Substack

Monday, June 22, 2020

Limits to technology in development

There is only so much that technology can do with addressing state capacity weaknesses or public service delivery failings. There is no digital pathway out of poverty or for chronic development challenges.

A reminder of this comes in a good article by Reetika Khera on the problems of excessive technologisaiton of NREGS. When it was piloted in Rajasthan at the turn of the millennium, its implementation was simple,
At that time, a public worksite would be opened in the village, and everyone who showed up was allowed to have their name included on the muster roll (an attendance sheet). At the end of the fortnight, some work measurement (if only notionally) would be done, and the muster roll would be sent to the block office to process payments. The “mate" (a worksite supervisor) or panchayat secretary would collect cash from the block office, and, on the designated day, names would be called out at the worksite for receiving a cash payment and a food coupon.
And this is what technology did,
From e-muster rolls (eMRs) and e-fund management system (e-FMS) and GPS-enabled biometric attendance to geo-tagged NREGA assets, MoRD became the site of a techno-orgy. The less tech-ready the rural area, the more fanciful the tech-fix that was imagined for it. Before first-generation issues with NREGA had been resolved, it was loaded on with more objectives and more technocracy. Workers could no longer just show up for work: e-MRs, with pre-printed names of workers who had ostensibly demanded work, became the norm. NREGA 2.0 was soon burdened with Aadhaar. Aadhaar, in its early years, needed ready-made databases to boost its enrolment numbers. Workers in the NREGA database were sitting ducks: vulnerable and anxious about losing their entitlements, they flocked to enrol. In spite of clear instructions from the Supreme Court, MoRD found ingenious ways of making Aadhaar compulsory. For instance, when “demand for work" was entered in the software at the block office, it could only be registered for workers who had submitted their Aadhaar numbers. Job cards without an Aadhaar number linked to it were cancelled...
The next victim was the payment system. Payments were held up if Aadhaar numbers were not linked to the job card and/or bank account. MoRD surreptitiously made it compulsory to link NREGA bank accounts with Aadhaar numbers. Instead of deploying technology to make things easier for labourers, technocrats started using it to make their own lives easier by centralizing control. Administrators no longer needed to visit villages to monitor its implementation: they merely peered at the NREGA real-time portal, where all was well.
Her suggestion is to revive features of the original NREGS,
One, large worksites should be opened proactively in each gram panchayat without waiting for anyone to apply for work. Such worksites should remain open for the coming months. Two, anyone who shows up to work should be allowed to work. The requirement of demanding work formally must be relaxed. For this, uniquely numbered muster rolls issued to sarpanches and panchayat secretaries, without pre-printed names, can also be considered. Three, states could be allowed to pay wages as cash-in-hand, bypassing the banking system. Many are horrified at the suggestion that we revert to cash payments (it has been projected as th
The main cause of corruption), but there are important reasons to consider it now.

The payment system for many welfare schemes, including NREGA, is beginning to resemble a popular visual trope from India—a tangled web of electricity wires. Instead of simply linking each job card to a bank account to transfer wages, today, four numbers might be needed: job cards, Aadhaar, bank accounts and mobile numbers. As a result, transaction failures are high
In another area of development, education, the promise of technology appears to have hit a roadblock in the form of access. With schools lying shutdown due to the pandemic, supporters have see an opportunity to not only manage schooling in this time but also mainstream Edtech.

But even before we come to issues like content and student-teacher engagement, the issue of internet access (never mind its quality) should be addressed. And, contrary to conventional wisdom, this is far from settled.
Just two states have at least 40% rural households with internet access.

The access challenges are well known and large as captured in this article,
According to a global survey by Pew, only 24 per cent Indians have access to smartphones. Not only does India lag among the list of emerging economies, there is also severe disparity in the ownership of smartphones between the genders: 34 per cent of men own smartphones as opposed to 15 per cent of women. The situation is no better when it comes to internet connectivity. “Many of my students are unable to access notes, audio clips and so on because they don’t have enough money to recharge their data packs,” says Rumi. According to a National Statistical Office (NSO) survey, only 23.8 per cent households in the country have internet connectivity and only 10.7 per cent have access to computers. So while students like Shruti are engaged in e-learning via apps such as Zoom and Microsoft Teams, teachers and students such as Rumi and Sangeeta are struggling to get by with low-resolution cameras and WhatsApp audios. The divide isn’t only between the rich and the poor, it exists between genders and states and is also determined by several socio-economic factors.
Further, like with all else, shutting down schools hurt children from the lower income categories, who are also those with lower learning levels and who are least likely to get any parental support in their learning process. In simple terms, lockdowns hurt the children from poorer income groups and causes permanent income loss.

It is imperative that primary schools open at the earliest, since early learning loss transmits forward and prolonged lockdowns are increasing the likelihood of permanent learning and income losses.

Yet another area where supporters proclaim a new dawn is fintech. I have blogged earlier here and here about its limitations. But here comes more evangelism from Navin Kukreja, who argues for "going completely digital",
By going completely digital, lenders will be able to lower their cost of operations. The benefit accruing from this can be eventually passed on to customers in the form of lower charges, fees, interest rates, etc. Digitisation should lead to better quality data becoming available to lenders. The more agile among them will be able to use it to offer better and more customised products to their customers. Digitisation will also enable borrowers from remote towns and cities to get access to credit from banks and other large lenders. One roadblock in the path of financial inclusion for credit products has been the need to be physically present to service borrowers. With digitisation, lenders with the right risk appetite and risk models will have the reach to deliver their products without the need to have branches and employees in specific locations. This could prove to be a turning point in bringing a significant part of the population under the umbrella of formal credit. Collections, too, could be digitised. Until now there has been a tendency to use “workforce in physical location” as the primary means to collect.
This assumes that everyone is linked to the digital system in such a way that their actions can generate digital trails which are sufficient to both make them credit-worthy enough as well as enable access to its use. In this context, The Economist has an entire survey which highlights the emerging limitations of artificial intelligence. 

No comments: