Tuesday, July 15, 2014

Leveraging conditional cash transfers to develop local mobile money ecosystems (part 3 of 3)

By IMTFI researcher Erwin A. Alampay

Understanding the potential impact of using mobile money for cash transfers to the poor requires understanding how beneficiaries make use of their money, and how their community ‘behaves’ as a result. For instance, do they consume all the transfers, or would they retain some as savings? How quickly do they spend their money, and for what? Are there gender issues with respect to decision-making regarding its use?

Market vendors tend to follow CCT
beneficiaries on day of CCT release
For instance, it has been observed how small markets/ local vendors appear near the places where you can obtain cash when Conditional Cash Transfers (CCTs) are being released. We noted how some of the items being sold were non-essentials (e.g. balloons) (see Picture 1). Gusto & Roque (2014) say that one advantage for entrepreneurial ventures such as this is convenience, saving them transportation expenses and time. However, they noted (like we observed) that some goods sold might be "unnecessary."

Roque & Gusto (2014) estimate that around half of conditional cash transfers are spent on food, a third on education, and about 6-10% on healthcare. If there were savings, they were usually reserved for education related purposes. At the same time, they noted that the more popular form of "savings" was "non-monetary" in form.

How might mobile money alter this? An experiment in Niger showed that there was greater diversity in goods purchased among beneficiaries who received their cash transfers through mobile money (Aker, et. al. 2011). They speculate that this may be the result of greater "privacy" which in turn affects intra-household dynamics in decision-making concerning how money is used.

At the IMTFI annual conference held last December, our group speculated on how such a scenario might play out (see Photo 2). One can also assume that if carried out in the Philippines a similar change could arise, especially if the space for transfers becomes "virtual." It would then be less public (and more private), and hence there would be no physical space for the local community’s vendors to congregate.

At present, there are some system "losses" in terms of use of money for "non-essentials" (see Photo 2). Not only is the release of the transfer known to the households, but the entire community is aware of the event. On the other hand, if the transaction is purely electronic, then there is a stronger possibility for such goods being used primarily for more targeted and developmentally beneficial use.

It should be mentioned that early on one of the potential benefits people considered in the use of m-money was a reduction in reliance on cash as a transactional medium (Porteous 2006). Indeed, this would be beneficial in the delivery of CCTs if a larger share of the transfers could be kept in m-money form, since this would reduce the amounts that need to be cashed out. This would simplify CCT logistics, reduce the need to transport huge amounts, and provide for a more secure way of saving money, while also expanding the network of people who would be more willing to accept or undertake m-money based transactions.

Output from the IMTFI workshop
illustrating expense “losses”
This brings me to two points: First, we can work only with merchants (especially the ones that provide basic goods and services) that most beneficiaries would buy from (e.g. rice vendors; hospitals, schools; vegetable/meat/fish vendors) to make these transactions convenient and easier. This would require working with retailers in the area to build their capacity to transact with mobile money. Second, one can view restrictions on CCTs through spending in m-money-only transactions as potentially beneficial, since this could encourage saving by reducing the possibility of spending the money on frivolous (unnecessary) goods.

This way, making it easier to use m-money to transact with certain merchants can be helpful in directing consumption towards more development-oriented uses and complementing the primary objective of conditionality. It will reduce the need for cashing out, and at the same time stimulate the local economy through the consumption of local goods and services.

We need to recognize, however, that this kind of change is not easy. There should also be incentives for both beneficiaries to encourage adoption of non-cash based transactions. For beneficiaries, for instance, the consumable value for m-money based transactions can be higher than the actual value (e.g. extra credit; discounted services). For the merchants, the volume of money channeled to CCT beneficiaries on a regular basis hopefully will be enough incentive for retailers to get interested and sustain the change.

However, limiting how CCTs can be consumed also poses a danger of reducing people’s freedoms and creating "unwarranted" or false demand on certain commodities that beneficiaries do not consider crucial, thereby affecting local prices. To address this, a system can be designed where the "savings" (m-money unconsumed over a period of time) are freed up for use on any other consumer product, encouraging/incentivizing savings among CCT beneficiaries.

