Revisiting a Botched Revaluation Process

Five months of silence means everything is fine, right?

Belfast performed a real estate revaluation in 2024. This is the process that is meant to ensure that your property taxes are fairly determined. Even though these are required to be performed every 10 years, Belfast waited 20 years to perform one. Even though our city is required to view each property in-person every 4 years, and even though properties in Belfast had not been viewed in 20 to 30 years, our revaluation made no plans to at least belatedly meet state property valuation visit requirements. Even though the state lays out a thorough process for how revaluations should occur, and even though Camden and Rockland followed this process, and took it seriously (it’s a two-year process), Belfast’s officials believed they could shortcut the standard process and stretched this process out over several…months. I don’t want to write a book about what went wrong (and I’m sure you don’t want to read it either), so today I’m just going to highlight one area that went astray — land valuation.

Revaluations are also known as mass appraisals¹. If you’ve ever purchased a home with a bank loan, you paid for an appraisal. The bank requires you to purchase this service to ensure that the value of the home is accurate relative to the loan they are providing you. They do this to prevent fraud and misuse of their services. It also hedges against the lender losing money if the borrower defaults. A mass appraisal is an attempt to value numerous properties at once. A town assessor is essentially its appraiser.² Belfast’s valuations for buildings and land use a classic appraisal methodology called the cost approach. In order to generate usable cost data, market sales of land are analyzed³. Once that is done, cost tables can be created and applied to properties in the town.

The authoritative text for new real estate appraisers is a book called, rather dryly, The Appraisal of Real Estate, by the Appraisal Institute⁴. This book is one every appraiser and every assessor will rely on for fundamentals of their profession. There are 30 chapters in it. Chapter 28 is titled “Statistics in Appraisal”. This is the only chapter that covers what appraisers need to know about statistics; the text covers concepts like how to find the mean, median and standard deviation of a set of data, regression modeling and other basic statistical concepts. They do this so that appraisers and assessors will at least have a basic grasp of where the data they are using comes from. Appraisers and assessors are not statisticians. They just need to know the basics to avoid misusing data.

Belfast’s revaluation process produced a set of land schedules. There are 25 zoning districts in Belfast, and 13 land value tables — land values are roughly tied to combinations of zoning districts⁵. Let’s keep in mind that this is a town with about 4,500 parcels, with only around 3,300 homes (the 1,200 difference is due to parcels of vacant land), and that zoning districts reflect policy, not market demand. Real estate markets don’t care about zoning codes. They care about view, access, buildability, amenities, and comparable sales. Tying land values to zoning creates arbitrary groupings that mask true property differences. That’s one problem.

One piece of land, a 0.55-acre lot with no special features, previously valued at $43k in 2023, was valued at $200k in 2024. In 2023, Belfast’s valuations were certified by the state as being 92% accurate. By 2024 they had dropped below 70%, which was the trigger for a revaluation. Even with a 22% quality decrease, a revaluation that increased a piece of land 460% over its old value should have been a giant red flag about the reliability of Belfast’s cost approach and the data used to generate it. After all, if the data was 70% reliable, this would lead to an expected new land valuation of $62k. This is precisely why revaluations take two years — because addressing data‑quality issues takes time⁶. One of the key requirements of any revaluation in Maine, per the state’s own training guides⁷ for assessors is this:

In practice, assessors will value residential property using the cost approach and confirm the valuation through the market approach.

This means that when you have such a large divergence, assessors must apply the market approach to diagnose errors. Had this been done, all the comparable sales of land between 7/1/21 and 6/30/24⁸ would have shown nine comparable sales with a range of $30k to $99k⁹. High-end sales were in-town lots and low-end were places with high traffic or elevated levels of economic/external obsolescence¹⁰. So how did Belfast come up with a land valuation of $200k on the land schedule for a property where there was not a single sale even remotely near that value? Bad data collection, sorting and failure to do quality checks¹¹.

