Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=

# The Limits of Government Outsourcing: Property Tax Re-Assessments in India

Authors:   
Discussant: None  
Video: https://www.youtube.com/watch?v=mC4ESywpIDc&t=11928s


## Talk (03:18:48 – 03:28:45)

[03:18:48] >> Now it's time for >> I was >> You're going to be meaner to me probably.  
[03:18:58] >> All right. Hi everybody. So um thanks for having me here today. Today I want to talk about um this paper kind of looks at this classic question for those of us who think a lot about government services.  
[03:19:11] Should the government directly provide the services themselves or should they outsource to the private sector and normally when we think about the theory that we always think about well the private sector is way more efficient but you it's very hard to contract with them  
[03:19:24] on certain aspects that the government might care about. And so the classic example is prisons. Like the private sector might be better at clothing and housing inmates, but you know, civil rights and a few people die along the way. Like they might not care as much.  
[03:19:37] And so we ask a different question. Um are there cases or types of tasks or types of situations where the public sector might actually be more efficient?  
[03:19:48] And so we're looking at this in the context of property taxes in India. And so um in Chennai, India, the government believed that lots of properties were very undervalued. They wanted to go in and do big reassessments um and raise  
[03:20:02] revenue. Um and so uh think about property tax. Basically, you have to do it seems like it should be simple, but you have to it turns out to be a little bit more complex. You have to go in and measure the property area. You have to get permission to enter the building,  
[03:20:15] and it might be multiple units. Um and you need to figure out what portions are residential versus commercial which is taxed at a much higher rate. And so importantly this is a common task done by both the government and the private sector. So you should be able to get  
[03:20:30] these kind of people in because the government you know assess taxes and the private sector will do this measurement for like sales of property or for mortgages or loans or all other things and this is very common to do this kind of outsourcing. Okay.  
[03:20:45] So we studied this in Chennai. Um and sorry I'm kind of I don't have that much time so I'm going through it very quickly but we basically randomized um so Chennai is a giant giant city. It has um wards which are basically  
[03:20:59] neighborhoods. We randomize the wards into whether or not they're going to get assessed by government assessors versus the private sector. And for the private sector we additionally recruited firms under two types of contracts. a fixed contract where they get a payment for  
[03:21:11] each property visited versus an incentive based contract where they get a greater share of the revenue. Um and so the project started within six weeks.  
[03:21:21] The government unexpectedly halted the entire exercise um due to political sensitivities. It was doing very well.  
[03:21:27] It was raising a lot of revenue um as you'll see from the costbenefit analysis. And so but there was a lot of uncertainty around this cancellation on the contracts. Firms didn't know when they would get paid. they eventually got paid um 18 months later. So we're going to understand what happened throughout  
[03:21:41] this whole exercise um measuring everything through administrative data a bunch of and a bunch of our own surveys.  
[03:21:48] Okay. So okay so first key result is that the government assessors were able to they got through more properties um they assessed more properties than the private sector even though the government assess like the government  
[03:22:02] workers had stuff to do so they were actually spending less of their time on the assessments but they actually did a lot more um and this led to them finding um so tax bills were higher so that's tax demand so that's the bill you get  
[03:22:14] the um uh the bill you get but also then tax collection ctions were also higher as well. So this raised um both raised more revenue but the government raised more am revenue than the private sector.  
[03:22:27] Now you might worry well are these government um assessor basically going in and inflating bills so they could get a lot more revenue. So we conducted an independent survey when we sent in our own people to go in and do property measurements and do assessments. And  
[03:22:41] what we actually find is that the government was much more accurate as well. It's just that um they were much better um and part of this is that it turns out they're much better at these more complex properties. They're much better when they're regularly shaped.  
[03:22:55] They're much better at assessing and finding um commercial space, especially within these mixeduse properties. Um when the properties are bigger to begin with at baseline, which is a sign of the complexity, they're also better at finding property and such. And so just  
[03:23:09] overall, they seem to be much more efficient. So then the question you would ask is, well, why? This is kind of counter to the way we typically think about these um things. Okay. Okay. So um  
[03:23:21] I you I hope you believe me. We don't think it's really due to a massive amount of soft information about what properties are higher value. Nor is it about these incentives. So we had done this you know very fancy thing of you  
[03:23:34] know randomizing the the incentive contract versus a more fixed conflict and contract. While it's small sample size as we'll show you, they look basically the same. But instead, what we think basically happened were uh two two  
