Some Prefatory Remarks to the N.Y.U. Real Estate Institute discussion, Oct. 25, 2001

Economic theory focuses on labor and capital, yet the largest category of tangible assets is not industrial plant and machinery earning profits, but real estate, and its primary objective is to make capital gains. Most new entries into the Forbes or Fortune lists of the richest men consist of real estate billionaires, or similar individuals coming from the fuels and minerals industries or natural monopolies. Those who have not simply inherited long-standing family fortunes have gained their wealth by borrowing money to buy assets that have soared in value.

The Federal Reserve Board publishes an annual balance sheet of the economy’s assets and liabilities showing real estate to comprise two-thirds of America’s tangible assets. Land represents most of this real property – upwards of 60 percent, depending on what assessment methodology is used. This explains why most capital gains are land-value gains. They have been spurred largely by credit creation, for on the liabilities side of the balance sheet mortgage debt absorbs 70 percent of private sector bank loans. Not only the savings banks and S&Ls but also commercial bankers are essentially in the mortgage finance business. Their activities are largely responsible for asset-price inflation.

I refer to the object of this price appreciation as “land” because it does not represent profit on capital investment as economists use the term. Developers spend most of their effort finding locations or neighborhoods whose prices are likely to rise. Examples in New York City include the upgrading of downtown as a result of the World Trade Center in the late 1960s and, a generation later, gentrification of the surrounding Tribeca area, as well as Times Square since the late 1990s,.

Ever since antiquity most surplus income and wealth has been invested in land. As modern economies grow richer and even as they “post-industrialize,” most of their surplus is invested in real property as the most desirable asset. It is financed by lenders eager to find a market for their savings and credit-creating powers.

The reason why developers are willing to pay their mortgage lenders so much of their rent roll (often their entire net rental earnings) is that they hope to emerge with a sizeable capital gain for themselves. This gain is essentially a land-price gain resulting from increasing the value of properties in burgeoning neighborhoods that are upgraded. This often entails rezoning, either of farmland on the outskirts of cities or gentrification of the core to high-income high-rise residential development.

Yet when one tries to get a sense of proportion by looking at the official nationwide real estate statistics, a kind of schizophrenia seems to be at work. The objective of most investors is to pick a property in the right location where it will appreciate in price at a higher pace than the interest rate that must be paid to acquire the property. In that sense their main concern is with rising land values – that is, the values that do not accrue as a result of earnings on capital (the rents that typically are pledged to lenders as interest payments on the loans taken out to by the properties) but are economy-wide asset-price appreciation in specific categories.

I will direct my comments to the national level. I recognize that this perspective is different from that of the local level, but I gather that the other participants will be giving their views on this aspect of matters. My own work over the past forty years has dealt with what factors raise or lower the nation’s overall real estate prices – rising income and savings levels, interest rates and the financial sector’s supply of mortgage credit, as well as changes in the tax laws and related market-shaping institutions.

Anyone trying to trace these price trends – and above all, how values for land and buildings change at different rates – comes up against a serious empirical problem at the outset: The official statistics either are lacking or seem to produce nonsensical results. The appraisal problems associated with determining how much value to attribute to land sites and how much to buildings and capital improvements are so serious that the U.S. Government agency delegated to publish such estimates – the Federal Reserve Board – has stopped doing so.

There are two approaches to land and building appraisals. Both start out by estimating the property’s market value. The land-residual approach then subtracts the value of buildings, leaving what is left over to signify land. These building values may be estimated in terms of their replacement cost (which usually produces a very high estimate, leaving little land value) or their depreciated value (which gives an unrealistically low building estimate, inasmuch as maintenance and repairs save most buildings from deteriorating through wear and tear). Using the depreciated value leaves more of a residual to signify land. In between, there is a third way to value buildings – by their “historical costs.” This is a more synthetic intermediate measure.

The building-residual approach starts by valuing the land, leaving the difference between the land price and the property’s market value to represent buildings. The first step in this approach is to construct a land-value map is employed for the overall district or city. This map has fairly smooth contours for land. Most of the variations in property prices thus are for the structures.

ORDER IT NOW

This approach rarely is used, however. For most cities and districts assessed land values vary drastically from one parcel to the next. The problem is especially apparent in the case of parking lots or one-story “taxpayers” in neighborhoods that are heavily built up. The low-rise or vacant land sites tend to appreciate as much as (or in many cases, even more) than the improved properties around them. Obviously this price appreciation cannot reasonably be attributed to rising construction costs. In fact, inflationary building costs should make such parcels less valuable, not more so, for older buildings have a great cost advantage The Fed’s land-residual appraisal methods do not acknowledge the possibility that the land itself may be rising in price.