Belfast is now 100% certified — meaning our valuations are, on average, correct. The problem with averages is that if you have $999,999 in your bank account and I have $1, we have an average of $500,000. You can have your average valuation be accurate, but how you get there can be wildly skewed if you have bad data. So, for every $1 million in valuation that is overapplied to modest homes, there’s another $1 million that is underapplied to very expensive homes. This is the Reverse Robin Hood Taxation Effect¹², something serious enough to have a real name (Vertical Inequity) and a whole bunch of really smart people like the nice folks at the Lincoln Institute of Land Policy in Massachusetts warning communities about it and explaining to them how to fix it (better data, better modeling). I’ve said that we should all be mad about this, but it seems that we’re not. This is now a demonstrable fact in Belfast, and it has really severe effects such as:

1. Overtaxing older modest rental properties (rents go up).

2. Overtaxing modest homes (working class gets pressured).

3. Creating a de facto tax policy that discourages affordable housing construction (small-scale developers less willing to build non-subsidized market-rate housing).

4. Opens up Belfast to lawsuits over unconstitutional taxation policies (failing to meet the definition of “just value” in the Maine Constitution¹³ for taxation purposes).

Water Views are Worth…Nothing

Over the course of about one year in my neighborhood two “contractor special”¹⁴ homes sold. One sold in 2015: a one-level ranch built in 1967 with 1,500 sq. ft. of living space and a 1.5 car garage. The other in 2016: a one-level ranch with 1,600 sq. ft. and a 1.5 car garage built in 1974. These homes could not be any closer in original style, design, function or condition. The 2015 sale went for $325,000 and the 2016 sale went for $135,000. The difference? One has a sweeping view of the bay (and nearly has waterfront access¹⁵) and one has good views of Route 1 traffic. The real estate market valued this difference at nearly $200,000 back then (before inflation took hold). This is pretty common sense, right? Belfast is a coastal community, and there are plenty of people moving here for waterfront and water view homes, and they’re willing to pay for the privilege.

Water view lots are considered desirable and sell for more money. It stands to reason that homes with water views should have associated valuations reflecting this — and back in 2016 our tax rolls did reflect this, even if the value gap back then was still somewhat unfair and out of tune because Belfast’s revaluation was already past-due then at 10+ years. In 2016, the water view home had a land valuation of $154,600 and the traffic-view lot was valued at $33,400, a difference of $121,200. A revaluation 20 years after the last one should have brought these land values more in line with reality, which today would probably reflect a $250,000 to $300,000 value gap.

When I received the valuation file, you know what it said? The water view home now had a land valuation of $214,100 and the modest home’s land value was $199,600. Both homes had the exact same base land valuation of $225,000, only adjusted for a small size differential — the water view differential was now gone. When I asked the assessing department about this, I was told they did not adjust for water views in their revaluation process during 2024. This is a huge failure. You cannot have a fair revaluation if you are not going to consider the actual market value of plots of land with water views. This isn’t even just unfair, it’s unthinkable in appraising, because it is an idea with no basis in reality.

An appraisal would likely value the non-water view lot around $60k and the water view for at least $350k, which totals $410k. Their skewed valuations presently add up to $414k. If both houses were knocked down and the lots sold on the open market this year, the impact on Belfast’s state-required sales ratio test would be negligible. I discuss this to demonstrate how a city can get its valuations so wrong and continue to meet the bare minimums of the certification process.

The effects here are pronounced — the owner of a modest home is spotting a wealthy homeowner $2,100 in tax savings. If you haven’t figured it out already, my family owns the modest home here. So, we’re personally subsidizing people that can afford to pay market rates for water views. Of course we’re angry about this. We all should be, because a constitutional failure to apportion taxes correctly is a failure of justice.

Statistical Failure

Belfast’s revaluation process appeared to consist of just two entities: our assessor and a single contracted appraiser. I will repeat, appraisers are not statisticians, and they must be extra careful to understand how to use data appropriately. Had they brushed up on the Appraisal of Real Estate’s statistics chapter, they would have learned about the Central Limit Theorem which, when properly applied, requires anywhere from 15 to 30 pieces of data to have statistically meaningful results. Mass appraisal standards, recognizing the limits of market sales data, says that 5 pieces of data is adequate¹⁶, but that 15 is desirable. So, depending on which part of the scale you want to land on, for each of the 13 categories in our land schedules, Belfast would need 5 land sales within the accepted time-period to meet the bare minimum, and 15 to have the data be considered reliable. That’s anywhere from 65 to 195 land sales, minimum.

I asked for the land sales data used for the revaluation and looked through it. There were 24 qualified land sales during the time period, total¹⁷. Belfast did not even have enough data to justify two land schedules, let alone thirteen. The reality is that 24 sales cannot possibly capture the market nuances of dozens of neighborhoods. With such a small sample, any outliers or peculiarities can skew the model, and large extrapolations must be made to fill gaps, leading to unreliable valuations.

But we went through with the revaluation with bunk data anyway. When I called for a thorough revaluation, it’s not because I like spending money — it’s because a proper process with public information sessions, laying out transparent data and methodology, checking your work, asking for feedback, and hiring experts to fix vertical inequity issues takes time and money. The Belfast City Council abdicated their duty here by soliciting and signing off on a non-public revaluation process at minimal cost.

The City Council has swept the valuation issue under the rug. Valuations have not appeared as a single agenda item in the past five months, and the last time it was discussed (in October), the only concerns that any councilor raised with the process were the plight of waterfront property owners (I kid you not¹⁸). This means there are no intentions to fix our broken valuation system and no plans on fixing the inequities in our system. You can tell me that we’re a housing-friendly city that looks out for our working class residents but be warned — I’m going to point you to our valuations, look you in the eye and ask you “Do actions speak louder than words?”. Because no community that cares about equity and fairness would continue to double down on this broken system. It’s time to do valuations correctly.

---

Prior posts about valuations in Belfast:

1. I know a lot about appraisals because my mother was an appraiser for some time and used to own a real estate appraisal business that employed several appraisers — My father & I worked in this family business for several years.

2. While there are different tests and licensing standards for appraisers and assessors, the fields have a high-level of overlap. They way to become qualified to become an assessor is by being an appraiser first. Assessors use appraisal valuation methods.

3. We’re not getting into how buildings are valued today.

4. Now in its 15th edition. I quote from the 13th, because that’s what I have access to.

5. with some differentiation for waterfront land.

6. Nothing made the Founding Fathers more angry about British rule than unfair taxation, right??

7. per STATE OF MAINE MAINE REVENUE SERVICES PT 101 Introduction to Property Tax Assessment.

8. This is the largest period of dates Belfast was allowed to use, per accepted valuation methodology. Comparable lot sizes would be 0.25 to 1.0 acres.

9. comparable sales, for this exercise, are lots between 0.25 and 1.0 acres, non-waterfront, non-downtown. Data available upon request.

10. This is coded appraisal-speak for “too many poor neighbors”, such as a lot that’s flanked by two run-down single-wides.

11. If you’ve been following this newsletter, you already know which homes have ridiculously elevated land valuations attached to them: Modest homes. The problem with the data was that it was not segregated enough and not supplemented with additional data or appraisals.

12. I just made this term up, but it’s appropriate. Maine has two tests for certification of valuations - the other is based on divergence from the statistical mean. There is no regressivity testing, but failure to correct regressivity can put a community in hot-water, constitutionally. At a later point, I may demonstrate, with statistics, how a community can pass its two-part test and still have horribly skewed results that under value expensive properties and overvalue modest properties.

13. Maine Constitution Art. IX, § 8. Taxation

14. In real estate, a "contractor special" refers to a property that requires significant work or repairs and is priced below market value to appeal to contractors or investors.

15. Waterfront properties are valued even more. This property is attached to a sliver of land that is technically owned by a property owner in Searsport.

16. Five is adequate in large homogeous neighborhoods, such as a housing development of 500 homes, all situated on 0.25 lots and all built within ten years of each other. Suffice to say, this part of Maine has a lot more variety and needs better data than the bare mimimum.

17. The state has a process for insufficient data. Belfast should have realisticly limited land valuation to a few schedules - waterfront, water view, in-town, downtown, everywhere else. Where data falls short they can extend the time period and commission appraisals to fill in the gaps. Yes this costs money. That’s why Camden spent $225,000 and Rockland spent $300,000 on this process. No idea what Belfast spent (lack of transparency), but it surely had to be way less than $50,000. You get what you pay for.

18. On Tuesday, October 16. Fast forward to item 10.A on the agenda at 34:00 if you want to see the valuation discussion.