As such, the volume of conditional cash transfers, and more recently cash transfers flowing in for rehabilitation efforts for Typhoon Haiyan, can provide a rare opportunity to get away from what Mas and Radcliffe (2011) call a "sub-scale" trap, by providing a rare opportunity to increase membership in the network of m-money users. This can attract merchants because of the large number of customers and actual funds in the system. Hopefully, its increased use will develop greater trust in the system of m-money commerce and sustain its use in the local ecosystems.

Read the first post, "Use of m-money for conditional cash transfers in the Philippines: Part 1 of 3," and the second post, "Conditional Cash Transfers in the Philippines."

Monday, June 30, 2014

Conditional Cash Transfers in the Philippines (part 2 of 3)

By IMTFI researcher Erwin A. Alampay

According to the Department of Social Welfare and Development (DSWD), only 20% of the population receiving conditional cash transfers (CCTS) are operationally problematic. The constraints offer some clue as to how the use of m-money might be unfeasible in these locations. In particular, it will be unfeasible for m-money to work in areas where there are no cellular signals, which likely accounts for part of the difficult to reach 20%. Nonetheless, would it still be beneficial to consider the implementation of m-money in other areas?

Photo 1: Enumerators prepare tokens for the survey respondents
At present, there are various modalities for delivering conditional cash transfers to their intended beneficiaries. These different modes consider the availability of Landbank branches, or in its absence, possible partner rural banks, and postal services (PhilPost). The expense for the DSWD vary per option. For those using Landbank/cash cards, there is no additional costs except for the initial cost of providing cash cards. For partner rural banks, the rates vary between 22 to 40 pesos (about US$.50 - $.91). Using Philpost, a government-owned corporation, the cost is 50 pesos. If none of these are feasible, then other groups bid to provide the service. By bidding this out, the cost of using other private conduits has gone down from Php75 initially to only Php42.

In terms of using m-money as an alternative, the question is whether it can deliver to beneficiaries at a cost less than other modalities in areas not serviced by Landbank/cashcards, but where access to mobile services are present.

We considered the technical and financial feasibility question, by visiting an area where GRemit has operated, San Jose, Mindoro Oriental.  In the past two years, at least half of CCT funds that have been distributed there were delivered through GRemit, albeit not through a “pure m-money model.’ There is also operational m-money infrastructure in the area (e.g. GCash and Smart Money/Padala).

Technical feasibility
We surveyed CCT beneficiaries that GRemit was already servicing and asked about their interest, access, and previous experience in using m-money through cellphones. The majority of respondents (71% of 307 respondents) expressed interest in receiving CCT through the mobile phone. Relative distance to known claim/redemption centers for mobile money was significant with regards their willingness to do this. Those who were closer to a known cash-out center were also more willing.

In terms of mobile phone ownership, almost half (49%) of the respondents owned a mobile phone, and a majority subscribed to SMART (90%). Interestingly, ownership of a mobile phone was not statistically significant as to whether they were interested in m-money for CCT. This suggests that even those who currently do not possess a mobile phone are open to this option. This also means, that if this is to be universally rolled out, then the cost of providing mobile phones to the beneficiaries must be considered in design implementation.

Photo 2: Beneficiaries wait for their names to be validated.
In the background mobile stalls sell toys, clothes and food.
Overall, a third (34%) reported knowledge of receiving money via the mobile phone. People’s pre-existing knowledge of the use of m-Money was also statistically significant with regard to willingness to use it as a transfer conduit. Those who knew how to use m-money were more willing to have CCTs delivered through their mobile phone. Furthermore, those who had previous experience using mobile money were also those who reported interest in receiving their CCT in this manner. More of them were familiar with and had experience transacting with SMART Padala centers than with GCash merchant partners.  Partly this was because more of them were SMART subscribers, and there was a greater presence of such outlets in the area (see Photo 2). A third had already experienced using SMART Padala, and only 1.5% had used GCash. Although training non-users would be still be needed for implementation of CCT thru m-money, the existing experience in the community can be helpful in its acceptance. 

Financial feasibility 
To be financially feasible, cost efficiencies and potential savings (for both the government and the beneficiaries) should be demonstrable. The DSWD expressed preference for more frequent CCT releases at smaller amounts. However, if this is done, the operational costs would be greater because of manpower costs and expenses directly connected to its distribution and compliance monitoring. 

Would m-money provide a financially viable option in such a case? At current market rates, it was calculated that at smaller increments (cash transfers lower than Php2000), both GCash and SMART Money can be delivered at a rate lower than what is currently being charged by DSWD’s conduits (i.e. MLhuillier or GRemit). SMARTMoney charges for the sender are also lower than what GCash charges (Php10 vs Php40 respectively).  With amounts larger than Php2000 (about US$46), GCash becomes less viable, whereas SMART Money continues to be a viable alternative. At present, the maximum benefit provided on a bi-monthly release is Php2800.  This would cost Php15 for the sender if coursed through SMART Money, and Php60 if coursed through GCash.

Photo 3: Merchants hand out money to beneficiaries.
Barangay Central and San Agustin.
Costs on the part of the cash transfer recipients can also be compared. For CCT beneficiaries, their primary concern is the travel expense. However, transactions using SMART Money has additional charges, whereas GCash does not. It was also noted that the different cash out centers we interviewed in San Jose were not consistent in their policy as far as cashing out was concerned. This implies, that if CCT is implemented in this way, then m-money merchants must be duly trained/informed; and beneficiaries taught the correct deductions, if any. For smaller cash transfers, GCash and SMART Money, even with their additional charges, are competitive and even better than the current amount DSWD pays for delivering the amount per beneficiary, particularly for transactions below Php2000. GCash is not viable with bigger transactions (higher than 2000), whereas SMART Money becomes less attractive once money transferred is higher that Php2700.  But, given the large volume of money possibly coursed through this system, the government may want to negotiate for reduced rates.

Nonetheless, there would still be limitations on m-money viability. It would be dependent on the scale of availability of these services, and accessibility of their cash out centers in the areas being served and the transportation costs for accessing them.

Conclusions: Implications on CCT program design
Even as mobile phone coverage increases, there will still be areas where access to mobile phones is not universal. As such, an important consideration for CCT implementors is whether to provide this option only to those with mobile phones, and/or to provide mobile phones to beneficiaries as well.

Rolling out this program would be easier to implement in areas where cash out centers already exist. There, existing knowledge and experience in receiving cash transfers through phones are more likely, and community knowledge can be leveraged to help convince and train non-users. 

Financially, an m-money based CCT can be viable, particularly for small and frequent transfers. Further, since there are two kinds of mobile money platforms in the Philippines, another consideration is to which existing networks the majority of beneficiaries in an area are subscribed and the ubiquity of partners/merchants present there.  This can reduce barriers to adoption and generate positive interest in the proposed modality.

However, would the existing number of m-money cash out centers be able to absorb huge single day demands for cashing out? Can cashing out be controlled or reduced by retaining CCT in a non-cash (m-money) form within a local ecosystem? This will be discussed in my next blog: “Leveraging CCT to develop stronger m-money eco-systems in local communities.”

Read the first post, "Use of m-money for conditional cash transfers in the Philippines: Part 1 of 3"

Monday, June 23, 2014

Designing Financial Inclusion

By IMTFI researchers Anke Schwittay and Paul Braund

In a recent article in PoLAR: Political and Legal Anthropology Review, I use IMTFI’s Design Principles, published in 2010, and the IMTFI-funded project Following the Bean: Navigating Value Exchange and Vulnerability with Farmers and their Stakeholder (Melissa Cliver, Fellow 2009 and 2010) to examine the emerging practice of Humanitarian Design in financial inclusion. Drawing on anthropologists such as Lucy Suchman, Peter Redfield and Bruno Latour, I interrogate the ways in which professional designers who are using their expertise towards social ends have redefined the problem of development as a lack of creative ideas and innovation as well as inattention to systems and the absence of client feedback. Through this reconceptualization, humanitarian designers are positing themselves and especially their methods of design thinking, empathetic research, co-design and prototyping, as the experts best placed to address this problem. Design as an integrative discipline is seen as well-suited to solve the ‘wicked problems’ presented by persistent global poverty.

