Last week, our VP of Software Nick Downer sat down with UrbanThree's Ori Babar, PhD to discuss unequal property tax assessment. Research indicates homes with lower values are routinely overassessed, leading to a disproportionately high tax on lower- and middle-income property owners. Read a transcription of their discussion below.

NICK DOWNER: Hello, hello.  Welcome to the Regrid webcast, semi-regular series. My name is Nick Downer, and I'm the VP of Software over here at Regrid.

And I am excited to introduce Ori, who is here with us from Urban3. Ori, do you mind just doing a quick intro of yourself, and what you do over at Urban3?

A bit of background...

ORI BABER, PhD: Absolutely. So, thanks for that introduction Nick. My name is Ori Baber. I'm an analyst at the firm Urban3. We're HQ'd in Ashville, North Carolina. We're a land use, economics planning firm specializing in data visualization -data-driven storytelling and policy development. We work with clients all over the US and internationally.

I joined the Urban3 team nearly a year ago. But have been working with them on and off since about 2017. I was working for the city of Gainesville, down in Florida. Some city staff actually brought the Urban3 team in to do an extensive land use-economics study for us to inform a strategic plan that we were doing. And so that was how I was introduced to Urban3 and the rest is history.

My background is kind of an interesting interface between data science as the umbrella underneath which I work But then my specialized training is actually in the field of public health. That's where I kinda work across the planning and public health realm, using data science as the bridge between the two.

ND: And I have to say, I'm really stoked that we're able to do this and that you're able to join us here. Cause I feel like from the time that we first started talking together, about the over and under assessment stuff I was like, 'man we've got to do this on a public stage.' So that others can hear it too. Because I think on the Regrid side, it really gets to a lot of the things that we care about and speaks to the issues of the day.

But I think it's also really timely in that there is a lot of emerging research, as you obviously know. Around these sorts of under- and over-assessments, and how value and taxes impact peoples' lives, right? And I think, without getting too feel-good or cliche about it. I think there is a world where this data analysis can be really used to make average peoples' lives better.

On all fronts, we are happy to have you here. And the stuff you've been working on, which I'm excited to hand it over to you in a second really hits the intersection for us. Where it's like cool data, presented interestingly also has really interesting implications and how do you translate that into the real world?

OB: Thanks for the opportunity to do that. We've uncovered some interesting stuff here in our local county of Buncombe. And we've been somewhat on a local roadshow, so sharing it with local organizations. So I appreciate the opportunity to bring this to your audience, and to share our findings with you all.

ND: Absolutely. And I think with that, let's get into it. Can you set the scene here; and talk a bit about Buncombe county. Ashville's, I believe, largest city within it, right?

OB: Yep. So Ashville is the county seat. I'm in Ashville right now. Our firm is HQ'd here. In general, setting the seat as you said our county recently reassessed all of the properties in Buncombe county.

We've got about 130,000 parcels in Buncombe county. Went through a year lot process to reassign values to each of those properties in Buncombe county. I'll get into this in a little bit of the presentation but we've got some backdrop, some local resolutions that were passed. Our firm, specializing in this field of economics, decided to take some time and look at how this reappraisal process played out. And what that means for our neighbors, our community members, ourselves...

So we embarked on this journey that's really uncovered some interesting findings. That's the backdrop to this.

ND: Let's get into it - I just wanted to quickly note that before do, if anyone watching live has a question feel free to drop it in the chat, there on the right. We'll answer them as we can. With that, Ori please go ahead. I'm stoked to look at this sort of stuff.

OB: Nick can you see my screen?

ND: Yep all good!

The 101 on unequal property tax assessment in Buncombe County, NC...

OB: The title presentation, Unequal Property Tax assessment: specifically looking at Buncombe County. What I've been doing, to kick these presentations off is to start with a quick 'property tax 101.' Right? Cause we know that different jurisdictions, different states, have slightly different or nuanced ways they approach their property tax administration.

Here, locally, I'll run through the basics for North Carolina. We're fairly fortunate in the fact that we have a relatively straightforward property tax system - a kind of legislative system. At its most fundamental level really it's just the assessed value of a property multiplied by the local jurisdictions tax rate - So if this is the county's tax rate or the city of Ashville's tax rate - yield the property tax bill. That's how much an individual property owner would be paying annually in property taxes.

In North Carolina, the assessed value is supposed to be equivalent to the fair market value. So homes and properties are assessed at 100% fair market value - and I'll talk through the distinction between those two, throughout the rest of this presentation.

