How we tax property, with Lars Doucet

This week I'm joined by Lars Doucet. We had a somewhat rollicking conversation about property taxes, and specifically within that, the mechanics and political economy of how assessments are done, to determine the value to be taxed.

[Patrick notes: As always, there are some after-the-fact observations sprinkled into the transcript, set out in this format.]

Sponsors

Support proven charities that deliver measurable results and learn how to maximize your charitable impact with GiveWell. First-time donors get $100 matched. Go to givewell.org (and type in "Complex Systems" at checkout).

Check is the leading payroll infrastructure provider and pioneer of embedded payroll. Check makes it easy for any SaaS platform to build a payroll business, and already powers 60+ popular platforms. Head to checkhq.com/complex and tell them patio11 sent you.

Timestamps

(00:00) Introduction
(00:23) How property taxes work (Texas Example)
(02:45) The political art of avoiding tax rate blame
(05:53) Sources of real estate data
(08:08) Historical property assessment
(11:04) Statutory guidance vs. Actual practice on market value assessment
(14:25) Tax rate strategy and sandbagging
(15:17) Assessed value vs market value
(16:16) Assessment caps and Prop 13
(18:22) Sponsor: GiveWell | Check
(20:27) Data collection in the field
(22:54) Data collection methods
(25:08) Property valuation: Beyond location and correlative factors
(26:52) Depreciation of buildings
(27:37) Orthodox view of depreciation
(30:53) Real estate cultural differences
(33:59) Urban redevelopment and land value
(36:59) Small business realities and perceptions
(46:19) Property tax protests
(50:45) Predictive protests
(52:19) Accuracy vs equity testing
(58:50) Cook county assessor's office
(1:01:11) Lars's background
(1:05:38) What is GIS?
(1:09:52) Wrap

Transcript

Patrick McKenzie

Hideho everybody. I'm Patrick McKenzie, better known as patio11 on the Internet, and I'm here with my buddy Lars Doucet. 

Lars Doucet

Howdy!

Patrick McKenzie

So, Lars, like many people, I recently received a property tax bill in the mail. There was a number on it, but I don’t understand what that number means. Can you explain what’s going on?

How property taxes work (Texas Example)

Lars Doucet

First of all, it depends on which state you’re in. Property tax systems vary widely across the U.S. You’re in Cook County, Illinois—Chicago, if I’m not mistaken? Okay. I’ll have some specific comments about Cook County later. I’m in Texas, so I’ll explain how it works here, which is similar to how it’s handled in many other places, with some local variations.

In Texas, your local government unit—called a central appraisal district—handles property valuation. This is a unique type of local government, fully sovereign, much like a school district. Its sole responsibility is to assess the value of your property. They don’t send you a tax bill but rather a notice of assessed value. In big letters, it will say at the top, This is not a bill.

This notice includes a number, which represents the central appraisal district’s opinion of your property’s market value. And that’s all the central appraisal district does. It doesn’t collect taxes or set rates. Instead, other local government entities—like your city, school district, county government, local water district, or special utility district—use that number to calculate the taxes they charge you. Collectively, these are known as taxing entities.

Now, here’s where it gets interesting: the notice of assessed value serves a critical political purpose. It redirects all of your frustration about being taxed to the appraisal district, which has no power to tax you. Their only job is to determine the market value of your property so that the taxing entities can levy taxes based on it. The taxing entities, on the other hand, have no authority to determine your property’s value—they can only tax based on the number provided by the appraisal district.

This separation of duties is actually pretty clever. The central appraisal district essentially acts as a "whipping boy" for elected officials. It’s a fascinating example of how local governments have co-evolved. The people with the authority to tax you don’t have the power to determine what your property is worth—they have to work backward from the appraisal district’s valuation.

The political art of avoiding tax rate blame

Patrick McKenzie

Alright, so it seems like what you’re describing is what [previous Complex Systems guest] Dan Davies calls an accountability sink. I’m an elected official, and let’s say we all had to vote on the city council for the city’s tax rate this year. That rate funds the park district, the police, and the fire department.

We decide we want to set it at 10 basis points—or whatever the number ends up being; I’m not even sure—but let’s say it’s 5% higher than last year. Naturally, I don’t want to be the one who gets yelled at for that increase. My play to avoid blame might involve some influence over the people who come up with the numbers—or is it something else?

Lars Doucet

We call it a “millage rate”; 1 mill is 1/1000th of the property value.

Here’s how it typically works. Your elected official will get up on stage and solemnly declare that they haven’t raised the millage rate in 10 years. And that might be true! The public, unless they’ve gotten savvy to what’s really happening, might naively think, Oh, they haven’t raised taxes.

Then the elected official will blame the appraisal district: That mean old appraisal district raised your property values! But the appraisal district didn’t raise your values—the market did.

The thing about appraisal districts is that they’re legally required to assess properties at market value. We can dive into how they determine market value if you want—there’s a lot of debate around that, and there’s even a whole mechanism for disputing their assessments, which exists by design. But at the end of the day, if the appraisal district doesn’t assess properties accurately, they can fail audits or tests by state oversight authorities, and they’ll get fired.

So, the appraisal district’s job is to assess your property at market value. That’s what local government units base their taxes on. Now, let’s imagine what would happen if local government units were responsible for determining property values themselves. That was actually the case in Texas up until the 1980s.

Back then, you’d receive five different tax bills in the mail—one from each local government entity—and each would have a different valuation and a different millage rate. There was no separation of incentives. A local government could simply inflate your property’s assessed value to generate more revenue.

That’s why Texas, like many other states, requires property taxation to be fair and equitable. It’s written into the state constitution. You can’t assess property arbitrarily. You can’t decide, I don’t like Democrats, so I’ll assess their property at twice the value. Or, I don’t like Republicans or this minority group, so I’ll inflate their assessments to shift the tax burden onto them.

Patrick McKenzie

And I assume part of the concern here is that small-town or local politics, which can sometimes be incredibly venal, might lead to situations like someone getting $10,000 knocked off their tax bill because they’re buddies with the mayor or play golf together.

Lars Doucet

Right, exactly. And it makes headlines every year when a mayor, a tax assessor, or someone similar has a property valuation that’s suspiciously lower than all their neighbors—or lower than it was the previous year without a good explanation.

That’s why transparency is so important. All these valuations are public information. And today, we have more tools than ever to analyze this data. We can map it out, run statistical tests, and determine whether valuations are truly fair and equitable.

Sources of real estate data

Patrick McKenzie

For non-specialists in real estate—and I’m certainly not one, though I heard about it at the dinner table growing up—this public information includes a lot of the historical data that goes into tools like Zillow. Or am I off base here?

