Ep 69: Low-Income Housing and 'Crowd Out' with Michael Eriksen

Subsidized affordable housing development reduces costs for lower-income households directly. It also reduces costs indirectly, by increasing the overall supply of housing — or does it? Michael Eriksen joins to discuss the issue of “crowd out” in affordable housing production.

Show notes:

• Eriksen, M. D., & Rosenthal, S. S. (2010). Crowd out effects of place-based subsidized rental housing: New evidence from the LIHTC program. Journal of Public Economics, 94(11-12), 953-966.
• Cummings, J. L., & DiPasquale, D. (1999). The Low‐Income Housing Tax Credit: An analysis of the first ten years. Housing Policy Debate, 10(2), 251-307.
• Click here for Pathways Home, our eight-part series on homelessness.
• Baum-Snow, N., & Marion, J. (2009). The effects of low income housing tax credit developments on neighborhoods. Journal of Public Economics, 93(5-6), 654-666.
• Diamond, R., & McQuade, T. (2019). Who wants affordable housing in their backyard? An equilibrium analysis of low-income property development. Journal of Political Economy, 127(3), 1063-1117.

Transcript

Shane Phillips  0:05  
Hello! This is the UCLA Housing Voice podcast, and I'm your host, Shane Phillips. After a short break, we are back with Michael Eriksen, joining us to talk about the Low Income Housing Tax Credit, and some of the pros, cons, and nuances of it and other supply-side housing policies compared to demand-side affordability programs, like housing vouchers. LIHTC, as it's commonly known, has been America's largest source of subsidized affordable housing for the past 35 years, so we're excited to finally spend a bit more time talking about it on the podcast. 

The research angle in this episode is something I've been interested in for quite a while, which is the degree to which publicly funded housing development adds to the overall stock of housing, and how much it simply crowds out privately funded unsubsidized development. 'Crowd out' is a real phenomenon, that diminishes the return on and benefits of public investments of all kinds, including housing, but it can also be invoked as a scare tactic by those who simply don't support government spending, regardless of its benefits, and that can make it difficult to know how concerned we should be about it in any given situation. What we learned in this conversation with Professor Ericksen is that crowd out is essentially inevitable, and that some amount is probably good. We also learned that depending on the geographic scope of the analysis, subsidized units may crowd out an equal number of unsubsidized units — and that would not be good. It would mean we're spending billions of dollars and missing out on one of the biggest benefits of supply side policies, which is the indirect benefit of lower rents for people who don't live in those subsidized units. But there's good reason to believe that when we zoom in to smaller geographies, or when affordable housing is more deeply discounted relative to market rate prices, the crowding out effect isn't nearly as strong. Those findings may tell us something about where and how to prioritize our affordable housing investments, and that is very valuable information indeed. In this episode, we'll get into all that and much more, and in much greater detail. This is just our first in-depth foray into the Low Income Housing Tax Credit, so stay tuned for more in the future. 

The Housing Voice podcast is a production of the UCLA Lewis Center for Regional Policy Studies, with production support from Claudia Bustamante, Jason Sutedja, and Gavin Carlson. If you have a question or show idea, send it to me at shanephillips@ucla.edu. And if you liked the show, give us a review or a five-star rating and tell your friends about it on social media. We really appreciate it. With that, let's get to our conversation with Michael Eriksen.

Michael Eriksen is professor of economics and director of the Dean V. White Real Estate Finance Program at Purdue University’s Daniels School of Business, and he's with us today to talk about the Low Income Housing Tax Credit and the extent to which LIHTC projects crowd out construction of unsubsidized rental housing. Mike, thanks for joining us, and welcome to the Housing Voice podcast.

Michael Eriksen  3:22  
Thanks for having me. I appreciate being here.

Shane Phillips  3:24  
And my co-host today is Mike Lens. Hey, Mike.

Mike Lens  3:27  
Hey, Shane. Good to see you, Mike, as well. I was a little nervous over the weekend when I thought about doing this paper on a podcast, until I realized that I've regularly been teaching this paper in class, including today. So I'm good. All the listeners out there, I'm prepared, even though, you know, a few days ago, I didn't realize I was.

Michael Eriksen  3:49  
Good to hear. 

Shane Phillips  3:50  
I mean, this is a good sign. It's a classic. 

Mike Lens  3:52  
Absolutely. But Mike is not old.

Shane Phillips  3:57  
All right, so we always ask our guests for a tour of a hometown or another place that they've lived and want to share with us and our listeners. Where did you pick Mike?

Michael Eriksen  4:05  
Yeah, so I'm originally from Redmond, Washington, which is a natural choice, that's home of Microsoft. And like my father's optometrist, my mom's a nurse, kind of kind of see that city grew up really through the 80s into kind of the tech hub that Redmond Washington kind of Seattle became which is quite fascinating. But the area like I really want to describe which I actually think guys my research better is that as an undergrad, so I had a chance to go to the Indian reservations, like the Spokane and the Colville Indian reservations in Washington State. And what kind of got my desire for this topic was in those communities just kind of see, they're really, you know, less than an hour from metropolitan areas. You really got to see differences in poverty. You know, a lot of times I think we think about their inner city poverty and we think you know, what that looks like in a lot of those communities, you know, as you drive into town, that absence of structure would be apparent, you would see some more like dilapidated housing was always interesting to me, it was really, you know, a lack of science and kind of when you thought which came first, right. And when we think about literacy and some of these challenges, there was very little Wayfinding, you would be hard to ever know you're on a reservation. But it was really a unique opportunity, because I got to work with really high school students to just convey my own love of science and math. And, you know, a lot of these groups, right, I had never really met people that went to college, right, maybe except their teachers, and I ended up kind of an economics program was kind of looking at these areas. And I think what was unique, and I think somebody to learn from is, you know, increasingly these areas due to gaming or to court, like the settlements, they had resources, right, but there still wasn't a clear, they were essentially still in poverty. And it really, I think, you know, as we think sometimes in our fields, right, it's not just about money. I mean, money thinks, you know, some problems go away. But there's really a culture and kind of what what we think about where we live and grow up, that are going to has really, like informed my life today.