The first takeaway from Buncombe county is the assessed value should equal the market value - I use the word should here, and we will take through why. The other thing is the idea of fair market value. So if we're going to reassign value to properties in Buncombe county, not every property transacts or gets sold in the market every given year. So what ends up happening is the sales that do occur, that data is then used to extrapolate the data for similar homes, for other homes in the neighborhood, in the area.

And that's the process known as 'computer-assisted mass appraisal.' Which takes sales and market data and helps determine those assessed values in the county. A quick, hypothetical situation here: Say we have a home in 2017 that was assessed or assigned a value of about $300k.

In Buncombe county, that assessed value will carry forward every tax year until a new reassessment happens. So for a four-year tax period, the owner would be paying taxes on a $300k home. That's the assessed value, right? This is different than market value. Ashville, Buncombe County, our real estate sector is really continuing to take off.

So that same home in 2017 may have sold on the market for $300k, but in 2018 you may be able to fetch a little bit more for it on the market. A little bit more, a little bit more. Such that by 2021, let's say that home is now able to fetch $350k on the market. So then when this reappraisal happens in 2021, instead of being reappraised at $300k going through this process of reappraisal bumps the value, the assessed value up to $350k. It's more fair market value.

That's generally the process we are looking at here. That we've looked at and assessed in the particular work we've done recently. Kind of dial that in, here in Buncombe County, here in North Carolina. Counties are required to do this reassessment at least every eight years. At least Buncombe County has adopted a four-year schedule. So our last reappraisal happened in 2017 and then again earlier this year in 2021. What we effectively have done is look at the change in assessed value between 2017 and 2021.

These are the three big takeaways: from this 'introductory property tax 101.'

I mention that there is some background context to this work, specifically Buncombe County, so that you all that are tuning in, may be familiar with Buncombe County or Ashville for a number of different reasons; it's a great tourism hub. Most recently, we've gained some local, or national, recognition for being among the first jurisdictions to pass reparations for black residents in Buncombe County and in Ashville.

Kind of led to some of the initial conversations around reparations. This was about mid-2021. And then we did this county-wide reassessment.

Urban3, effectively, we saw the reassessment come out and we asked the question, "How does reassigning value to all 130K parcels in Buncombe county, how does that relate to these objectives that we have committed to as a community, specifically around race equity."

That's the backdrop to the work; We started this work in an exploratory type way. We call this a pick-and-poke analysis. We find some properties we are familiar with, we either own or our neighbors. And we look at how this reassessment process played out. I'll start with one of those properties we started with. So this house in Brown is actually my boss's house, Joe Minicozzi, the founder of Urban3.

We looked at his property, and we also took his neighbor's, the Lordmans. The Lordmans are long-term residents of this particular neighborhood, this is the historic Mumford neighborhood. African American family, military veterans, living on a fixed income. Unfortunately, experiencing a couple of chronic health conditions that limit their mobility.

So our first comparison was to look at Joe's duplex versus his neighbor's. I'll show you some of those numbers right here. We'll look at 68 Courtland the Lordman's, this is Joe's neighbors, and I'll show you, kind of work through a couple of different examples. We'll look at three metrics that right off the bat we wanted to look at to be able to compare this reassessment process.

We look at land value per acre, the Lordman's is about $500k. We look at building value per square foot, $127 per square foot. And then the percent change is the amount the valuation changed between 2017 and 2021.

So we see the Lordman's increased in value by about 117%. Keeping that in mind, the Lordman's are here in the Mumford neighborhood. We decided to pick out another property: This is equivalent to moving closer to the beach if you're in a coastal community or close to waterfront properties.

This is Kimberly Avenue, another very well-established neighborhood. We're looking at a home that faces the Grove Park Country club, this is a luxury golf club. Right there on the 9th hole. This is the house. This is the view. It's right along a street that has underground utilities, 10-12 foot sidewalks, and obviously a pretty gorgeous view.

The house. The view. We can look at the same sort of metrics or indicators for this particular property. What we see is land value per acre on this property is $447K per acre. Building square ft is $107 per square foot. We see about a 106% change.

We can put those side by side back with the Lordman's and just look across these two. Land value per acre - the Lordman's are valued at about 50 grand more per acre in just the land value. The Lordman's home is assessed at $20 more per square foot.