If every property were built up only last year, the problem would be simple enough. The land acquisition price would be recorded, along with the construction costs. Together, these two measures would add up to the property’s value. But of course not all structures are newly built. Many were erected years ago, even as long as a century ago.

How does one decide how much a building’s value grows or shrinks in proportion to the property’s overall value? Does one multiply the building’s cost by the rise in the construction price index since its completion? If a property is sold at a higher price (as usually is the case), is it because the building itself rises in value, or the land site? If the property is sold at a lower price, as sometimes occurs, what is responsible for the decline – the building or the land?

This is where the judgmental problem occurs, for construction prices keep on rising even when property prices crash. If it is agreed upon that any explanation of land/building relations should be symmetrical, then the same appraisal methodology should be able to explain the decline of property values as well as their rise. The methodology also should be as uniform and homogeneous as possible. By that, I mean that similar land should be valued at a homogeneous price, and buildings of equivalent worth should be valued accordingly.

If these two criteria are accepted, then I believe that economists would treat buildings as the residual, not the land. Yet just the opposite tends to be done. At the end of this paper I will suggest some reasons why this is the case.

The Fed’s quandary with regard to its 1993 and 1994 U.S. land value estimates

Statistics on overall U.S. real estate values are published by the Federal Reserve Board in its Flow-of-Funds statistics, Table Z.1 (Balance Sheet of the U.S. Economy). For many years these estimates were broken down between land and buildings for a number of categories. These categories included residential, non-profit, government, corporate and non-corporate nonfinancial real estate. Since 1997, however, the conceptual problems underlying the allocation of value as between land and buildings have led the Fed to stop publishing comprehensive real estate statistics. 1994 is the last year for which it has estimated economy-wide land values.

The problem was that the Fed discovered that its methodology produced nonsensical results – a negative value of $4 billion for all land owned by nonfinancial corporations for the year 1993. This number resulted from imputing land values by subtracting the estimated replacement cost of buildings from overall property market prices. This “land residual” method left little room for land value, for replacement values continue their rise even when overall market prices decline, as periodically occurs. In such downturns the replacement value absorbs nearly all the market value of corporately owned real estate.

In operational terms, government statisticians multiply the original cost of buildings (or, in cases where buildings are sold, their assessed share of the property at the new transaction’s price) by the annual rise in the Commerce Department’s construction price index. The fact that this price index tends to rise steadily seems to explain the rise in property values by wage inflation and the rising materials costs. On this logic real estate prices seem merely to keep up with inflation. There is no hint of unearned gains or a free lunch.
In view of real estate’s dominant economic role, it is ironic that no attempt has been made to create better statistics. The Fed’s methodology undervalues the land by as much as $4 1/2 trillion. As matters stood in 1994, for instance, the Fed estimated the U.S. economy to hold some $20 trillion in real assets (excluding human capital, for which no official statistics are published). The land’s value was calculated to be $4.4 trillion, and building values $9 trillion. However, an estimate based on historical values resulted in a shift of land and buildings suggests that land rather than buildings represents two thirds of the nation’s overall real estate value.

A partial selection of real estate statistics continues to be published. The Fed estimated that at the end of 1996, for instance, households and non-profit institutions held $11.4 trillion in tangible assets. Nearly 80 percent ($8.2 trillion) of gross household wealth took the form of real estate (whose $3.6 trillion in mortgage debt represented nearly two-thirds of the household sector’s liabilities). Non-profits held $0.8 trillion. Real estate also accounted for nearly half ($3.4 trillion) of the $6.9 trillion in tangible assets owned by non-financial corporate business.

Where did all the land value go? The answer is that the reproduction costs of buildings and other capital improvements were estimated by multiplying the original purchase price or new construction price by the annual rise in construction costs, as estimated by the Commerce Department. These costs typically rise at an average annual rate of 3 percent. In downturns, however, these values dip below the estimated replacement-cost trend of structures. The Fed estimated that it would cost more to rebuild the corporately owned structures than the overall market price of their properties would justify. On the basis of this calculation land was assigned a negative value so as to account for the excess of replacement cost over the market price.

ORDER IT NOW

One obvious problem with this land-residual approach is that many buildings simply would not be rebuilt in their existing form. For one thing, occupancy use in their neighborhoods has changed. In Lower Manhattan and many other central city areas, industrial loft buildings, factories, and even high-rise commercial buildings on Wall Street have been gentrified into residential real estate. After New York City’s near bankruptcy in 1975 over 40,000 manufacturing spaces in Manhattan were shifted from commercial to residential purposes. A similar conversion of industrial structures to residential or high-density commercial use is found in most large American cities. And as formerly commercial and industrial districts have changed their character, site values have skyrocketed. Many sites are worth more when they can be delivered vacant, and more yet if the buildings are torn down so that developers can start with a rasa tabla.

