IFII Blog Series: Advancing inclusive digital financial services to empower women in Indonesia

Indonesia is a standout country in terms of gender-equitable financial inclusion. Reports from the Global Findex and SNKI Financial Inclusion Insights state that although there is an average gender gap in account ownership of 8 percentage points in developing economies, no such gap exists in Indonesia. At the same time, there is still a large scope for overall expansion, as over 44 percent of Indonesian adults remain unbanked. To ensure Indonesia continues to experience gender-equitable growth, efforts geared towards financial inclusion must take into consideration women’s needs, in particular among poor and digitally illiterate populations.

On June 9, 2020, the Government of Indonesia (GoI) launched the National Women’s Financial Inclusion Strategy (SNKI-P), aimed at promoting access to finance for Indonesian women in a way that accommodates women’s diverse needs, interests, and backgrounds. Among others, SNKI-P’s mission is to provide a comprehensive and gender-responsive program to improve access to formal financial services, while ensuring all Indonesian women are equipped with financial skills (including digital finance). How can insights from Indonesia’s past financial inclusion experiences inform policies to support this mission? In this blog post, we summarize some key lessons from J-PAL SEA research conducted as part of its Inclusive Financial Innovation Initiative.

Understanding determinants of women’s financial inclusion

Social protection is important for women’s access to financial services, and digital readiness is likely to become more important as digital finance grows. As Indonesia moves to expand financial inclusion, it is important to understand why people do or do not own accounts. We utilized data from the 2018 Financial Inclusion Insights survey (FII) with a machine learning algorithm called Random Forest to identify key predictors of account ownership.

We find that features related to demographics, socioeconomic status, and economic engagement rarely rank highly as predictors. Gender was also not one of the top predictors of account ownership. Instead, the top predictor of female account ownership is whether the woman receives government assistance: specifically, women who report receiving government assistance are more likely to have an account, all else equal. This suggests that the Government of Indonesia’s push to digitize assistance is a driving force in women’s financial inclusion.

We also find that owning a mobile or smartphone and the ability to perform phone-related tasks are important predictors of account ownership for both men and women. In the future, digital engagement and digital literacy are likely going to become even more important for financial inclusion.

Source: FII 2018 Survey

Indonesia has a gender gap in digital readiness, which if not addressed may lead to future gender gaps in financial inclusion. As digital financial services (DFS) continue to expand in Indonesia, digital engagement and literacy will be key for connecting women to new technologies like e-money. However, we see a significant gender gap in phone capability between men and women. While 45 percent of men are digitally ready—that is, own a smartphone and can use it to download apps and surf the web—only 38 percent of women are.

Source: FII 2018 Survey

Inclusive digital financial products must take account of women's life stages and needs

Use cases of DFS will be different for women who are running the household, leading their own business, or working for a firm. How included are these groups now, and prepared are they to take up digital tools?

The use rate of financial services is significantly lower among women not engaged in income-generating activities. Our analysis of FII data indicates that married women are more likely to report having a dominant role in financial decisions, as compared to married men. This is in line with previous research suggesting Indonesian women have authority over household finances, regardless of their contribution to the household’s income. Despite this, married women whose main occupation is taking care of the home or who only work temporarily/seasonally are significantly less likely to use financial services as compared to those who own a business or are in permanent employment. As such, there is still scope for expansion to ensure financial inclusion covers women who do not engage in income-generating activities.

Targeted digital financial education may be needed to ensure financial inclusion for all women. We observe a lower rate of digital literacy among business owners. Relative to business owners, women who are in full-time employment are 18 percentage points more likely to report the ability to perform basic tasks (i.e. place calls, navigate menu, send SMS). This gap is even larger at 38 percentage points for complex tasks (i.e. search internet, download apps, financial transactions). This discrepancy is likely to affect knowledge and use of DFS, especially with newer innovations such as card-based and app-based e-money products. One way to address this capability gap would be digital financial education and opportunities to build smartphone skills targeted to female entrepreneurs.

Source: 2018 FII survey

Efforts by the Government of Indonesia to drive financial inclusion for women, such as through government assistance, have resulted in significant achievements. As DFS continues to expand, digital engagement and literacy is only going to become more important. To ensure Indonesia experiences continued expansion in financial inclusion for women, significant gender gaps in phone capability must be addressed. DFS innovations which contribute to financial inclusion will also take into account women’s life stages and needs, including female entrepreneurs and those not engaged in income-generating activities.

This blog post is based on efforts funded by the Bill & Melinda Gates Foundation. The findings and conclusions within are those of the authors and do not necessarily reflect positions or policies of the Bill & Melinda Gates Foundation. The mission of IFII is to ensure that digital financial services can drive economic development while lifting up women, low-income groups, and marginalized communities.