Lars Doucet

I can tell you exactly where Zillow—and property tax assessors, for that matter—get their data. 

Patrick McKenzie

Oh wonderful. Why speculate when I can just have you explain.

Lars Doucet

There are basically three sources of real estate information. Let’s start with what Zillow does. They use what’s called an automated valuation model (AVM). That’s their proprietary estimate of a property’s value. Redfin has something similar, and so do most realtors.

When a realtor comes to your door and says, “I think your property is worth between X and Y based on comparable sales,” they’re offering a valuation. Now, the key difference between a valuation and an appraisal is that an appraisal is official and legally recognized, often for financial or lending purposes. Think of it like the difference between asking a friend to look at a document versus having it notarized.

An appraisal has a special legal status—it’s used by banks or for loans. A valuation, on the other hand, is just someone’s opinion. It might be a very informed opinion, even better informed than the legal appraisal, but at the end of the day, it’s still just an opinion.

For both valuations and appraisals, though, you need data to work from. That’s where this gets interesting. The data comes from several places.

One major source is your local property tax assessor. In most counties—and sometimes in all counties—the property tax assessor’s office is responsible for gathering a lot of this data. Assessors might drive around neighborhoods, collecting details about every property.

This can be done in collaboration with other local government units, like the city recorder or whoever is responsible for permits or deeds. The specifics of who handles what often vary by county or state, but someone is out there recording: That’s a brick building. That’s a Cape Cod. That’s a townhome. That’s a detached house with a three-car garage.

Historical property assessment

Patrick McKenzie

And we have hundreds of years of history with property taxation, right? They used to count the number of windows and doors on a house. That method substantially predates electricity—might even predate the Industrial Revolution, if I’m remembering correctly.

Lars Doucet

Yes, exactly. We could go down that rabbit hole if you want, but briefly, there have been all kinds of theories about how to tax property over the years.

In England, for example, they famously decided that the number of windows a house had was a good proxy for how big and fancy it was. So they created a flat property tax based on the number of windows. Naturally, any tax creates a response, and people started bricking up their windows to avoid the tax.

If you visit historical homes in England today, you can still see houses with bricked-up windows. Those windows were sealed off during that tax period specifically to evade the tax—a fascinating relic of that policy.

Zooming back to the modern era, let’s talk about other groups that provide data on property characteristics. One major source is realtors. When a realtor lists a house for sale, they advertise it on the Multiple Listing Service (MLS). The MLS is essentially a cooperative system shared by realtors.

To access comps (comparable sales) from the MLS, you have to contribute your own listings to it. It’s a give-to-get system. In some states, MLS data is public and accessible to tax assessors, while in others, it’s private.

A third set of data comes from other local government units, third-party vendors, or even the federal census. These groups also collect property characteristics, but much of the data that realtors use often piggybacks off the assessor’s original data collection efforts.

Now, here’s an interesting wrinkle: about 12 states in the U.S., including Texas, are what we call real estate non-disclosure states. In these states, real estate transaction data isn’t directly shared with property tax assessors.

This doesn’t mean your data is private or that nobody knows what your house sold for. In fact, it often means everyone knows what your house sold for—it’s just made slightly harder for the assessor to access.

In practice, what happens is that assessors end up buying this data from private vendors. So, congratulations: your data isn’t really safe, and your tax dollars are now going to private companies so they can sell your local government’s data back to itself.

Patrick McKenzie

Oh, I love America some days. [Patrick notes: This is wry sincerity, not sarcasm.]

Lars Doucet

This system also has the effect of increasing the effective property tax rate on middle-class homeowners. Enough people in that group voluntarily provide their information that the only properties left out of the picture, statistically, are the very high-end homes.

And at that top end, it’s not just a matter of whether a house is worth $5 million or $50 million—that’s a huge tax difference. The result is that the absolute top tier of property owners disproportionately benefits, and it’s middle-class taxpayers who shoulder the burden.

Statutory guidance vs. Actual practice on market value assessment

Patrick McKenzie

You mentioned earlier that the statutory guidance is that assessments must reflect market value—which, as we know, is something all actors in society always rigorously follow.

But in practice, does the assessment actually track market value? For instance, in the first year after someone buys a house, when we have the freshest quote on the fair market value because it sold just two weeks ago, how likely is the assessment to match that number?

Tax rate strategy and sandbagging

Lars Doucet

First, we need to unpack the concept of market value itself. What even is market value? It’s a specific definition, typically outlined in state constitutions, property tax statutes, or assessment handbooks.

Market value, in the context of property taxes, is what an unconstrained buyer would pay for the property if:

1. Both the buyer and seller had equal access to market information.
2. Neither party was under duress or had a special relationship—like a father selling his house to his son for $5,000, which happens more often than you’d think.
3. The sale didn’t include bulk purchases (e.g., buying multiple parcels at once), bundled items (e.g., a house and a boat), or atypical conditions (e.g., a house sold with all the furniture).

In short, market value is based on what’s called an arm’s-length transaction.

The second key aspect is the valuation date. Unlike forecasting future value, assessment is about predicting the past—specifically, what a property was worth on the valuation date, which is often January 1 of the given year.

Assessments are finalized a few months later, say February or March, depending on the jurisdiction. If you’re on an annual valuation cycle, the assessment reflects the property’s value as of January 1—not what it might be worth in the summer or later in the year.

To estimate this, assessors use sales data. Sales are the ground truth for determining market value. Assessors look at validated arm’s-length sales and calculate the central tendency—essentially the median value for properties of a similar type.

Some buyers might negotiate a great deal and pay less, while others might overpay for sentimental reasons (e.g., “This house reminds me of my childhood home”). These outliers are balanced out in the central tendency, which becomes the basis for assessments.

In addition, some states use what’s called a target ratio. This is the ratio between the assessed value and the sale price, calculated as assessed value divided by sale price. For example:

• If the property is overvalued by 2x, the ratio is 2.0.
• If it’s undervalued by half, the ratio is 0.5.

Many assessors intentionally undershoot slightly, aiming for something like 95% of market value (a target ratio of 0.95). This accounts for potential errors and also tends to be more politically palatable.

Patrick McKenzie

Is that kind of sandbagging—keeping the assessed value a bit lower—more of a political concession, or is it about leaving some wiggle room in case taxes need to go up next year while denying that taxes are actually increasing?

Lars Doucet

It’s an interesting question because there are a couple of layers to this stack. Let’s break it down.

Assessed value vs market value

First, there’s the distinction between appraised full market value and assessed value. The terms "appraisal" and "assessment" are sometimes interchangeable, but their usage varies greatly by state. And if you use the wrong term in the wrong crowd, people can get pretty annoyed. For example, in Texas, they’re called appraisal districts, and the people working there are appraisers.