The fact is that most people wouldn’t want to reproduce these buildings as they are. For instance, Manhattan’s highest-rent district until the World Trade Center bombing of September 11, 2001 was Tribeca, the downtown area where artists and other low-income individuals (including myself) moved into properties that had lost almost all their value in the late 1980s after New York City almost went bankrupt and industry began to move out. Office buildings were turned into residential properties. The building in which I lived rose in price from $40,000 in 1986 to $120,000 in 1980 and $4,000,000 in 2000. There is no way this increase is explainable by rising building costs. The building itself steadily deteriorated. All that increased was its site value.
Today, that property near the World Trade Center suddenly has plunged in price. The site’s value has changed without any reference to construction costs. And it is the site that determines the property’s value – or as real estate agents explain to prospective buyers, the three keys are “location, location, and location.” So we are brought back to the role played by land-value gains in the strategy pursued by investors and developers.

A percentage-composition chart of the Federal Reserve statistics shows that while these loft conversions were occurring prior to the 1993 downturn, land’s share of the total remained remarkably constant. Indeed, they are so stable that the data seem cooked to remain at 20 percent of the overall value of value of the real estate owned by non-farm nonfinancial corporate business. Much of this property represented industrial plant that had become obsolete. When property prices turned down in 1993, the entire decline in value was attributed to land, as hypothetical replacement costs continued to climb. The estimated proportion of land to building values plunged precipitously, even as cities were becoming postindustrial and old factory sites were turned into high-rise developments.

For real estate owned by households and partnerships (the latter being the preferred legal instrument for holding residential apartment buildings and office buildings), the Fed has estimated much higher proportions of land to buildings, but these estimates also tended to overvalue buildings relative to land. Every time a property changes hands at a higher price, building assessments are raised proportionally – and begin to be re-depreciated for these higher valuations, regardless of how often the buildings already have been written off.

The problems inherent in the land-residual method of real estate appraisal are most apparent in the cyclical downturns to which real estate markets are prone. While overall prices fall, replacement costs rise inexorably. Suddenly, land estimates plummeted, sharply changing the historical land/building proportions (Chart 2). The upshot was that in 1996, after completing its calculations for the year 1994 in which land prices stagnated while replacement costs continued their inexorable rise, the Fed stopped publishing land estimates.

This avoided one source of embarrassment, for it can hardly be true that all the corporate land in America really had a negative value in 1993. Suppose somebody same to you and said, “I’ll give you $4 billion. But there’s a catch. Along with the $4 billion in cash, you will have to accept ownership of all the land owned by every nonfinancial corporation in the United States.” Most people no doubt would conclude that they were being given assets much more valuable than $4 billion and would jump at the offer. The Fed’s statistic would be dismissed as a comic exercise showing yet again how otherworldly economists tend to be when not kept under the oversight of more realistic judges.

To give the Fed economists their due, they evidently came to the conclusion that their statistics were fatally flawed, and stopped reporting a number of categories of real estate. The September 1997 balance sheet made a start along new lines by including a calculation reflecting the original (historical) cost of buildings. This gave land a positive value. But nationwide totals no longer were compiled, so as not to report figures in which the Fed’s own economists had little faith. There no longer is a line labeled “land,” nor does the Fed publish a residual number for market value less the historical cost (or even the replacement cost) of buildings. Instead of making better land estimates, the Fed simply has dropped what had become a political and statistical hot potato.

On the other hand, this leaves us in limbo. By “us,” I mean macroeconomists and business analysts concerned with explaining the FIRE sector’s role in the economy as a distinct sector.

Land and building values over the course of the business cycle

If the appraisal controversy as between the land- and building-residual methods is framed in terms of business cycle analysis, then when the cycle rises and falls, the difference must be in the land, not buildings. What people are buying are not reproduction costs, whose fluctuations over the course of the credit cycle are relatively minor. They are buying site value in limited supply, akin to a natural monopoly. Most of all, they are buying the right to resell their property at a higher price at some future date as prices are bid up by what they expect to become an increasingly affluent economy fueled by an abundant supply of mortgage credit.

The land-residual approach appears to work as long as a fairly constant proportion of land to buildings is maintained. Statistically, this can occur only when prices are rising. In a thriving real estate market, appraisal methods typically use a rule of thumb in allocating property resale prices as between land and buildings to reflect their pre-existing proportions. As prices rise, most building values are assumed to account for between 40 and 60 percent of the property’s value. Thus, building values grow along with the property’s overall sales value. This appraisal practice is made to appear plausible as the pace of asset-price inflation tends to go hand in hand with rising construction costs, and hence in the theoretical replacement cost of buildings.

