Closing Keynote — public infrastructure and the people it records

Transcript

[music] [music] Okay, thank you so much. Um, when I started working on Aadhaar, uh was the first time in my life that I wished I was u tech techy upperccast male in India. Uh and since I can't be that I think this is the closest I can get to at least gate crashing uh conference of the tech bros. Um so this work actually it's quite helpful that it follows uh Subashish's talk because he has already uh talked a little bit about some of the issues that I want to cover.

I'm not uh from the tech background. I'm trained as a development economist and my focus has been quite different. The only reason I was interested in Aadhaar uh [snorts] which is uh the sort of digital uh infrastructure that I'm going to talk about uh is because they wanted to integrate it with uh welfare systems in India and uh I'll try to explain both the big brother and the Kafka aspects uh as I go along. Uh most of you will know what Aadhaar is. Uh it's a biometric ID. It's supposed to be made unique by the use of three biometrics. uh the face uh the photograph uh iris and fingerprints.

Um and initially they said there would be three ways of enrolling. One would be to bring two IDs. Uh there would also be an introducer system for people who don't have any existing IDs and then there was a third route which has hardly been used. Now this is a very quick timeline uh and perhaps I don't need to go into all these details so that uh there's time for other things but it's important to note that it sort of began in a legal vacuum. Uh there was no law backing this kind of project even though the implications were so widespread both uh for the right to life as well as for the right to privacy. uh and yet it began without any regulatory framework

in place and quite quickly as Subashish also mentioned it ended up in court and over 2012 to 20178 17 is when the right to privacy judgment came and 2018 is when the Aadhaar uh judgment came. Uh the all these matters were deliberated um in court and in the public etc. And then we won one battle uh which was the right to privacy law uh judgment where a seven judge bench said that it was a fundamental right. Uh but we lost the other one which was the Aadhaar judgment where the majority opinion said that it was valid uh in all

in most aspects. There were some caveats here and there but by and large it was given a green signal. So uh what is my objection or what was our worry? Uh with respect to Aadhaar um basically when the program began it was being projected as a an intervention that would improve uh welfare delivery uh and you know a lot of my work had been on documenting all the uh leakages from these welfare delivery mechanisms um and so that's when I got interested in Aadhaar is because I wanted to see exactly what it was going to bring to the table as far as uh welfare delivery is concerned. Um and so there were

different aspects to this welfare claim. Uh as you'll see the claims that Aadhaar was making kept changing because they get kept getting challenged on each one of them and so they had to reinvent themselves but let's stick with the welfare one which is where most of my work lies. They said it would help improve inclusion that you know the a number of uh poor people or people who need state support [snorts] are left out because they don't have valid ID documents. The second one of course uh which was that it will help reduce corruption and corrupt practices in the system. And then later after these two got caught out they started focusing on portability that right now at right now

meaning in 2010 a lot of welfare delivery is placebased and that people who migrate from the village to the city may not be able to claim their benefits in the city and that Aadhaar would bring portability and financial inclusion was also in the mix. I'm going to focus mainly on the first two because it's just to give you a flavor of uh what was wrong. Uh so actually uh the problem is that in fact there was no welfare um and I'll explain why I'm saying that uh but the fact is that they needed to make Aadhaar seem useful to people and therefore as this person who has written a book on Aadhaar said they needed to

create the push uh a push for the pull uh that there's uh no need in general when government is doing something there's no need for the system for uh for the system to convince people to queue up. But with Aadhaar, there was no explicit compulsion to get in the queue. And so what the government did is they created an impression that if people did not queue up for Aadhaar, they would be denied their access to subsidized grain and other forms of public support. And that's why they needed the link with uh welfare rather than actually uh making any welfare claims. So um so just taking on the first one uh why do I say that Aadhaar could not have done very much

for improving inclusion into uh welfare programs. Now one thing they said is that people lack ID documents and that's why we uh you know that's where Aadhaar will come in that it will give an ID document to people and therefore [snorts] they will uh be able to access their benefits. But in fact what happens in the Indian case is that the exclusion is primarily because the government predecides its budget on welfare irrespective of the demand or the need for it. So if you have already decided that you're only going to support 20% of the population with subsidized grains and in fact you find that 60% of the

