Column by Richard Bedard: Nassau County reassessment model flawed

The Island Now

Thousands of local homeowners are furious. After a mass reassessment, Nassau County says their house is worth tens of thousands of dollars more than its fair market value.

They’d be even more furious if they knew what I know: that many of their neighbors are getting big property tax breaks of 10 percent, 20 percent, even more than 40 percent. Call them the silent winners of this Reassessment Mess.

I know all this because I took a deep dive into the county’s property database. I’m a journalist and former Bloomberg News editor. Early in my career, I wrote a three-part investigative series about property tax unfairness in southeastern Pennsylvania.

My research showed that the reassessment that County Executive Laura Curran boasted would fix “corrupt” values simply installed a new set of “corrupt” values. But these will be harder to uproot because they have the county’s blessing.

Let me give you an example. The reassessment sought to determine the market value of properties as of Dec. 31, 2018. On May 1, 2018, a house sold in Sands Point for $1,999,000. You would expect its market value to be very close to that amount by year-end.

You would be wrong, according to Nassau County.

The property was reassessed at $1,125,000. That’s a nice 44 percent tax break.

I did a study of more than 200 single-family homes sold between roughly July 1, 2017, and Aug. 15, 2018.

I threw out some sales that didn’t appear to be valid, arms-length transactions. Then I looked at what the properties would have been worth on Dec. 31. This required a small-time adjustment to the original sale price; Nassau County itself usually did this calculation for me.

After that, I compared the sale prices, which are true market values, to the county’s assessed values, which claim to be market values. The two sets of numbers should have been practically identical.

They were not. What I found surprised me.

For 43 percent of the properties, Nassau County’s reassessment missed the fresh, valid sale price by more than 10 percent. Sometimes it missed high, and these homeowners are among those now loudly complaining.

But one dirty secret of this reassessment: sometimes it missed low, and those homeowners are happily silent.

You may wonder how the county could have messed up the market value of recently sold homes.

For instance, how could a house that sold for $719,000 on Feb. 9, 2018, right down the street from me, be determined to have a market value of $586,000 10 months later? Did the market collapse for that particular property as others rose in value?

Of course not. Nassau County botched these assessments because it insisted on blindly using something called an “estimation model.”

As best as I can tell, the model was initially fed recent sale prices for hundreds of homes in each particular market. Then, apparently through sophisticated statistical analysis, it abstracted a set of rules about how much certain characteristics added to a property’s value.

Take a fireplace. The county’s model figures it increases how much your Port Washington property is worth by 2.01604 percent. So if the market value of your home is $500,000 without one, the fireplace adds $10,080.

A big problem should be obvious. The estimation model has no idea what your fireplace looks like. Yours might add only $5,000 of value. Or it might make your home $15,000 more valuable.

The model seems exquisitely precise, but it’s actually just throwing darts. It never sets foot on your property, and it will never make an offer to buy your home based on its estimate.

Curran does say the county also considered sale prices of neighboring homes to validate its assessments. I suspect she’s referring to what are called “sales comps.”

One way to try to derive a home’s market value is by comparing it to similar properties that have sold recently. In fact, for almost all single-family properties in my data set, the county did use sales comps.

But here’s another dirty secret: they weren’t used in an independent way.

If they had been, they would have routinely nudged the model’s estimated values up or down by tens of thousands of dollars. The reality: the sales comps generally rubber-stamped whatever the model found.

In fact, for 96 percent of 202 properties I examined, the final assessed value was whatever the model came up with, rounded down to the nearest or next-nearest thousand. The probability of this occurring with truly independent sales comps is about as likely as your toaster spontaneously jumping out the window.

So the estimation model wasn’t very accurate. And, by and large, the sales comps just rubber-stamped its guesses. Even so, someone could argue that using a model makes sense to value properties that haven’t sold for 10 or 20 years.

But all the properties in my study had fresh, valid sales. The market value of these homes is obvious: it’s how much they sold for, with a slight time adjustment. Further, this is a defensible assessment if a homeowner tries to appeal.

What will happen next is indisputably unfair.

On appeal, recently sold homes that were overassessed will almost certainly have their assessments reduced. Older homes that were overassessed may see reductions, though they’ll face a tougher fight.

Meanwhile, thousands of properties across Nassau County that are underassessed won’t have their assessments raised. This is the built-in unfairness. It’s most glaringly obvious with fresh, valid sale prices that should match the county’s assessed values, but don’t.

That’s because Nassau County used an “estimation model” that sometimes hit a home’s value, like a semi-skilled dart player, but more often than not, missed – and sometimes quite badly.

 

Share this Article