ORDER IT NOW

The anomaly occurs when real estate prices fall, for real estate prices are highly volatile, while construction costs rarely dip more than slightly, if at all. When real estate prices turn down, they often plunge below the reproduction cost of buildings. Hence, the residual (“land”) rises and falls much more sharply than do building replacement costs (which are estimated as rising at a fairly steady pace) and overall property values.

The result is a curious asymmetry. Building prices seem to be responsible for the rise in real estate prices, while land prices are held responsible for their decline. When the fall in property values intersects the rising reproduction-cost trend, the land residual turns negative. Because this land value often represents the owner’s equity, this decline may prompt heavily indebted owners to default on their loans or even to walk away from their property, leaving it to revert to the bank or other mortgage holder. In this sense the financial system itself is based largely on real estate, as the economy learned in the S&L deposit insurance crisis of the late 1980s. Real estate prices reflect the supply of property (including a fixed value of land) as compared with the fluctuating supply of mortgage credit, which in turn tends to be a function of the economy’s overall liquidity.

It may help clarify matters to think of “land” in the broad sense of comprising all elements of property value that cannot be explained in terms of capital investment and its profits. This category includes the site’s locational value. This site value is the essence of long-term planning by real estate developers, after all. To be sure, an examination of the economy-wide figures shows property prices to be determined by broad macroeconomic factors such as the availability of mortgage credit. Real estate is the major recipient of bank credit, and price waves or cycles are determined largely by the supply of mortgage loans and their interest rates.

Stated the other way around, the costs of reproducing buildings and structures are “sunk costs.” Price trends are determined mainly by today’s market demand as supplemented by mortgage credit, not by yesterday’s supply prices. If anything, buildings, plant and equipment wear out and depreciate, but their obsolescence is offset (and indeed, usually more than offset) by the rise in their site value. This rise in turn reflects the desirability of real estate in general as an investment vehicle that has become the prime recipient of the modern economy’s savings flows and new credit creation.

After 1990, for instance, when commercial banks found real estate to be largely “loaned up,” they reduced the rates paid to depositors, not having an alternative use for these deposits. (Stock market lending and loans to finance financial derivatives were just beginning to take off.) Depositors accordingly began to shift their savings into money-market funds that were invested mainly in bonds, and then into mutual funds invested largely in stock. The allocation of America’s savings thus shifted away from banks and their real estate lending to other forms of investment. This shift was largely responsible for the stock market’s takeoff, substantially in advance of the real estate takeoff.

One would think that land prices would play a central role in business cycle analysis, if only because a large share of stock market values consists of corporately owned real estate. Since the late 1940s “concealed value” in the form of real estate carried at outdated book values that reflect low acquisition prices was a major factor behind corporate raiding, mergers and acquisitions. Aggressive firms employed their accountants pouring over the stock exchange’s 10K reports seeking for such hidden values. But macroeconomists lacked the statistics needed to follow how land prices – that is, the “non-building” aspect of real estate value – were affected by the business cycle. This made it difficult to provide meaningful analyses of corporate net worth.

The real estate cycle is essentially a credit cycle. Traditionally, land acquisition has been the object of such savings and new credit. Site values are the economy’s “credit sink” as well as its ultimate “savings sink.” This is why real estate values rise as a multiple of gains and declines in the economy’s surplus. More loanable funds result in more mortgage lending that provides more credit to real estate developers, speculators and homebuyers, enabling them to bid up prices. However, borrowing to buy buildings is discouraged by the fact that when interest rates rise, more of the rental income must be paid to lenders. Profits fall as business upswings approach their crest. What declines is the economic surplus available for saving.

How can business cycle statistics be collected without land values as a prime indicator, in light of land’s dominant role in the economy’s asset structure and as the economy’s savings and credit “sink.” After all, it is the inflation of real estate asset prices that provides real estate owners with the collateral to justify further borrowing from banks to acquire yet more property. And by the same token, as Japan has learned, falling land prices extinguish the collateral that backs the banking system’s savings, leading to financial insolvency.

Just as real estate lending fuels land speculation, so the withdrawal of such credit leaves property markets to decline, sometimes with a crash, as occurred in Japan after 1990 when its financial bubble burst. Should this rise and fall be attributed to buildings or to land? It seems to me that inasmuch as the price rise and fall is homogeneous, applying to parking lots as well as to skyscrapers, we should attribute it to land. This achieves logical symmetry for the downturns as well as upturns in the real estate cycle.

The real estate industry’s vantage point

Real estate developers have defended the land-residual method of assessment, especially for major new urban projects. They point out that if they were not given a subsidy (which typically takes the form of a tax abatement), they couldn’t afford to build major office buildings or high-rise residential developments. In such cases it does indeed cost more to build the property than is reflected in the overall surplus price. Statistically, fiscal subsidies such as New York’s J-51 tax rebates appear as a negative value for the land.