population needs subsidized grain then you are de facto creating a system that is going to be exclusionary no number of ID documents can help you get into the system. So that's the second uh point on this slide. So the [snorts] caps were one important source of uh exclusion and then of course there's some element of misidentification that inclusion errors and exclusion errors that some people who are not entitled to these benefits they manage to wiggle into the system either through better contacts or through genuine mistakes. [snorts] The other problematic thing about the ID uh people don't have IDs and therefore this is going to be a very important thing is actually quite banal to even

have to say it as I already mentioned to get Aadhaar the government wanted people to come with two ID documents. This is a you had to bring a proof of ID and you had to pro bring a proof of address and then you could enroll for Aadhaar. So if Aadhaar was giving an ID to those who didn't have an ID then why did they have to bring two IDs to get this ID and [snorts] uh so we initially we didn't know uh but I did file an RTI at one point a right to information application with the government and we found that hardly anyone had used the introducer system which is what was there for people without existing docka so it was 0.03% 03%.

Um so even that was it [snorts] was quite remarkable that they managed to get away with that you know bamboozling people with that kind of circuitous uh argument. The second thing that I wanted to deconstruct a little bit is the corruption claim and I'm going to give just one example and go into a bit of detail so that people understand because it is very widely understood that Aadhaar is helping to reduce corruption in these welfare programs. I'm just giving one example but we have done work on other programs as well. Now these are the different types of fraud that exist in uh this public distribution system which provides subsidized grains to a very substantial part of the population.

The first one is eligibility fraud which is basically you're not eligible and you manage to get yourself on the list. The second is identity fraud which can take various forms but the most well-known ones are ghosts where either the person doesn't exist at all never existed and you know somebody is on the list for rations [cough] or it's a dead person someone who did exist but their name continues even after they're dead because the government doesn't have the wherewithal to remove these names or to what do they they call it purification now purification of the list. That purification is now happening with the voter list. The other form of identity fraud is that I'm only supposed to get

one ration card and I managed to get two. Again, through fraudulent means or by accident. The third form of fraud is quantity fraud. is when I go to the ration shop to purchase whatever it is that I'm supposed to get from them, um I am forced because of the power that the person has over me to sign his register saying that I did in fact get my full entitlement of say 35 kilos. But after I've signed the register, they just give me 20 kilos out of that. So that's what we call uh quantity fraud is at the point of sale you get cheated out and but you're forced to verify a full purchase

and the last two are overcharging and quality of grain. one is you know I'm supposed to get it free but he charges me 5 rupees for it and [snorts] uh the quality of grain there's some adulteration that you know in a once 100 kilos of grain you might add 5 kilos of rubbish and just mix it in and so actually you're making quite a bit of money because each person is dealing with quintilles of such grain now if you look at uh whether Aadhaar with what it brings to the table for these different kinds of fraud uh [snorts] with eligibility it cannot help because it's not my biometrics that can you know if I put my fingerprint in a machine it's not going to tell anyone whether I'm rich or

poor need in need of uh this kind of assistance or not it can help with identity fraud uh but there is a caveat for that later if there's time I'll get into it so if indeed the Aadhaar database is dduplicated and it is unique every entry is unique And if your public distribution system is universal and if you ask every person to link their uh Aadhaar number with themselves in the registry for the food rations then you can use it to dduplicate and get rid of this kind of identity fraud. The remaining three you can't again Aadhar can't really help.