Shane Phillips  6:05  
I actually, I spent a few years in Redmond in I think, the 90s When I was a kid, and one of my grandparents lives on one of the reservations outside of Marysville Washington. And they're not indigenous, it was just kind of an open area. And it was it was much more suburban before transitioning to more of a rural area. But yeah, it's kind of interspersed all throughout Washington state. You were talking more about Eastern Washington, but all over the place on Western Washington do I mean, Seattle itself is named after a chief for the tribe there? Absolutely. So this conversation is going to focus on a study that you published with Stuart Rosenthal, in the Journal of public economics titled crowd out effects of place based subsidized rental housing, new evidence from the light tech program. The study actually dates back to 2010, almost 15 years ago now. And you were saying this was initially being written way back in 2006, almost 20 years ago. But in some ways, this is just an excuse, I think, to talk in more general terms about the Low Income Housing Tax Credit, usually referred to as light tech, because we've yet to have an episode dedicated to it, despite doing almost 70 now. Wow, is that true? I think so. Yeah, we've, we've talked about now we've done like one or two on vouchers, we really haven't done one on light tech. Wow. Okay. But this is this is a program that's been around a while since its inception in 1987. passed by in 1986. Tax law, or implemented through that, light Tech has been the largest source of subsidized housing affordable to low income households in the US, producing an average of about 100,000 rental units each year. Part of the reason we chose this particular study to discuss it is because it's addressing the very interesting subject of crowd out, which is the idea that public sector investment can take the place of our crowding out private investment, rather than complementing and adding to it. Credit concerns apply to a whole bunch of areas really just about anywhere where the government makes investments. But we're going to be sticking with housing here. And I'm going to give my pitch for why policymakers should pay attention to it why the audience should be interested in this. So governments can help low income households afford housing in many ways. But when it comes to the uses of public funding, the different approaches can all basically be categorized as either demand side or supply side subsidies. demand side subsidies increase the financial resources of people looking for housing. And in the US housing choice vouchers that pay a portion of rents. For some not all, very low income households are the most common form of demand side subsidy, supply side subsidies lower the cost of housing directly. And that's often done by subsidizing a large portion of the cost of developing new homes. Freed from having to earn enough to pay off the full cost of development, the units can be rented or sold at below market prices. So that's how light tech works in like a very small nutshell, there's there's so much more to it. But one of the main arguments in favor of supply side subsidies, like light tech, and one that I make in my book is that they sort of do double duty. On the one hand, they make housing affordable to low income households just directly by providing it to them. And that's true of demand side programs like vouchers in sort of a different way. But on the other hand, supply side subsidies also add to the overall supply of housing, which can have a stabilizing effect on prices across the whole housing market. Vouchers only help people pay for housing that's priced below the median rent in an area. And that amount of rent is usually not enough to cover development costs. So it's more difficult to make the case that those kinds of subsidies increase the overall housing supply, that they're somehow motivating the Construction of additional housing that wouldn't be built otherwise, we can talk about maybe some nuances to that. Sure. But all of this hinges on the idea that publicly subsidized housing construction is additive to private unsubsidized housing construction, and if subsidized development is crowding out private development, so that the total supply of housing is no greater or not much greater than if you were to eliminate these construction subsidies. Or if you just moved over the funding to demand side subsidies, like vouchers, then the case for supply side subsidies, I think gets quite a bit weaker. Mike and Stuart, you found that there may be a lot of crowd out happening with the light tech program. But my read of this work is that there's plenty of room for nuance and for different interpretations here. So that's what we're going to get into. So Mike, with all that said, Before we dive any further into your study, let's just take a step back and talk about the Low Income Housing Tax Credit and its history. The program was created under the Tax Reform Act of 1986. And it was implemented the following year, we've been building subsidized housing in the US for much longer than that, though. So what preceded lie tech and what motivated Congress to adopt this new approach and 80s.

Michael Eriksen  11:16  
I mean, I think it's always important before we talk about any programs, kind of understand the context. So predating this was this kind of idea of public housing. And I think the context of that was really the great like depression, the 1930s, there's really a, you know, a growing amount of slums. And I think a lot of I still think was well intentioned and actually more effective than people think was, we really did a good job of clearing some blight and giving opportunity to people. Really, though, when we talk about traditional public housing, it really is not one program, it's probably at least 37 different programs. The idea there, and I think, you know, when the government is the developer, they decide, you know, where it is, how to construct it, how capital intensive how tall buildings that are, and then how to operate it, what rents do we charge, right? This is sort of this idea that where the government really pays for this initial like that, you know, like an outlay for the expenditure. And kind of the idea, and I think this is, again, pretty well intentioned, it wasn't necessarily a super use exclusively a low income housing project, at that point, it was going to be much more of building closer to market with the idea that there's capital constraints, the government has a low cost of capital, they can, you know, tax us do bonds, you know, go into debt, but that would be a lower cost and a private developer and the capital they would have to have. So we would push out the supply would push down rents. And again, I, I personally feel traditional public housing is somewhat got an unfair set of circumstances that made it less successful. But ultimately, I think, you know, a big growth in the 60s. But ultimately, I think, a series of decisions that put income limits on it, and the income limits themselves restricted the rents, which basically made them where we couldn't only build the project, but we couldn't maintain them for the rents they collected as well. So we kind of knew this, in the 70s, we started to see this, really, there's been no new public housing since 19. Like the 80s, as we know, you know, there's been a lot of tear downs in the 90s, what the immediate replacement was, the 70s was what we call kind of housing vouchers today, or housing choice vouchers. But kind of the new idea, I think with vouchers, there was, again, context matters. In the 1970s, there was kind of a hollowing out of America's cities. And there's a lot of vacancy. And if anything, not only was there kind of decreased demand, but at the same time, insufficient incomes, all those things that that we want to do so at least in the 70s. And I think even maybe up until now in some instances, it's really just, you know, a voucher could kind of supplement that demand, there's existing units, existing vacancies, and it's far more cost effective to basically put houses into those existing than building new, something

Shane Phillips  13:50  
I hadn't really thought much about, but is obvious. Now, when you say it is when vouchers were kind of growing as as an approach for subsidizing housing. This was a time when we were just building a lot of private rental housing through the 70s 60s 70s 80s, that just a lot of housing was being built nationwide. And the problem was not, you know, high rents and low vacancies, it was high vacancies and people not being able to fill up these units. So I can see why demand side subsidies make a lot more sense in that context. We just happen to not be so much in that context in many places these days. Yeah,

Michael Eriksen  14:26  
that context matters. I think a lot of ways. I think the context, the tax credit program, this kind of in these two different programs. How does the tax code benefit rental housing, and really starting 1982 The Economic Recovery Tax Act passed these generous like depreciation allowances didn't limit the passive income and a lot of things that made including my own father, in like Seattle, the 1980s become a landlord. Right, he purchased some properties for the tax purposes and my brother had no comparative advantage. I think you'd even say that now of really owning these properties. And, you know, ultimately, when end up happening was they started to realize this created some kind of weird incentives. Again, my father's optometrists, he should be spending more time seeing patients and trying to be like, you know, like a landlord. So really the Tax Reform Act of 1986. At this point, it was kind of well known that we were going to remove the subsidies for rental housing. And by the way, if you look at the numbers, you know, the 80s was a great time for rental housing in general. And I think what people knew, which I think was true was starting in 1987. With the new tax law, there's gonna be this huge drop off and rental housing, and people were really worried about affordability. So of all people, Senator Bob Packwood, he was infamous for other reasons, created the Low Income Housing Tax, really, as a temporary solution. It really had, you know, a sunset in it, it was a temporary outlier, kind of per person. And it really, it wasn't until the 1990s, that it really became permanent. But even then, it wasn't overly generous. You know, the number of units that supplied in the early 90s, late 1980s was pretty small. But a group that, you know, a lot of my study really looks at the 90s, which is a growth period. But we really significantly increased per capita starting in 2001. You know, a lot of late 2000s was very kind to the program. But most recently now, right, we're really thinking about almost quadrupling, the subsidy, bipartisan support. So you know, I think this podcast is really timely. And suddenly, because I think this is the top of a lot of people's conversations of, you know, should we increase it? Should we decrease it, and I look forward to talking more about it?

Shane Phillips  16:27  
Yeah, I'm gonna actually put in our show notes and article from 2010 by Jeanne Cummings and Denise de Pasquali. It's titled The Low Income Housing Tax Credit and analysis of the first 10 years. And I remember reading this several years ago, sorry, this is published online in 2010. But it's originally from 1999. And just reading about how people thought about it at that time, especially, you know, when it was initially temporary, when they just thought this was going to be regular private developers taking advantage of this. And it's become almost a whole separate industry of development with very specialized developers. That's just not how it was really conceived of at the time. So definitely encourage people to check that out.

Michael Eriksen  17:12  
Know, for sure. And you know, and I think, you know, something, which is important, as context is, one of the great things about the Lankan housing tax credit is we have some of the biggest corporations in America, suddenly very interested in housing like provision, right? If you look at who's absolutely lobbying, to increase the tax credit, right, it's some of our country's largest banks, some of the wealthiest people, and it's really created this powerful lobby, which is looking out to supply and increase the supply of affordable housing, actually, some sense think it's a feature as much as the glitch. So

Mike Lens  17:42  
yeah, it's a bullet proof policy right now politically, in some ways, right, which is totally wild for housing subsidy in this country.

Shane Phillips  17:50  
Yeah, yeah. I mean, it's it's very frequently criticized for its structure, the fact that it's a tax credit means there's a lot of middlemen involved. You know, it's everyone gets to take their cut, including the banks and the syndicators and everyone else as it moves to the process. But it has also created this lobby that has maintained it, whereas other programs have faltered, have gone away, have been cut back. And like tech just keeps growing. And at the end of the day, it is building housing for low and very low income households. But maybe the impact is not quite so positive or great as we imagine. So I want to give you Mike the chance to say a bit more about crowd out beyond my very oversimplified explanation, in the case of housing, like tech in particular, but maybe just housing generally, what is the mechanism by which subsidized housing is believed to or hypothesize to crowd out private on subsidized housing development? Yeah,

Michael Eriksen  18:51  
so it kind of comes from two places, as in the literature, and I think what's important about this study is you're really focused on credit. There's a lot of other benefits that, you know, we will talk about, but kind of a fundamental issue of the local housing is is it actually increasing the supply of units, and it's not just public units, but public and private. And the essence of the problem is, from more of a demand perspective, is substance. There's only so many households out there. And I think at the most fundamental level, if we're building units were would have those households seek housing otherwise, and as economics are kind of viewed as aggregate demand shift, we're kind of pushing the market down. Now, again, it sort of depends about the characteristics of that market. But I also think there's a fundamental issue of developable land. You know, if we really look at crowd out a single parcel of land, you know, if the government's going in there and building something, that means for sure, right, at least for the next 3040 years, the private market can build on that as well. So the two different dimensions kind of have a supply demand sense. Which begs the question, okay, well, when we build 100,000 units a year through the taxpayer program, how many net units do we have another way to view it is How many fewer units were both developed or maybe demolished. After we build this program,

Mike Lens  20:05  
that certainly helps frame this credit challenge for me. And I guess, thinking about the role of the subsidy. So without the subsidy, it's unlikely that developers are going to be able to build housing and be profitable if it's targeted towards people below a certain income range, right. And so that additional supply, if it's targeted towards the lower income range, I think, as Shane pointed out at the outset, that could have two positive effects, right, the one positive effect is that you're producing something that the market is not going to produce for lower income households. And that's going to directly benefit people an indirect effect. The second piece might be that we're raising overall supply. Unless of course, there's this like big crowd out, or if this crowd is close to 100%. And I think what we're getting to is like, there's always some crowd out. So it's not necessarily whether there's crowd out, but kind of how much crowd out there is, again, going back to this first piece of the crowd out, like, we might be happy, or some people might be happy about some level of crowd out because you're making a city or a housing market more affordable to lower income people directly. And whether or not you like or dislike private development, you are favoring lower income housing over broader income housing. But the question, of course, is really how much of a program like live tech is crowding out private development. And if every 100 units of subsidized housing crowds out 80 or 90, then we might be paying a lot of money, and not getting that supply benefit. But if it only crowds out only 20 or 30, on side subsidized units, you know, then we might be getting quite a lot more overall supply, in addition to this more targeted supply to lower income housing. So is it fair to say crowded can be too high and too low? And there's maybe some sort of sweet spot that we're looking for? Yeah, no,

Michael Eriksen  22:09  
I mean, I think that's actually, you know, a sophisticated viewpoint in the sense that it really comes down to what are we crowding out? You know, the earlier studies looked at the national level, like the state level. And again, I mean, in some regards, and I think you could say this, that there is some probably core inefficiency, and it's potentially quite large, about producing new units. Now, if we're producing different types of units, then what the private market would have provided, if those units are in different neighborhoods, and the private market wouldn't provide it. If maybe, you know, it's closer to transportation and jobs. If it's at different rents, you know, they look back at this result as well. It is where is that crowd out? Right? There's a geographic dimension as well. And I think it's somewhat naive of me as a PhD student, almost 20 years ago, as economists, we always try to create one estimate. And I think my view would be is it really depends what geography you look at. Right?

Mike Lens  23:03  
Right. Yeah. And

Shane Phillips  23:04  
I just want to clarify, sort of why we think some amount of crowd out is good. And tell me if this is a correct interpretation. Essentially, if you're building more housing than the market would otherwise provide, you're hopefully bringing down rents across the market. And so as a private developer, if you're seeing rents fall, rather than climb or climb more slowly, you're less likely to build something in the future, because you're worried, you know, it's just not going to earn enough profit to be justified. So you've built these subsidized units, you've brought rents down across the whole housing market. And that necessarily is going to reduce private construction, it's sort of the cost you pay for building more than what the market will build and bringing down rents. I think this is really what we're getting at when we say some amount is maybe not only inevitable, but desirable. It's just once it gets into those higher ranges, you're not getting the supply side that indirect benefit much at all. And

Michael Eriksen  24:03  
Olson, I think, has really done some excellent studies looking at cost effectiveness. And the GAO has looked at this as well. And I think we're gonna look at all these programs. And I think, again, this is what I would say, in some sense, our studies too narrow, right, we need to look at kind of the all encompassing 30 years of housing, you know, how do we actually for the targeted recipient? How do we decrease their housing costs? One of the core things, is there some housing features that we want these households consume more of? In some markets? You know, I really think the one because if tax rates very effective at doing what the private market would ever do, and again, that should be celebrated. Maybe one thing I want to add is, we always talk about the targeted subsidy. And you know, again, one of the features I think, as long housing packs, right is as we expand out the supply of housing, it's not just the tenants in those units that benefit. It could be the person across the street, it could be rent markets expanded a little, maybe there's this kind of spillover benefit, which I think is important. The nuance of crowd out if we Look at long enough is, depending on the market, it could very well be that now that landlord abandons that unit because rents are no longer enough or above kind of operation costs and property taxes. So.

Shane Phillips  25:11  
So a really important point that you make in this study is that the amount of crowd out depends or we believe it depends on who the private unsubsidized and publicly subsidized developers are providing housing for, the closer those two markets are, the smaller the distance between rents in one market and rents in the other, the more likely you are to see crowd out. So could you say more about that, and why you and Stuart believed that Latech might crowd out more on subsidized housing production than earlier publicly funded housing programs? Yeah,

Michael Eriksen  25:43  
I mean, I think an important part whenever we look at this, and you know, the earlier studies by Michael Murray and others who are retargeting by the program matters. And I think it's important to think about what the supply elasticity is actually happening in the markets. So in some markets that are more highly elastic, we really think about Cleveland, Ohio, I think being a great example of this, there's sort of a fixed number of households, it's very different than then, let's say Los Angeles. But ultimately, I think what's important with a tax payer program is, is that when we talk about, it's being targeted to you know, low income households, the metric use there, 60% of the area median income. And I think this number is kind of abstract for a lot of people. But for 2023, a three person household in Cleveland would have to earn less than about $48,000, to be eligible, which, which isn't necessarily, you know, that's not the homeless population, the same time, the rents being charged with just the tax rate, it would be $1,200 a month. So the person that the margin who's benefiting, like, you know, this is our teachers, or firefighters are probably not our police. But for the most part, right, these are people that they could afford a house and a rental house and the absence and now we're getting, there's still an opportunity that we can shift where these people live, we can have them consume a different type and quantity and quality of housing. Now, there are situations where the homeless can live in these. And I think, you know, that's, again, really an important point is what I discussed was the 60%, kind of the maximum. But in most states, there's a competitive bidding allocation, that bids are considerably lower. And again, I think lower,

Shane Phillips  27:17  
meaning you have to rent it at lower prices to people who have incomes that are even lower, maybe below 50%, of area median income are below 30%, that kind of thing. Yeah,

Michael Eriksen  27:27  
for sure. So like, you know, like the California, which is really kind of my dissertation I looked at, when you really look at the competitive process, right? Somewhere around 40 to 30%, is kind of, you have to bid your rents down there. Now, in Cleveland, at least, you know, somebody who's earning 30 percents, that's probably about 25,000, which is probably actually is probably that it'd be tough to live in Los Angeles, or even Cleveland, again, maybe still not homeless, but maybe more in a crowded situation, a less, you know, favorable housing situation. And I think that's an important feature as well, that, you know, as we look at this program, what are the maximum incomes? What are the maximum rents? And then what are they in reality once they get bit down? Yeah,

Mike Lens  28:05  
and I just wanted to add a little bit there, as you connected this to housing for people experiencing homelessness are first off a plug for our pathways home series that we just finished recording. This is our first episode, not on homelessness for quite a few. And we also will probably talk about how the live tech program has evolved over time. And we might get into how it does play a role in helping to create housing specifically targeted for the homeless population. You know, like in California, and specifically in Los Angeles, light tech becomes a necessary layer of financing on top of, you know, some more direct targeted subsidies. You know, that's another way that light tech gets used for a even deeper subsidy or even lower income population or more housing insecure population, for sure.

Shane Phillips  29:01  
I agree. And so just to like, state this in a few sentences, am I correct in saying that, you're more likely to have crowd out in markets where the gap between the rent of a new like tech unit and the rents in an older apartment building across the street are more similar? Because they're just more direct substitutes for one another? Yeah,

Michael Eriksen  29:23  
no, I mean, I think I'd probably say more about the, what the housing characteristics are as well with elasticity. As we know, in LA, the you know, the counterfactual is often no new housing. So in some sense, we would at least expect in Los Angeles to really supply and elastic markets that we really would see the lowest amounts of crowd out.

Mike Lens  29:41  
Right. And I think this this connects to a couple of fairly controversial concepts in in the housing discourse, right. You know, first is this idea of market segmentation, right? That even in the same metropolitan area, like Los Angeles, the, you know, high income market are the new housing the luxury housing market is completely separated from the lower income housing market, that they're very distinct things, right. And so to somebody who is hearing this and thinking like either, I want my life tech to completely crowd out the private market, or I don't care if like tech crowds out the private market, they're likely to say, Oh, I live in Los Angeles, where the only new housing is for rich people. And so it's completely separate from the housing that is built for lower income people. And that's why we need like tech. And that's what that's doing. And then, you know, of course, that gets into, I think, at least a second concept, which is filtering or the research that we've talked about on this podcast with folks like Evan mast on migration chains, which to dumb this down and be very quick is like this growing evidence that when we build housing, at higher income levels, people are more able to move into the head, there's a new price range for them at these different levels of the market. And that just tells us that these things aren't perfectly connected maybe across high and low income, but they are more connected than we might think. And so in that respect, people should think that crowded is more relevant than maybe they do. And I

Michael Eriksen  31:20  
think, you know, this idea of filtering down and, you know, certainly controversial, right, you know, let them eat cake, you know, definitely controversial, what I'm increasingly worried about, and I think, Mike, you started to kind of get it this is, I think we definitely see filtering historically, what I'm always curious about, to what extent are we putting kind of unnatural impediments to filtering, you know, again, if the tax audit program, I just clearly think the more that we target lower income households that would have never rented and be able to afford a unit, I think it's more effective. And again, the thing that I keep thinking is, you know, housing services, few private developers I've ever met have decided to just voluntarily put housing was service designed for a daycare or larger units, some of these things that we really could kind of be a heavy hand, and make sure are supplied, that might not come up from a unit count, but most vulnerable population can benefit.

Shane Phillips  32:13  
Something I hadn't really appreciated until the last few years traveling around to some other states is, I have a very California centric perspective where I think of light tech as being this like very deeply discounted double market housing, but you go to other markets around the country. And the gap between market rate for even a new apartment and the below market price in a light tech building are very similar to say nothing of the rents in an older, you know, 40 year old apartment building versus a new light tech unit, the light tech might actually be more expensive by some meaningful amount. And so again, I think it's just, I want to make sure people are aware that that dynamic exists, that this is going to be very context dependent on the market itself. So let's get into the design of your study. And I gotta say, the study design and results in here in this article, very tough to parse for a guy without a PhD, aka me, and as neither an economist nor a statistician, so if I can ask you to put on your econ 101 undergraduate lecture had not the not the 310, or 501, or anything, not the PhD lecture course. Could you explain your approach here to measuring crowd out by light tech developments?

Michael Eriksen  33:30  
Yeah. So I mean, I think there's kind of two innovations, existing studies kind of looks either at very highly aggregate kind of national level, you know, ignores these kind of micro movements that I think we have. So you know, previous studies you're talking about? Yeah, yeah, that's a previous study. So So one of the things that we did when we started do this was started thinking about the what's the geography of crowd out. So if we're at the most highly aggregated, it really comes down to how many new households are created, because the results of this program, that is the Michael Murray is viewpoint, that even if we get the state level, I mean, again, we see almost near 100% crowd out in those regards. We should expect higher crowd out the higher we go on that

Shane Phillips  34:09  
idea that as you zoom out further, the crowd out grows larger. Can you explain that? Because I'm the way I've been thinking about it, this is almost the opposite, where the further you zoom in, the more certain credit becomes down to the parcel level where if you build a light tech subsidized development there, you by definition, cannot build a unsubsidized project. So what am I missing in the kind of space between the parcel and the whole state or the whole country? Yeah,