ORDER IT NOW

This argument is valid as far as it goes, but it is based on a microeconomic logic. Precisely because real estate development is largely about improving the value of well-located sites, this logic makes sense from the real estate investor’s vantage point. But even here, macroeconomic factors play a role. Developers seek to leverage their own equity by borrowing funds. Many are willing to pay out most of the property’s available rental earnings in the form of interest in order to raise such financing. As I have noted above, the availability of mortgage credit is a “non-building” factor, and hence an environmental “land” dynamic.

Another factor also comes into play when real estate investors seek to maximize their returns. They understandably want to maximize their ultimate earnings by minimizing the local property taxes and federal income taxes they are obliged to pay. To do this they engage in political lobbying, mainly by pointing to the social importance of encouraging new construction to provide for the economy’s growing residential and commercial occupancy needs.

In this endeavor the real estate industry is backed by the financial and insurance industries, which recognize that the revenue that is collected as rent tends to be paid out as interest. Together, the finance, insurance and real estate industries that make up the FIRE sector have mounted a public relations effort and hired lobbies to depict property values as residing in buildings, not the land. Precisely because their objective is to generate capital gains from rising site values, for instance, the FIRE sector has led the campaign to lower capital gains tax rates below normal income-tax rates.

In making this attempt, the real estate industry depicts rising property prices as simply enabling real estate investors to break even. The logic is that recapture of the principal value is not a profit, and hence should not be taxed. An investor should be permitted to recoup the original investment’s replacement cost, and pay interest only on the gain. According to this logic, property values rise not because of asset-price inflation or rising rental charges, but because of the rising replacement costs for structures already built. The return to real estate investment thus is not an unearned “free lunch,” such as land-value gains often are depicted as being. Rather, rising property prices reflect the cost inflation for putting up new buildings.

In sum, developers seek to explain their gains in a way that provides an argument why they should not be taxed on their capital gains. If their property rises in price, they are simply keeping up with cost-price inflation. This means that they merely break even, in terms of the cost-adjusted value of their properties. After all, it would cost them more to buy a new building to replace the one they are selling. In this view there is no asset-price inflation of land values explained by the supply of mortgage credit on the “demand” side of the equation, but rather the old-fashioned wage and commodity-price cost inflation on the “supply” side.

This was the logic that the oil industry put forth for many years in arguing for the depletion allowance. Assuming diminishing returns for mineral reserves as low-cost supplies were the first to be exploited, oil and gas producers argued that it would cost more and more to find new sources of supply. On this basis they were allowed to deduct about 25 percent of their revenue as a depletion allowance, to provide them with the income to go out and find new supplies.

The effect was to make the oil and gas industries – and indeed, mining in general – tax-exempt. To the extent that the real estate industry (and also stock-market investors and securities owners in general) would be able to “index” the cost of their investment to a construction-price index, their capital gains would be rendered tax-exempt.

The upshot is somewhat ironic. An industry whose investors deal primarily with the development of land sites wishes to minimize the statistical treatment of land, in order to depict their capital gains as resulting from cost inflation and hence the reproduction costs of buildings. And the Fed economists followed this logic until it lost its semblance of reasonability when the real estate market turned down in the early 1990s.

Real estate lobbies recognize something that has been known from time immemorial: What is not seen is less likely to be taxed. Hence, what is not quantified for public policy-makers to see clearly – in this case, land and site values or the ability of investors to ride the wave of asset-price inflation – may avoid taxes and hence leave property owners with a larger after-tax return.

These considerations help explain the real estate industry’s preference for land-residual statistics at the national level, while individual investors seek site-value gains at the local level. I believe this is why it has been so hard to get Washington to produce national land value statistics.

The upshot is to leave land and other investments whose return takes the form mainly of capital gains (that is, those that rely mainly asset-price inflation fueled by the credit cycle) as the most seriously neglected area in today’s economic thought. The economy as a whole is taking its returns less in the form of profits (which are eaten up increasingly in the form of interest payments), and more and more in the form of capital gains. This is what makes the FIRE sector’s economic behavior structurally different from that of manufacturing and other industries whose returns are based on direct tangible capital investment.

Building values and depreciation (“capital consumption”) allowances

Property owners also have another reason for preferring to explain property values in terms of the reproduction costs of buildings rather than rising land value. Like all other investors, they want to minimize their tax liability – income tax at the national level, as well as state and local property taxes.

The income-tax liability of real estate investors may be minimized in two ways. The most general is through the tax deductibility of interest. The working assumption here is that interest charges are an inherent business expense, not simply the result of a business decision taken by investors to leverage their equity. For interest to be an inherent business expense, interest-bearing debt would have to be a factor of production, which it is not. Properties would yield their rent regardless of how they are financed. Still, every industry is permitted to charge its interest expenses as a cost, and the issue is far to be large to be covered here.