"Following the Bean" presents a good example of such design expertise. The original proposal was for the creation of a visual financial management tool for farmers of a coffee cooperative in Oaxaca, Mexico. Through the process of fieldwork, which encompassed ethnographic observations, story boarding, behaviour journal models and participatory design workshops, Cliver and her team discovered that farmers had a greater need for cash. They also discovered that coop members’ various ways of spending money stood in a special relationship of how that  money was earned. Income from coffee and corn, for example, covered everyday expenses, while extra money made from enjoyable activities like selling flowers was spent on enjoyable things like butter and hard-earned cash from the US was used for serious work like house construction. What was referred to as savings, namely putting money aside for an undefined future as opposed to for specific occasions, did not have a paired cash stream, expressed in the observation that ‘there is never enough money left over to save.’ The designers then conceptualized a new income stream for savings into a concept called Send the Change, where a US consumer would buy the coop’s coffee and round up the change, the difference of which would be sent electronically into a special savings scheme. Unfortunately the team was not able to test the concept’s prototype in the field. In the article, I show in much greater detail how to team carried out its work and the opportunities and constraints they faced in the process.

"Following the Bean" was just one project whose findings informed IMTFI’s Design principles, which were published as part of the first annual report. For IMTFI, design was a way to make the findings of its researchers actionable. You can see all of the principles here. In my article, I synthesize them as principles corresponding to codes (related to social status and rank), convertibility (related to different value scales and conversions among them) and cycles (related to temporal rhythms and obligations attached to them). I argue that these principles constitute poor people as innovators whose calculative and other logics inform humanitarian design, which on the other hand also contributes to disciplining them through financial practices. The result is a particular form of hybrid knowledge which is characteristic of what Daromir Rudnyckyj and I, in the special PoLAR issue of which this article is a part, have called the Afterlives of Development. This knowledge seeks to enable the coexistence of calculative and cultural rationalities in financial inclusion.

Last but not least, I am also interested in how design and anthropology (which are coming together in the emerging field of design anthropology) can inform development practice. I argue that both hold complexity in view rather than trying it render it technical. Design and anthropology offer development novel ways of looking, listening and learning, working with an experimental approach that questions the very assumptions that most development interventions take for granted. Rather than assuming what people need, humanitarian designers are wondering whether they are even asking the right questions. Having said this, we cannot overlook that humanitarian design practitioners and expertise have Western origins and in that regard present a continuation with orthodox models of development. The challenge is to find, in the words of Bill Maurer, “collateral, collaborative praxis” whereby anthropology, design and development can be drawn into relation and begin to create alternative figures of development. I think IMTFI is a great place to explore these questions further and  I would love to hear your thoughts on this, or on the article itself. You can email me at a.schwittay@auckland.ac.nz

Preview Anke Schwittay's upcoming book, New Media and International Development: Representation and Affect in Microfinance, coming out September 2014.

Tuesday, June 17, 2014

How to Increase Formal Savings for the Papad-Makers of Dharavi Slum

IMTFI researchers Mudita Tiwari and Deepti KC report their research results on the Financial Access Initiative (fai) blog

In Dharavi, Mumbai, the largest urban slum in Asia, groups of women make papad, crispy lentil dough wafers, for Lijjat Papad Company, one of the world’s largest papad retailers. Lijjat requires any woman who works for the enterprise to first open a savings account, and to encourage savings, the company deposits a small proportion of the women’s earnings (2 rupees of every 32 rupees earned) directly into the savings accounts, adding a bonus during the Diwali festival.


Go to FAI blog post for full text.

Monday, June 9, 2014

Afford Two, Eat One: Understanding Motivations for Spending and Saving

By Jan Chipchase, Founder, Studio D Radiodurans

Most of you reading this are sophisticated banking customers: you expect instant access, online apps, notifications, analytics and not a little delight. But imagine not having access to any of that. Imagine having to carry all your wealth on your person, or hide it in your home.

When people talk about “banking”, they often lose sight of its core value proposition: take money out of circulation and keep it somewhere safe where it can be accessed at a later date; access to credit; and to pay for things. After those three features, the benefits of banking are largely incremental. But to not have these three things, makes life very tricky. Rudimentary banking services (so-called financial inclusion) can help pull people out of poverty.

Myanmar is a country with very low formal banking penetration, but changes are afoot: it will soon have its first country-wide 3G network; there is significant inwards investment, an impending development of more stable, flexible financial policy and a stock exchange in 2015 that will stimulate domestic financial activity.