[03:23:49] sort of major factors why you see the government doing better than the the private firms that were selected into the project. One is sort of broader procurement and contracting issue and the other one is issues around authority.  
[03:24:02] Okay. So first um you know this is I could talk for a very long time about what happened um but you this is a standard procurement procedure that happens all the time follows you know official government of India rules you  
[03:24:15] put out um tenders the government negotiates with the lowest price bidder um there were also there were all these information sessions it was advertised really well all of this lots of firms were potentially interested so about 23 firms attended at least one meeting to  
[03:24:29] understand more about the process but very few firms bid Now, why did very few firms bid? Well, they complained about the low prices.  
[03:24:37] Um, you know, you can see from the graph, the orange um the sorry, the blue is the non- biders and the orange is the biders. Many people thought that they should have p been paid much more to do this at a higher quality rate. Um, and they also complained a lot, which you  
[03:24:52] know, not surprising given that the government canceled the project and can pay people for a really long time. They complain about the fact that oftentimes they didn't get paid by the government in the past and like this unless it's like really worth their while, they weren't going to do it. And so you kind  
[03:25:06] of see that then in terms of the kind of firms that are selected in, they tend to be bigger. They tend to be the firms that work with the government. It might be that these are the firms that could take on the risk of these kind of government contracts. And then as a result, it maybe it's not so surprising  
[03:25:20] that they hired workers that were less skilled than the government um workers, less experienced. They paid them very little. They didn't seem to monitor them much. The private workers expended a lot less effort. They spent very little time at the properties. Um you know, and at  
[03:25:35] the end of the day then it maybe it's selecting in the way the procurement is done and the contract uncertainty is selecting in these firms that are not necessarily doing a a great job. Okay.  
[03:25:46] So the other issue is that let's say the firms did want to do a good job, they were having a lot of issues with authority. So if it were me and somebody came to my house and they wanted to enter my house and I would think it's like a giant scam and I would not let them in. And I think this was part of  
[03:26:00] the problem that the private sector assessors were facing. There was a greater share of refusals. There was more the owners were more likely to say let me, you know, can I negotiate with you on this? Um, and they were very hesitant to go back when there was a  
[03:26:14] problem. Even though if there was a problem, there was a whole procedure to set up to go to a manager and to go to the government and all this other stuff.  
[03:26:20] But it just there's a lot of issues about, you know, this is a task where you're going into people's houses. It's you're you're um determining their taxes. I think uh a certain amount of um the government workers have a certain amount of authority where the private  
[03:26:34] sector guys did not and made it challenging. Um, and I suspect this is probably more challenging for some of the more complex properties. Okay.  
[03:26:43] Oh, okay. So, I'm going to be perfectly on time here. Okay. I feel like I'm talking very fast. Um, so this was wildly widely amazingly um beneficial to the government um regardless of what they did like the amount of revenue that  
[03:26:58] they brought in relative to the cost of this whole exercise especially, you know, since you know um I mean with the private firms they didn't pay them that much. you know, you can they can make a huge amount of money off of this. Um, in the end though, because the government was much more efficient at doing this,  
[03:27:13] you see much higher returns of of keeping it inhouse than if you g sent out to the private sector. But you could also imagine there's different cases.  
[03:27:21] So, for example, maybe the government workers don't have time to do all of this because they've got other tasks to do. You might say, well, the government workers are much better at these very complex cases. Um so if you allocate them like the larger properties at baseline um versus you know for the  
[03:27:35] smaller kind of less complex properties the government and the private sector tend to do a little bit um they're they're more similar. Um you could actually just get a much bigger revenue increase by just efficiently allocating  
[03:27:47] which property to who. Okay. So okay so you know as I said you know a common view is that the private sector is more efficient and there's all these you know people might die in the prisons or something as a result of that but in this case we actually find sort of the  
[03:28:02] opposite that the types of firms that are kind of attracted for these kind of contracts um tend to be less efficient than the government workers um and we think it's not just like a band-aid of like adding in a little bit more of an incentivized contract. It needs to be  
[03:28:17] something larger about how we do procurement and contracting to be um to be thinking really about how to improve um the kind of private sector firms we have. Um and that also you know maybe there are certain types of tasks where  
[03:28:31] it is very adversarial with citizens or where you really believe somebody entering your house is going to be a scam whereas you know this is why the government does these tasks to begin with and maybe they just have an advantage in doing that because of the nature of these kind of tasks. And am I