Just want to call your attention to the two pictures of these homes and ask whether or not you would expect the Lordman's house to be priced at $20 more per square foot. When you just look at these two pictures. And the percent change between 2017 and 2021. This house on Kimberly Avenue only increased by 106%. Again, the Lordman's increased by 119%.

And you can also look at how much they are actually paying in just county property taxes alone. We do this on a per-acre analysis just to be able to compare apples to apples. The Kimberly Ave home paying 9 grand in property taxes. Courtland pays about 10 grand in property taxes annually.

Let's look at one more property - before I do that, while we're in the process of doing this pick-and-poke analysis, this house on Kimberly Ave actually sold for $765K. If you look at its assessed taxable value of $495K, right this is effectively an indicator of its market value - right, [its market value] is considerably higher than its assigned taxable value.

So it's being taxed effectively at 60% of its sale price, which I'll talk about more in a second. Let's look at one more property. So here's that Kimberly Avenue house, if we go across the golf course, up to the top of this mountain, this is effectively beachfront reality up here. Right down the street from the Governor's Mansion, this is the property we're looking at: It's got a tennis court. Quite a large single-family home.

So let's look at the same kind of indicators across the board here.

Land value per acre for this mountaintop, tennis courthouse is $83K per acre. Kimberly Ave at $447K and then Lordman's at nearly $500k per acre. And if we just look at the difference in the number of pre-taxes paid per acre, we talking about 4 grand per acre for the mountain top tennis court home, 9 grand per acre for the Kimberly Avenue home, and an astonishing 10 grand per acre for the Lordman's home in Mumford.

So right off the bat, I mentioned that was kind of the pick-and-poke approach to this work: where we just look at a handful of properties and just doing that alone led us to feel like there were some curiosities that may need further exploration.

So we decided rather than just doing this piece by piece, let's take this countywide. This is in our wheelhouse - it's kind of what we do on a day-to-day basis. We took the entire county and looked specifically at the residential properties for this analysis. And providing again just some backdrop, we decided to do a sales ratio analysis. I mentioned before that our statutes require that homes be assessed at 100% of their fair market value.

If we look at a sales ratio, which is basically just dividing the assessed value by its market value, its most recent selling price, gives you the sales ratio. Here, a 1-to-1 ratio means that we're being assessed at fair market value.

There are situations where a home may be overassessed - where the assessed value is greater than the fair market value - so your sales ratio would be greater than one. In a situation where a property may be under-assessed, this is when the assessed value is lower than the fair market value. The big distinction between being over-assessed and under-assessed is if you're being over-assessed you're effectively paying an unfair, higher amount of property taxes. Then, if you were to be under-assessed you're effectively receiving a form of tax subsidy in which you're only paying a fraction of the amount of property taxes that you should if you were being assessed at fair market value.

With this in mind, we did a comprehensive, countywide analysis.

A countywide analysis...

We looked at transactions over the last twenty years and compared the assessed value to its most recent sales price. What we ended up finding is that homes on the lower end of the value spectrum were being overassessed fairly routinely. And that homes on the higher end of the value spectrum were being consistently under-assessed.

The tipping point is right around $225K. Such that if you owned a home $225K and under, generally what we observed is these homes are being overassessed. And vice versa. If you owned a home over $225K, generally you were being under-assessed. And so I'll talk through the implications and the ramifications of this disparity we see.

But what we are also able to do is track the change in the value of these, of this initial cohort of homes over a 20 year, like two-decade period essentially. We've bent the homes by their starting value - so what we found is that the lowest value homes, to begin with, in 2001 increased by nearly 323%.

While the highest 20% of homes only increased by 138%. This is a year-over-year compounding trend that we see. And the problem that we're experiencing in Buncombe county, and it's a very pronounced and real problem not only here in Buncombe county but I'm sure that everyone can kind of relate to this situation,

When we see the value of homes increasing this dramatically, and when we compare that to the change in median income in Buncombe county, our median income has only increased by 36%. CPI at the same time frame has only increased by nearly 50%. I think this really illustrates one of the mechanisms that are at play here when we talk about housing affordability when we talk about Buncombe county.

But what's even, perhaps, more astonishing, is when we take the buildings off the property, and just look at the change in land value across time the same cohort of homes and we just look at the change in just the land that's underneath them. We see that just the lowest 20%, that land increased by nearly 700% over a 20-year time frame. While the highest value homes only increased by 153%.

