Getting There

1. Flight to Pune: 11:25 – 13:00
2. Bus to Satara (ST, from depot next to railway station): 14:00 – 16:40, Rs 78
3. Car drive to Mhaswad (म्हसवड): 17:00 – 19:30

I knew nothing of Mann Deshi when I arrived. My associates from our partner organisation (a large bank) briefed me as we drove to Mhaswad. They said Mann Deshi’s founder Chetna Gala Sinha was a remarkable person: a strong woman and a great achiever. By the time I left two and a half days later, I was to learn much more about Chetna, her organisation, microfinance, and the peculiarities of running a technology operation in a rural area.

Mhaswad is a large village of over 30,000 people. It is a long way from the Satara bus terminus. In big cities, we’re used to short distances taking a long time because of the traffic. The scale is completely different here. Mhaswad is about 100-150 km from Satara. It takes 2-3 hours to travel despite the near complete absence of traffic. Driving around to visit multiple locations in nearby areas in a single day is clearly not practical.

We realise the full impact of this later with our Karnataka kiosks: any software installation that requires a tech support person to visit the kiosk can be rolled out at the maximum rate of two kiosks per person day. With 800 kiosks and 30 support staff, that’s two weeks full-time work for what is otherwise merely a 10 minute installation process.

The only cellular operator covering Mhaswad and surrounding areas is BSNL. This means Airtel’s Rs 600 GPRS scheme is unusable here. We had been planning to use GPRS for connectivity in kiosks that were far too remote for any other form of connectivity, but being reduced to a single, generally uncooperative provider did not bode well. (As a result, the kiosks we now operate in Karnataka use VSAT.)

During the ride, we found that we could ask any random person on the road and be given directions to Chetna Sinha’s place. She’s a public figure in Satara.

We arrived to meet Chetna and her family in time for dinner. My associates did most of the talking, negotiating for their organisation and learning about Mann Deshi’s operations. The barrage of information was a bit too much for me given my limited understanding of microfinance, but I managed to glean a few things and, over the next couple of days, built a more complete picture.

A microfinance bank deals in small loans, amounts that are too small for most regular banks to consider economically viable. Owing to the nature of the demographic, the bank can only afford to loan for productive uses and must closely monitor the borrower’s progress. Interest rates are high, between 12 to 24%, but this is not a true picture because, in a well run microfinance operation, operating costs are typically lower than a regular bank’s. It only appears high because the loan amount is so low. The average loan is in the Rs 10-30,000 range.

Chetna said street vendors are among their most reliable clients, and also the biggest demographic (close to 50%). From experience, she only lends to women as their repayment rates are higher. Here’s a broad overview of how the loan giving process works: Mann Deshi loans to women’s self help groups at 12% interest per annum. They in turn distribute among themselves at 24%. Mann Deshi also loans to individuals for about 15%. With a remote group where a representative has to travel to meet them, it goes up to about 18%. (I may not have remembered the figures correctly.)

One of Mann Deshi’s most successful schemes is the Gold Loan, started two years previously. They sold 5kg the first year, 10 the second, and predicted 20 in 2006. The scheme was initially available during major festivals and since has become standard all year round. Mann Deshi buys gold from Solapur (not pure gold, forgot rating) and gives in 5-10 gram units to customers as a loan.

I was visiting to discuss Chetna’s ideas for setting up single-person information kiosks in areas where the bank did not have any presence. They would preferably be located in the kiosk operator’s house to save on rent. They would also serve as loan repayment centres, thus becoming a source of revenue for the operator and reducing the cost of loans to remote customers. One obstacle to doing this is the Reserve Bank of India (RBI)’s regulations, which mandate that only bank employees may have access to financial records. As the kiosk operator is an independent agent, they cannot be informed in anyway of how much is due from each customer.

We discussed additional services that could be offered at these kiosks. She thought electricity bill collection at the kiosks was a good idea because people waste a lot of time and about Rs 15 in travel to pay the bill. This could be a revenue earner if we can figure out how to make a deal with the Maharashtra State Electricity Board (MSEB). MSEB is not very keen on kiosk-based bill collection because politicians award bill collection jobs to about 200 (per politico?) otherwise unemployed youth and do not want to see their boys losing out. Unless MSEB’s accounting process is privatised, there doesn’t appear to be much hope here. Chetna frequently lamented the fact that the government in Maharashtra is hard to deal with. They’re not as interested in e-governance as Karnataka is.

After the meeting, she arranged for my accommodation at a nearby guesthouse. Chetna’s house is spacious, on the edge of the village, with a farm attached. She grows corn there. She has two sons, both in their early teens (another son was not present when I visited). I guessed that put her age roughly in the early forties. Her husband Vijay was also present but did not speak much.

