by Lindy Davies
From a presentation at the New York Federal Reserve Policy Breakfast on Nov. 20, 2013, along with Josh Vincent and Matt Stillman, for the Land Values Working Group
I was tempted to title this talk “New York City’s Resource Curse.” The phenomenon known as the “Resource Curse” tends to occur when a jurisdiction is endowed with an extremely valuable natural resource, and over-reliance on this resource leads to various forms of dysfunction: institutional corruption, industrial stagnation, etc. We usually think of Nigeria. But, we also can identify this process in New York City, where investors’ over-reliance on one very valuable natural resource does indeed lead to great problems. The population and great economic dynamism of New York City make its land staggeringly valuable. When that land value goes overwhelmingly into private hands, land is used very inefficiently, and the city is obliged to shake down its producers in order to pay for public infrastructure; the city is (doubly) self-cursed.
The opportunity to gain private profit from public investment — and collective productive contributions — enables landholders to collect large paydays that they did nothing to earn. However, the loss to the whole community ends up being far greater than the individual land speculators’ profits.
Consider this example: four separate tax lots on Park Avenue South were bundled together and sold, in 2006, for $31 million. The land under one of them, the single-story bank building on the corner, sold for $4,496 per square foot; but it had been valued by the city at $288 PSF. The market value of these four lots was obviously enhanced by their status of being four contiguous underdeveloped lots in a highly desirable location — but this was not referenced in their tax assessments. This block had looked this way for quite some time — as evidenced by a vintage billboard painted on the neighboring 16-story building, advertising the latest typewriters and adding machines. Though the $31 million was a windfall for the landowners who had, essentially, warehoused those lots for decades, the economic loss to the city is more accurately represented by the $200 million that was ultimately spent on the 19-story Gansevoort Park Hotel that now stands on the site. This has been a desirable neighborhood for quite some time; the 16-story building next door was built in 1917, and remodeled in 1979.
Without doubt, the amount of economic activity, jobs, revenue, etc. generated by the new hotel is far greater than what was being generated by those four small buildings.
Two Points Are in Common Agreement:
1. Land rent is a superior source of public revenue, because it is undistorting, and promotes efficient use of sites.
2. New York City’s property tax system is an unwieldy, inefficient mess.
Our first point is formalized in the Henry George Theorem: To the extent that public infrastructure investment is efficient and population is optimal, land rents will equal the expenditure on public goods. We can extrapolate from this to assert that land rents generally amount to the return on public infrastructure investment; if the investment is inefficient, the rents will be lower. There is general agreement among economists about this. But, two widely-held misconceptions have worked against cities basing policy on it: 1) that the aggregate amount of land rent isn’t enough to provide significant revenue, and 2) that it is seen as politically unfeasible. Point 1 is, as we will see, utterly false. Point 2 seems a daunting problem. However, once we see the tremendous benefits the city can gain from adopting such a policy, it could start to look more doable.
New York City’s assessment rolls state its aggregate taxable land value to be $344.1 billion (out of a total taxable real estate value of $821 billion). Taxed at 3%, that would yield $13.8 billion in annual revenue — less than the $19.6 billion the city plans to raise in general property taxes, and way less that the total projected city tax revenue of $45.3 billion. These figures seem to confirm the conventional wisdom that land values aren’t enough to make a significant dent in a big city’s public revenue needs.
But there are many reasons why the land figure is low. Now, New York City contains a great many types of real estate, and this real estate is unfortunately assessed and taxed in a variety of ways. Therefore, to understand how low NYC’s land assessments are, and how they got so low, we must look at each major category in the city’s tax system. We will use the city’s published assessment and sales figures, and make a back-of-the-envelope estimation of the land value that’s hidden in each one — taking care to use safer, lower figures whenever we can. What we come up with is a snapshot of New York City’s true aggregate land value. We overlook a great many local nuances and perturbations. However, the great size of the database, and the large number of available sales figures, give our aggregate estimate enough support to make it useful.
In terms of overall value, Class 1 parcels are assessed essentially correctly, both in Manhattan and in the rest of the city. However, the assignment of value to land and to buildings is suspect. The average class 1 building in Manhattan is 3 stories. For those parcels, the mean reported BV/LV ratio is 1.05, and the median is 0.78. If such parcels were to be sold at arm’s length, without historic status or other restriction, their highest and best use would be a far bigger building than they currently have. Manhattan Class 1 parcels sell for an average of $2,545 PSF; those with buildings whose value is recorded at less than half the land’s value sell for $2,733 PSF. The average land value of these parcels is reported to be $1,346 PSF. Essentially, these parcels sell for the same price per square foot whether they have a building on them or not. So, we can calculate an underreporting of Class 1 land value in Manhattan of $14.8 billion.
Class 1 lots that have little or nothing built on them are significantly undervalued. In Manhattan there are very few of these (all fully vacant lots in Manhattan have been removed from Class 1). In the Outer Boroughs, Class 1 parcels sell for an average of $180 per square foot of land area, whether they have buildings on them or not. The mean reported land value for Class 1 parcels in the Outer Boroughs is $98 PSF, which means there is an underreporting of land value for Class 1 parcels in the Outer Boroughs of at least $179.2 billion (we must understand that there are a great many of such parcels; excluding fully tax-exempt parcels, the total number in the Outer Boroughs is 672,815).
The above figure is probably low because premium-value locations are most likely underreported to a greater degree. For an example, we can look at Brooklyn’s Community District 2, which includes Brooklyn Heights and Downtown Brooklyn. Lots in that district on which the building value is less than 20% of the land value (we will henceforth refer to this as “BV<20”) sell for an average of $440 PSF. We compared this with similar lots in the Outer Boroughs in districts cited in the “Mapping Poverty in New York” study as having the greatest numbers of poor people. Typically-sized (2-3000 SF) BV<20 lots in these depressed areas sold for the Outer Borough average of $180 PSF.
Because the assessment treatment of condos and coops is so different from that of rental buildings, we cannot make sense of Class 2 figures unless we consider owner-occupied apartments and rentals separately. We start with class 2 rental parcels.
Citywide, class 2 parcels, excluding condo units, sell for about 220% of their reported market value. (The percentage goes up as the BV/LV ratio gets lower.) Because class 2 buildings are income-producers, there is a strong incentive to over-report their building values (to increase the amount of depreciation on federal taxes). For example, in Manhattan Class 2 rental parcels:
|Parcels||Avg. SF||Price PSF||MV psf||Price % of MV|
|Above, 2-5 stories||1,433||2,271||1,682||726||232|
In Manhattan, parcels whose building value is less that 20% of their land value must be considered vacant in economic terms — yet note that they sell for more, per square foot of land value, than do parcels with various kinds of buildings on them. Thus we see that in Manhattan, the actual land value of Class 2 rental parcels tends to be more than twice the reported value of land and buildings together. If we estimate the actual average land value for Class 2 parcels in Manhattan to be $1750 PSF, that adds up to an underreported land value for these parcels of $118.4 billion.
Larger class 2 rentals in Manhattan, those 6 stories and over, have a reported PSF land value of $567. The $1,750 figure is conservative for these parcels because they tend to be in better neighborhoods. We make no correction for this, thus ensuring that these estimates cannot possibly be too high.
In the Outer Boroughs, BV<20 lots tend sell for an average of $247 PSF. The average price for all Class 2 rental parcels on the Outer Boroughs is $238. But, if we exclude the 11% or parcels over 10,000 SF in area, the figure is $350. So, on Outer Borough Class 2 lots, less than 10K in area, the underreported land value figure is $30.1 billion.
The larger lots are left out of these calculations because they skew the data considerably. For Outer Borough lots over 10K in area, BV<20 lots sell for a whopping $425 per square foot (when you can find one; there were only 17 recorded sales of such parcels since 2006), while those over 10K SF with buildings on them (the average is 6 stories) sell for a mere $68 PSF.
Class 2 Condos & Coops
Citywide, class 2 condos sell for 5x reported MV; this relationship is consistent across all value levels. Taken together, these two factors yield $22.2 billion for Manhattan and $4.1 billion for the Outer Boroughs, or a total of $26.3 billion in underreported land value for class 2 condos. There are also a few class 1 condos. (Citywide, 13% of residential condo units are Class 1, of which only 1% are in Manhattan.) As with the rest of class 1, they are assessed by means of comparable sales. They sell for an average of 20% more than their reported MV. Their BV/LV splits are probably wrong, but to clearly determine that would take further study. However, the fact that there is a market for Class 1 condos (and that they are much more accurately assessed) shows that there is no practical reason why all condo units couldn’t be assessed as single-family housing units, rather than by comparing them with rental units that are worth far less.
It is hard to compare price and market values for coops with the data we have. However, an IBO report from this year estimates that both coops and condos sell for an average of five times their reported market value. If use the BV/LV as published, we get an estimated underreporting of land value for coops, citywide, of $47 billion.
Citywide, the average reported market value for a Class 4 parcel is 3.6 million, compared with $568K for Class 1. However, only 3.9% of Class 4 parcels (excluding condo units) have a reported MV of over $10 million — and 75% of those are in Manhattan.