Generally speaking, the appraised full market value is the platonic ideal of what a property is worth. It’s the unbiased, informed expert opinion of the property’s market value, irrespective of tax policies, exemptions, or reductions.

Then, you get into the political side of things. Local government assessment units are also responsible for adjudicating exemptions and write-downs. For example:

• A home owned by a disabled veteran might have an exemption.
• Another might qualify for a homestead exemption.
• A property could be in an opportunity zone.
• There might be universal building exemptions or other specific tax breaks.

After applying all these exemptions and reductions, you get the assessed value, which is usually the taxable value. This is the number to which the millage rate is applied to calculate your tax bill.

To address your question directly, there are often legal caps—called assessment caps or circuit breakers—written into state law. These caps limit how much the assessed value can increase annually, regardless of how much the appraised full market value rises.

So, while assessors still calculate the market value to know the property’s true worth, what you’re actually taxed on is the assessed value, which includes all exemptions, write-downs, and caps.

Assessment caps and Prop 13 

Patrick McKenzie

Isn’t that similar to California’s Prop 13, where property tax rates are essentially locked for life as long as you stay in the same house?

Lars Doucet

Yes, exactly. I’m not a lawyer, and I don’t have the statute in front of me, but Prop 13 effectively enshrines a kind of established landed gentry in California law.

The basic idea is that anyone who owned property in California as of 1978 (the year Prop 13 passed) has their property valuation essentially frozen at that level. Property values under Prop 13 are allowed to increase annually, but only by a fixed, very small percentage—far below actual market value growth.

As a result, while those properties have gained massive value since 1978, their taxable valuations haven’t kept pace. Many of these properties are now worth millions on the market, but their owners pay property taxes as if they were still middle-class homes from the 1970s.

There’s another wrinkle: Prop 13 allows children to inherit this tax advantage in some cases. However, if the property is sold to a new owner, its taxable valuation resets to current market value.

This means California assessors still calculate the true market value, but they also have to track the much lower Prop 13 value. The result is that many long-time property owners effectively have a lifetime exemption from meaningful property tax increases.

Patrick McKenzie

Meanwhile, someone living next door—whose only "crime" is not being born in California—can own an identical house from the same builder, built on the same date, with the same appliances, and end up paying tens of thousands of dollars more annually in property taxes.

Lars Doucet

Exactly. If you were to perform a horizontal equity test on California’s property tax system, it would show gross horizontal inequity.

Horizontal inequity is a statistical measure used in property assessment. It examines the variation in valuations for similar properties in the same category. Ideally, that variation should be zero—or very close to zero—depending on how tightly clustered the valuations are.

In theory, similar properties should be taxed equitably, meaning their valuations are consistent with each other. That’s the classic sense of equity—not in the modern sense of diversity, equity, and inclusion, but in the old-school sense of equal treatment.

Data collection in the field

Patrick McKenzie

So, calculating all this stuff seems like a lot of work.

We've got the appraisal district. We have people at it. Some of them are driving around in cars all day, counting windows, maybe guesstimating square footage. Asking the planning department if anyone has like put in a new deck because – oh, well, there's a question for you: If I put it in a new deck, I think platonically speaking, my house, is worth more than before I put the deck in because otherwise I wouldn't have put the deck in that seems like a, you know, poor choice for Homo Economicus.

Does that feed back into, uh, the assessment office in any way?

Lars Doucet

Yes, it generally does. In most jurisdictions, you have to get a permit for that kind of work. So, anything requiring a permit typically feeds back to the assessment office.

That said, practices vary wildly—there are over 10,000 assessing jurisdictions in North America, maybe even 14,000. It’s a huge number, so there’s probably somewhere this doesn’t happen. But in most places, assessors have access to permit data, which gives them notice of updates to properties.

Now, not everyone files permits for their home improvements. To address that, many assessment offices subscribe to services like EagleView. EagleView flies planes and takes aerial photos, so they can spot changes like a new pool or an added structure.

Patrick McKenzieI

love this. It reminds me of something we’ve discussed before: hedge funds used to task commercial satellites—basically the civilian equivalent of spy satellites—to track Walmarts. They’d count the number of cars in the parking lot to get a read on Walmart’s quarterly performance before anyone else.

Now it seems like the government is in on the game, using similar methods—not spy satellites, but commercial satellites—for property assessments.

Lars Doucet

Not even satellites—planes.

Patrick McKenzie

Planes? Wow. [Patrick notes: I was mildly surprised because, while I certainly saw table-sized books of aerial photography used for real estate purposes growing up, I had assumed that as the Elon Musk Industrial Complex came online the last few years, the cost of imaging must have declined precipitously.]

Data collection methods

Lars Doucet

There are four main tools you can use for property imaging: drones, planes, balloons, and satellites. Balloons might be the future—they don’t need propulsion and are therefore cheaper. Satellites are another option, but unless you pay top dollar, they often don’t provide the high resolution you need.

The primary goal with these visuals isn’t to obsess over things like decks or updated interiors. Assessors don’t usually have great visibility into what’s actually been done to a house. And frankly, buyers often say they care about those things, but the reality is different.

We all know the three rules of real estate, right? Location, location, location. It’s so important that it counts for all three. That’s the Holy Trinity.

Property valuation: Beyond location and correlative factors

Beyond location, there are five additional characteristics that drive nearly all the value of a property:

1. Building size
2. Land size
3. Building quality
4. Building condition
5. Building age

Sure, you might have a nice deck or patio, but almost everything about your house—from the number of bedrooms and bathrooms to how nice it looks—correlates strongly with one of these six variables (location included).

American real estate tends to be fairly homogenous, so statistically, you can model property values using just those six characteristics. Modern mass appraisal software is incredibly effective at tracking the central tendency of the market and providing accurate predictions without needing exhaustive details.

If you really want to refine your model, you can start looking at additional features, but you don’t need to capture every single characteristic. For example, you don’t need to count the number of bricks in a house or differentiate between 15 types of siding when five broad categories will do.

The key is that the data must be accurate. I always recommend assessors prioritize data quality over granularity. You don’t need 40 unique fields, because most of them will auto-correlate to the core variables anyway.

Patrick McKenzie

I’ll throw out a thought and see if it goes anywhere.

People have many rights, and we also have many intuitions about the assumptions we should and shouldn’t make regarding them. Houses, on the other hand, have sharply fewer rights—they don’t speak up in court very often.

Because of that, many housing features might not be direct drivers of value. Instead, they correlate strongly within this massive, multidimensional matrix of housing quality in the U.S. For example, if a house is above a certain size, it’s almost a given that the interior has had significant work done—because anyone who can afford a house that large likely also invests in its upkeep. Yada, yada, yada.