A Day with Mann Deshi

10 AM: A car arrives to pick me up and take me to Mann Deshi’s headquarters. Chetna’s in a meeting when I arrive. The bank officially opens at 10:30, so most staff haven’t arrived yet. Someone gives me a stack of brochures to read. They detail a fair bit of what was covered in the previous day’s discussion. One mentions that Mann Deshi is now a microfinance case study at Harvard University’s JFK School of Government.

After her meeting, Chetna introduces me to her branch managers and Anita, her computer operator. I boldly attempt to speak in unbroken Hindi and am complimented for it. Turns out they are all Marathi speakers and Hindi is as foreign to them as it is to me. Anita shows me their tech setup. They have dialup internet access from BSNL Sancharnet, the only ISP in the region. We get online so I can demonstrate an application we have developed for use at kiosks, but the page won’t load.

11 AM: We get into a car for Waduj (also Vaduj), a nearby (50km away) town where Mann Deshi is expanding operations. During the ride, I explain that the kiosk initiative is a serious affair at Comat. We’re not just tech providers for our partner. We run the kiosks ourselves, offering their services among a much larger array. Chetna mentions being sceptical of a partnership with them because they tend to assume too much. She says one of their officials frequently says that they have a rural finance initiative going with Mann Deshi when in reality nothing has worked out so far. She’s concerned that this bank (which is on a rapid growth trajectory) may come in as a benefactor but eventually elbow them out of the business. It is a concern I’ve since heard from the bank too, about tensions when dealing with other banks in areas they are expanding into.

Chetna also explains how she was first introduced to Sriram Raghavan (Comat’s president) through Ashoka Fellows, and how she had been impressed with the speed with which he responded to email. What she thought would take a week would get a response before she even closed her session.

A picture of Chetna Sinha shaking hands with Kofi Annan while a group looks on hangs in every branch (at least, in three of the four I’ve seen today). It’s large, about 4x2½ feet.

12 PM: Chetna leaves Anita and me in the Waduj branch to attend a meeting in Karad. They have no power in Waduj, so Anita uses her laptop to get online. Dialup again. I explain the online application’s process. The connection is very slow. I doubt a scanned form upload will work on dialup. If we reduce the resolution, it’ll be too illegible to read without having an electronic copy to cross-verify with. In our tests, an 80 kB image timed out, while 20 kB uploaded successfully. Until this, we had been considering having customers fill out loan application forms on paper at a kiosk, then to scan and upload the form, while keeping the form at the kiosk until a bank representative arrives to verify it during loan disbursement. It becomes clear now that this won’t work without a good internet connection.

There is a State Bank ATM right across the road. ATM machines don’t use dialup, so surely there must be some form of reliable connectivity in these parts? Both IDEA and BSNL have cellular service here, but not Airtel. Waduj is the Khatav taluk HQ, so is somewhat important. Maybe they have some fibre optic cable running by? Will it be affordable? I later learn that most ATMs in India use VSAT for connectivity. Being satellite-based, VSAT is more suited for broadcast-type communications where a central station beams data to field locations. Having these locations send data back consumes precious, limited bandwidth and is therefore expensive. ATMs, however, are uniquely suited to VSATs. Their data requirements are very small, they need the rock-solid reliability offered by VSAT, and most customers don’t mind the latency of a few seconds as data is beamed thirty five thousand seven hundred and eighty six kilometres up to a geostationary satellite and all the way back down again.

VSAT-based connectivity for a kiosk with its general purpose internet access requirements, on the other hand, is an option only for the very serious and very desperate. The rest may look elsewhere.

1:20 PM: Anita is now demonstrating the online services to another person. I’m sitting by watching. The power goes out while she’s still waiting for the logged in page to load. The inverter is unable to support their two computers, so Anita leaves for lunch. I figure we may have a case for using low power thin clients that won’t overload a UPS. They do not use a printer regularly. In Waduj they have an Epson FX-1170 dot matrix and no scanner. In Mhaswad, a HP LaserJet and a scanner. The computer here runs Windows 98. It must be ancient. The one in Mhaswad runs Windows XP.

The branch is busy. There’s a constant stream of customers. They occasionally run out of space, so waiting customers come to sit in the office where I am.