So I think this really begs the question, what's at play here that we would see such a dramatic increase at the lower end of the spectrum? And a fairly moderate increase on the higher end of the spectrum? These are the two numbers: total value, land value, bend by quintile. What we also pulled during this period was the POdata. In the process of getting values reassigned, a property owner has the opportunity to submit a formal appeal and basically say, 'hey I think the value my property has been assigned is off. I have more information about the property or I'd like to take a closer look.'

And what we ended up seeing was that homeowners of more valuable homes were far more likely to engage and then be successful in the appeals process. The appeals process tends to move in the direction of deflating the value of the home.

I think what I've shown so far illustrates, perhaps, a trend in economic bias. There are burdens being placed on lower-income families, lower-valued homeowners. Like I mentioned in the beginning, we as a community made commitments to advance race equity in Buncombe county.

We also wanted to look at this from a race equity perspective. What I'm showing here is just a simple distribution of home values by homeowner race. So I'm just showing black or African American in blue and then white homeowners in green. Effectively, what this shows is that white homeowners tend to own homes that on average are 20-40% more valuable than black or African American homeowners in Buncombe county.

When we layer on the J-curve, the under- over-assessed, who is more likely to feel the impacts of the over-assessment of lower-value homes. By and large, this is our communities of color in Buncombe county. And we see this is a pretty stark indication of a racial bias or a racial disparity in the property tax assessment.

Urban3's analysis and findings...

I mentioned we did this county-wide, so I want to show some quick graphics that we can talk through. But this is just the change in land value, throughout the entire county, parcel by parcel. We're able to layer on change in building value on top of that. And then zoom in and look at this neighborhood by neighborhood. Here we are looking at a neighborhood that's just on the northside of downtown Ashville.

What I'm showing is a change in land value from 2020 to 2021. Light green is low percent change, dark green is high percent change. And then I've extruded by that same percent change. So a higher peak indicates a higher percent change.

This was really kind of an astonishing graphic to put together for a number of different reasons, and I will walk you through why. One, I think the big thing that caught our eyes that we've got down here in the Montford area we've got neighborhoods, we've got homes that are immediately adjacent to each other that are demonstrating very different changes in land value, because of this reappraisal process.

And we know, living here as residents of the Montford neighborhood, we know this is a larger geographical footprint it's a curiosity to see such stark changes. The other thing that was really kind of interesting was that all of these parcels were assigned the same percent change in land value.

You can see all these clusters elevating at the same change. That led to additional questions about the mechanics of the reassessment process and what these mean if you will. I will talk through that in a moment.

Our office is here as indicated by the star. If we were to walk across the highway, into that neighborhood I was talking about, down Flint Street, and just look at two of these homes side by side one, that's seen a dramatic increase versus one that is not increasing.

These are those two homes. Side by side. 96 Flint St, 98 Flint St. With an arbitrary neighborhood boundary in-between. On one side of this arbitrary boundary, we see a 316% increase in land value.

And on the other side, we see a 102% increase in land value. And what I'll point to, is that this arbitrary boundary is actually an artifact in the map, that was used to assign value. This line here indicates a tax neighborhood boundary.

Which is effectively a grouping of homes produced to develop the assessed value.

If you're like me, I know I'm moving through this kind of quickly, but I want to leave some time for conversation on the tail-end. We got to this point and we asked ourselves, how are some of these disparities or anomalies happening?

And what we did, is we grouped our findings into three vague categories:

One is algorithmic bias, one is human bias, and then we're dealing with situations in which we have structural inequity from decades of historical policy that has been implemented based on racial or class constructs.

I'll talk through -

The source of the data...

ND: Hey Ori, I have a question for you - just before you dive into the biases, can you just talk a little bit about the source of the data we are talking about with Buncombe?

OB: Yeah, great question. We're fortunate we had a relationship with our local tax assessor. So we were able to request a lot of this data directly from our local tax assessor.

I should also mention this is part of the race equity work we - that our county had been engaged in. There was a push to build more transparency in the data process.

So an open data portal here for our county that we were able to pull some of this data from. By and large, most of this came from a direct request to our county assessor. That's basically the data we are working with. Both the parcel geographies and then the tax roll data are coming from our assessor.

ND: Cool, thanks!

OB: Then, like I said, diving into the questions of why and how does this happen - I will caveat this and say this might be somewhat of an oversimplification but I think it does help communicate some of the deficiencies in the maps that are currently being used.

How does property over assessment occur...

