Substack

Thursday, May 21, 2020

The targeting challenge in delivering welfare services and limits of digital solutions

India's digital identity program, Aadhaar, has achieved several successes. It has facilitated financial inclusion through simplified no-frills savings bank accounts under the Jan Dhan Yojana (JDY). It underpins the entire architecture of the Direct Benefits Transfer (DBT) program which has doubtless increased the efficiency of welfare services delivery.

Perhaps the most important transformation from Aadhaar may be its role in becoming an anchor that underpins financial transactions and allows the flowering of fintech in India. The UPI is rightly being hailed as a transformative development. 

But it has not addressed several other important issues, which, while it was never likely to have resolved, had become part of the narrative around Aadhaar. Targeting and delivery of public services are two examples. 

Targeting has been one of the biggest problems with the delivery of welfare programs across the world. The ongoing Covid 19 pandemic has only reinforced the challenge with targeting, especially in the case of migrants. It was one of the major public narratives that Aadhaar will help address this problem. As India has proceeded with the likes of no-frills bank accounts (JDY) and DBT, this narrative may have become entrenched.

However, like several other narratives, this too has little basis. Aadhaar is an identity validator - it validates a person. Validation comes after the identity (or eligibility) is established.

Aadhaar cannot identify whether he's eligible for something. That eligibility depends on whether the person meets the requirements of the particular program. This, almost always, involves some form of physical survey and attestation before being declared eligible. Once declared eligible and tagged with Aadhaar, the administration of that person's account for the particular program becomes simpler. Here too, if there is a dynamic dimension to eligibility (people move out of poverty, by say getting a formal or government job), then too the administration of the person's account becomes a problem.

In other words, Aadhaar is relevant only for administering program beneficiaries identified as eligible through other means, and that too where the eligibility is static. The problem of targeting remains.

Take three of the biggest examples in India - PDS, JDY and farmers.

In case of PDS, the biggest challenge is the issue of identification of eligible households. There has been much progress in this area, including with the latest Socio-Economic Census Survey. But even with these, exclusions are significant. The exclusions cover not only those unidentified but also those identified and not covered under the PDS for various administrative reasons. Some studies point to the extent of under-coverage under PDS being as high as 100 million.

While the actual number is most likely lower, as multiple independently done studies by reputed institutions/researchers here, here, and here show, the numbers are large and very significant. A fundamental problem is that the underlying statistical considerations on PDS are based on the 2011 census.

This problem does not figure in the entrenched narrative about PDS. For far too long, PDS reform has been about eliminating wastage and leakage. Commentators look at any PDS reform as one aimed at enhancing efficiency. The attention of young bureaucrats in the field is aimed at "weeding out the bogus ration cards" using the wonders of Aadhaar-enabled digital technologies. You get an award for reducing leakages due to inclusion errors but not for reducing exclusion errors and expanding coverage. While this should be done, a greater or at least equal priority should be to identify those excluded genuine beneficiaries.

One immediate fiscally-neutral policy response would be to mandate that District Collectors would be allowed to retain the total number of ration cards. They should be incentivised to "weed out the bogus cards" and allot them to the excluded. Needless to say, even this policy can create its set of distortions and will need to be revised in 2-3 years of implementation.

In case of JDY, again the challenge is the eligibility of those who have opened the no-frills JDY account. In the absence of robust eligibility screening mechanism, it suffers from both exclusion and inclusion errors. Sample this,
According to official statistics, roughly 200 million Indian women (47 per cent of adult females) have a PMJDY account... Official statistics do not tell us how many of the 200 million female PMJDY account holders belong to poor households... A 2018 survey (using) a Grameen Foundation methodology where answers to 10 questions about a household’s characteristics and asset ownership are scored to compute the likelihood that the household lives below the poverty line... tells us that roughly two-thirds of adult women — just over 325 million in total — are living on less than the UN-recognised poverty criterion of $2.50 per day. In normal times, nearly nine out of 10 of these women say it would be difficult to pull together Rs 6,000 within a month to deal with an emergency. So, even if we go by government statistics and assume that PMJDY accounts were opened only by these poor women — a generous estimate — then over one-third of poor women or 125 million women, do not have a PMJDY account. However, we know that some better-off households also have accounts. The 2018 survey numbers suggest that 75 per cent of PMJDY account holders are poor. If we instead allocate the government’s count of PMJDY accounts to poor women based on these 2018 survey numbers, then roughly 175 million poor women lack PMJDY accounts.
In fact, as the article shows, even when a JDY account is notionally opened, it still does not ensure access,
A nationally representative survey from 2018, the Financial Inclusion Insights Survey, asked respondents whether they have a bank account and, if yes, whether it is a PMJDY account. Roughly 80 per cent of female respondents stated they have a bank account, but only 21 per cent said they have a PMJDY account. What drives the gap between government and survey numbers? Likely some combination of dormancy, account duplication in the system and the lack of knowledge among women about the type of account they hold.
And all this is even without the biggest challenge of them all, easy access to physical cash-out or digital transaction channels so as to be able to regularly utilise these accounts. A cash-transfer mechanism does not achieve the objective of ensuring genuine access for those once enrolled. There is the issue of accessing the transfers under JDY, utilisation of the no-frills account, replenishing the gas cylinders after the first one, and so on.