There are many varieties of real estate in Class 4, so it is hard to zero in on types. Here is a list of some Class 4 categories (all excluding condo units and fully tax-exempt parcels):
|Parcels||Avg. SF||Price PSF||LV psf||Price % of MV|
area 2K-5K SF
|BV < 20,
area 2K-5K SF
MV > $10 million
It seems reasonable to assume from the above figures that the average land value of Class 4 parcels in Manhattan is at least $2,400 PSF (not 340), and that the average land value for Class 4 parcels in the Outer Boroughs is $175 PSF (not 28). Thus, Class 4 parcels in Manhattan underreport land value by an average of $2,120 PSF, and Class 4 parcels in the Outer Boroughs underreport land value by an average of $147 PSF. That leaves us with an underreporting of land value for Manhattan, Class 4 parcels of $140 billion, and for Outer Borough Class 4 parcels of $109.9 billion.
Class 4 condos: Citywide, the 28,949 Class 4 condos have roughly the same BV/LV ratio across the board, but sell for over 3 times their reported MV. Assuming the BV/LV splits to be correct, that yields a total of $13.1 billion in underreported land value for class 4 condos.
In Other Words:
New York City’s true land value is at least 297% of what the assessment rolls say it is. (All of these calculations exclude parcels that are fully tax-exempt, but do not take various partial exemptions into account.)
We should note here that although we are saying the New York City’s land values are underreported by at least this amount, it doesn’t follow that its real estate value is underassessed by this much. Much of the missing land value is attributed to buildings. However, vacant or grossly underused lots are indeed underassessed.
Our calculations tell us that the full taxable land value of New York City is at least $1.007 trillion. A tax rate of 3.16% on that value would yield $32.3 billion — enough revenue to replace current taxes on ALL buildings, personal income and sales.
De Blasio Steps in the Right Direction
Mayor De Blasio supports switching vacant lots citywide from Class 1 to Class 4. For many years, in the entire city except for Manhattan below 110th Street, vacant lots situated next to a Class 1 parcel were designated as Class 1 — which made no sense at all economically, because there were no homes on them to gain the benefits of tax relief for homeowners. In 2011, this designation was removed from upper Manhattan; there are now no vacant lots in Manhattan designated as Class 1. This did, indeed, appropriately raise the tax burden on warehoused lots, and it is likely that this has played a role in the ongoing revitalization (and gentrification) of Harlem. De Blasio would extend this reform to all five boroughs. Currently, in the four Outer Boroughs, there are 15,737 such lots (Class 1, vacant, 2000 SF or more in area). Their total area is about 10.2 square miles.
However, it is vital to understand that under-used land creates a far greater economic loss to New York City than vacant land, simply because there is so much more of it. For example, the four buildings we referred to earlier, which were bought for $31 million and immediately torn down, had, according to the city, a combined building value (not land value) of $4.66 million in 2006, the year they were sold. New York City has thousands upon thousands of such economically obsolete buildings, which are almost universally overvalued on the assessment rolls, and which, if they change hands, can be tax-depreciated anew by each new owner.
All these incentives create a city that abounds in “taxpayer properties” — where housing, office and retail space are all more expensive than they need to be, and where public infrastructure is used inefficiently. When too much value is attributed to the buildings, the effect is to subsidize the under-use and warehousing of land, and to penalize new construction. The disincentive to new buildings is stronger than it might appear at first, because the tax on the building is an addition to the building’s fixed annual cost, the biggest part of which is usually the project’s finance cost. One consequence of this is the familiar phenomenon of overbuilding — creating an oversupply of pricey, upscale housing and exacerbating the scarcity of affordable housing. If the building tax is reduced (or eliminated!) the annual cost of a given-size building is less. Thus, modestly-sized buildings that would be unprofitable under current tax conditions would become profitable once the building tax was removed.
Up-taxing vacant lots is only a first step. The real need is for a full reassessment of New York City’s real estate, using the building-residual assessment method, in which a site is valued as though it were vacant, and any additional real estate value is ascribed to the building. This is widely agreed to be the accurate way to assess the true market value of land. Economist Mason Gaffney wrote the following in response to a question from this writer:
You… need to start again at square one and use the “building-residual method….” The only construction cost that counts is the current cost of building a new optimal unit on vacant land. Subtract that from the anticipated DCF of said unit, and what’s left is site value.
Current construction costs are available in various published commercial indexes. Historical costs are irrelevant; depreciation rates are arbitrary; obsolescence is capricious and largely unknowable and the result of many factors. “Locational Obsolescence,” often overlooked, results from the rise of the underlying site value in its best alternative use, so you can’t even know obsolescence without first valuing the site as though it were vacant.
Many assessors and appraisers claim they can use any of three different methods. That gives them great latitude to argue for a wide range of values. Our estimates should begin with the postulate that there is only one basic correct method. Within that method there are many details to fill in, but outside that method there is naught but confusion.
A reassessment using this method would show, I believe, that New York City has well over a trillion dollars worth of land value. Once that truth is demonstrated, a new conversation about sensible public revenue policies can begin. It would be a discussion of the economic benefits New York City would enjoy if it were to eliminate 100% of its municipal wage, sales and building taxes while simultaneously taking the profit out of land speculation. That seems worth talking about.