For more information about IFII, contact [email protected].

Baca blog ini dalam Bahasa Indonesia.

Since the launch of branchless banking agents (agen Laku Pandai) in 2014 with a mission to accelerate financial inclusion among Indonesia’s largely unbanked population, this channel has been used to deliver basic banking services to clients in remote areas and to provide government-to-people (G2P) transfers. Branchless banking agents are deemed to be a more cost-effective way to deliver banking services to rural areas compared to providing an ATM or building a local branch. 

Despite the promising opportunities of this channel, there remain numerous challenges to promoting banking services to the unbanked population. To achieve Indonesia’s financial inclusion goals, it is imperative to gather evidence on how to improve the effectiveness of branchless banking agent networks and enhance agents' performance.

To contribute to the important work of digital financial services (DFS) stakeholders in Indonesia, the Inclusive Financial Innovation Initiative (IFII) hosted a learning collaborative webinar, “Banking the unbanked: The effects of agents’ financial incentives and transparency in increasing the take-up and usage of financial products.” This webinar featured a presentation from Grace Retnowati and Raunak Kapoor from MicroSave Consulting Indonesia (MSC) on Indonesia’s agent network landscape, and IFII invited researchers Gianmarco León-Ciliotta (Pompeu Fabra University) and Firman Witoelar (Australian National University) shared learnings from a randomized evaluation conducted with IFII invited researcher Erika Deserranno (Northwestern University), which tested the impact of changing the level and transparency of agents’ financial incentives on the take-up of branchless banking products. This webinar also provided a forum for DFS stakeholders to discuss ideas or pressing issues in advancing financial inclusion in Indonesia.

Despite significant progress, numerous challenges limit the impact of Indonesia’s agent network

Since first introduced in Indonesia, the number of agents has been increasing rapidly, surpassing the stagnant growth of ATM and bank branch office expansion. In 2018, there were more than one million branchless banking agents providing basic banking services such as account opening, cash-in-cash-out (CICO), money transfers, bill payments/top-ups, and G2P disbursement. On top of branchless banking agents, there were also more than 300,000 electronic money agents in 2018 serving CICO, money transfer, and bill payments/top-ups. In 2019, fintech companies (including e-commerce, ride-hailing, and other fintech) reported 6.2 million agents serving cash-in, money transfer, bill payments/top-ups, digital loans, and other digital financial services. 

Raunak argued these numbers are sufficient to serve Indonesia’s population. However, the unequal distribution of agents across the country means that a large number of clients in remote areas remain underserved. He outlined another fundamental problem: the economic viability of branchless banking agent networks. The lack of consumer demand for agent services, combined with insufficient service fees from the consumer, becomes a barrier to gain revenue to sustain the networks. In order to make the network work more effectively, Raunak proposed that we need to drive more people to use the services.

Incentives might be one way to increase agents’ performance, but we need to ensure that they lead to more take-up of financial services

In the second presentation, delivered by Firman Witoelar and Gianmarco León-Ciliotta, the researchers highlighted the urgency to produce evidence on how to improve agents' performance to attract more clients. They suggest that agents’ efforts to convert people to adopt the services might be hampered by potential clients’ lack of awareness and knowledge of branchless banking products. Low trust in the products, agents, and banks may be another constraint, making potential clients rely on observable cues to help them decide whether to adopt a new financial product.

The researchers tried to answer this critical policy question through conducting a randomized evaluation measuring the impact of providing much larger financial incentives to agents and adjusting the level of transparency around incentives information for the public on the take-up and usage of branchless banking products. 

The evaluation results showed that higher financial incentives could be used to boost branchless banking agents’ efforts in finding new clients more proactively, therefore increasing take-up of the financial products they are promoting. However, in a context with low levels of information and trust in financial institutions, we need to be careful about the transparency of agents' incentives.

When information about the incentives is public or known by potential clients, higher incentives (and the subsequent higher agent effort) did not translate to higher take-up or usage of financial products compared to the lower incentive amount. Researchers believe this lack of impact occurs because learning about agents’ high level of incentives to promote a product affects potential clients’ perceptions of the reliability and trust of the product, the agent, and the bank, therefore reducing their demand.

Potential ideas to improve agent network expansion as a tool for financial inclusion

Higher financial incentives for agents could be part of a strategy to increase financial penetration in rural areas; however, ensuring financial sustainability and improving clients’ trust are key. During the interactive discussion, the speakers highlighted several recommendations to improve service quality and maximize the agent network's impact. 

