See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319562834

The effects of changes in land value on the value of buildings

Article  in  SSRN Electronic Journal · January 2017

DOI: 10.2139/ssrn.3025058

CITATIONS READS
0 528

2 authors:

T. Nicolaus Tideman Florenz Plassmann

Virginia Polytechnic Institute and State University Ohio University
116 PUBLICATIONS   1,614 CITATIONS    63 PUBLICATIONS   692 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

An Evaluation of the Benefit of Using Two-Stage Election Procedures View project

All content following this page was uploaded by Florenz Plassmann on 24 January 2018.

The user has requested enhancement of the downloaded file.

 The effects of changes in land value on the value of buildings
Nicolaus Tideman
Department of Economics, Virginia Polytechnic Institute and State University
Blacksburg, VA 24061
ntideman@vt.edu
Phone: 540-231-7592

Florenz Plassmann*
Department of Economics, State University of New York at Binghamton
Binghamton, NY 13902-6000
fplass@binghamton.edu
Phone: 607-777-4934

This version: January 11, 2017

Abstract

We show that increases in land value generally lower the value of existing buildings by amounts
roughly proportional to the square of the difference in size between the existing building and a
new building of optimal size. Our model suggests that the sight of deep discounts for space in
new buildings in a growing city is not necessarily a sign that developers have overestimated
demand; it might rather reflect their anticipation of continued urban growth. Our model offers an
intuitive way of dividing property values between the land and the structures on it, which is
relevant for establishing the appropriate base for taxes on land rent, for identifying the correct
income tax deductions for depreciation, and for assessing appropriate amounts of property
insurance.

JEL classifications: R14, R41, R52

Keywords: residual; transportation improvement; urban growth

*
Corresponding author.
1. Introduction

How does a change in land value affect the value of existing buildings? We develop a simple

model of property value that shows that events that increase the value of land generally lower the

value of existing buildings, and that the reduction in building value is approximately proportional

to the square of the difference in size between the existing building and a building of optimal

size.

Our result is counterintuitive. Consider a negative external shock—say the closure of a

subway station—that permanently reduces the attractiveness of a near-by site with a new five-

story building. The site now supports only a four-story building, so the value of the land falls.

Because the rent per square foot of building falls, the value of the existing new building falls

below its construction cost. Thus, the negative external shock directly reduces the value of the

land as well as the value of the existing and now inefficiently large building.

It is tempting to argue that a positive external shock—say the opening of a new subway

station—leads to the opposite result. However, while such a positive external shock increases

the value of the land, it generally lowers the value of the existing building.1 Our argument is

straightforward. Property value is determined by market demand. Land value is defined as the

opportunity cost of leaving unimproved land undeveloped and does not depend on the land’s

actual use. The value of any existing building is therefore residual, that is, it is the difference

between the property value and the land value. When an exogenous shock makes land more

valuable, it increases the optimal size of buildings. Because properties with inefficiently small

buildings do not take full advantage of the opportunity provided by the land, the increase in land

value reduces the residual value of the now inefficiently small buildings. Thus, a negative

external shock decreases the value of existing and now inefficiently large buildings directly,

while a positive external shock decreases the value of existing and now inefficiently small

1
We evaluate the changes in value that occur when people learn that a new subway station will open, rather than
any that occur when the anticipated opening of the station happens.

1
buildings indirectly, because the increase in land value exceeds the increase in total property

value.

A key point is that the value of existing buildings ought to be derived as the residual:

property value minus land value. There is a common practice among economists of modeling

land value as the residual, that is, as the difference between property value and the cost of

constructing efficient buildings. This approach is appropriate for assessing the value of land that

has recently been improved in an efficient way, or the value of vacant land that could be

improved in an efficient way. Furthermore, depending on whether one knows the value of the

land or the value of the building with sufficient accuracy, one can determine the value of the

other component of property value as a residual. However, buildings on sites that experience

exogenous shocks are subject to idiosyncratic economic depreciation that depends on the

changing value of the land they are on, making it inappropriate to model the value of land as a

residual when the shock has altered the property value and the value assigned to the building has

not been adjusted appropriately.

