Aadhaar - Biometric Mismatch

Last week I found myself in a bit of a fix while trying to book a ride back home late at night.
Standing on a lonely roadside waiting for a confirmation on my Uber ride, I found that I no money to pay for it - there was no cash in my pocket, and my e-wallet account balance was below the minimum INR 350.

My account with Uber is linked to PayTM, a popular e-wallet platform in India. Uber gave me two options for topping up - through my credit-card or directly through PayTM. The latter had been an easier process but this hit a wall with the message which said that my account was not yet KYC compliant.

Know Your Customer (KYC) is now compulsary for all e-wallet accounts as per the new rules set by the central bank. Introduced with the objective of reducing misuse and money-laundering, KYC requires submission of id proof -  details of Passport, Tax PAN or the Adhaar universial id number.  I was under the impression that Adhaar was the fastest way of fulfilling KYC norms. I had done it earlier for my bank accounts and for a JIO mobile connection. It had taken just a few minutes to get an online confirmation.

However, the process was quite different for the PayTM. As soon as I sent my 16-digit number, I got a message saying that a PayTM representative come an meet me personally for a confirmation. A confirmation? Why is additional confirmation needed when, according to UIDAI's own procedures, the number could be used to confirm my identity with their central database? Anyway, since there was no hope of completing the KYC standing by the roadside at night, I went back to Uber and transferred some money to my PayTM account using my credit card.

A few days later, and after a series of SMSs, a representative of PayTM's "partner" turned up at my door with portable fingerprint scanner plugged into his mobile phone. He passed me his mobile and told me to type in my Adhaar number. Soon I got a message on my mobile with a code and a URL with the message - "By providing this code to our agent, you agree to become a full KYC customer of PayTM Payments Bank and confirm acceptance." You have no time to check the fine-print so the agent gets his code.

After this, a mouse-like device is used to scan my thumb-prints. One by one, the scanner moves from my thumb, to the pointer and index, until all 10 fingerprints are covered. For each and every scan he gets a message (from where? UIDAI?) saying that all the authentications had failed!

The look of amazement on my face prompted the agent to console me - "Aise hota reha hai...fingerprint badal jaate hei" (This keeps happening, fingerprints change over time). WTF?? I had heard about farm workers losing their fingerprints to hard labour but my fingers were anything but callused, or even unclean!

How can UIDAI authentications fail in urban areas? A quick internet search reveals that mine is not an isolated case. While authentication failures have been quite common on rural areas - due to incorrectly captured fingerprints, poor internet connectivity or a change in biometric details because of old age or wear and tear - it is now increasingly common in urban areas as well.

The Adhaar UID is no doubt backed by the laws of probablity and complex algorithms but this experience has placed me firmly in the ranks of the Adhaar skeptics. Failing to get an Uber taxi ride due to an Adhaar biometric failure hardly makes a difference to me, but to think that millions depends on this flawed system for their rations is just unexcusable.


----------------------------
Other Unanswered Questions:

* Now that the private sub-contractor to PayTM has all my fingerprints scanned and saved, what are the chances of misuse?


----------------------------
REFERENCES / LINKS

* https://www.businesstoday.in/current/policy/linking-aadhaar-with-paytm-deadline-ends-today-to-complete-kyc-for-e-wallets/story/271698.html
* Scroll on KYC problems - https://scroll.in/article/870873/indias-e-wallets-are-struggling-to-weather-the-storm-set-off-by-mandatory-aadhaar-linked-kyc-norms
* How to link PayTM with Adhaar - https://www.thehindubusinessline.com/info-tech/how-to/how-to-link-paytm-with-aadhaar/article22866742.ece
* Fingerprint authentication failure - https://scroll.in/article/857274/now-even-the-fingerprints-of-urban-indians-are-failing-during-aadhaar-authentication
* Medianama rebuttal to N.Nilekani's claims - https://www.medianama.com/2017/04/223-nandan-nilekani-aadhaar/

MPesa India: Wrong Time, Wrong Place?

According to a recent article, there are 271 different mobile money services offered in 93 countries worldwide. Among these, the most successful one is Kenya's M-Pesa.

Launched in 2007 by Vodafone for Safaricom and Vodacom, the largest mobile network operators in Kenya and Tanzania, M-Pesa has spread quickly, and by 2010 had become the most successful mobile-phone-based financial service in the developing world. By 2012, it had about 17 million M-Pesa accounts registered in Kenya alone.

M-Pesa's key advantage is that it does not need a 'smartphone'. Apart from being relatively expensive, smartphone tend to be battery guzzlers which need to be charged frequently, quite unlike the older mobiles which are cheap and far more durable.