ORDER IT NOW

What is more to the point with regard the land- vs. building-residual methods of property appraisal is the treatment of building depreciation. This is perhaps the most unique tax advantage enjoyed by the real estate industry. Investors are able to depreciate their buildings based on their assessed acquisition price. They are able to do this at rising prices for the same building, even when prior owners already have depreciated these structures once or even many times. There is no limit as to how often a building can be depreciated. What matters is how often it changes nominal hands.

The result of the re-depreciation of buildings (a practice for which Lowell Harriss has suggested the felicitous term “over-depreciation”) has been to make commercial real estate investment largely exempt from having to pay income taxes. (Homeowners are not permitted to charge depreciation on their owner-occupied residences, but only on what they rent out.) On the other hand, it may be argued that although buildings wear out or become obsolete, they are kept in functioning condition by maintenance and repairs. These typically consume about 10 percent of the rental value. For business owners, these expenses are of course tax-deductible expense. But their explicit purpose of this expenditure is to maintain the building’s value intact, so that it can survive year after year while its site value rises.

The tax laws governing depreciation thus turn largely on how much value is assigned to buildings relative to the land, which is not depreciable. Like manufacturers, real estate owners are permitted to count part of the revenue they obtain over and above their current expenses as a return of their capital investment, as distinct from a taxable return on capital. No income taxes are levied on this portion of their revenue. That is only fair, for an investor who buys a $100 bond only pays tax on the interest, not on the original $100 principal. Likewise, industrialists are able to recover their initial investment in plant and equipment without being taxed. Their “sunk cost” gets reimbursed, so that they get their capital back by the time the equipment wears out or becomes obsolete.

For real estate, the economics are different. Whereas machinery rarely can be re-depreciated, this is not true of buildings, as long as they are kept in proper repair. When Congress passed the tax laws and turned to the fine print, they accepted the argument from real estate investors that if they are going to spend $1 million on a building, they want to get this sum back without having to pay tax on it. Furthermore, their objective was to allocate so much net revenue to this recovery of the building’s capital cost that nothing is left over as taxable income over and above direct operating costs and interest charges. Depreciation allowances are maximized by permitting investors to write off their capital investment as rapidly as possible.

Once the building is written off, investors have a tax motive to sell the property and buy another. The sales price obviously will be higher if the new buyer can begin depreciating the building all over again.

And if the building is sold at a higher price, the building assessment usually is raised. Suppose a property is sold for twice the price the owner paid for it. The local appraiser is likely to come by and say, “I see you’ve sold your building for $2 million. Under my rule of thumb, I appraise the land as half this value, and the building as half, so that gives you a $1 million dollar building.”

Under this rule, the building that formerly was priced at $500,000 is now re-depreciated at a price that builds in this $500,000 gain. A substantial portion of the rise in value is treated as a depreciable building value, not as a non-depreciable site-value gain.

The land-residual method of real estate appraisal thus gives a statistical illusion of capital being used up in production at rising transfer prices. But building values rarely are used up or reduced in practice, except when the property market collapses in the wake of a financial bubble.

The real estate industry argues that this fiscal practice is reasonable, as rising raw materials prices, wages and other construction costs have increased the replacement cost of buildings. New owners thus appear justified in getting a rising tax credit for their structures, even when this credit is so large as to make the flow of rental revenue exempt from income taxation. Generous depreciation allowances are supposed to enable sellers to buy new buildings on a par with the ones they are selling. The principle at work is that they should keep their investment value “intact,” in the sense of being immune from cost-price inflation.

This logic was pressed to an extreme by the 1981 Reagan tax reform. Under the double declining balance method permitted to real estate tax accountants, the new depreciation schedules permitted landlords to write off the building’s entire assessed value within just 7 1/2 years. The hypothetical rate at which buildings were estimated to wear out was so large that the write-off covered all the net revenue over and above operating costs. This meant that commercial investors as a whole did not have to pay income tax. Owning a building became much like owning an oil well and benefiting from the depletion allowance.

This tax favoritism was defended in Congress on the ground that it was in the public interest to provide a special inducement to the real estate industry to build more homes and office buildings. But in practice, many of New York’s most desirable buildings are over 100 years old, and have been depreciated many times over the course of their existence. The bottom line is that setting the depreciation rate so high that commercial real estate investors are tax-exempt gives them a fiscal favoritism that no other sector enjoys, now that the oil and mineral depletion allowances have been terminated.