In the next section, we offer several reasons why establishing separate values for the two

components of real property is relevant. We have found that a diagram conveys the intuition of

our argument best, so we present a graphical argument in Section 3, assuming that the marginal

construction cost per square foot of building increases in discrete jumps at additional stories. In

Section 4, we use our diagram to analyze the effects of various types of exogenous shocks on the

values of land and buildings. We accommodate building characteristics other than size in

Section 5, which allows us to develop a mathematical model in Section 6, assuming a continuous

and differentiable marginal construction cost function. We use the mathematical model to show

that profit-maximizing developers in growing cities have an incentive to build at rates that appear

to be overbuilding. Section 7 concludes.

2
2. Reasons for establishing separate values of buildings and land

Because it is generally not possible to separate buildings from the land they are built on, one

might reasonably ask why establishing separate values for the two components of real property is

appropriate. In this section, we offer three reasons why this task is more than just a conceptual

exercise, explain how the appropriate establishment of building values yields an insight into the

literature on land valuation, and make an observation on urban growth that follows directly from

our separation of land and building value.

First, buildings depreciate in value over time, while land does not depreciate. When

taxpayers are allowed to deduct the annual reduction in the value of their capital improvements

from their income, they must separate the value of their buildings from the value of their land.

For example, publication 946 of the IRS tax code emphasizes that land cannot be depreciated

(IRS, 2016, p.6). For property originally held for personal use and later converted to business

use, the depreciable basis of the building is the lesser of the building’s market value and its

adjusted cost basis (IRS, 2016, p.12). Thus, owners must establish their buildings’ market

values, separately from the values of their land, to determine their allowable tax deductions.

Second, buildings can lose their value by being damaged or destroyed, while land is—

generally—indestructible. To identify the appropriate basis for property insurance for mortgaged

properties, owners and mortgage lenders must establish the values of the to-be-insured buildings

separately from the value of the land.

Third, economists agree that taxing land avoids many of the drawbacks related to taxing

capital and labor. In most communities that assess real properties for the purpose of taxation, the

assessor establishes separate values for land and buildings, even if both components of property

value are taxed at the same rate. Our model indicates that buildings depreciate faster in

communities with increasing land values. Thus, assessors who account only for the ordinary

depreciation of buildings are therefore likely to underassess the value of land, causing

communities to underestimate the revenue that they can raise from taxing land at a higher rate

3
than buildings. In communities in which land and buildings are taxed at different rates to raise a

specified amount of revenue, the failure to depreciate buildings fully results in additional

deadweight loss from the tax on buildings.

Establishing the separate values of buildings and land for the purposes of depreciation,

insurance, and taxation is particularly important when the land and the building are owned by

different entities. For example, all land in the Australian Capital Territory (ACT) is owned by

the ACT government, and all properties are leaseholds with 99-year leases. The Australian

federal government converted these properties in the early twentieth century from freehold to

leasehold with the goal of raising sufficient funds from the land-leases to build the public parts of

Canberra. Establishing properties as leaseholds has the same effect as the imposition of taxes on

the rent of land that owners of freeholds must pay, and in both cases the value of the land needs

to be established separately from the value of the buildings on it. Leasehold properties are also

common in England, Wales, Canada, and China. Even in the United States, land and the

buildings on it may have separate owners. For example, when the Empire State Building was

sold in 1951 for $51 million, the buyers immediately re-sold the land for $17 million to

Prudential Insurance Company, which leased the land, at an initial annual rent of about $1

million, to the building’s owners.2 Thus, the parties to the land transaction had to establish

separate valuations of the Empire State Building and the land underneath it.

By treating the value of buildings as residual, we resolve a puzzle in the recent literature

regarding the growth rate of land value in the United States. Sirmans and Slade (2011) analyze

sales of vacant properties, finding that, between 1995 and 2005, land value increased by a factor

of about 3.5. Similarly, Nichols et al. (2013) determine land values from sales of properties that

were either vacant or had buildings slated for demolition at the time of sale, finding that land

prices increased about fourfold during those years. In contrast, Davis and Heathcote (2007),

Davis and Palumbo (2008), and Davis (2009) determine land values as residuals, taking account

2
See Ennis (1961).

4
only of the ordinary depreciation of aging buildings. Davis and Heathcote (2007) and Davis

(2009) find that land values increased by a factor of only about 2.5 between 1995 and 2005,

while Davis and Palumbo (2008) find that average land values no more than doubled between

1999 and 2004. Our framework suggests that the authors of the last three papers have

underestimated the increases in land value, because the external shocks responsible for these

increases have caused existing buildings to become inefficiently small, with values below those

predicted by ordinary depreciation. Our model therefore supports the finding that, between 1995

and 2005, land value in the US has quadrupled, rather than just doubled.