You can go to a Russian shop, put your fingerprint or iris can and he can still cheat you out of the grains that you're supposed to be getting exactly as before. And similarly with overcharging and quality of grain. So [cough] the way it was integrated with the public distribution system is that first people were told to get in the queue to enroll for Aadhaar. Once they did that then and they did that mainly because of the arm twisting that I was talking about that they it was an impression was created that they would not get rations if they didn't do it then people were told you have to do this linking presumably for the dduplication of the database and then finally every month

now people have to biometrically authenticate themselves and so what happened is that people fell through the cracks not very surprising uh this is just one per so we did a survey And these are just snippets from that survey. I'll just show you the quantitative results in a little while. So she's a woman who tried several times to enroll uh and somehow she has not been able to do it even though she has all the documents that she needs to have. uh this young girl 11 yearear-old she died uh in 2017 because her family was not successful in linking their Aadhaar to the PDS database and so they were struck off as ghosts in the system

because any uh ration card without an Aadhaar against is is deemed to be a ghost or a duplicate and the roles are purified in this manner and so her family was one that suffered and within two months of them being struck off uh she died of hunger. Uh yet another woman who is now not very mobile and she hasn't been able to get her grains because she can't physically walk to the shop uh and authenticate herself. So in earlier times she would have just asked somebody to get her ration for her. Uh but that would have been seen as identity fraud. You know that person A is taking person B's but actually it's just someone helping her neighbor. um

this person has biometric issues. So he tried uh for 3 months he kept going back and then finally he gave up and he said you know there's no point my fingerprints don't work and so now he didn't uh get his rations anymore. So what we did at that so these are all uh people from the study that we did in 2017 and so we picked uh villages it was a random sample of villages from that had Aadhaar linkage we call it Abba sorry I should have said Aadhaar based biometric authentication is Abha and so we picked both online and offline villages and the key finding is that there was quite a lot of pain transaction costs had gone up for the

people who were being subject to this technology. The transaction failure rate was five times as high in the online villages. Um those who were not transacting was not because they were not interested. Everyone was almost everyone was interested in getting this. In terms of time taken, uh it was as opposed to 80 minutes. Now they were taking 107 minutes. Uh half of them had experienced issues related to the POS machine. um more than half said it had become harder and many of them almost twothirds wanted this system to discontinue.

On the other hand, so you might want to see if you know if all this uh pain is worth it. Were there any gains in terms of reduction of corruption in the system and so for that we look at the purchase entitlement ratio. how much of what they're entitled to were they getting. And what [snorts] we found is that there was hardly anything to show for this. Uh before and after the introduction of the POS machine and you compare online and offline villages, people were getting 93 to 94% of their entitlement.

Um on the other hand what we also found is that the pain was not only for people who were buying the grains but even from the point of view of the seller there were increased transaction costs. Earlier they were able to sell everything within 4 days. Uh but after the introduction of this it had almost double I mean it had become almost 2 weeks. So they were also because the system wouldn't work right because when you go to the shop you have to put your fingerprint in the server the UIDI server as well as the food server of that state has to work and your biometrics have to be exactly as the ones that were stored at the time of enrollment and so you could fall through the cracks at any of these steps. Um so

there is you know often when I talk about Aadhaar and I bring up all these problems with enrollment linkage authentication people will say oh but you know let's try and fix these problems surely they are fixable and I would say yes let's fix these problems if there was something to gain at the end of all this if after doing all this you're still going to get as good or as bad a performance as before why are you imposing this technology uh and it's also expensive expensive. Yeah. I mean not it's not just transaction cost, it also costs money. [clears throat] Uh [cough] the [clears throat] only interesting piece of this whole chain, the only

interesting piece of technology in this whole chain is actually the POS machine, the digital trail that you know shows how much is supposed to have been given because at least that creates a trail that can be audited by people like us later on and that doesn't need the biometrics. It doesn't need Aadhaar. just needs the POS machine. Um but the government was not very open to this. This is more macro data. This is using secondary data sets and it looks at the problem of corruption in the public distribution system over 1999 2000 to 2223.