In March this year a team from Myanmar-based Proximity Designs, frog and strategy consultancy Studio D Radiodurans mapped the changing financial landscape in Myanmar. Over the two month project — funded by the Institute for Money, Technology and Financial Inclusion — we explored the diverse financial landscape for the poor in Myanmar and uncovered the nuances of income and loan cycles. We mapped behaviours around and attitudes to savings, investments, loans and transactions. We also explored the duality of development, how the poor balance their culture and beliefs with the advancement and globalisation of Myanmar, and how it has impacted their current lives and their outlook for the future. It’s a journey that takes in betel sellers, monks, motorbikes, goats and a lot of gold, with not a little of the afterlife.

The report identifies thirteen findings and twenty one insights, as well as a number of opportunities for future products and services. It aims to provide a foundational reference for organisations wishing to develop products and services for financially constrained consumers in Myanmar.

“defaulting on the loan would place a heavy burden on them and 
their family not just in this life, but also in the next.”

Some of the findings map to what is known in other markets, albeit with characteristics that are unique to Myanmar. For example motivations for not defaulting on a loan vary by culture, person, context. In Myanmar, a devout Buddhist borrower defaulting on a loan would be placing a heavy burden on themselves and their family not just in this life, but also in the next. We also learnt the significance of the novitiation ceremony in the life of a devout Buddhist Burmese, an event that results in them spending as much as US$1,700 at once, even though they earn less than US$10 a day.

Monday, June 2, 2014

Making Sense of Mobile Money in Urban Ghana: Personal, Business, Social and Financial Inclusion Prospects

By IMTFI researcher Vivian Dzokoto and Elizabeth Appiah

The goal of this study was to explore personal, business, and social money-related practices that have emerged with increased patronage of Mobile Money (MM) in Ghana. Of particular interest was the impact of MM on the urban poor who so far appear to be the sector of the population least aware of and the least likely to use MM in their daily lives.

The goals of the proposed study were to (i) investigate MM uptake patterns in year 3 of its re-introduction to Ghana, (ii) to explore the social and cultural interfaces between MM and existing money behaviors, including savings and money transfer practices among Ghanaians of different socioeconomic classes, and (iii) to investigate of the internalized (cognitive) representations of MM that Ghanaians develop. The study focused on the segments of the Ghanaian population and behavioral practices that were perceived as included and excluded from the MM adoption process. Research to answer these questions was conducted using surveys, spending diaries, interviews, and analysis of secondary data.

Transfer Money to Loved Ones Instantly with Mobile Money*


Study 1: Adoption of MM into Personal Financial Practices, A Quantitative Inquiry

College Student Monetary Preferences Study: Second only to treasury bills in overall preferences, Mobile Money was preferred over notes, coin, and card-based options.  Cash was preferred over card-based options. High value notes were preferred more than lower denomination notes. Notes were preferred over coins.
College Student Spending Diary Study: Cash was the predominant form of payment for daily expenses while Non-cash transactions made up 2.86% of reported purchases and MM accounted for less than 1% of transactions.
Mobile Money Use Population Survey: 14% of our sample reported using MM at least once in order to receive a money transfer and/or send a money transfer, pay bills, make purchases of goods and purchase airtime.
Industry Data: There was a significant uptake in MM use based on provided data.

Study 2: Adoption of MM into Personal Financial Practices, A Qualitative Inquiry

Interviews with Individuals in the MM industry: Interviewees cited regulatory, agent, educational, pricing and profitability issues as barriers to MM uptake, but respondents were optimistic about the potential for growth in the MM industry in Ghana.
Interviews with Consumers (both users and non-users of MM): Participants expressed an overall preference for cash over MM, and displayed a lack of trust of MM, yet basic knowledge of MM was higher than in previous years.
Interviews with Early Adopters: First use of MM occurred in a situation where the individual had an urgent need to send money to someone in another part of the country. 9 of 10 participants used it again.
Interviews with Retailers: Retailers predominantly used cash over MM in commerce because of lack of trust (network problems and concerns about fraudulent activity), yet knowledge of MM among retailers had increased.
Mobile Money in the Church: The nature of MM (intangible and mobile-phone based) made it undesirable for incorporation into church activities (including funerals and weddings).