Our model implies that, in a growing city in which developers anticipate land becoming

more valuable over time, developers will construct buildings that are larger than is currently

optimal. Thus, the sight of deep discounts for space in new buildings in a growing city is not

necessarily a sign that developers have overestimated demand, but might rather reflect their

anticipation of continued urban growth. Our model emphasizes that a building’s optimal size

must be measured over the building’s whole lifetime.

3. A graphical framework for dividing property value between land value and

building value

Consider a parcel of square feet of land that holds a building with stories and square feet,

with 0 if the parcel is vacant. Express the building’s size (average height) as the

property’s floor-to-area ratio (FAR), that is, as the ratio of square footage of building to square

footage of land: / . Let V denote the property’s value, so that the property value per

square foot of building is / and the property value per square foot of land is

⁄ .

We assume initially that buildings differ only in size, but we relax this assumption in

Section 5. A building is of optimal (profit-maximizing) size if is

5
 (1) greater than or equal to the marginal cost of constructing, maintaining, and

demolishing the last story of building that was built, and

(2) less than or equal to the marginal cost of constructing, maintaining, and demolishing

the next (not-built) story of building.

It is reasonable to assume that the marginal cost of constructing a building increases with the

height of the building, because of the increasing costs of providing vertical transport and

supporting a heavier building.

marginal cost, marginal value

per square foot of building

m f g
e h
d j
land value per b c
square foot of
land
x

building value per

a k square foot of
land

0 square feet of building per

square foot of land
(Floor-to-Area ratio, FAR)

Figure 1. The division of property value per square foot of land between
land value and building value

Figure 1 shows the case when the marginal construction cost increases by the same

discrete amount for each story, leading to a marginal cost schedule that is a step-function with a

6
discrete step of the same size at every integer value of .3 The horizontal axis measures the

property’s FAR, and the vertical axis measures marginal construction cost as well as the property

value per square foot of building. If, for example, the property value per square foot of building

is , then the parcel’s optimal FAR is —that is, a four-story building—and the property

value per square foot of land is (the rectangle 0 ). The area 0 underneath

the marginal cost schedule describes the cost per square foot of land of constructing a building of

size , or the building value per square foot of land if the building is new and is built efficiently.

The area above the marginal cost schedule and underneath describes the value per

square foot of land of the opportunity of constructing a building of size , or the land value of a

square foot of land.

A parcel’s land rent is defined as the opportunity cost of leaving the parcel unused for

one period when it is initially vacant, and a parcel’s land value is the present discounted value of

all future land rents.4 Because “opportunity cost” is defined relative to a parcel with a building

of optimal size, neither land rent nor land value depend on the actual building size. The actual

building size might differ from the optimal size if the marginal cost of adjusting the actual

building size in response to changes in exceeds the marginal benefit of adjustment. For

example, the marginal cost of reinforcing the foundations of an existing building to carry the

weight of an additional story might exceed the marginal value of an additional story, even though

it would be efficient to construct the additional story for a new building. In terms of Figure 1,

assume that the value of a square foot of building is , implying an optimal building size of .

The parcel’s land value per square foot is the area above the marginal cost step function.

If the parcel is vacant, then this area describes the parcel’s property value per square foot of land.

3
We assume constant jumps at constant intervals for expositional simplicity only. Alternative marginal cost
schedules change our results only in causing the effect of a size difference to depart from the quadratic relationship
that we describe in Section 4.5. We consider a continuous marginal cost schedule in Section 6.
4
See Tideman (1999), p.110. The definition of rent is analogous to the definitions of wages and interest in perfectly
competitive markets.

7
 Now consider an existing building on this parcel of size . The property value per

square foot of land is , or the rectangle 0 . Because the parcel’s land value is

independent of the parcel’s actual use, its land value per square foot is still the area

above the marginal cost step function, or the area 0 above the marginal cost step function

plus the rectangle . Thus, the building value per square foot of land is the area 0

below the marginal cost schedule, minus the rectangle . It is efficient to not increase the

size of an existing building from to if the marginal cost of increasing the height of the

existing building exceeds the marginal cost of increasing the height of a new building by more

than . It is optimal to tear down an existing building and construct a building of optimal

size—assuming that the tearing down can be done at no net cost—if the existing building is so

small that its value computed in the way described above is negative. If, for example, the size of

the existing building in the previous example is rather than , then the property’s value per

square foot of land is , or the rectangle 0 . Because the parcel’s land value per

square foot of land is still the area above the marginal cost step function, the building

value per square foot of land—the difference between the rectangle 0 and the area

above the marginal cost step function—is negative.5 Thus, it is optimal to tear down the

inefficiently small building of size and replace it with one of optimal size if the tear-down

can be done at no net cost. If demolition costs are positive, then it is efficient to tear down the

building as long as demolition costs are less than 0 . If demolition costs are

negative—that is, if there is net positive salvage value to the building—then it is efficient to tear

down the building sooner.

Our analysis of the demolition decision corresponds directly to the framework used in

Rosenthal and Helsley (1994), who estimate land prices in Vancouver, B.C. by analyzing the

sales of properties with single-family homes that were subsequently redeveloped. Demolition

5
It might be tempting to argue that an existing building of suboptimal size cannot reduce property value, and that
building value per square foot of land should be defined as the building’s marginal construction cost, or 0 .
However, the combination of the requirement that land value be defined independently of use and the fact that
/ makes it necessary and appropriate to define building value as a—possibly negative—residual.

8
costs were about 2,000 Canadian dollars, or about 1.7% of the average price of redeveloped

properties in their sample. Rosenthal and Helsley use information on each property to estimate

its price as either or , depending on whether the property was vacant or developed at the

time of sale, impute the corresponding alternate value, and identify properties that ought to be

redeveloped as those for which . Thus, their approach also uses the idea that it is

optimal to demolish a building if the building reduces the value of the property, that is, if the sum

of building value and demolition cost is negative.6

Our framework complements the model of a monocentric city in Brueckner (1987) that

explains why building value per square foot of building differs across locations, with the height

of buildings decreasing with increasing distance from a city’s center. Agents bear higher

transportation costs as they move further away from the center, leaving them with less disposable

income to spend on housing and non-housing goods. Because the prices of non-housing goods

do not vary with distance, the price per square foot of building must decline with increasing

distance from the center to maintain the attractiveness of more distant parcels—for example, the

price might decline from at a distance of 5 miles to at a distance of 6 miles from the

center. Since the agents in the model have equal incomes, those who live further away from the

center therefore consume more building space. Because the marginal building cost function does

not vary with distance, a lower price per square foot of building requires that land value per

square foot of land decline with distance, that is, the land value per square foot of land of the

parcel at a distance of 5 miles from the center equals the area , while the per-square-foot

land value of the parcel at a distance of 6 miles is given by the smaller area . The decline

of the price of a square foot of land relative to the cost of a square foot of building induces

producers to substitute land for capital, leading to lower FARs at larger distances from the

center. Those agents who live farther from the center therefore live in larger houses with fewer

stories that consequently have a lower building value per square foot of land—building value

6
See also Brueckner (1980) and Brueckner (2000).

9
0 at a distance of 6 miles is smaller than building value 0 at a distance of 5 miles from

the center.

Brueckner models an equilibrium situation that incorporates all long-term adjustments of

building sizes that occur in response to exogenous events. As shown, our framework also

describes the relationship between building value per square foot of building and optimal

building size, but in addition, it provides information about the short-term changes in land and

building values that are caused by exogenous shocks. We examine the short-term effects of such

exogenous events in the next section.

4. The effect of exogenous events on land value and building value

It is necessary to distinguish between the effects of unanticipated and anticipated events. The

occurrence of unanticipated events leads to discontinuous adjustments in stock value (the selling

prices of land and buildings), when people adjust their valuations in response to revised

expectations of future flows (the rent of land and of buildings). In contrast, while the occurrence

of anticipated events may lead to discontinuous changes in rent, it leads to continuous changes in

the values of land and buildings, since discontinuous changes would imply unbounded rates of

return over the discontinuities, which is not consistent with market equilibrium. Because we are

concerned with discontinuous changes in value, we analyze the effects of events that are not

anticipated. Thus, when we consider the effects of transportation improvements or changes in

zoning codes, we analyze the changes in value that occur at the time when people learn of the

transportation improvement or of the change in zoning code, rather than the continuous changes

in value that occur at the time of implementation.