Lars Doucet

Exactly. Large buildings almost always fall at least in the "above average" category, if not higher. Typical quality classifications include categories like poor, fair, average, above average, good, very good, and excellent. Organizations like Fannie Mae and Marshall & Swift use rubrics to define these grades.

Big houses tend to be higher quality, and higher-quality houses are usually in better condition. Houses in very poor condition tend to be old—until they reach a certain age. At that point, the fact that they’ve survived implies they’re in better condition than you’d expect for their age.

Patrick McKenzie

They’ve gone through the Great Filter. If they’ve made it through the other side without being torn down, it likely means the owners have invested in updates or the house has historical significance. Either way, those factors suggest it’s much better than the median house of the same era, most of which were demolished decades ago.

Lars Doucet

Yes, there’s a kind of Lindy effect at play here. The longer a building survives, the more likely it is to continue surviving. This is also why all the depreciation formulas are wrong.

Depreciation of buildings

Depreciation, in theory, follows a straight-line formula based on a building’s useful life. But that’s entirely fictional. Buildings do depreciate over time—they’re depreciating assets—but the reality is more complex. A 200-year-old building, for instance, should have theoretically depreciated to nothing long ago. The fact that it’s still standing suggests it’s a very high-quality and well-maintained structure. Otherwise, it would have rotted away decades ago.

In practice, a building’s lifetime isn’t a fixed endpoint but an asymptote—it stretches out the longer the building survives.

Orthodox view of depreciation

Patrick McKenzie

To ground this in the orthodox definition: depreciation applies to capital assets, including buildings. It’s based on the assumption that an asset has a useful working life. Under straight-line depreciation, if the useful life is 20 years, the asset loses 5% of its value each year.

There are other depreciation formulas, but the timelines we use are arbitrary. They’re a creature of statute, designed to balance the political economy of depreciation timelines and the administrative complexity for taxpayers and tax offices.

Reality, of course, doesn’t follow statutory constraints. For example, the depreciation timeline for a factory stamping out plastic chairs and one producing silicon wafers would logically differ, but legislation might not account for that nuance.

Lars Doucet

Oh, absolutely. And it’s not just legislation—appraisal handbooks have their own arbitrary standards. These handbooks guide practices like the cost approach for valuing buildings.

When using the cost approach, you’re valuing a property as the sum of land value and building value. For the building portion, there are tables that outline how depreciation is supposed to work. These tables are “good enough” but far from perfect, and depreciation factors often overwhelm the entire valuation.

Depreciation is essentially a multiplier between 0 and 1. If you get it wrong, the impact on valuation is massive—whether you’re rounding closer to 0 or 1 can make or break the assessment.

For buildings, appraisal handbooks typically assume an effective lifetime of 50 to 60 years. But unlike human beings, who have a clear biological limit (no one reliably lives past 120, and even 125 seems like a hard cap), buildings don’t have such a wall. They can keep going indefinitely as long as they’re maintained.

In the U.S., another factor is that we don’t build much new housing. In many districts, the average home was built in 1950, 1960, or 1970—well past the supposed “expiration date” in the cost handbooks. Yet, these homes are often in average or better condition because they’ve been renovated multiple times. They’re still standing, functional, and valuable.

Real estate cultural differences

Patrick McKenzie

It kind of blows my mind walking around neighborhoods in Chicago, where a house built in 1950 would be considered ridiculously young due to various factors.

Then I apply my "Japan brain" to this. In Japan, for a variety of historical and cultural reasons—some of which you’d learn if you did an East Asian Studies degree—there’s a strong market tendency for families purchasing single-family homes to construct a new home rather than buy an existing one.

It’s somewhat analogous to how other cultures might view buying used underwear—not something people typically want to do. That joke might be a bit off-color, so let me rephrase: in Japan, historically, people buying single-family homes have expected to purchase the land, knock down any existing building, and then have a developer construct a purpose-built home tailored to their needs.

This has some interesting implications:

1. The value of a real estate purchase in Japan is typically concentrated in the land, not the building. That’s the opposite of what most Americans would expect.
2. The stock of single-family homes in Japanese neighborhoods tends to be much younger than in comparable American cities.

Lars Doucet

Exactly. And we also need to consider another type of depreciation—functional depreciation.

I’m not here to justify any specific depreciation numbers someone might assign, but let’s consider what happens in a teardown scenario. Say I see a lot I like, but I want a different house there. I buy the property, tear down the existing house, and build a new one.

Effectively, what value did I place on the house I purchased? Zero—or arguably less than zero, since I was willing to pay demolition costs just to achieve my goal of an empty lot to build on.

This is one way we measure land value: by looking for teardown sales. These are sales where someone buys a property, and shortly afterward, the building is demolished, and either a new structure is built, or the lot is left vacant.

Functional depreciation occurs when the existing house, regardless of its age or condition, isn’t what the market wants anymore. This phenomenon is more common in commercial real estate, but it happens in residential real estate too.

Urban redevelopment and land value

There’s a paper—I believe it’s by Clapp and Lindenthal, though I’d need to confirm—that outlines a modified depreciation method to capture this signal. The idea is that when entire subdivisions start being sold off and redeveloped, it’s a sign of functional obsolescence. The market value of those homes effectively drops to zero overnight because the market has decided, We don’t want these houses anymore.

At that point, buyers are bidding solely for the lots. It’s like ordering a hamburger and throwing away the bun because you’re on a keto diet. You valued the burger patty at the full price of the hamburger, but the bun was worthless to you—so you discarded it. Similarly, buyers in these scenarios are paying for the land and disregarding the value of the house entirely.

Patrick McKenzie

What’s the narrative for a community in a case like that? Not in the sense that an anti-gentrification activist might ask, but in terms of economic life—what’s happening when a subdivision built outside a city in 1970 suddenly makes sense to mass bulldoze and replace?

Lars Doucet

The likely answer is that the city has experienced significant economic growth. Incomes are up, wages are up, and perhaps the population is growing. People want better housing than what’s available in the existing housing stock.

Developers see an opportunity: they can buy for 1, spend 1 on improvements, and sell for 3. That’s what the market is signaling—it’s telling you that redevelopment is profitable. Developers generally won’t build unless they believe they’ll be able to sell. Of course, sometimes they miscalculate and end up stuck, but in cases like this, the houses flip, they get sold, and people buy them.

What this redevelopment tells us is that, at that moment, the property’s value was almost entirely in the land.

Land valuation is interesting because the most common method involves looking at vacant lot sales. The challenge, though, is that vacant lots can have a sampling issue—they tend to be what’s left over, often irregular in shape, size, or usability. You need to ensure you’re comparing apples to apples.

