1002/smj.4250060305 Peprah, J. A., Oteng, C., & Sebu, J. (2020). Mobile money, output and welfare among smallholder farmers in Ghana. SAGE Open,10(2), 215824402093111. https://doi.org/10.1177/2158244020931114 Poi, B. P. (2006). Jackknife instrumental variables estimation in Stata. The Stata Journal: Promoting Communications on Statistics and Stata,6(3), 364–376. https://doi.org/10.1177/1536867X0600600305 Raghunathan, S. P. (1995). A refinement of the entropy measure of firm diversification: Toward definitional and computational accuracy. Journal of Management,21(5), 989–1002. https://doi.org/10.1177/014920639502100510 COGENT ECONOMICS & FINANCE 19 Stock, J. H., Wright, J. H., & Yogo, M. (2002). A survey of weak instruments and weak identification in generalized method of moments. Journal of Business and Economic Statistics,20(4), 518–529. https://doi.org/10.1198/ 073500102288618658 Stock, J. H., & Yogo, M. (2005). Testing for weak instruments in linear IV regression. In D. W. K. Andrews (Ed.), Identification and inference for econometric models (pp. 80–108). Cambridge University Press. Theil, H. (1967). Economics and Information Theory. North-Holland. World Bank Microdata Catalog. (2023). Mobile cellular subscriptions (per 100 people) –Ghana. https://data.worldbank. org/indicator/IT.CEL.SETS.P2?locations=GH Xie, H., Huang, Y., Chen, Q., Zhang, Y., and Wu Q. (2019). Prospects for agricultural sustainable intensification: A review of research. Land 8(11), 157. https://doi.org/10.3390/land8110157 20 R. ATTA-ANKOMAH ET AL.