4.1 Events that reduce land value and reduce the value of existing buildings

Consider the closure of a subway station, which reduces the demand for building space on near-

by sites. In terms of Figure 1, assume that the initial value of a square foot of building is ,

10
implying an optimal building size of . The value of a property that holds such a five-story

building is the rectangle 0 . Suppose the reduction in the demand for building space

reduces the value of a square foot of building to , thereby reducing the property’s value to the

rectangle 0 . The land value per square foot is the area above the marginal cost

step function—a reduction by the area —and the building value per square foot of

land is the area 0 underneath the marginal cost step function—a reduction by the rectangle

. Thus, by lowering the demand for building space, the closure of the subway station

directly reduces the opportunity cost of leaving the site vacant as well as the value of the existing

building.

4.2 Events that reduce land value and increase the value of existing buildings

All events that impose binding limits on the permissible uses of land reduce the opportunity cost

of leaving a parcel vacant and therefore reduce land value. Consider the adoption of a zoning

restriction that limits buildings to a size that is smaller than what property owners would build

without the restriction. Consider a parcel in a place where the value per square foot of building

is , implying an optimal building size of . If there is no restriction on erecting a building of

size on this parcel, then the parcel’s land value per square foot is the area above the

marginal cost step function. The adoption of a zoning code that limits the maximum building

size to reduces the land value from to above the marginal cost step function,

or by the rectangle .

This insight is supported by the empirical work of Groves and Helland (2002), who

examine the effect of the adoption of a residential zoning ordinance in Baytown, TX in 1995.

They find that in the two years after adoption, the average price of properties well suited for

residential use increased by about 0.8 percent, while the average price of properties better suited

for commercial use decreased by about 3 percent. Thus a zoning ordinance that prevents the

construction of commercial buildings lowers the value of the land that is best suited for

commercial use.
11
 Bertaud and Brueckner (2004) examine the effect of existing restrictions on maximum

FAR in the city of Bangalore, India, concluding that these height restrictions have (1) reduced

land values near Bangalore’s center where the restrictions are binding, and (2) caused the city’s

area to be about 17 percent larger than optimal, leading to the addition of a four-kilometer-wide

ring to the city’s optimal eight-kilometer radius. Their general equilibrium model emphasizes

that binding height restrictions that depress land values at one location will lead to greater

building heights and higher land values at other locations, absorbing the demand for housing that

cannot be satisfied in the restricted areas.7

Brueckner and Sridhar (2012) extend the analysis to 101 Indian municipalities and

calculate that the building height restrictions have an average welfare cost of about 106 million

Rs per city, which—after adjusting for differences in average income and household size—

translates to about $1 billion for a US city with a population of about 750,000 persons.8

Brueckner et al. (2017) undertake a similar analysis of the effect of building height restrictions in

more than 200 Chinese cities. They find average land price elasticities with respect to FAR of 75

percent for residential land and of about 60 percent for commercial land, with the highest

elasticities in cities with binding building height restrictions and therefore relatively small

buildings.

A zoning restriction that grandfathers in existing buildings whose size exceeds that of

permitted new buildings increases the value of these existing buildings. Consider the value of a

building of size prior to the zoning restriction, which is the area 0 underneath the

marginal cost step function. The zoning restriction does not affect the marginal value per square

foot of building at this site and therefore leaves the property’s value per square foot of land,

unchanged. The zoning restriction must therefore increase the value of the building by the

reduction in land value, that is, by the rectangle . This rectangle represents the profit that

7
See also Arnott and MacKinnon (1977).
8
The authors emphasize that their welfare calculations consider only the higher commuting costs that are the
consequence of urban sprawl, and that they ignore any aesthetic benefit of smaller buildings as well as the costs of
any infrastructure updates that increases in a city center’s FAR might require.

12
the owner of the existing building receives from the privilege of having the use of a building that

is larger than a new building can be under the zoning restriction.