How a land-value contour map would avoid statistical distortion

The macroeconomist needs to ask where real estate statistics reflect changing land use, and hence the added locational value that occurs with rezoning, such as permitting a residential use of formerly commercial and industrial loft buildings. Where are the changes wrought by public transport facilities and other government infrastructure, parks and proximity between living and working spaces, or for that matter the processes of gentrification and urban decay? These public improvements in site values are determined mainly by location, not construction-cost increases.

ORDER IT NOW

An almost random distribution of building values occurs as a result of when the market sale or appraisal was made. The previous pages have shown how, when properties are sold, building values tend to be upgraded and overstated at the expense of land. Land also tends to be undervalued as a result of its not being sold. Real estate that does not come onto the market, such as large family estates and government land, tend to be especially undervalued. This problem may be cured by constructing a smoothly contoured land-value map for each county, as land is homogeneous. With such a map the difference between property’s market price and its assessed land value would reflect the value of buildings as the residual.

The advantage of the building-residual method of property appraisals is its recognition that what rises is essentially the site’s land value. Buildings wear out and a neighborhood’s use of buildings tends to shift from industrial and commercial to residential occupancy, but asset-price inflation buoys an overall rise in the property’s value. This “capital gain” is not a return to “capital” in the form of profit. It is a land-value gain. If one grants that buildings depreciate, then what rises must be the site’s value. This is confirmed statistically by the fact that vacant sites fetch rising prices when they are put up for sale.

Conclusion

My approach to the appraisal issue is that of a financial economist trying to explain the economy’s changing shape and motivations. As a new kind of economy has emerged over the past two decades, the most savvy investors have sought capital gains more than profits. Price/earnings ratios have played a shrinking role in determining stock market prices. Institutional investors have projected upward price trends by calculating the rising flow of pension fund and mutual fund savings into the market. The investors who have succeeded best are those who have planed their portfolios to live off capital gains rather than dividends and interest.

A realistic set of national statistics would explain the shift toward “total returns” reflecting these asset-price gains. Integrating such capital gains into national income statistics would explain America’s low saving rate. Some critics worry that this rate has turned negative when foreign inflows of savings are segregated out.

But though the U.S. economy on balance is saving its income, it certainly is enjoying rising asset prices! The major asset for some two-thirds of the population is their home. Real take-home earnings for most of the population have drifted down over the past two decades, but home values have soared. While the ratio of household debt to income has increased sharply, much of this debt is “carried” by rising real estate value. When these capital gains for households are taken into account, financial analysts find less cause for concern. Most household debt is home mortgage debt, and it is secured by residential real estate.

The recycling of savings into new mortgage lending has fueled an economy-wide inflation of asset prices – real estate prices for land, homes and commercial properties, as well as stock market and bond prices. It was to show that despite their low net savings rate, the net worth of Americans has grown by enough to carry their rising debt burden that I began to calculate residential, commercial and industrial real estate values.

The real estate industry opposes the building-residual method of land appraisal primarily because it results in lower estimates for buildings relative to land values. This provides less leeway for depreciating buildings, and hence for economizing on taxes. It also limits the scope for indexing capital gains to cost-price inflation, and for the logic of taxing capital gains at a lower rate than wage and profit income. For if what is rising in value is mainly the land site, then the property owner appears as a passive beneficiary enjoying a free lunch. He rides the crest of asset-price inflation while doing little to increase the value of the building beyond picking a good location and making the normal maintenance and repair expenditures.

For investors to get favorable tax treatment for their capital gains, it is essential that the rest of the world does not believe that there is a free lunch of the sort that occurs in speculative bubbles – Japan in the late 1980s and the United States in the 1990s. If no free lunch exists, then it will not be taxed, to say nothing of being given special tax treatment as being in the public interest.

The real estate industry thus appears to be of two minds. To pick the best locations for acquisition or development, investors need to forecast trends in the nation’s property and credit markets. This requires sound statistical projections of how the availability of mortgage credit helps buyers bid up site values. On the other hand stands the political interest of promoting the idea that it is not really the land’s site value that is rising, but construction costs. If these costs explain the property’s price rise, nothing left over for passive land-value gains.

Using the land-residual method undervalues these land gains, even to the point of giving land a negative value in some years. Stated the other way around, the land-residual method over-values buildings, and gives the impression that instead of depreciating, properties are rising in value because construction costs are rising. But of course, rising construction costs lead to changes in the use of sites, and also of building practices.

I understand the real estate industry’s position that it wouldn’t build without being given a substantial subsidy. This is what they’ve gotten used to, after all, thanks largely to their active role in public affairs as campaign contributors and civic boosters. Most large cities give these subsidies to spur construction. But such subsidies should be seen as such, not disguised as building value.

To explain the overall economy it is necessary to draw a reasonable picture of the lines of causation at work. Real estate, construction and financial investors take a microeconomic perspective as to what they need in order to build. But a macroeconomic perspective calls for distinguishing between whether property owners are increasing their asset values by making direct investments to increase output or merely by sitting and waiting for their holdings to appreciate in value because of economy-wide forces.

As we all know, academic economics has become a highly mathematized discipline. But what is the point of all this mathematical treatment if it remains only a set of symbols rather than being given a real-world statistical sense of proportion? As computer systems analysts put it, “Garbage in; garbage out” (GIGO).

ORDER IT NOW

The 2001 Nobel Prize was awarded to economists who recognized the asymmetry of market knowledge. It would seem that this exists especially in the real estate sector. Investors know that the name of the game is capital gains, but the prospects for favorable tax treatment are enhanced by claiming that it is the buildings – that is, depreciable capital – that rise in price, not the land’s site value. The upshot is that developers use one set of statistics and logic to calculate their “total returns” on their properties, but support a different logic for use by government statisticians and Congressional authors of the nation’s tax laws, whose support from the FIRE sector depends largely on their not understanding its essential dynamics.

What I am urging in this paper is that some macroeconomists somewhere create a set of statistics that can explain the total returns for real estate as the economy largest sector. A truly functional format would give a sense of proportion to the motivations of investors and their degree of success (or failure), while tracing the linkage between land prices, property values, and the economic and financial dynamics that affect these prices.

The Q ratio serves as one measure of the financial bubble, as it compares the market value of the stock market (or real estate market, or other asset prices) to their book value or replacement-cost value. Long before James Tobin assigned the letter “Q” to this asset-price ratio, financial analysts were using it as a means to judge how much capacity an economic sector had to take on more loans. The operative question was, how much was mortgage credit and stock market credit fueling a financial boom that increased the prices at which assets were being transferred above what it would cost someone to simply create these assets afresh.

For real estate, the issue concerned how the volume of savings flowing into the mortgage market was fueling the land-price boom, turning the financial bubble into a real estate bubble.

Introduction: My approach is macro-economic, from a financial view.

For the past 40 years I have been working on how much credit the mortgage market can absorb.

70% of bank loans are mortgage loans.

Good news: The volume of savings is soaring. As interest is paid, the money is recycled into new mortgage lending.

In addition, retirement savings are flooding the economy with liquidity. If Social Security is privatized, this will provide more money.

The question is, how much money can the real estate market absorb?

Around 1990 it was loaned up, and the banks lowered their deposit rates. That was when money began to flow into the stock market.

This inflow of funds raised asset prices.

As more mortgage money is available, real estate prices also are bid up.

The question is – how should economists assign the rise in price: to LAND, or BUILDINGS?

The argument for buildings is based on construction-cost prices. But we are dealing with the demand for property from the demand side.

Should we say that what is not “buildings” is land?

This is what real estate brokers say when they say, “Location, location, location.” It is what developers try to achieve in building in an area where they believe real estate prices will rise.

What rises here are land prices.

Economists and especially the financial industry need a price index. But there are no figures.

And the Fed has stopped publishing them.

But of course, real estate prices also turn down over the course of the business cycle. The real estate cycle itself is basically a byproduct of the financial cycle.

It is when prices turn down that the real statistical problems occur. Land prices are wiped out, because construction prices continue to rise, while overall prices turn down.

Does it make sense for land to have a negative value?

What is the linkage between real estate’s rental cash flow, and property values?

Just as P/E ratios rise in the stock market, so property/cash flow ratios rise. RE is a non-profit sector.

It is kept this way by favorable tax laws. So, the real estate industry wishes to assign value to buildings.

Summary:

How does America avoid a Japan-type of situation, where mortgage obligations exceed values.

I promised earlier to send Alfred Marshall’s statement favoring the building-residual method of valuing land separate from buildings. Here it is.

“… the aggregate site value of any piece of building land is that which it would have if cleared of buildings and sold in a free market. The ‘annual site value’ … is the income which that price would yield at the current rate of interest.” — Marshall, Alfred, *Principles of Economics*, Chapter XI, “Marginal Costs in Relation to Urban Values.” (at p.441 in the edition I copied from, which I think was 1920. It is the 2nd page of Chap. XI.).

All of Chap. XI is well worth study. We might even want to post it on a website.

Anti-Georgist writer George Stigler made an effort to paint Marshall as anti-Georgist, based on a personal confrontation at Oxford, years earlier, using oral statements attributed to Marshall during a disorderly public meeting, but not published by him. Marshall as a theoretician obviously had many ideas supportive of George’s. (So did Marshall’s successor, A.C. Pigou; and Pigou’s student, Ramsey.) There was a great temperamental difference: Marshall was conservative. It was wrong of Stigler to turn that into a basic difference of viewpoint.

Mason G.