A P.S. on Assessments
All this talk about how NYC’s assessments are fouled up should not be taken as an indictment of the professionals who determine and publish New York City’s real estate assessments. All indications are that these people do their jobs with a high degree of integrity and professionalism. Unfortunately, however, they work in a bizarre world in which the law requires them to generate profoundly inaccurate data. Most of these inaccuracies are consequences of the 1981 law that mandated current assessment practices. The sources of error that stem from that law are threefold:
1) Class 2 and Class 4 parcels are assessed not according to comparable sales, but rather according to capitalized income. Unfortunately the concept of opportunity cost is not incorporated into these assessments; the income stream that is capitalized tends to be that of the current use. The cases in which this makes some sense are those that are used to justify the practice: residential buildings have leases of various age, and rent-stabilization regulations lock in rents at certain levels that can vary widely with the age of leases. Thus, the income potential of a Class 2 apartment building can vary in ways that don’t neatly correlate with location value; in this case, valuing a building according to its income stream can appear to be more fair. But, of course, the process of arriving at these valuations is complex and invites various kinds of abuse. (Commercial leases, though they are not regulated, are also of various ages and are therefore subject to a lesser degree of the same sort of limitation.) None of this, however, has anything at all to do with the productive turnover of capital in an economically dynamic city. We have seen that huge amounts of residential and commercial real estate are not put to economic use — not until land owners decide these investment properties have “ripened.”
2) Coop and condo units are assessed as though they were rental units. This is economically absurd. These units consistently sell for five times what the city declares them to be worth. This leads to great distortions. Two-thirds of New York City’s people live in rented housing, so you’d think there would be a healthy demand for it — yet every year, as population increases, thousands more rental units are lost (to coop/condo conversion) than are built. The value of condos is enhanced by the fact that mortgage interest is tax deductible, whereas rent is not. Finally, condos and coops have long received a further property tax abatement (though this is controversial, and may not be extended indefinitely). Additionally, the distortion built into the market by the tax on buildings, which adds to the annual cost of a building, creates pressure for new buildings to be overbuilt. This has led to developers taking advantage of transferable development rights to create the sort of “pencil towers” that have sprung up around Manhattan in recent years.
3) The building value/land value splits are wildly inaccurate. These are reported by the city every year, and cited as though they had some sort of factual basis. However, even in the case of Class 1, in which assessments of overall value are correct, the BV/LV ratios are capricious. Reported land values are often very much lower for vacant or grossly underused lots than for neighboring developed lots, though locational factors are obviously identical. In the other tax classes, which are valued according to income stream, there is an obvious problem: how are lots valued if their cash flow is far lower than equally-situated lots that are put to their highest and best uses? Or, to reduce the question to absurdity: how can an assessor capitalize the income stream of a vacant lot? I have not been able to find any explanation of the rationale(s) by which New York City’s property assessments are divided between building values and land values.
Ana Champeny, “The Coop & Condo Tax Break Has Expired, Giving Albany Chance for Long-Promised Fix”, Independent Budget Office Fiscal Brief, January 2013, http://www.ibo.nyc.ny.us/iboreports/coopcondo2013.html
William Batt, afterword on the Henry George Theorem in The Self-Supporting City by Gilbert Tucker, Robert Schalkenbach Foundation, 2010
billdeblasio.com, “Unlock Vacant Properties and Direct New Revenue to Affordable Housing”, http://www.billdeblasio.com/issues/affordable-housing
Community Service Society, United Way of New York, Mapping Poverty in New York: Pinpointing the Impact of Poverty, Community by Community, 2007-2008, http://www.cssny.org/publications/entry/mapping-poverty-in-new-york-city-pinpointing-the-impact-of-poverty-communit
Mason Gaffney, After the Crash: Designing a Depression-Free Economy, 2011, Wiley-Blackwell (information on the effects of building taxes; a personal email from Prof. Gaffney is quoted with his permission).
Andrew Haughwout, James Orr, and David Bedoll, “The Price of Land in the New York Metropolitan Area”, Federal reserve Bank of New York, April-May 2008, http://www.newyorkfed.org/research/current_issues/ci14-3.pdf
New York City Department of Finance, Property Tax Guides for Class 1, and for Classes 2 and 4, http://www.nyc.gov/html/dof/html/property/assessment.shtml
Assessment and sales figures are those published annually by the New York City Department of Finance. The sales figures used are from 2006-present. Calculations and sorts were done using a Lotus Approach database application prepared by the author. I also used NYC’s online GIS map at http://maps.nyc.gov/doitt/nycitymap/, the city’s PLUTO shapefile data, and Google maps.