This insight is supported by the empirical work of Glaeser et al. (2005), who show that,

in the United States, the ratio of housing prices to construction costs has increased considerably

between 1950 and 2000; they suggest that housing supply regulations might be responsible for

this change. Similarly, Quigley and Raphael (2005) provide empirical evidence that Californian

cities with stricter regulations on new construction have higher ratios of housing prices to

construction costs than cities with less regulation.

4.3 Events that increase land value and reduce the value of existing buildings

All events that increase the opportunity cost of leaving a parcel vacant increase land rent and

therefore land value. Such events generally reduce the value of existing buildings.

In terms of Figure 1, assume that the value per square foot of building is , implying an

optimal building size of . The parcel’s land value per square foot is the area above the

marginal cost step function. Assume that the parcel holds a building of size , so that the

building value per square foot is the area 0 . The property value per square foot of land is

the sum of these two areas, that is, the rectangle 0 .

Now consider the opening of a new subway station that raises the marginal value per

square foot of building to , leading to an optimal building size of . The parcel’s new land

value per square foot is the area above the marginal cost step function, an increase by

the sum of the rectangles and . Because the building size is still , the new

property value per square foot is the rectangle 0 . Subtracting the land value from the

property value yields the building value per square foot of land as the area 0 minus the

rectangle . As long as the marginal cost of increasing the height of the existing building by

one additional story exceeds , it is optimal to keep the existing building at the now

inefficiently small size .

13
 The opening of the subway station increases the opportunity cost of leaving the site

unused and therefore increases the value of the land. If the parcel were vacant, then it could be

sold for an amount equal to the area above the marginal cost step function. Because the

increase in the value of the property is smaller than the increase in the value of the land, the

existing four-story building falls in value. Thus, by increasing the value of the land, the new

subway station indirectly reduces the value of the existing and now inefficiently small building.

This insight is supported by the empirical work of Cervero and Kang (2011), who

analyze land-use conversions in Seoul, South Korea that occurred between 2001 and 2007,

encompassing the time of establishment of exclusive median lanes that nearly doubled the

operating speed of bus rapid transit by 2004. They find that about 4 percent of the 52,000

parcels with single-family homes they examined that were closer to median-lane bus stops than

regular bus stops, were converted to—mostly—larger multi-family housing, and to a lesser

extent to condominiums and mixed-use (residential and commercial usages combined within a

single building). Thus, the transportation improvement lowered the values of several close-by

single family buildings by enough to make it profitable to replace them with larger buildings of

different use-types.

4.4 Events that increase land value and increase the value of existing buildings

Increases in land value do not reduce the value of existing buildings when the site is overbuilt.

Consider a parcel with a marginal value per square foot of building of , holding an

inefficiently large building of size . The parcel’s land value per square foot is the area

above the marginal cost step function, while the building value is the area 0 . The building

is inefficiently large, because the marginal cost of erecting the fifth story, the rectangle ,

exceeds the story’s marginal value, the rectangle , by the rectangle . A new subway

station that raises the marginal value per square foot of building from to makes a five-

story building optimal and therefore increases the per-square-foot value of the existing five-story

14
building by the rectangle , in addition to increasing the parcel’s value per-square-foot of

land by the sum of the rectangles and .

The values of land as well as existing buildings might also increase during economic

booms that increase the marginal value per square foot of building as well as the construction

cost of new buildings. In terms of Figure 1, consider an economic boom that raises the marginal

value per square foot of building from to as well as the construction cost of each story by

a rectangle of size , thus shifting the entire marginal cost step function upwards by the

distance . Because in Figure 1 the distance is half the distance , the boom increases

the per-square-foot value of land as well as the per-square-foot value of the existing building of

size by ∙0 each. Because the marginal benefit of having a building with a fifth story

equals the marginal cost of building this fifth story (the rectangle ), the existing four-

story building as well as a five-story building are both efficient.

4.5 Determining the additional economic depreciation of a building on a site that has become

too valuable for the building

If the increase in construction cost per story is constant (that is, the steps in Figure 1 are of equal

height), then the relationship between increases in land value and the building’s corresponding

additional economic depreciation is approximately proportional to the square of the number of

stories that are missing for the current building to be of optimal height. To understand why,

consider a property on which a value per square foot of building of makes it worthwhile to

erect a five-story building. For an average such property, is exactly halfway between the

marginal costs per square foot of a five-story and a six-story building. Figure 2, which replicates

Figure 1 for the sake of legibility, shows the parcel’s land value as the area above the

marginal cost schedule.