In areas with no vacant lot sales, appraisers use techniques like paired sales analysis. This involves finding physically identical homes—common in cookie-cutter subdivisions—that are in different locations. By comparing groups of these homes, you can determine the locational premium: the additional value a property derives from its location, assuming all other physical characteristics (age, quality, size, etc.) are the same.

The goal of all this analysis is to ensure properties are assessed accurately and equitably. Property taxes are one of the most hated forms of taxation, largely because they’re so visible.

Ironically, most people pay far less in property taxes than they do in income taxes, but income taxes are structured to be less noticeable. You might get a refund, which feels generous, and your employer withholds the payments before you ever see the money.

Contrast that with property taxes—you get a bill in the mail with a big number on it. Anyone who has ever worked freelance and dealt with income tax payments directly can tell you it’s a completely different experience. Freelancers often come to hate income taxes just as much as everyone else hates property taxes.

Small business realities and perceptions

Patrick McKenzie

This drives a significant part of the political economy of small businesses in the United States.

Theoretically, small business owners should be paying estimated taxes throughout the year. In practice, however, many don’t—including myself when I was a small business owner. Instead, they leave it until the end of the year, write a big check, and then think, Wow, small business owners pay so much in taxes, while all my friends with W-2 jobs don’t pay anything at the end of the year.

It creates a sense of injustice, even though those W-2 employees are paying taxes—it’s just being withheld automatically. This perception fuels a lot of political attitudes, but we won’t dive into the partisan implications here.

That said, contrasting income taxes to property taxes: I’ve never had the experience of calling up the IRS and saying, Hey, the W-2 says $80,000 in income, but I’d like to bargain it down to $70,000.

[Patrick notes: Fascinatingly this is how it has historically worked with Japanese small businesses.

To simplify, Japan has two methods to file taxes for a small business, a so-called “white” return, designed for smallest-of-the-small businesses which might have very informal recordkeeping, and a “blue” return, which is designed for more professionalized but still small businesses. Japan has attempted to transition more to blue returns over the years. (The return types are named for the color of the physical printed paper they used to be filed on. My annual “blue” return certainly looks black and white with marigold highlights to me, but for continuity’s sake, it says “blue return” at the top.)

But the way white returns worked was you’d take your extremely fragmentary records to the tax office, and your calculations, and have a conversation with the local tax officer. And that conversation would go something like: “Our store had sales of 10 million yen (~$100k) this year.” “We agree that your store had sales of 12 million yen.” “But unfortunately we only had sales of 11 million yen.” “Oh that’s a shame. What were expenses like?”

The first time the local tax office attempted to talk me through this process, there was a substantial cultural gap, because I was a foreigner fluent in Japanese who did not expect that a government employee would attempt to guide me through socially acceptable amounts of tax optimization.

There is a phrase in Japanese tax accounting which translates to 9/6/4. The government successfully identifies ~90% of the actual income of regular employees, ~60% for small business owners, and ~40% for farmers. Of course, choices regarding Japan’s taxation policy are properly the business of people who hold the franchise in Japan, and thus I could not possibly comment on how Japan ended up there.]

Lars Doucet
Right. It’s like saying, I think the market value of my income is less than you assessed me at.

With income and sales taxes, the market value is straightforward—it’s right there in the transaction. That’s what makes them relatively easy to tax.

Some people suggest replacing property taxes with a real estate transfer tax instead. The problem with that is political: nobody would ever want to sell their home. Transfer taxes are easy to assess because you just look at the sale price, but they can significantly disincentivize transactions.

For property valuations, one of the key factors in getting public acceptance is having a clear process for protesting assessments. Texas is an especially interesting example here. It’s a state with an all-Republican state government, which has been in place for decades. They’ve actively campaigned for property tax abolition for over a decade.

Texas is famously anti-tax, yet it has some of the highest property taxes in the nation. Why? Because we’ve essentially eliminated every other form of taxation except sales tax. That’s why we have such high property taxes—to fund schools, police, fire departments, and other essential services.

This contrast is one reason so many people are moving from California to Texas. California and Texas have diametrically opposite tax policies. In California, you’re taxed heavily on what you earn, but real estate taxes are relatively low. In Texas, it’s the reverse: you get to keep most of what you earn, but you pay higher property taxes instead.

Patrick McKenzie

If I can make an observation: with respect to people’s primary residences, every realtor in the world will tell you that buying a home is an investment—but it isn’t. It’s consumption.

So, when the property tax base is primarily residential, it behaves more like a consumption tax. And it carries similar dynamics in terms of progressivity.

Take the example of a consumption tax on food. Everyone needs to eat, but the dynamic range of how much people eat or spend on food is relatively compressed. At the lower to middle end of the income scale, people spend a larger percentage of their income on food. At the higher end, even if people spend more in absolute dollars on luxury items like champagne and caviar, there’s only so much of it they can consume.

Lars Doucet

There are limits to consumption—only so many whole songbirds you can eat.

Patrick McKenzie
So, as someone moves from, say, the upper middle class to a higher economic strata—think of someone who’s done very well in Silicon Valley and relocates to Texas—they may find their housing consumption effectively capped. As a result, the percentage of their resources spent on housing might be lower than the typical percentage for the broader population.

While they’re paying more in absolute dollars, their relative tax burden has a regressive effect, similar to a regressive income tax. Does that match your intuitions, or am I off base?

Lars Doucet
I think that’s broadly correct. To frame it another way, Republican parties in Texas have campaigned on abolishing property taxes for years, but they’ve never done it. The reason is simple: what’s the alternative? Raise sales taxes? Raise income taxes? Everyone hates those even more.

We’ve already run the experiment of eliminating property taxes—it’s called California. There, you get high taxes on everything else, and you end up with a landed gentry.

Texas, by contrast, has some sensible laws around homesteads. For example, I believe there’s a law stating you can’t lose your homestead in a lawsuit, though I’d need to confirm that.

Patrick McKenzie
For those without a real estate background, what’s a homestead? Is it related to the Homestead Act from U.S. history, or is it something different?

Lars Doucet
A homestead is the house you own and occupy—it’s your primary residence. This is relevant in tax terms because homesteads qualify for a homestead exemption, which is the most common tax exemption people are familiar with. It’s meant to incentivize homeownership.

In Texas, they recently raised the homestead exemption, which is often how property tax relief campaigns play out: “We’re going to abolish property taxes!” followed by, “Okay, we raised the homestead exemption slightly. Mission accomplished.”

Homestead exemptions can be complicated to adjudicate because they’re tied to the homeowner, not the property itself. For example:

• If you own a house you rent out, neither you nor the renter gets a homestead exemption for that property.
• Married couples are entitled to only one homestead exemption. If they separate but remain married, only one spouse gets the exemption, and the assessor has to determine which one.
• Assessors even need to investigate potential fraud, like whether a couple is divorced or just separated, to ensure they’re not double-claiming exemptions.