The results of this series of studies revealed the following:
1. Cash is King in Ghana: Cash is still the main form of payment for day-to-day purchases. Large payments generally involve the formal banking sector. Cashless payment forms have not yet began to dominate the payment scene.
2. College students prefer Mobile Money over cash, but not over treasury bills (for saving).
3. Mobile Money knowledge and use has increased, but MM has not become a major means of payment for goods and services, or savings.
4. MNOs have increased MM products available to the public, slowly creating a MM ecosystem. However, apart from money transfer, this ecosystem is largely targeting the middle and upper class.
5. Barriers to MM uptake remain: Information gathered from interviews indicates the persistence of regulatory, partnership, and educational barriers that hamper the growth of the MM industry in Ghana.

In conclusion, the answer to the question of how Ghanaians from different socio-economic backgrounds are making sense and use of mobile money in urban Ghana in their personal, business, social lives is a simple one. The use of MM is increasing over time, and the commercial settings in which MM can be used is slowly growing due to the development of new products and business partnerships, but cash remains the major means of payment in urban Ghana. Also the majority of the MM products (apart from money transfer) are aimed towards the middle and upper classes to the exclusion of lower income groups. For instance, willingness to use MM especially in markets was low at the time of data collection. It is therefore hoped that as the MM ecosystem grows, new products that benefit lower income segments of society will be developed.

Socially, MM is gradually establishing itself as a means by which individuals can fulfill their financial obligations to extended family members in financial need. Apart from sending remittances, however, these series of studies indicate that MM has not widely permeated the social sphere, and thus has so far not had a salient impact on social life (e.g. churches, funerals, weddings). Whether or not this will change over time remains to be seen.

*Transfer Money Instantly to Loved Ones with Mobile Money. [Photograph]. Retrieved July 3, 2013, from: http://m.myjoyonline.com/pages/business/news/201304/104850.php.

You can read Vivian Dzokoto's full report here.

Tuesday, May 20, 2014

One researcher's thoughts on money and metadata based on fieldwork in Ethiopia

By Mesfin Fikre Woldmariam

Mobile money has become a buzz word since the success of M-PESA in Kenya. But whilst the topic is hot among development and poverty reduction workers; it has so far not gathered enough attention from academics. A simple review of literature in the area of mobile money reveals that there are many articles that associate mobile money with development, poverty reduction, security issues, adoption issues, acceptance and use by customers. But, there are no adequate pieces of work that relate to the issue of system design. 

As a researcher I tried to unpack what cash money really is. I am an information systems expert and my analysis of mobile money is primarily pragmatic and depends on my readings across different domains, particularly sociology and money, anthropology and money, behavioural economics and money.  I get knowledge and information from these domains and combined with my own areas of expertise I am able to reflect on what mobile money or digital money should look like and the nature or elements of the kind of system that can handle mobile money. Here are my reflections: 

First what is money and how it is changing?  Even though there are different categories of money from social and anthropological points of view, in this writing I am interested in looking at cash as money and thinking about how its changed forms affect system design. To me cash money is simply information with some specified metadata, like color, image or icons, numbers, and some other hidden security controlling means (see the figure below). People agree to accept this information as money because they know they can give it to others without concern. What makes this piece of paper (cash money) and or coins is the information (metadata inscribed on them). 

Ethiopian currency.

This is to say that if we remove these metadata, the remaining piece of paper cannot be considered as money. And thus, these metadata are making money valuable and make people develop trust and confidence. But with the digitization of money, in its current state of research and development, these metadata are excluded from existing platforms and solutions. In the current platforms or solutions, money is represented as a simple positive rational number of the form say 2.89 USD, 0.89 USD, 247 USD etc, excluding the metadata as well as the different money denominations. For example, in Ethiopia our currencies are denominated as (5 cent, 10 cent, 25 cents, 50 cents, 1 birr, 5 birr, 10 birr, 50 birr, and 100 birr notes). This indicates that we cannot pay (get paid) for example exactly 12 cents, 11 cents, 9 cents, 9.87 etc. However, we know that with the digitization of money, it is possible to accommodate any amounts and thus, unlike in the case of cash based transactions, in digital transactions making changes is not an issue. That is the bonus of current computing and mobile money technologies. But such money representation with positive rational numbers and removal of money’s metadata elements faces challenges when it comes to people that are illiterate. For example, illiterate users know their balance by counting the material money. They do not know what numbers 20, 40, 12 etc mean. Rather, they know these figures when they are physically handed them and able to count them by hand. The following paragraphs outline some of the design implications of my ethnographic research I did in 2012 in Ethiopia among open air market participants. 