2:10 PM: Back to demo with Anita. They’re confused whether the slash in the URL is ‘/’ or ‘\’. Guess that’s one of the things that becomes easier as they use the web. The page doesn’t load. IE attempts to load MSN Search, also fails, and thus loses the URL that was typed. Anita now tries opening “bank.comat.services”. I correct the URL again. This time it loads. They’re not sure how to navigate the tree view. Click on the +/- signs or the names? When nothing happens on clicking the name, they try the “+”, which works. In the date selector popup on the application form, I have to explain that they must hold down the “«” to see a list of years. Anita understands that only the fields with a ‘*’ mark are mandatory. She also remembers how the application type selection works. However, when she clicks “Calculate” and the amount is shown as 0, she’s confused. I look at the form and notice the birth date (on basis of which the amount is calculated) is not specified. The form should have validated and raised a flag on “Calculate”.

Alongside, Anita is demonstrating how to search for “Mann Deshi Mahila Bank” in Google. Next, they try “Mehandi Design”. Anita takes Google’s suggestion for the alternate spelling “Mehndi Design”.

The form has now loaded. Anita explains that they must first select Print, print the page that comes up in a new window, then select “Print Successful” in the original form. The upload form that comes up is sorted in an arbitrary manner. Anita can’t make sense of it and selects the “Upload Docs” link at left, which opens an upload form that (as far as I can tell) has no correlation to the form that was filled in. The user interface clearly needs more attention.

When the demo’s over, I sit down with the staff to look at their record keeping. They use some software from Pushkar Computer Group. It runs on DOS with Novell Netware 3.x (release year 1999). The software appears to be written in Cobol—it crashed once and I noticed a Cobol error message. It is English-only. Names in paper records are entirely in Marathi. In the computer, they’re in English. The operator says she has no trouble with this disparity. Her transliteration, from the few records I saw, appears to be accurate. She didn’t seem to speak English though. When showing me a record, she would occasionally point at a person standing in the room “look, there’s the customer” and reel off their history. The bankers clearly know their customers on a one-on-one basis.

The customer’s passbook is a booklet that is hand-updated, although the software can also print reports. The bank keeps two record books. A large one with custom-printed header on each page is for savings. One page per customer, with a photo stapled to the corner. If the page gets full, it’s continued as a new page (doesn’t seem to have happened yet). The loan record book is an ordinary school notebook, without photo. It tracks repayments. The photo itself is in a loan application file that includes the application form, Rs 100 stamp paper, attestation from the zamindar, fingerprint of the applicant on several pages, identity proof such as a copy of the ration card, and a quotation for the goods for which the loan is being taken out. There may have been other documents that I cannot recall at this time.

The day’s transactions are printed and cross-verified with paper challans — which are required to be filled out by the customer to perform any transaction. There is no online sync of accounts between branches. I suppose Mann Deshi’s typical customers are not prone to moving between areas and needing the ability to operate accounts from another branch, but Mann Deshi does have a concern with reaching out to customers in areas where they have no branch. Their overheads have pushed up the interest rate from the typical 12% to nearly 18% (if I noted the figures accurately the first night).

Collection agents carry handheld machines on their rounds (called “pigmy machines”). These are roughly the size of the larger credit card swipe machines at shopping malls. They have a brand label “Balaji”. The agent types in the account number and the amount that was collected, and it displays customer name on its two line LCD and prints a receipt (dot matrix) with the bank’s name (pre-printed), account number, amount collected yesterday, today, and the total. The machine has a serial port from which the computer is updated.

Branch staff don’t all know Hindi, so they had trouble explaining, often being reduced to uttering keywords and pointing at things. I found this fascinating. They can read and write the Roman alphabet, can transliterate Marathi names into English, use English financial jargon, but otherwise don’t speak a word of the language.

4 PM: Anita is on the phone with computer support, trying to sort out some problem. She’s speaking in Marathi, but from the spattering of English tech jargon, I can make out she’s a competent computer user. She tells me she’s comfortable with all of Marathi, Hindi and English. She’s clearly a dependable support person for the bank staff.

At around 5 PM, Anita and I take a bus back from Waduj to Mhaswad. The bus stops at an Indian Oil station on the way. There is no power, so driver and conductor get off and take turns at hand cranking the pump. I take pictures.

We reach Mhaswad around 6:30 PM. I have a slight headache from the bumpy ride and decide to walk around a bit. The village turns out to be rather small, not more than a kilometre from one end to the other. At its center, where Mann Deshi is located, roads are narrow and buildings are stacked right up to the edge. Most are a single floor, with an occasional two or three storey structure. Some of these houses seem ancient. Their walls are stone and doors wooden, inlaid with aging decoration you’d typically expect on a temple door. There is a functional drainage system. Pigs wander freely.