This adds administrative expense and complexity, which taxpayers ultimately fund.

Patrick McKenzie
And you’re saying property taxes are fundamentally tied to both land and buildings.

Lars Doucet
Right. Property taxes are levied on land plus buildings. The problem is that taxing buildings discourages development—we get fewer buildings on the margin. This is where the economist Henry George comes in. He famously argued for taxing only land, not buildings.

If Texas truly wanted to abolish property taxes, one way to approach this would be a universal building exemption. This would exempt buildings from property taxes entirely, offering more tax relief to the median homeowner than a homestead exemption.

It would encourage development, simplify assessments, and shift the tax burden more equitably. The big losers would be those who own undeveloped downtown lots—like parking lots in Houston—which should be converted into housing or something more productive.

A universal building exemption would also eliminate the tangled mess of exemptions we have now. Some jurisdictions have exemptions for vacant land, commercial properties, and all sorts of specific cases. In one major American city I recently reviewed (not Chicago), nearly half of all commercial property was fully exempt from taxes. Since most of the real estate value was tied to commercial properties, homeowners ended up making up the difference.

Replacing these complex exemptions with a simpler, universal building exemption would reduce administrative burden, encourage development, and provide meaningful tax relief to average homeowners.

Patrick McKenzie
And presumably there's also some, large amount of differential I suppose.
In one sense, there’s differential willingness to pay property taxes, and consequently, differential predisposition to haggle over the totally not a bill that will become a bill when it arrives in the mail.

Property tax protests

Lars Doucet
Oh, absolutely. Do you want to talk about protests and defenses against property tax assessments?

Patrick McKenzie
Sure. Let’s frame this. I’m a homeowner in the United States, where roughly 60% of the population owns homes. That group includes people who are very sophisticated about dealing with the government and those who aren’t.

So, when a homeowner protests their property taxes, are they typically navigating the process—figuring it out as they go along—or is there a cottage industry to help them?

Lars Doucet
Yes, and it’s not a cottage industry—it’s a tech industry. There’s a whole crop of startups called property tax agents. These companies have noticed that most homeowners don’t protest their property taxes pro forma every year, even though they arguably should.

The way these companies work is simple: they automate the protest process for homeowners. They file protests on behalf of the homeowner, aiming to reduce the property taxes, and then take a cut of the savings as their fee.

Assessors tend to dislike these agents—not because they’re trying to overvalue homes, but because these companies essentially spam the system. The property tax agents often file protests with minimal effort, clogging up the process.

That said, I’ll give you some practical advice on how to effectively protest your property taxes. Assessors don’t want to overvalue your home because they can get dinged by their oversight committees for doing so. They’d much rather assign the correct value, though they won’t willingly undervalue it either.

Here’s the strategy:

1. Pull three to five comparable sales (comps) for properties very similar to yours.
2. Ensure these comps are legitimate sales—not, for example, a $5,000 house sold to someone’s nephew.
3. Calculate the median value per square foot for the comps, then multiply that by your home’s square footage.
4. If your property’s assessed value is higher than that number, present this evidence to the assessor.

If you provide credible evidence, assessors are more likely to work with you. The best assessment districts have a huge backlog of protests to clear, so they prioritize cases where the homeowner is serious and has evidence. These districts operate like a sales funnel: at every step, they give people opportunities to exit the process.

If you bring solid evidence, you’ll likely get a reduction without wasting their time. The property tax agents, on the other hand, usually do the bare minimum to secure the easiest reduction possible. They won’t necessarily fight for the maximum reduction you could achieve.

From the assessor’s perspective, the volume of these low-effort protests can be incredibly frustrating.

Patrick McKenzie

It’s like my Charge More negotiation advice except applied in reverse.

[Patrick notes: I have long advised software people, who are frequently abominable at pricing themselves or their products, to always ask for more than either their future employer’s initial offer or more than they’re comfortable with (when they price products). This is on the theory that asking is free, worst case scenario is you’re back where you started, and in expectation this will shake out substantial money.

This is the lowest cost and dumbest negotiation strategy which can be called a negotiation strategy. And so I’m analogizing it to the lowest cost and dumbest protest strategy which is still a protest strategy, which is “reverse” in the sense that its “I want you to Charge Less.”]

Lars Doucet
Exactly. What often happens is that these property tax agents file a ton of protests, and the assessment district responds with, Great, show up for your informal hearing and make your case.

But then nobody shows up. The agents don’t send representatives for each protest—they have hundreds or even thousands of cases to manage in a day. So the district just marks the case as resolved, because the protester didn’t follow through.

In principle, there’s nothing wrong with protesting your property taxes. In fact, property taxes have more built-in accountability than almost any other type of tax. As you mentioned earlier, you can’t haggle with the IRS about your income tax.

The issue arises when people try to get special treatment that places their property value below market value. That creates equity problems for everyone else in the neighborhood.

Accuracy vs equity testing

Here’s another interesting point: most successful property tax protests aren’t based on accuracy—they’re based on equity. People aren’t protesting because the average percent error is off; they’re protesting because Bob’s house was assessed at one value, and mine was assessed higher, even though our homes are basically the same.

This perception of inequity drives protests. It’s predictable to the point where you can forecast where protests will come from with surprising accuracy—and without needing fancy tools like machine learning.

Patrick McKenzie

So, hypothetically, a software company could sell appraisal districts a kind of Minority Report for protests. They’d say, Of the following 30 homes, you’ll get protests from about 12 of them this year.

Predictive protests

Lars Doucet

Exactly. I haven’t developed software like that specifically, but I know it exists. And it’s not as nefarious as it sounds—it’s not pre-crime.

What typically happens is more straightforward: certain types of properties or neighborhoods tend to generate protests because they’re systematically over-assessed. If you can identify those patterns, you can adjust assessments across the board to hit a sustainable target. It’s more about improving accuracy and equity than anything sinister.

Patrick McKenzie

This is nice. The optimistic view is that many government organizations in the United States lack a tight feedback loop—whether from user feedback, tracking metrics, or analytics—and as a result, their realized state capacity falls short.

What you’re describing sounds like a software upgrade for state capacity: a system that flags potential bugs probabilistically, shows you what they look like, and then feeds that back into the process to reduce bugs in the future.

Lars Doucet

Exactly. But it’s not just about flagging people likely to protest—it’s about ensuring accurate and equitable assessments for everyone. These are testable metrics. Let me explain briefly.