We're just looking at, right now, the distribution of home values here in Buncombe county. Our homes, tend to center at about $250k to $300k. That's our average home value. We've got fewer homes on the lower value end and fewer homes on the higher value end. Effectively what this means is that if we look at the two tail ends of this distribution, since we have fewer homes in the low/high end, it's more difficult to establish certainty.

We have a little bit more margin of error that happens, we have fewer transactions and the margin of error tends to be higher than in the middle where we have more properties to work with, we have more transactions going on.

I'm illustrating that here, with, sorta, more wiggle on the tail ends, is how we refer to this. The second driving factor here that we've seen is a lot of the math that is used to assign value works off of central tendencies. Mean, median, averages. Which effectively pulls values in, towards the center. We call this 'regression to the mean.'

Such that lower value homes get pulled towards that $300k center, and our higher value homes actually get pulled down towards the center. Just in the process of working through some of the math that's used. These two in combination, exert upward pressure on our lower value homes and downward pressure on our higher valued homes.

The other thing is if we have more uncertainty on the tail ends, we might expect to see appeals happen more frequently on the low and the high. People saying, 'Hey assessor, I think you got my value off a little.'

You'd expect to see both of those happening, but as I mentioned, we're seeing participation in the appeals process highly skewed by the higher end of the value spectrum. Which is again placing additional downward pressure on the higher end of the spectrum.

We're asking the question, do those simple mass mechanics produce some of these over- and under-assessments trends that we're seeing? Perhaps. But there may also be situations where human bias is involved.

There are situations in which assessor staff have to go out and make value decisions. Have to use discretion or judgment. A very simple example of this is making distinctions between quality grades on different housing properties.

We have six housing grades, from luxury to poor. Each one is assigned a quality number, which then affects the math used to determine its assigned value. I can pick out the difference between a luxury and a poor. But making these nuanced decisions between average and fair, superior and average may have compounding implications if there's a judgment that needs to be happening here.

Our assessor staff is a fairly small and nimble team but is responsible for assessing 130K parcels. We've got quality, but we've also got condition adjustments. Whether or not a property is renovated or good, fair, normal, poor. All of these have implications on the math that's used to assign values.

Then the last one, this big piece I want to spend a second unpacking is the process of extrapolating value from one market transaction to other homes in the area. Here in Buncombe County, and many of the locations we are working in now, use a convention to group homes into tax neighborhoods. And they'll use sales that exist within these tax neighborhoods to reassign value to homes within that same neighborhood.

The issues that we see here - this is a patchwork quilt of tax neighborhoods on the North End of the city of Ashville. These tax neighborhoods don't mirror any of the neighborhoods we're familiar with. We see them as being somewhat arbitrary. At this point in time, uncertain about how they are delineated and produced.

Asheville, like many cities across the US, invented/introduced red-lining in the '30s and did so for a 30-year period.

A geographically based form of specifically racial discrimination in which homeowners of color might have found themselves in neighborhoods that were redlined and finding themselves unable to engage in lending, mortgages, and effectively, unable to sell or transact their home in the marketplace.

We're still using these neighborhoods delineations: the question I'm reflecting on right now is whether or not these tax neighborhoods are inadvertently recycling some of the discrimination that may have been baked into the geographic fabric of our communities from these decades of racist policies and practice. This is an area for us to think of better math, better geographic approaches, that were not subject to the discrimination we've seen in the past.

I want to touch on some of the highlights we've taken this work and presented it here, locally, to community organizations, to elected officials, to staff. We've seen some positive movement come from all of this. The county has established an ad-hoc advisory commission, to go back and look at the whole reappraisal process.

To provide recommendations for how to do this process better. Especially considering the commitment to raise equity. Several of our local cities, Ashville, Woodfin, and Buncombe County, have passed homeowner tax relief programs.

I'd love to talk through my thoughts on these but I'll leave my editorial comments for, maybe, our discussion.

Then, real quickly, taking you through our next steps process.

From here, we are fortunate to be working with a coalition of local partners, and we've received word that we've gotten some funding to expand this research to an 18-county region in western NC -

Which I am very much looking forward to - this work will kick off in 2022 and be about a two-year effort to take this work that we've done kind of at a pilot scale and take it even deeper. But then also expand the geographic scope to a wider area, and hopefully shed some like on state policies that may be supportive in undoing the disparities we're seeing in the tax administration process.

With that, I will say thanks. I know I ran through that quite quickly but I'm looking forward to our conversation next. I've included my email ori@urbanthree.com. Feel free to reach out outside of this webinar, I'd be happy to chat more.