Mhaswad is too small to be classified a town, but it’s not a rundown country village either. It’s clearly prosperous, featuring several urban conveniences. I even saw something resembling a supermarket (the board said “bazaar”). Mhaswad may be an example of a place that is not urban solely because it has failed to achieve critical mass. It is too small for some of the trappings of urbanity to be economically feasible here. To use a fictional example, the only cellular operator here is BSNL. If Airtel too strode in and set up service, the two may find that there are not enough users here to make both sustainable. Airtel therefore skips the place, BSNL remains sole provider, and now has no concern for quality of service because there is no competition. The government then steps in as regulator, discounting infrastructure providers such as cellular operators, wishing Mhaswad a fair shot at becoming larger and making its growth self sustaining. How are they discounted? By taxes collected in big cities.

That was fictional. I made it up. It may not be true, but it does give a sense of the undercurrents when setting up operations in rural areas. The infrastructure isn’t there yet, but everyone’s hoping your operation will go some way in helping that infrastructure come to exist — the same infrastructure you need for your operation. Operations must necessarily bootstrap off whatever scant resources are available because when that growth comes, they will be at the centre of it, and hence stand to gain the most.

Winding Up

On my last day, I meet with Chetna to describe what I’ve observed and record her ideas, then leave for Phaltan to survey other potential kiosk sites. Later in the afternoon, I will take a bus to Pune, spend the night there, and board a flight to Bangalore in the morning.

Chetna says branches must send their reports every Friday, because the consolidated data must go to RBI (once a month, with status every alternate Friday and daily balance). Anita and Mr Durgesh say they had some trouble with HyperTerminal, which they could use to transfer the data by phone from the various branches (using ZModem, not TCP/IP). It’s not working, so they make a monthly CD. The binary data is not used for consolidation — that is done by hand in Excel from printed records. Currently balance status updates are done by voice phone call.

The banking software’s developer, Pushkar Computer Group, is based in Mahabaleshwar. An earlier provider was based in Karad, but they moved to Pune and have been reluctant to provide service, so Chetna switched to Pushkar. Pushkar’s software may be English-only, but Chetna says that is good. RBI regulations say records must be in English or Hindi, but financial jargon in Hindi is heavily sanskritised, so English works better. The locals are comfortable with English jargon.

Mhaswad’s primary occupation is farming, followed by shepherding. There is also a large migrant population — they work in sugar cooperatives as threshers or as construction labourers in Mumbai part of the year, and return here during the monsoon to tend to farms. Waduj has a lesser migrant crowd than Mhaswad. Chetna says Mhaswad’s growth is not as much as neighbouring areas like Waduj because it receives lesser rainfall. However, it used to be a revenue village during the British Raj and so continues to have a large marketplace. The water supply problem is also solved here (I did notice pipes sticking out of roads and going into houses; water piping is all underground). Because of this, some people prefer to stay here and visit the villages in the daytime.

We discussed her kiosk plans. I suggested low power computers with solar-backed UPSes. She was very interested, but also worried about support as the provider had no presence in the neighbourhood. She mentioned the Computer on Wheels project in Baramati, said the person there (forgot his name) had given her the idea of smart cards, which she’s interested in because her women don’t want passbooks. I’ve been hearing of Baramati several years now. It’s time to visit and see what’s going on. Chetna thinks it’s significant. Plus the person there has Sharad Pawar’s backing. Pawar sees it as a vehicle to get funding from World Bank and others.

History Lesson

Anita gave me a draft copy of the Harvard study, written by Pamela Young under the supervision of Guy Stuart, Associate Professor of Public Policy at the Kennedy School of Government. Once back in the comfort of home, the study helped fill in the missing details.

Chetna Gala was born and raised in Bombay in a Gujarati Jain family. When Indira Gandhi was re-elected to power in 1980, she joined the anti-emergency movement and via the Chhatra Yuva Sangharsh Vahini, travelled throughout western Maharashtra to organise the poor. During this period she fell in love with fellow activist Vijay Sinha, married in 1986, and moved to Mhaswad.

The two remained active in the farmers’ movement, even spending time in jail for their “agitations.” Chetna succeeded in forming dairies, getting the government to improve infrastructure, and encouraging women to stand for elections to the village councils. Despite several successes, however, she felt the need for a more permanent, institutional approach.

In 1992, Vijay and a few other activists started a credit union that Chetna later joined. The credit union could lend but couldn’t take deposits, so they decided to form a bank, petitioned RBI for a license, and eventually received one in 1997. Six months later, the Mann Deshi Mahila Sahakari Bank (Women’s Cooperative Bank of Mann) opened for business. According to an RBI official, it was the first of its kind in India. The bank had a rocky start but became profitable in just two years.

In 2002, Chetna became a World Fellow at Yale University. She moved to the US for a year with her family. As part of the preparation for her absence, the bank got itself an internet connection and email address. When she returned, she brought back international assistance and attention. Mann Deshi was now a success story worthy of study and emulation by others.