In India, 61% of Indians own a basic mobile phone while only 17% of own a smartphone. Why, then, has M-Pesa not caught on in India?

Perhaps the answer to this question lies in a closer look at the Kenyan mPesa model.

The idea behind M-Pesa has its origins in 2002 when researchers found that the Kenyans had been using talk-time on pre-paid mobiles to make small payments across the country. People were purchasing everyday necessities - provisions and services - and paying for it by adding talk-time to the service provider's mobile phone. Thanks to code written by a student, this was expanded into a payment-transfer service.

Vodafone was the first to spot a big opportunity here and purchased the code. It then improved upon the model and created a wide network of agents across Kenya, selling this as a service. Within 10 years, M-Pesa transactions have grown so fast that they account for 20% of Kenya's gross domestic product.

Yet, when Vodafone tried to replicate the same M-Pesa model in neighboring Tanzania, it bombed. What was so different between Kenya and Tanzania?  An IFC report suggests the following reasons:

• Demand - urban migrants sending money back to family is not as prevalent in Tanzania and since there is less crime and insecurity in Tanzania
• Access to Financial Services - Only 9% of the population has access to formal financial services and 54% don’t use any form of financial services. When compared to Kenya, we see that 19% of the population there has access to formal financial services and only 38% don’t have access to any form of financial service
• Strength of Economy and Banking System - Kenya also has a stronger economy, a higher GDP (USD 890 per capita in Kenya versus USD 520 in Tanzania). 1.38 bank branches per 100,000 inhabitants in Kenya versus 0.57 in Tanzania
• Geography and distances - Tanzania is a large country with a dispersed population so quickly reaching dealers in rural areas was difficult
• Competition - Tanzania as Zantel’s competing service - ZPesa -- Tanzania also faces competition from informal money transfer channels such as the use of airtime as a currency
• Business Model -  Vodafone decided that the service would be offered to local subsidiaries on a license fee model instead of the shared revenue model that was agreed with Safaricom in Kenya. 
• Population and Customer Base - Tanzania has a population of 40 million - dispersed across Vodacom’s market share in Tanzania (41%) is significantly less than Safaricom’s in Kenya (79%) which translates into a base of 13 million customers for Safaricom as opposed to only 5.9 million for Vodacom
• Technology - M-Pesa is delivered in Tanzania using USSD which does not require any application to be stored on the SIM card. The user dials a short number to receive a set of menu options.
• Laws & Regulations - no national id in Tanzania -- cumbersome Anti-Money Laundering procedures -- 
• Lack of preparation - no market survey -- pilot test only covered one aspect: The main focus of the pilot was to test the USSD gateway since the USSD channel was not used in the Kenyan implementation. 

So the key factors that prevented this useful technology from crossing borders were - population dispersion, crime rates, financial literacy and mobile penetration.

In 2013, mPesa was launched by Vodafone India, in collaboration with ICICI Bank. From the look of it, this seems to be based on the menu-driven USSD model that was used in Tanzania, and not the faster, more efficient SIM-based model that was a huge success in Kenya. Over and above the challenges faced by Vodafone in Tanzania, the Indian market was already groaning under severe competition by the time Vodafone stepped in.

Perhaps the biggest advantage it still retains is the fact that unlike other leading e-wallet services like PayTM and Oxigen, M-Pesa users do not need to have a smartphone or 3G/4G internet connectivity, to use the service.

Howver, this advantage has also been diluted with the launch of the national Universal Payment Interface (UPI now renamed  BHIM). ICICI Bank too is offering this as a service independent of Vodafone.

So, is it just a matter of time before M-Pesa turns belly-up in India?

------------------------------------------------------
REFERENCES & LINKS

* mPesa India FAQs - https://www.mpesa.in/portal/customer/FAQ.jsp
* Smartphone penetration in India (2014-2019) - https://www.statista.com/statistics/257048/smartphone-user-penetration-in-india/
* mPesa launch in India - 2013 - http://gadgets.ndtv.com/telecom/news/vodafone-india-launches-m-pesa-mobile-wallet-with-icici-bank-355406
* IFC Case Study mPESA in Tanzania - http://www.ifc.org/wps/wcm/connect/3aa8588049586050a27ab719583b6d16/Tool%2B6.8.%2BCase%2BStudy%2B-%2BM-PESA%252C%2BTanzania.pdf?MOD=AJPERES
* Padmanabhan, Vishnu (2016) - http://www.livemint.com/Opinion/GOqw0yvSZqFbIB5Oxd1J2O/Transforming-the-digital-payment-infrastructure.html