15
marginal cost, marginal value
per square foot of building

v r f g
e h
y n d
land value per b c
square foot of o
land
u x

building value per

a square foot of
k
land

0 square feet of building per

square foot of land (Floor-
to-Area ratio, FAR)

Figure 2. The division of property value per square foot of land between
land value and building value

If the property holds a five-story building, then its value is the rectangle 0 , and the

value of the building is the area 0 underneath the marginal cost schedule. If the property

holds a four-story building, then its value is the rectangle 0 , and the value of the building

is the area 0 underneath the marginal cost schedule, minus the rectangle , which is the

additional depreciation of an inefficiently small four-story building on a site that supports a five-

story building. If the property holds a three-story building, then its value is the rectangle

0 , and the value of the building is the area 0 underneath the marginal cost schedule,

minus the two rectangles and .

Because we assume that is exactly halfway between the marginal costs per square

foot of a five-story and a six-story building, the area of is three times the area of .

Thus, the additional depreciation of a three-story building on a site that supports a five-story

16
building is four times the additional depreciation of a four-story building, and the additional

depreciation of a two-story building on a site that supports a five-story building is the sum of the

three rectangles , , and , or nine times the depreciation of a four-story building.

In short, for the average property, the additional economic depreciation of a building of

suboptimal height is proportional to the square of the number of stories that are missing for the

current height to be optimal.

The general relationship between the number of missing stories and the additional

economic depreciation is given by

, (1)

where is the cost of building an additional story (equal to the rectangle 0 ), and f is the

fraction of the step size by which exceeds the cost of the last story built. Because f = 0.5 for

the average property, for which is exactly halfway between the marginal costs per square

foot of a five-story and a six-story building, the additional economic depreciation of a building

that is three stories below optimal height is given by the area above the marginal cost

curve, or 4.5 times the area of 0 , or 3 3 = 4.5 .

Similarly, in the extreme case in which is exactly equal to the marginal cost of

erecting the sixth story, rather than being halfway between the marginal cost of erecting the fifth

and sixth story as in Figure 2, the optimal building has either six or seven stories. Thus, the

value of a five-story building is reduced by a full-step rectangle of the same size as rectangle

0 , while the value of a four-story building is reduced by three such rectangles, and the value

of a three-story building is reduced by six such rectangles, or 3 3 1 =6 .

17
5. Accommodating ordinary depreciation
Buildings differ from each other in many more ways than just size, leading to a series of figures

like Figure 1, each of which represents buildings with a different set of characteristics. One

might combine these figures into a single figure by either incorporating the joint value of all

characteristics into an adjusted, or effective, marginal value per square foot, or by converting the

different characteristics into their square-foot equivalents, leading to an effective FAR, . The

latter approach is equivalent to the common practice in economics of adjusting for differences in

quality by treating amounts of goods of varying quality as representing different quantities with

the same price.

Assume that a new four-story building has all the characteristics that are optimal for a

building at this site at the time of construction. Because construction and insulation techniques,

building codes, and general space requirements change over time, older buildings generally do

not have the same characteristics as new buildings. We accommodate such differences by

adjusting an older four-story building’s actual square footage from to in Figure 2.9 The

property value per square foot of land is the rectangle 0 , rather than the rectangle

0 that describes the value of a new four-story building. The fact that the property holds

an older building does not affect the land value, , because land rent is independent of land

use. However, the building value per square foot of land is the area 0 minus the rectangle

, indicating that an aging building absorbs the entire decrease in property value that

results from the reduction in the building’s attractiveness compared to a new building. In the

following section, we use the concept of effective FAR to develop a mathematical model of land

value and building value.

9
Colwell and Ramsland (2003) analyze the degree of technological change that needs to occur before it becomes
efficient to upgrade older buildings.