Purposes of money’s metadata: 
Images or icons: The icons or images on money bills have different purposes like: as national identity (collective national values). For example, Ethiopian money bills have pictures of coffee plants, farmers, a map of Ethiopia, historical buildings, the signature of the governor, and a tractor machine. Some countries also inscribe photos of elite individuals and their sacrifice for the nation, for example photos on USD. In Ethiopia, the icons and images inscribed on money bills enable illiterate users to identify money bills.  
Security tools: These are features inscribed on money bills to identify real money from forgery. The problem with these tools is that, when the bills get old, these features usually fade away and may not be visible, which makes differentiating forgery from valid money bills difficult. 
Serial numbers: These have also a very important role. When all or parts of these numbers are lost (faded away), individuals do not accept. Such money bills have to be taken to banks for replacement, as per the respondents from my field study. Otherwise, they become worthless, as people, particularly illiterate users will not accept for change during transaction.
National identifier: This is written in both English and Amharic and identifies the legal issuer (governor). 
Color: All Ethiopian money bills are color coded, which enable illiterate users to identify different bills. Color and images on money bills are used for counting and computational purposes. For example, illiterate individuals know the sum of 10 birr and 5 birr will give 15 birr and yet do not know how to spell these numbers. When they are also asked to pickup money bills of say 50 birr from a lump sum of bills with different denominations, they easily identify them through their color. Thus, it could be said that color of money bills is a means to identify them. 
Economic value: These are numbers written in terms of roman numbers as 1 birr, 5 birr, 10 birr, 50 birr, and 100 birr as well as Ethiopian numbering systems and the Amharic language. Birr has denominations of 1, 5, 10, 50, and 100.

Characteristic features of Ethiopian money bills of five and ten birr.

The material nature of money bills also has an added value for illiterate and visually impaired people. Illiterate people make some simple mathematical computations (additions, subtractions, multiplication, and divisions) by moving money bills here and there as they cannot accomplish these through writing numbers on paper or calculating machines. For example, in order to make payment or receive payments people count money bills and in order to count them, people usually sort and arrange them according to the denominations (from smallest money bills to the largest) and then hold the stack in one hand and count with the other hand. In this context digitization can make illiterate people frustrated, unless there are solutions for this issue. 

One thing we learned is that current mobile money platforms and solutions did not consider these metadata in their design. My reflections and proposed solution for this problem will be presented in two forthcoming papers to be shared on the IMTFI Blog once they are published.

Money organization
From my open air market study I have also observed and understood that illiterate merchants and customers distribute their money among different bags or pockets. They also give different labels or names like bag for sales from coffee, sales from salt, and sales from other materials. When they need a change, they try to look into the respective bag. In fact if there are no changes in the necessary bag, they take from another bag and return the money later. It is kind of a “loan” from the other bag. I call this “distributed cognition”. They remember from which bag they took change and want to replace the amount they took. Thus, it appears that from a technology design perspective, designers need to be aware of such cognitions and practices and its implication for design. 

Problems with cash 
I also observed that even though the material property of cash helped illiterate people count and know their balance as well as the difference between different currency notes, (based on its color), it has some limitations. It can get old and stick together, part of it can be torn and people are not willing to accept for change, some individuals also make forgery money and easily cheat illiterate rural people. It also creates difficulty for making changes, if there are no changes. But, with the upcoming mobile or digital money systems, even though it appears that the issues of forgery, sticking together, and changes can be addressed, the issues of operationalizing technologies for these issues can not be an easy task. Addressing one of the issues will come at the expense of another.  

Final thoughts
Finally, based on the nature of money digitization and money handling experiences of illiterate people in developing counties, I feel that new technologies need to have capability like audio in order to  embed money’s metadata, and enable individuals’ to experience “physicality within digital environment”. I also recommend interested readers to consult the work of (Balen et al 2009) for more reflections about money digitisations as related to ease of usability, security, and auditing.  In this case I want to make a note that the work of Balan et al. (2009), focuses on literate people while my reflection is in the context of illiterate people who have different money practices. 

For further details on the project see here for the final report. You can also read Mesfin's working paper from 2012: "Understanding Social Relationships and Payments Among the Poor in Ethiopia" or email him with any comments and suggestions: mesfinfw@gmail.com