Michael Eriksen  34:38  
I mean, I kind of think of it like a you right? So when we talk about the national level, again, it really comes down to how many new households that this program create, and to the point that we're kind of doing low moderate household income, right. From a national sense, we would expect, you know, let's assume everybody in the program would have rented a house somewhere, right? You can imagine a situation where there's a lot of households trapped in a bad neighborhood, even maybe a job market without opportunities, if we build more units and kind of one part of the city, right, or one part of the country, there's this fundamental shift, right. And I think this kind of gets to this fundamental identification problem, which is that ultimately, you know, although there's a good share of nonprofit developers here, there's a lot of for profit motives directly or indirectly thinking about what happens to these units in the future. Right. So if I'm a developer, I want to make sure there's sufficient demand for my units and Shang, the point you made and, and a lot of housing markets taxed at rents are actually above market. So now these could be higher quality they could have, again, there's different features that we talked about are important. But developers again, I just don't think they're randomly picking neighborhoods. And I think if you, if you would look at this, and I think this is one of the reasons why we might kind of erroneously see to less crowd out, as I'm trying to say is, the developers are purposely targeting neighborhoods or parts of the cities where more private market development happened within the absence of the program, right? It's gonna look like there's less crowd out and displaced than there actually is. So if we just take kind of a naive view, and just say, Okay, well, let's just plot in this part of the town, this I think, one mile ring, how many tax rate units there were, compared to market unit, we see about a 30% crowd out, but we're curious again, is Okay, is there something unobserved and kind of thinking about what would have happened counterfactual without this and kind of our innovation, there is basically a political instrument of who voted for the sitting governor in 1988. It was it was something you know, we're always throwing straws as academics to try to show causal effects. And, and this one surprisingly, worked pretty well. The story they're important in this context is when the program was created in 1986. And when he 87, a lot of states had to think from nowhere, what exactly we would do, and state capitals tend to have a lot of low income housing tax credit population. So this is something if you if you just look at cities across the country, Columbus, Ohio, Indianapolis, Raleigh, right? I mean, all these places, even Sacramento, right tend to have more per capita, and kind of drawing on that intuition. When they set these geographic set asides and ADA and kind of figured out, well, who's gonna get how big of the subsidies, we at least found, if you're a city, or you're part of the city, voted for the sitting governor, when they're doing this in the 90s, you receive marginally more tax allocations than if you did it. And again, that's sort of the intuition create this word, say, exogenous shift. And ultimately, at that point, you know, we almost found a one for one decrease in the amount of new units and demolitions, at least of those areas that have this exogenous shift at the margin. And

Shane Phillips  37:49  
you were talking earlier about some of the earlier studies looking at this either in like very large aggregate at the state or national level. And also, you mentioned sort of a one mile radius around a project location or census tract or what have you. Your analyses at the geographic levels you did were sort of in between each of those, right? It's like you did all the way up to the MSA, the metro area level, also did the counties which in many cases are smaller than the MSA, and then also a 10 mile ring around every census tract within these locations. Right. So those are three different levels of geography you're looking at. Yeah, yeah, for sure. And so moving on to the results here, let's talk about what you found. I think we should avoid worrying about regression coefficients and just translate the findings into percentages, because I think that's, that's ultimately what connects with people. So at different geographic levels, what share of low income housing tax credit development was being offset by reductions in private on subsidized housing production? Yeah.

Michael Eriksen  38:50  
So I mean, I think in most of our results, or at least the results that, you know, I tend to think about as I think we generally find kind of a lower bound without controlling this political shift share of about 30%, which means for every 100 tax rate units, we get about 30% class private market units, generally, the earlier literature at the state level, you know, really found closer to 100%. And as we kind of got larger and started checking about this, and again, we we had more refined data, more specific periods, right, we found something closer that was closer to 100%. Once we kind of grew big enough now an important caveat. And I think there's a pretty large confidence interval, or so kind of saying that's the average for every market in the country. We are discussing good reasons why it could be larger or smaller. What's unique and kind of, you know, helped our study, at least a little bit is tax units are really built on a per capita basis. So we aren't just looking at California, we just start looking at it one or two big cities. But because of that, then we might maybe overstate the size of the crowd out maybe in those markets as well. So I think it's summary. It's a relatively large and I think it's important to realize and I think have some of the conversations of what our study would suggest. She

Shane Phillips  40:02  
said a lower bound of about 30% and an average of about 100. I hadn't thought of this while reading the study, but is over 100% crowd out impossible, like in theory where, you know, you build 1000 Light tech units, and instead of building the 5000, on subsidized units, you had the previous five years, you only build, you know, 2000. So you've you've lost 3000 on subsidized units and only built 1000 subsidized or is 100%, just the upper bound for one reason or another.

Michael Eriksen  40:33  
When I talk with developers and think about one, I tend to think 100 is the upper bound. Now, again, thinking about if it's 30, you know, what's the correct verse and Drax one of the kind of indirects that I could at least think of why it could be larger than just the pure offset is, again, we're taking developable land, or maybe the private market would have built something different on that lands. You know, one of the challenges, I think, with low income housing tax credit. Now, I think the research probably suggests the opposite, is a lot of times we're fighting hard to get it approved. So it could be that a private market developer would come in, and they would build a seven storey tall apartment building. But if we're really worried as a tax rate developer, you know, maybe we only build a three storey tall building, right, because it's tougher to get through the regulatory process. Now, in some arena, I think the results suggest the opposite. I think, if anything, the subsidy really is as a percent of construction costs, and the development costs. So if anything, I think and I think Bri Lang has some great stuff, showing that we actually probably build more capital intensive projects, you could make an argument, uncertainty, developers never like uncertainty, you could make an argument again, that we have different effects from the supply side, it's not just the aggregate demand shock, but there's only you know, so much lumber in the metro area. And somehow, right, we're increasing the burn unit cost of, you know, plywood, right? We can This is where we can get this kind of multiplicative effects. I think they're unlikely. But still, does it mean that they could exist?