Accuracy is measured using sales ratios: how close your assessed values are to actual sales prices. You can run all kinds of predictive algorithms to match the sales data, and many of them will look great on the surface.

The problem is that I can create a "perfect" algorithm without using any computers: just value every property at exactly its last sale price. This is known in the industry as sales chasing.

Sales chasing is tempting for inexperienced assessors. If a property sold for $100,000 or $237,000, why not assess it at that value? The issue is that it ignores the broader market tendency. It also leaves you with no plan for valuing properties that didn’t sell. Since not everything sells every year, you end up with neighborhoods full of wildly varying valuations based solely on which properties sold recently.

To counteract this, you need equity tests. Accuracy measures how close you are to sales data, while equity examines valuation consistency across similar properties.

For equity testing, group properties by neighborhood and physical characteristics. Within each group, the variation in valuations should be small. This ensures properties with similar features are assessed similarly, avoiding unfair discrepancies.

Goodhart’s Law states that when a measure becomes a target, it ceases to be a good measure because everyone starts gaming it. To mitigate this, you use two cross-metrics—accuracy and equity—that are difficult to optimize simultaneously. Checking both helps keep assessments honest.

Here’s how I approach this:

• I deliberately run garbage models, such as random valuations or assigning every property the median price.
• Then I “cheat” by injecting actual sales data into these garbage models to simulate sales chasing.
• I analyze whether the statistics I’m using can detect that the model is garbage despite its apparent alignment with sales.

This kind of testing helps advance the statistical methods we use, and I’m working on publishing papers about it.

The simplest way to validate assessments, though, is to plot all the values on a map. Once your assessments are provably accurate and equitable—matching comps and showing consistent valuations within neighborhoods—you map them out.

If you see a giant anomaly, like a property valued at half the rate of its neighbors, you investigate. Maybe it’s justified because the house burned down last year and hasn’t been rebuilt. Or maybe it’s because the owner is politically connected, which raises equity concerns.

Patrick McKenzie

I’ve got a question for you. The idea of putting things on maps on a computer screen is slowly making its way into various government offices in the United States.

What’s the last year we’d expect no appraisal office to have maps on a computer screen? And back then, what did the physical setup of an appraisal office look like? Was it a giant map on the wall? Filing cabinets stuffed with papers?

Lars Doucet

I’m not entirely sure—it depends on how far back you want to go.

Right now, though, there’s a wave of retirements happening in local governments. Everyone is short-staffed and scrambling for replacements. It’s a prime opportunity for software to move into this field.

Interestingly, assessors’ offices are ahead of the game compared to other local government units. In the last five years or so, many districts have adopted open data portals. Most of these portals seem to use the same open-source web stack, often with Esri tools integrated.

If I work with a random assessing district and Google the district’s name plus “open data portal,” I have better than a 50% chance of finding a portal with almost everything I need, including a map. It might not be the most polished interface, but it’s there.

Some more advanced districts let you look up individual properties on demand, compare them to others, and access detailed data. I think this trend will only accelerate as new staff and systems come online. Assessors’ offices might innovate faster than other local government units because of this turnover.

Patrick McKenzie

I know a little about the political economy of taxation and taxpayer preferences, but I know more about how to get people to adopt software.

If this industry went from zero to widespread software adoption in a few years, there must be a story behind it. It probably wasn’t a "software meteor" hitting overnight. Was there an outside actor? A new law? What explains this step change in adoption?

Lars Doucet

It wasn’t just assessors’ offices—it was local governments in general. Many of them share the same software suite. If you Google "open data portal" for one city and compare it to five others, you’ll see they all use the same stack.

This shift likely started at some point, though I’m not sure when. Someone with 30 minutes on archive.org could probably trace the history of the stack. It’s probably tied to some legislation and an opportunistic open-source consortium—maybe even government-led—that made this happen.

Speaking of open-source tools and local governments, you’re in Cook County, right? The Cook County Assessor’s Office is well-known in the assessing industry.

Cook county assessor's office

Patrick McKenzie

For the benefit of people who don’t read the Chicago Tribune or follow industry news, can you explain why that is?

Lars Doucet

Sure. First, I want to clarify that I’m talking about their modeling department specifically. This is separate from Chicago’s broader reputation for local politics or its ongoing disputes—I’m not deeply familiar with those.

The modeling department at Cook County’s Assessor’s Office is notable because it has one of the most sophisticated GitHub repositories for automated valuation models (AVMs) I’ve ever seen. It’s incredibly advanced. [Patrick notes: You should have seen me at this point; my jaw nearly unhinged. Yes, the Cook County Assessor’s Office Github is indeed a thing that exists.] 

That said, as a Cook County resident, you might have more insight into the day-to-day property tax situation. From my perspective, there seems to be a significant gap between the sophistication of their modeling tools and what actually happens on the ground.

Patrick McKenzie

I actually know almost nothing about the Cook County property tax system, but I don’t usually get a homework assignment during these episodes. I’ll take this as one and hopefully follow up sometime.

[Patrick notes: While my dad gave me an excellent education in all things real estate while I was growing up, mostly just by geeking out about his special interests, he felt family finances were not something children should ever worry about, and so I got nothing with respect to taxes. Then I left Chicago for 25 years and haven’t read about the system in the interim. Which would probably have been better to do prior to purchasing a house here, but it has been a busy year since immigrating to my hometown.]

Lars Doucet

Fair enough. I just want to be careful here—I don’t want to stick my foot in it and say something substantive about specific assessments or politics.

First, I never want to slag off an assessor, especially not on public airwaves. Second, I know these are often politically sensitive issues. If I say I enjoy working with one group, someone might say, Well, they overvalued my house by four times! and get upset.

One thing I emphasize is that you can have a great assessing department that has no control over the political situation but ends up getting blamed for it. That’s the caveat I’m trying to convey here.

And since I know Chicago is full of potential landmines for me to step into, I’ll just stick to objective observations: the Cook County Assessor’s Office has the most sophisticated open-source automated valuation model repository on GitHub I’ve ever seen. I’d love to geek out about it—it’s genuinely fascinating.

Lars's background

Patrick McKenzie

Awesome. I’ll check out that repository.

So, you’ve written extensively on these topics. I think the first thing I read from you was your entry for the Slate Star Codex essay competition. Want to share the history of that?

Lars Doucet

Sure. It was a book review contest—you read a book and write about it. I reviewed Progress and Poverty by Henry George, which was the most popular economic treatise of its time. It was written during the Gilded Age.

In the book, George asks a compelling question: why is it that as material progress improves, poverty persists—and sometimes even worsens? It’s not just that some people get left behind. There seems to be a paradoxical relationship between material progress and a certain kind of poverty.