Uh and what we found I mean this is my earlier work which is that corruption was a very serious issue in the public distribution system. The line at the bottom is the allindia leakage ratios and you could see that they were around 50% uh in the early 2000s and have now declined to about 25%. uh Bihar is the line on top where things were especially bad. Uh and I think even uh the Subashish has actually attended one uh public hearing that we had conducted. [snorts] But you can see that around 20112 there was a sudden improvement a drop in the leakage ratios in the state of Bihar. But most interestingly what from this point of

view is that Aadhaar compulsory integration of Aadhaar came after all these improvements uh were already taking place on the ground. Some of these improvements had something to do with technology but there was a lot else uh going on and post introduction of Aadhaar it's more or less as before it's not I mean this is more macro data so that's why I thought it's interesting to see and [snorts] like I said that once they get it start once they were [cough] [clears throat] getting caught out on these claims they started talking about other benefits of Aadhaar like counterterrorism or elimination of middlemen etc. But in fact, I think you

can catch them out on most of these claims. And so then it raises this question, why were they so keen to create this false impression about the usefulness of Aadhaar on the basis of what I would say are lies at the end of it at least in the welfare domain. Was it data mining? Uh was it surveillance? And I think perhaps from the state point the state's point of view there is uh some interest uh even if they can't say it openly in creating some kind of surveillance architecture and I think on the welfare side perhaps they were thinking that they would switch out of inind benefits to cash transfers and that in their heads the expansion of banking access to

banking would not have been possible without Aadhaar but I don't see why they would think that way that if you know so Indians were able to get bank accounts before why couldn't they get it even afterwards through other means uh through other KYC uh means uh I have quite a lot to say but maybe I should just um so I think one problematic thing is that when the government claimed that they were saving money a lot of the time it was basically people like Santoshi who died and her family who were being struck off these roles and they were these exclusions were being passed off are savings. So even in the pension scheme uh social security pensions same same playbook you know you strike off

names because they didn't manage to link it to Aadhaar and then all the names that are struck off are deemed as savings that have occurred to the state through the purification of these welfare roles. Um and you know the World Bank has played its own role in uh in this propaganda and let me just uh the basically there was a headline saying that Aadhaar could potentially save 11 billion uh if it was to be integrated with India's welfare programs and when we looked for where that estimate came from it turned out it was a footnote and when you go to that document it's a policy brief by an anthrop

anthropologist who says that India's welfare budget is 11 billion that's it she has now not made any estimate of the potential savings and that just the entire budget was put in the world development report as potential savings from Aadhaar it was pretty egregious that they got away but we we actually went after them and they have now two versions of that same report online at least one is the corrected version where actually they do all kinds of acrobatics to um stick with the earlier claim. Um and then of course what has happened over a period of time is that the government has asked people to link it to everything and because there are demographic errors in the database. So

you know raika is different from ra and r i ka space ka is different from r i ka space ka. when they try to link these databases, a lot of people are just falling through the cracks and it's late to led to a big mess including in one case a death when one woman's pension was being credited to somebody else's account because in the process of linking her Aadhaar with her bank account something went wrong and there is very poor grievance redressal. uh people are queuing up all the time and paying money to correct all these errors and now they are also being

asked to every 5 years update their fingerprints etc because they change over a period of time. This is just data from a survey where we talked to people in Delhi about the difficulties that they were facing in um in getting their data uh corrected in the Aadhaar database. So [cough and clears throat] sorry I went backwards. So what [snorts] I'll do is uh I'm going to skip this because there's no time. But taking away from all this and this excessive digitalization digitization of all these different aspects of our life >> [snorts] >> uh to me okay I can see it's a huge convenience in my own life also. But in

this process all of us are leaving data sha are creating data shadows and leaving digital trails and all of that data is very personal and private and can be used in all kinds of ways. So you know what kind of food do I order is it red meat or is it healthy? If you integrate that data with uh you know what kind of job I have, do I have an active lifestyle? Uh all of this is very valuable information for say a company that wants to sell me a health insurance policy. You know, Ritika is has a desk job. She doesn't exercise. She orders red meat all the time. Bad person to

give a policy to. So there is uh this one uh economic philosopher Deborah Sats who talks and there are many but her work is especially interesting. She basically says that some markets are noxious. So think about organ trade, child labor, uh sex work etc. And in this paper that is separate, I say that you know the data uh the market for our private data, our personal data is also a noxious market. It shares the features that she lays down in her book for noxious markets.