In case of farmers, the big and insurmountable problem is to differentiate and identify tenant farmers. It is widely acknowledged that the biggest problem with PM-KISAN is that of identification, as it includes only landowners and mostly excludes landless and tenant farmers and sharecroppers. Aadhaar cannot solve this problem of identifying the excluded.

Any formalisation will run into complex political economy challenges since owners will be loath to recognise them. Land records maintenance by way of updating Adangal every crop season has long since fallen out of favour across the country. In short, there is no record that links the tenant/sharecropper to the land. Their identification therefore requires physical field verification surveys. And, given the dynamic nature of these relationships, they need to be revisited periodically.

The only option is to do what some states like AP and Telangana have done. Do physical survey and then recognise tenants and give them some document, and keep doing it every 3-4 years. Even with all its challenges and problems, it seems the only practical solution. It is a reminder that Aadhaar and digital technologies have not moved us one inch in the tenant farmer targeting problem!

Again, even if the identification and validation problems are solved, there is the real problem of access to the associated benefits.
In its 2019 report, the Reserve Bank’s Internal Working Group to Review Agricultural Credit estimated that despite numerous existing initiatives, at most, only 40 per cent of India’s small and marginal farmers are covered by formal credit... KCCs, a scheme first introduced in 1998, over 20 years ago, should concern us. The RBI’s Internal Working Group estimated that as of 2019, only around 45 per cent of all Indian farmers possessed an operative KCC and that given the existence of multiple accounts per farmer, the percentage is likely to be even lower. Nabard’s own NAFIS Survey 2016-17, reported that only 10.5 per cent of agricultural households were found to have a valid KCC.
All this puts in perspective the true gains from Aadhaar and DBT. For sure while efficiency gains have been aggressively reaped, what about the welfare loss from these various aforementioned problems? 

We should not be under any illusion that the problem of identification has somehow become any less important in the aftermath of Aadhaar. To put it in simple terms, Aadhaar has not moved the needle in any meaningful manner on the issue of identification. And it will not do so. It was meant to only validate identified beneficiaries of public welfare programs. 

And let's not talk about the other idea supporters often point to, the use of data analytics. We can safely say that while data analytics will doubtless help with weeding out certain categories of false positives in some programs, it will be of no help with identification of new beneficiaries.

Then there are the technical challenges with delivering the cash benefit through the Aadhaar enabled eco-system. See this account of the problems in case of NREGS. 

Ironically, there is a compelling case that in times of Covid 19, despite all the Aadhaar innovations and digital technologies, the good old NREGS job cards may be the most reliable (in terms of being dynamically adjusting) targeting database for rural areas,
There are... about 14 crore for NREGA job cards, and 12 crore or so for women’s JDY accounts in rural and semi-urban areas (assuming that the gender distribution of accounts is similar in rural and urban areas). For purposes of cash relief, the JDY approach turns out to fare poorly on several counts. First, JDY accounts are a mighty mess – the NREGA job-cards list is far more transparent and well-organised... a large proportion of JDY accounts (40% in March 2017, down to 19% in January 2020) went “dormant” as customers were unable or unwilling to use them... It is not clear what proportion of JDY accounts are operational today, in the sense that a bank transfer to these accounts will actually reach the recipient in good time. Second, cash transfers to women’s JDY accounts are likely to involve large exclusion errors... Third, inclusion errors are also likely to be larger in the JDY approach. Job cards are meant for rural workers, JDY accounts are for everyone... (studies) show... JDY beneficiaries tend to be better-off than NREGA beneficiaries... the probability of having a JDY account is more or less the same for poor and non-poor households. 
Aadhaar has doubtless helped improve the efficiency of transfers through DBT. But it has done precious little on addressing the issue of eligibility verification and helping enrol the excluded into government programs. It was never meant to. 

There are serious limits to any digital pathway to address targeting and access, leave alone poverty reduction. Acknowledging that may be a good first step. 

Update 1 (23.07.2020)

Rohini Pande et al on how PDS helped during Covid 19,
Our research team recently evaluated how Chhattisgarh’s public distribution system functioned through the lockdown and how rural households were faring in the state. Ration shops functioned well: Out of over 4,000 PDS shops we surveyed, 99 per cent were open through the lockdown and stock-outs were extremely rare. Of the over 3,900 households we surveyed in rural Raipur, 95 per cent reported receiving rations. But 20 per cent of the surveyed households worried they would run low or out of food in the coming weeks. Interviews with anganwadi workers revealed that households were eating fewer fruits and vegetables, and more rice and dal than before the lockdown. This is consistent with the NSS data that suggest free rations in Chhattisgarh helped households cover 15 to 33 per cent of their monthly food expenditure, depending on the ration card holder.

Update 2(01.11.2020)

Indian Express investigation on fraud with the DBT in scholarships for minorities in Jharkhand. The two sources of corruption being failures in verification during registration of eligible students, and fraud in fingerprint validations.  

No comments: