The Twenty Year House Value Study

The East Greenwich Town Council recently "conducted a 20 year analysis of median house values" from FY1998 to FY2017, including median tax bills.

Among the conclusions stated were that, over the period of the study,

This is the latest installment in a continuing series of communications intended to create the impression that the town is on an unsustainable path with respect to property taxes, and that drastic actions like level funding the schools are called for.

Like the resident mailer from last Spring, it is easy to tell at a glance that these numbers are wrong.

If tax bills had increased faster than property values, the FY2017 tax rate of $24.06 should be higher than the FY1998 tax rate of $26.22, but it is not.

So, why are the numbers wrong?

Flawed Computation of Tax Bills

East Greenwich property owners recently received an official notice jointly issued by Northeast Revaluation Group,LLC and the East Greenwich Tax Assessor.

Referring to the December 31, 2017 valuation, it states:

Please do not multiply this value by the current tax rate. Your assessed values do not reflect any exemptions to which you may be entitled

Yet, this is exactly how the supposed median tax bills were computed for both FY2017 and FY1998.

Ignoring exemptions makes tax bills look are higher than they actually are, and invalidates the computed medians.

Failure to Use Same-House Changes

A more subtle problem with the methodology involves the comparison of the overall median tax bills and property values twenty years apart.

The need to use "same house" metrics for this kind of analysis is discussed in the "flawed metrics" section of the Pro Bono Analysis page.

This is important because between FY1998 and FY2017, nearly 300 new homes with a median value in the $750,000 range were added to the tax base. An apples-to-apples comparison requires that you look at the FY1998 and FY2017 assessments and tax bills for each individual home, house by house, and compute the median of those changes.

Failure to do so produces inflated estimates of the actual increases in valuations and tax bills due to new construction.

Every analysis of median property values published by the Town Council has failed to do this.

Exclusion of 1998 Fire District Taxes (which are included in 2017)

Another obvious error is the exclusion of the Fire District taxes in the FY1998 numbers, which makes the increase in taxes appear larger than it actually was.

Inexplicable Exclusion of FY2018

It is not clear why FY2018 was excluded from the analysis. The FY2018 tax roll has been posted on the town website for months.

Including FY2018 would produce a slightly lower estimate of the growth in tax bills, because the FY2018 rate of $23.66 is lower than the FY2017 rate of $24.08.

Although the title would lead you to believe that 20 years of data were analyzed, in fact only two years of data (that happen to be twenty years apart) were used.

From the years 1998-2017, you can make 190 different pairs, and each gives a different estimate of the medians and growth.

There are much better ways to estimate trends that use all of the data, and they are much less amenable to cherry-picking the start and end years.

Coincidentally, FY1998 is the only year in recent history with a 4% increase in the tax rate within a revaluation cycle.

Getting it Right

Extracting the individual tax bills from the FY1998 and FY2017 tax rolls and matching the contents house-by-house produces a list of 3,619 single family (state code 01) homes that appear on both tax rolls.

Computing the percentage change in the assessed values and actual tax bills for individual homes produces the following median percentage changes:

These annualize to 3.8% and 3.4% per year over the interval.

As noted earlier, the result of any trend analysis based on only two points is likely to be quite sensitive to the choice of starting and ending years, so if you insist on using only two data points, it is a good idea to conduct a "sensitivity analysis" by examining the results for different choices.

If we repeat the analysis using FY1994 and FY2018 as the endpoints, the results are:

The annualized values are 4.4% and 3.3%, respectively.


The Town Council's study contained numerous flaws that render its conclusions invalid, while a more careful analysis found no evidence that the tax bills of single family homes are increasing faster than their assessed values (if anything, valuations are increasing faster).

I asked three economics professors if a situation where property val- ues were increasing faster than property taxes would be considered sustainable, and all three said essentially that it's a no-brainer, as long as you are able to come up with the cash to pay the tax bill it is sustainable.

A list of the 3,607 properties used in this analysis, ordered by per- centage increase in their tax bills, is posted here.