To illustrate, he compares two cities: San Francisco and New York. At the time, San Francisco was a frontier town, nowhere near as wealthy or sophisticated as New York. But despite its relative lack of wealth, San Francisco didn’t have the grinding poverty seen in New York.

George points out that in 1879 San Francisco, there was a limit to how poor you could be, whereas in New York, you’d find truly desperate poverty. He makes a prophetic observation, which I’ve half memorized:

"If San Francisco does not yet have as much desperate poverty as New York now has, is it not because San Francisco is yet behind in all that both cities are striving for? For who can doubt that when San Francisco arrives at the point where New York now is, there will also be barefoot and ragged children on her streets?"

Patrick McKenzie

Goodness, they could write back then, couldn’t they?

Lars Doucet

Absolutely. And what’s remarkable is that George’s prediction came true.

His argument boils down to this: land is the one truly scarce resource. It’s the last irreducible scarcity because everything depends on it. You need land to do anything, and location is inherently valuable.

Over time, all value reduces to location. We can compete away the marginal costs of everything else, but land remains—and owning the right land in the right place allows you to rent-seek from everyone around you.

George argued that both capitalists and Marxists were wrong. It’s not capital or labor that’s the central issue—it’s land. He believed we needed a serious conversation about land and proposed a land value tax as the solution.

I think history has proven Henry George right in many respects.

When I wrote that book review, I didn’t think much of it. But it ended up winning the contest, becoming wildly popular, and prompting readers to ask, Nice theory, but does it work?

That led to requests for follow-up articles. I wrote those as guest pieces, which eventually drew me into the field of mass appraisal. I realized that a significant obstacle to reforms like a land value tax is simply knowing what property is worth.

You can’t even advocate for such reforms unless you know the value of the land. Otherwise, you don’t know who the winners and losers will be. People have all kinds of wild assumptions about this.

In some places, like San Francisco, almost all the value is in the land. In Japan, where buildings are often torn down and rebuilt quickly, the value is also predominantly in the land.

But in suburban Sprawlsville, USA, land might only account for 15–35% of a typical property’s value. In such areas, a land value tax wouldn’t hurt anyone significantly—at least at an urban level.

So, that’s my origin story. It’s how I transitioned into statistical science from a background in computer science and video game development. I’d always wanted to explore statistics, machine learning, and data.

Developing real-time simulations in a spatial coordinate system was natural for me, so moving into GIS and mapping felt like a straightforward evolution of those skills.

What is GIS?

Patrick McKenzie

GIS is a topic I wish more people knew about. Can you give a brief explanation of what GIS is as a software artifact and why it comes up in so many government-related contexts?

Lars Doucet

Sure. GIS sounds more complicated than it is—it’s just computers plus maps.

In practice, GIS involves reading and writing shapefiles. A shapefile is a specific file format, originally developed by Esri (the makers of ArcGIS and other GIS platforms). Much like how "Kleenex" is often used generically for tissues, "shapefile" has become shorthand for spatial data formats.

There are other spatial file formats as well. Modern formats include Parquet, which is fantastic, and GeoJSON, which is not so great. Regardless of the format, GIS software deals with data in a spatial context.

Think of it as a spreadsheet where each row isn’t just a data point—it’s tied to geometry. For example, a parcel of land might include latitude, longitude, and a shape. You can then load that data into a mapping application and visualize it on a map.

GIS also involves special considerations like projections. For example:

• Are you using a Mercator projection or an equal-area projection? This matters if you’re calculating area.
• Are you calculating distances? Then you need an equal-distance projection.
• Is your data in feet or meters? The unit matters.

These details are critical for accuracy when working with GIS data.

Patrick McKenzie

The upshot is that governments, like the California government I’m most familiar with, employ more GIS experts than you’d expect. Some work in semi-centralized service bureaus, handling GIS tasks for various state agencies. Others are embedded in state-level departments like health, education, and so on. GIS programmers are everywhere.

Lars Doucet

Exactly. What’s interesting is that even in rural America—“Dinkoville, USA,” if you will—there’s more GIS work happening than you might think.

While there’s a chronic shortage of skilled GIS staff everywhere, the demand exists because every county or municipality needs to maintain parcel maps. Anytime someone:

• Splits a lot,
• Builds a Walmart, or
• Develops a new subdivision,

someone has to update the parcel data.

Nowadays, this work happens in the IT or records department of even small towns. Someone, let’s call them Jimmy or Jane, is entering shapefiles into a computer and keeping the maps up to date.

Patrick McKenzie

In the old days, would this have been done with pen and paper?

Lars Doucet

Exactly. It’s meticulous work, but it used to be all pen and paper. If you’ve ever read a survey deed, it’s written in this almost ye olde language. For example:

"Thence from the marker at this corner, proceed forth X feet, and thence…"

It’s essentially one giant run-on sentence encoding deterministic spatial instructions. The amazing part? That text can be parsed. With the right tools, you can create a parser that reads those instructions and generates spatial data.

Patrick McKenzie*eyes light up* That’s incredible!

[Patrick notes: I love these descriptions in both Tokyo and Chicago, and will take the liberty of posting a fictional one so you get the flavor.

LOT 7 IN BLOCK 22 IN THE MIDWEST LAND IMPROVEMENT COMPANY'S SUBDIVISION OF THE WEST 1/3 OF THE EAST 1/2 OF THE SOUTHEAST 1/4 AND THE SOUTH 25 FEET OF THE NORTH 1/2 OF THE EAST 1/2 OF THE SOUTHEAST 1/4 LYING SOUTH OF THE RIGHT OF WAY OF THE CHICAGO UNION ELEVATED RAILROAD, IN SECTION 9, TOWNSHIP 39 NORTH, RANGE 14, EAST OF THE THIRD PRINCIPAL MERIDIAN, IN COOK COUNTY, ILLINOIS.

In Tokyo they’re even better because the recitation of them at the closing ceremony goes into pedantic detail about how ownership of the private road north of the parcel is subdivided. Then you are read a phrase which rhymes with “And you shall place no structure nor permanent fixture atop your private road, as by ancient practice the right to pass unhindered is ceded to the people of Tokyo.”

What is the law, after all, if it is not a system of ritual magic.]

Lars Doucet

It’s a well-specified language with a consistent grammar. Well… mostly.

Patrick McKenzie

Mostly. As much as I’d love to dive into all the edge cases and the history of the legal construction of property in America, I’m flagging a bit.

Where can people find you on the internet, Lars?

Lars Doucet

You can find me on BlueSky or Twitter at @LarsiusPrime. My website is fortressofdoors.com.

Patrick McKenzie

Awesome. Thanks so much for the fascinating deep dive into these topics.

And to the rest of you folks, thanks for listening! We’ll see you next week.