And of course she leaves it at that. She says you know if we find that a market is noxious what should we do about it? Uh you know do you want to ban that market? do you want to regulate it? uh you know there are different options but I think for me right now is at least for people to understand understand that there is this noxious side to this kind of uh digitization and creation of so much uh you know so much of our public uh private life being public in this way and [snorts] I feel that even the DPDP or GDPR for that matter the GDPR which is a very good uh law as far as uh data laws go uh has not been very successful.

If you look at the scale of the fines that they have imposed, they are like 1% or less than 1% of the revenues of companies like Alphabet etc. So even if you have a good law in the Indian environment where you know implementation of that law is going to be very weak and people you know to pitch an ordinary citizen say like Santoshi's mother against a golad like the Indian state is not a fair battle at all. So even a good law is not going to be uh particularly helpful. And let me just wrap up even though there's a couple of minutes to say that the Aadhaar project I in my view has both undermined right to life by [snorts]

undermining all these welfare programs and welfare benefits and it's also undermining the right to privacy uh because it's making our lives so public in this way and basically because the law the DPDP law is so weak all our data is fair game for all kinds kinds of uh misuse. Let me give one example to explain what I mean. So in Pondicherry or Tamil Nadu, somebody downloaded one of these loan apps on their phone and within and didn't obviously didn't read the terms and conditions of the loan. I'm not even talking about the privacy uh matters. And so after a little while they started getting

blackmailed. There were a few suicides in Andhra Pradesh because they were being haranged by these loan app companies. So in this particular case what happened is that when they downloaded the app it asked for permissions to contacts and gallery which is a pretty standard thing and once they had that and they started heranging this person the way they started backmailing him or her I can't remember is by morphing these photographs and saying that you know in some compromised position and saying that we are now going to send this to all your contacts and that's when they in this case thankfully they reported to the police but in other cases people have ended up taking their own lives. So

this is the kind of thing you know where you think it's really innocuous uh private it is private on your phone but actually not really. Um so I think yeah I'll stop with that and let you ask questions if you have any. [clears throat] >> So thank you very much. Round of applause. [applause] And so when I announced you and your interesting talk, I totally forgot about mentioning her academic credentials. So Dr. Cara, it's a professor of economics at the department of humanities and social sciences at IIT in Delhi. So that's done for it. So, thank you so much for this interesting talk and you're welcome at a crypto bro

conference of course and it's it's so good to also have these other kind of topics to actually show us what the data we produce or that we work with what that actually means to the people of whom we get the data. So now we still have a little bit of time left for questions. So whoever is interested in in posing your questions please mind you keep the question as short as you would make it if you would have to type it into Mircat so that we have enough time for the answer please. Who has a question here? >> Two of them are from my institute. It's okay. Subhan Rashid.

>> Well, then >> this is very tangential to Aadhaar as a topic but on the graph you showed there was a steep decline in leakages of PDS in Bihar in 2011. Do we know why? Because it didn't seem like it was caused by Aadhar. >> Yeah. Yeah. No. So actually what happened is that in the state of Tamil Nadu they've throughout this period managed to run a pretty good PDS and learning from that Chhattisgarh implemented some reforms and having seen the political gains from fixing the PDS to the Chhattisgarh min chief minister even Nitish Kumar implemented those and so one of there was a whole r it's a battery of reforms one was to expand it

to go from giving only to 20 30% to going closer to 100%. The other was to reduce prices and what that did is it created greater voice from the point of pe people they said you know if market price is 10 rupees and government is giving for nine I only lose one rupee for each kilo I'm cheated off but if market price is 10 and the government is giving it free then for every kilo I'm losing 10. So you know measures like that and the other one was computerization. Computerization just of the database helped to keep track of things much better. Um in Chhattisgarh something they did which Bihar didn't do is they painted all the delivery trucks yellow and so everyone knew very quickly

that if a yellow truck is offloading grain at a random shop something is off. And the trucks also had a toll-free number which people could call. So there were whole battery of measures that help. >> That's very interesting. Thank you. >> All this is in my new book called freebies and rights if you're interested. >> Thank you very much. So thank you for this talk. Thank you for sharing your research with us. And if you're interested in any more questions and answers afterwards or in where you can acquire the [music]