Shane Phillips  41:55  
Yeah, yeah. Well, I do think, you know, something I hadn't considered also, I mentioned how, when you're subsidizing these units, and adding to the overall supply, at some level, you're reducing rents, market rents relative to what they otherwise would be. But you just brought up a valuable point as well, that you're also using resources, labor and materials and other things that may be limited in some sense. And so maybe over longer periods of time, they can adjust upward to account for that new investment. But in the short run, maybe not so much. I would agree. Yeah. So if journalists were writing just one line about this study, I imagined it would be something like, researchers find that low income housing development crowds out 100% of Private Housing, Construction, eroding benefits of government investments, I would like to hear you complicate that a bit. Since it's definitely not the whole story. It's maybe part of it. As you know, in the study, the competence intervals on these results are pretty wide. And so maybe we can talk about that a little bit. And even more importantly, from my perspective, they're really sensitive to the geography that you're looking at. So I guess, if light tech crowds out on subsidized development at a more local level, but is mostly adding to the total supply at, say, the metro area level, is that really a problem? Or what should we be thinking about as kind of different problems that may be arising from this, even if a loss of supply or crowding out of supply is not so much of an issue? It

Michael Eriksen  43:25  
kind of depends, you know, I still generally, I don't think it's very settled, you know, I tend to view probably at the Metro level, we would probably get more credit than we would, at the local level. This is very much in nuance. And and this is where I would challenge and again, say, Look, our study is very narrowly based, right? If you only care about literally the count of units, you know, I would challenge and and have challenged reporters in the past to say, you know, there's at least at minimum, right, there's a difference in the quality of units. This, you know, it could have be these households were consuming a lower quality unit, and now they're able to, you know, have something that has less housing problems. Now, we could still say, is the program effective at that or not? A personal passion for me more and more is where are these units and as we aggregate up, we generally find different credit levels going apart. And point is, is the ability to actually build affordable units. Now, they still could be the low moderate household types, but in areas where luxury housing would avail or more luxurious housing, right, if it would have been developed lands, we would have built my counterfactual always say is, from a development perspective, there's 40,000 square feet of multifamily housing development cost to build 10 4000 square foot units is pretty similar to you know, 40,000 square foot units right. And from a developer's perspective is pretty similar and, and obviously we wouldn't, we would expect different rents of what they would charge on a per unit basis. If the tax credit program is again shifting that where it would have been a larger rental unit and a desirable part with good schools higher opportunity Although we might find crowd out right at the end of the day, the fact that the program can nudge a developer to build those, you know, more dense housing, that's sort of the missed opportunities. And again, it's not just about units about what are the units? And how do we do it, I generally tend to find and, as you said, the tax rate program is pretty ineffective, actually, I think lowering rents for at least the recipients themselves. But that doesn't mean that there aren't other benefits, you could make an argument is 100% crowd out is really problematic. I would say, you know, one that was in the 1990s, you know, which you could check a little bit other programs change. But I also think, as we think in a bigger picture, we're perhaps looking at this too narrow.

Shane Phillips  45:38  
That actually leads into my next question, which is about how the results of this analysis might come out differently. If you were to do this study today, instead of the late 2000s, rents were quite a bit more affordable back in the 90s, which is the time that this study is focused on. And in many markets, at least, the gap between a new market rate unit and a below market light tech unit is now very large. In Los Angeles, a two bedroom unit reserved for a household making 60% of area median income rents for about $1,250 per month, whereas the going rate for a new market rate to bedroom is upwards of $3,000. La is a little bit of an exceptional case, because we have among the highest housing costs in the nation, and pretty average incomes nationwide. But does it sound plausible to you that crowd out is maybe less of a problem today than it may have been in the past? Or is this still maybe a big issue? And just there might be more regional variation than there used to be?

Michael Eriksen  46:39  
No, I mean, this is an excellent point. Really, our analysis really looks the 19 like the 90s. And I think the tax CERT program today, and it looks very different. One of the features of Resetti, which is kind of the chaotic environment of LinkedIn, 1980s, where I think governors are trying to figure out well, how do I navigate this program for my state, and one of the things that I was impressed with my initial study was how much some states really seemed to get it. They understood that credit was a concern, they did their best to at least be thinking about it, you know, something I think California does the textured allocation group does is that the rents you chatted about, you know, they try really hard to keep them below market. And you know, I think the state of California has this 10% Rule below market now that it's ultimately hard to maybe enforce over the long term. But I really think there's an effort there. And I think the state of California really gets a lot of credit, kind of no pun intended to really thinking about trying to do the most with the dollars they have. Right? And it could be how do we leverage it? You know, how do we think about different subsidies and the HOME funds, and again, I've been pressed by that. And then flipside, I would say, and I won't throw any other states under the bus. But some states and the other rides, it'd be clear, even in the mid 2000s, they had a hard time even telling me where their units were in the state. Right. One of the things that I think, Shane, I want to pick up on about, you know, as we think about how the program's changed is one, I really think we've done a better job of administering it, I think people have figured out that this is a valuable scarce subsidy. And we should really be careful in how we do it. But I think we have become more targeted. Let it be homeless shelters, treatment facilities. You know, I always like the scholar houses targeted lower income households going back to school, right, again, things that a private market would never do and, and really give an opportunity to create something different. You know, and I think you have that as a part of it. And I think the other part shame that you picked up on more and markets look like LA than they did in the 1990s. Right. So kind of my own research said, from tax credit, this gap between what we think and affordable city would be for someone kind of earning the median income, you know, it was at the top 50 largest cities in 2000, maybe only seven or eight were unaffordable. My own work now suggests like 38 to 40 of the largest 50s. Part of that is and again for crowd estimates, free and look at this, I tend to think crowd out in Los Angeles and supply elastic markets are are less than areas that are more supply elastic. And I think sadly, through zoning and other regulations, the rules become more supply and elastic over the last 30 years. So

Mike Lens  49:10  
and continuing this thread of what's changed or how some of these things have evolved? You know, I know you've continued to study light tech, and of course, you've studied other subsidy programs and other folks have been doing their research on on light tech and the broader subsidy picture. What's been happening in the field over the past 15 years, if you could pick, you know, a couple of exciting findings or important findings. And is there anything that we've learned that you'd like to see incorporated into existing programs? Or have we changed to light tech in better or worse ways? Over that time period? You know, how have those things evolved?