Speakers from MSC and J-PAL SEA’s IFII agreed that as providers face profitability challenges to manage agent networks, it might be beneficial to explore how to decrease providers’ operational costs of managing agents. MSC provided several alternatives that the Government of Indonesia could consider, such as allowing agents to work for various banks in rural areas, allowing a third-party agent network manager, evaluating the incentive and fee schemes based on agents’ appetite for operations, and ameliorating the G2P model through incentive schemes for providers to expand their service and deliver better quality.

Gianmarco specifically highlighted recommendations to increase clients’ trust towards branchless banking channels. Banks might introduce agent customer ratings, as seen on e-commerce or ride-hailing platforms, to provide more reliable and crowdsourced information to help clients to increase their trust in agents and the services they provide. 

It is also important to recruit agents who are known, trusted in the community, and able to promote branchless banking services' reliability. Banks also need to invest in financial products that are easier to understand and use. Grace added that financial products have to be developed to cater to rural and low-income clients' needs to ensure adoption. 

Another promising avenue, Gianmarco mentioned, is through increasing agent competition at the local level. He believes that this competition will improve agents’ efforts and improve the information and choices on the clients’ side, hence enhancing their trust in the channel.

DFS innovation might improve the efficiency and effectiveness of agent operations

Grace mentioned that technology could be used to improve customer onboarding processes. For instance, a better e-KYC (Know Your Customer) infrastructure can shorten and improve the process of new account opening, whereas fintech utilizes mobile-technology to reach users, especially in rural areas. Moreover, Gianmarco pointed out that while developing user-centered products is necessary, we need to build agents' capabilities to deliver the information about the products and build people’s abilities to use DFS.

Improvement in government regulation is also needed to maximize the impact of agent networks

In order to maximize agents’ roles in advancing financial inclusion, the speakers recommended the government improve regulations related to the agent network. Firman highlighted the importance of improving e-KYC regulation. Gianmarco added that the government needs to strengthen consumer protection regulations so that it may increase clients’ trust and improve community financial literacy by introducing relevant branchless banking products. 

Promoting well-designed financial products to serve rural and low-income clients may result in increased adoption and improved financial literacy in the community. MSC recommended the government open up regulations for fintech to recruit individual agents to enable them reach more clients, especially in rural areas. 

This article is based on the webinar “Banking the unbanked: The effects of agents’ financial incentives and transparency in increasing the take-up and usage of financial products” on February 3, 2021, featuring IFII invited researchers Firman Witoelar (ANU) and Gianmarco León-Ciliotta (UPF, Barcelona GSE, IPEG), and Grace Retnowati (Country Director, MSC Indonesia) and Raunak Kapoor (Senior Manager, Country Program Development, MSC Indonesia).

To learn more, read the content of the presentations and CICO Economics in Indonesia report by MicroSave Consulting Indonesia.

The COVID-19 crisis has created not only massive public health challenges but also an unprecedented economic shock for countries across the globe. Many people have lost their livelihoods due to the sudden slowdown of the economy and movement restrictions, setting back global poverty alleviation efforts. On top of that, emerging economies have more budgetary limitations to providing social assistance and are characterized by large informal sectors, making it more challenging to cushion the impacts of the crisis on the poor and the new vulnerable individuals.

Indonesia has massively expanded its social assistance programs to respond to the crisis. However, with more beneficiaries and a larger scale crisis, a new question arises: How can we strengthen existing institutions as a backbone to provide assistance to the crisis?

As a commitment to support the Indonesian government in mitigating the economic impact of COVID-19, J-PAL Southeast Asia (SEA) hosted a webinar on September 22 on “Social Protection in the COVID-19 Era: What can the evidence tell us?” featuring Indonesia’s Head of Fiscal Agency of the Ministry of Finance, Febrio Kacaribu, J-PAL SEA Co-Scientific Directors Rema Hanna (Harvard) and Benjamin Olken (MIT), and the Chief of the Policy Working Group of Indonesia’s National Team for the Acceleration of Poverty Reduction (TNP2K), Elan Satriawan, as moderator.

In his keynote speech, Mr. Kacaribu highlighted Indonesia’s investment in social protection programs amounting to IDR203 trillion, the largest allocation from the National Economic Recovery Program (Pemulihan Ekonomi Nasional/PEN). The programs are expanded to protect not only the current poor against major shocks, but also the growing number of new middle-income individuals and small businesses who have become vulnerable due to sudden loss of income. The government has increased the coverage and benefit levels of existing assistance programs while also introducing several new schemes to reach the newly vulnerable groups, including pre-employment cards, cash assistance for micro, small, and medium-sized enterprises, and wage subsidies for formal workers. Currently, the government’s priority is to ensure the effective delivery and implementation of these programs.