Michael Eriksen  49:50  
Yeah, no, no, first, I won't apologize for every study I haven't mentioned yet. I still won't be able to mention them all my own nuanced view is there's really this viewpoint and I think partly motivating This paper, our supplier tenant approach is best, right? We have these households that are unaffordable. And you know, people say okay, well, we only can have supply programs, we can only have tenant based programs vouchers. And the reality is is probably a mixture of both right? And it depends. What problem are we trying to fix? Now? What are you at some instances, more of one is better than the other. Right? So and supply elastic markets, I still think vouchers tend to have more bang for the buck, but really push comes to shove when you look at state governments. They're fundamentally different expenditures, right? I've chatted with multiple states over the years, probably at least 15, about how they run their programs. Most of these people sort of realize that, you know, the program could be improved. But it's not like if they suddenly improve the program and make it more efficient, right, that necessarily, they're going to get more subsidy or even more units. So taking a step back and realizing this is essentially a free lunch for a lot of state governments. And while they might like to transfer them into vouchers in Wisconsin as a good example, they don't have that opportunity. So I don't know if we ever go to a less targeted kind of, you know, where some states could do more owner occupied mortgage subsidies, or downpayment assistance, and others can do more vouchers. I think as we think about rural America, and how that looks different than urban America, and and it's not necessarily clear, the long domestic tax rate is the best to serve really quite severe affordability challenges fitting people that don't live in an urban area, Nate bump, snow and Justin Marion had this great feature of the program, this discontinuity at qualified census tracks. So once kind of a majority of your census track is eligible to live in low income housing, your developers now suddenly get a basis boost 30% increase in subsidy, it seems like we should be doing the opposite. So in the low income neighborhoods where we're trying to get more low income housing, I would love to say them to target let's say the middle third instead, right? So instead of the lowest income neighborhoods where we're already housed, the majority of our low income population kind of by definition, if we targeted the middle third, I just fundamentally think crowd out would be less. So I've never really understood the qualified census track. And, you know, I think people are doing the right way. And states are thinking about the Qualified Allocation plans to think about these areas of opportunity. But that's one that at least I've always sort of had in the back of my mind. The other one, and you know, as a prepared for our conversation today, is really a paper I finished last month. So when I originally wrote the paper we're talking about today, it was it was really about stigma. And it was as much about demand and kind of when we build low income housing, what happens to the neighborhood property values? How do we think about owner occupied property values? And, you know, it's easier to look at crowd out. But if we wanted to get the disability effects, what would we do and like Rebecca diamond has this really excellent paper that looks at the spillover effects of low income housing, on owner occupied property values, you know, has great econometrics and thinking about the spillover effects, and does a new econometric estimator. This was 2019. My work recently, and it shows us come full circle with my own PhD student, we're now looking at this ability of and I think maybe an original limitation of our own research was the fact that, you know, one density matters, she shows these huge negative effects in more moderate and higher income, like the areas and kind of suggests that we shouldn't be building low income housing, and there's neighborhoods that again, probably have higher opportunity. And again, these are more moderate income neighborhoods, as she shows these rather large effects. One of the first things that that we do is we really show this is really about density. And really look at the nation of the number of units and kind of where our apartments, there are more low income households where there's more dense housing on average. So first thing we look at is okay, what happens if we look at density. The other thing that we looked at, I think, an interesting conversation as to be thinking about what happens with any apartments, I think their own study really was limited to be only looking at Low Income Housing Tax housing, because we had that data, I was able to assemble a really my PhD student was able to assemble a great database of a large amount of apartments that rebuilt from really the mid 1990s, up through 2016. And what we show is actually, you know, at the end of the day, and the majority of markets, the effect of low income housing tax credits are pretty much the same as market rate. So as you're worried about these kinds of what will the neighbors think, but to combine those two results, you know, if we build apartments, let it be subsidized or not, where there's other apartments where there's already some existing density and we can actually raise property values. And I think that's an interesting nuance, both from a kind of a pro density side, but even a low income housing tax credit side is, you know, this important effect where I think people like density and I think, you know, given that you aren't the pioneering property, I think more apartments actually lead to more amenities right, more likely to have restaurants more likely to have public transport. Asian options. And on the margin that, you know, I think this is an important finding that, you know, hopefully, a wins at least more people from a NIMBY JMB optimism I like,

Shane Phillips  55:11  
yeah, I definitely the studies of the impact of subsidized housing development or low income housing development, in particular on property values, really highlight this tension. We have this this indecision we have in this country, and I think in many places between, do we want housing affordability? Or do we want inflated home values and asset values, and just when those two things are right in competition with each other, and it gets, it's just most apparent where we're actually, on the one hand, we want affordable housing to provide affordable housing, we want it to be affordable, aka, low cost, not worth much. Ultimately, we also want it to boost the property values of everything around it and make them less affordable, ultimately, or the very least not make them more affordable. And I don't know what to say about that. That's just that's part of why we have all of these challenges, because we haven't really grappled with the trade offs of those two priorities.

Michael Eriksen  56:14  
I would agree. And I think at the end of the day, you know, it kind of this is why I sort of like the mixed approach, right? vouchers, increased demand, and it's a good way to, you know, for the truly needy households afford stuff they couldn't run otherwise, or some mixture of getting the tax rate to build those units, and then allow households to access those neighborhoods and those units. You know, I think that's a good mix. And hopefully where our policy goes as a country.

Shane Phillips  56:37  
All right, Mike Eriksen, thank you so much for joining us on the Housing Voice podcast.

Michael Eriksen  56:42  
No, like I really enjoyed listening to my nine year old son, Nate. He's into podcasts and pretty excited just to talk to you about my research. So thank you for the opportunity. 

Shane Phillips  56:52  
Welcome to our youngest listener. We love you, Nate. You can read more about Michael's work on our website, lewis.ucla.edu. Show notes and a transcript of the interview are there too. The UCLA Lewis Center is on a bunch of social media, I'm on Twitter at @shanedphillips, and Mike is at @mc_lens. Thanks for listening, we'll see you next time.