“We want the fund that has been budgeted to be received by the intended recipients faster so it can help their livelihoods.” – Febrio Kacaribu (Indonesia’s Head of Fiscal Agency of the Ministry of Finance)

Lessons from existing evidence to support Indonesia’s social protection programs during COVID-19

To support the Indonesian government in rapidly expanding social protection programs, Rema Hanna and Benjamin Olken presented lessons from existing evidence to improve the design of the social protection system to be more flexible in addressing the ongoing and upcoming challenges. Four critical aspects of social protection programs were highlighted:

Financing

With the uncertainty of the crisis, designing policy to respond to the dynamic public health situation can be costly. The government needs to provide adequate social protections to accompany the stay-at-home recommendation. This is necessary not only for a humanitarian reason (helping people survive) but also to reduce the need to go outdoors to generate an income, making social distancing or lockdowns more effective.

Moreover, social protection can be designed to adjust to the public health situation. One idea is to automatically link social protection to region-specific health or lockdown measures. For example, in Indonesia, the social protection policy can be tied to the green, yellow, orange, and red categories of COVID-19 health situation. Hence, when a region faces more severe health and lockdown situations, it would automatically trigger a more generous social protection policy. This enables direct compensation and provides a cushion if a region needs to go back into a tighter social distancing policy.

Targeting

With a new group of people facing a sudden shock, determining the right beneficiaries becomes more challenging. Improving the beneficiaries database can help us be more flexible and prepare better for future crises. While a proxy means tests (PMT)—one of the most common indirect measures of income—is useful for identifying those living in long-term poverty, it might not accurately identify the new vulnerable individuals.

The notion of community-based targeting, or giving communities a fixed number of slots and allowing them to identify beneficiaries, was found by J-PAL affiliated researchers in Indonesia to be a more effective means of targeting during the COVID-19 crisis. This method reduces exclusion error as the community might be able to better identify the impact of the recent shock among its members. Additionally, researchers found that the community targeting method led to higher community satisfaction. In line with this evidence, the Government of Indonesia has been using the community targeting method to distribute cash transfer sourced from the Village Fund (BLT Dana Desa) program at scale in response to COVID-19.

Another idea is to use self-targeting through an on-demand application, complemented by officials' verification through a PMT or bank or phone records. Evidence shows this method can reduce exclusion error as eligible people make themselves known and reduce inclusion error as it screens out ineligible, or richer, people.

In addition, the expansion of novel sources of data can be used to identify intended beneficiaries, using, for example, cell phones, electricity, bank account data, or lists of people laid off from formal industries.

Conditionality

Many social protection programs implemented during the non-pandemic period required conditionality. For example, beneficiaries will receive social assistance if they have completed certain health or education activities. Countries must consider how strongly they want to enforce conditionality during the crisis, considering its benefits and costs. There can be a trade-off between achieving intended outcomes and missing out on supporting the very poor, as they may struggle to meet the condition.

Many countries loosened conditionality due to capacity constraints with program expansion and providing conditioned-upon education and health services in the pandemic. To achieve similar intended benefits without having to enforce conditionality, evidence from Morocco suggests that labeling the cash transfer with a recommended, but not required, use had a similar, if not better effect, than a traditional conditional cash transfer. Labeling is cheaper to implement and can be applied during a pandemic situation to address limited capacity to ensure conditionality.

Delivery and digitization

Finally, the nature of the crisis has reinforced the need to move to digital enrollment and payments when disbursing assistance. Some topics under discussion in the policy space are, for instance, how can we have a telephone or web-based service for enrollment in social protection programs and how to create a payment method that does not require people to go to public places to pick up assistance.

Despite the enormous benefits of digitization, studies conducted by J-PAL affiliates in Andhra Pradesh and Madhya Pradesh, India; and Indonesia illustrate the challenges in designing and implementing digital enrollment and payment programs. For example, it can be difficult to enroll beneficiaries in new digital systems, and incentives for providers to enforce the implementation may be lacking. Hence, as digital enrollment and payment systems are expanded to deliver assistance that will last even after the crisis, paying attention to the program's details, such as the implementation strategy, incentives for providers, or to whom the programs are targeted, will matter considerably.

This article is based on the webinar titled “Social Protection in the COVID-19 Era: What can the evidence tell us?” on September 22, 2020, featuring Indonesia’s Head of Fiscal Agency of the Ministry of Finance, Febrio Kacaribu, J-PAL SEA’s Co-Scientific Directors, Rema Hanna (Harvard) and Benjamin Olken (MIT), and moderated by the Chief Policy Working Group of Indonesia’s National Team for the Acceleration of Poverty Reduction (TNP2K), Elan Satriawan.

To read the content of the presentation, see “Social Protection in the COVID-19 Era: What can the evidence tell us?” by Rema Hanna and Benjamin Olken.