18
6. A mathematical model of land value and building value with continuous
marginal construction cost
Consider a parcel with square feet of land that holds a building with effective square feet,

and effective FAR / . Let be the marginal cost of constructing the next square

foot of building, with 0 so that marginal construction cost increases with building

size. The construction cost per square foot of land of a new building with effective FAR is

. (2)
∗
The building is of optimal FAR if the marginal building cost per square foot of building

equals the market value per square foot of building at the parcel’s site, or
∗
. (3)
∗
The parcel’s value per square foot of land, , can be expressed as the difference
∗
between the property’s total value when it holds a building of optimal FAR, ∙ , and the

construction cost of this building, or

∗ ∗ ∗
∙ . (4)

The value per square foot of land of a building of effective size that is located on a parcel
∗ ∗
with optimal FAR , or , , is the difference between the property’s total value per

square foot of land, ∙ , and its land value per square foot of land, or
∗ ∗
,
∗ ∗
∙ ∙
∗ ∗
∙
∗
∗ ∗
∙ if
∗
∗.
if (5)
∗
∗ ∗
∙ if

Thus, the value of a building of optimal size equals the cost of constructing such a building,
∗
while the value of a building with is below the cost of constructing a building of

effective size , with economic depreciation equal to

19
 ∗
∗ ∗
∙ if
∗
, . (6)
∗ ∗
∗ ∙ if

For a linear marginal cost function, each part of equation 6 describes a triangle—that is, the

difference between a rectangle and a triangle with the same base and height as the rectangle.

Thus, analogously to our result in section 3.3 for a discrete number of stories, since the area of a

triangle of a given shape is proportional to the square of its side, the excess economic

depreciation of a building that is not of optimal size is proportional to the square of the

missing/excess effective floor space.

6.1 The effect of a transportation improvement that increases

Consider the opening of a new subway station that increases the value of floor space in the

surrounding neighborhood. Equation 3 indicates that the increase in raises the optimal FAR,

as long as constructing larger buildings is not prohibitively expensive, or ∞. If

∗
so that the site was not overbuilt, then the increase in reduces the value of the

existing building because

, ∗ ∗
∗
∗
∗
∗
0. (7)

∗
However, if the site was overbuilt with , then as long as the new optimal FAR does not

exceed the building’s effective FAR, the increase in necessarily increases the value of the

existing building because

, ∗ ∗
∗
∗
∗
∗
0. (8)

Although, not surprisingly, the mathematical model yields the same result as our
∗
graphical analysis in Section 4.3, equation 7 obscures the fact that, when , the value of

the building falls because the increase in land value exceeds the increase in property value.10 In

10
However, equation 5 (from which equation 6 and hence equation 7 are derived) shows that this is the case in the
mathematical model as well.

20
contrast, Figure 1 emphasizes that the increase in affects building value only indirectly,

through the associated direct increase in land value.

6.2 The size of new buildings

Finally, we use our model to analyze the optimal size of new buildings.11 Assume first that the

property value per square foot of building is constant over time, in which case equation 3
∗
indicates that is constant as well. Our definition of effective floor space implies that a

building of age t has an effective FAR of

0 , (9)

where 0 is the building’s FAR at the time of construction, and is the building’s rate of

physical depreciation. Because a building’s value per square foot of land is below the cost of
∗
constructing a building of size when , a developer who must pay the
∗
opportunity cost of the land breaks even if and only if 0 . Thus, when the value per

square foot of building is constant, profit-maximizing developers build new buildings that are of

optimal size.

Now assume that the parcel is located in an emerging area in which the property value

per square foot of building is expected to increase, for the foreseeable future, at a constant rate of

i, so that

0 , (10)

where 0 is the property value per square foot of building at time 0. Equation 3 indicates

that, if the marginal cost function remains unchanged, then an increase in leads to an
∗ ∗
increase in optimal size, . For a building with , the continuous increase in

leads to a continuous increase in economic depreciation because

, ∗
∗
∙ ∙ 0. (11)

11
Brueckner (2000) provides a review of theoretical and empirical analyses of urban growth.

21
The first term represents the loss in value that arises from the loss in rent associated with a

building that is too small for its site when rent is rising, while the second term describes the loss

in rent that results directly from physical depreciation. Thus, buildings on sites that are not

overbuilt depreciate faster during economic booms when 0 than during times of stagnation.
∗
In contrast, the economic depreciation of a building that is larger than changes

according to
, ∗ ∗
∙ ∙ . (12)

The first term represents the gain in value that arises from renting out additional square footage,

while the second term describes again the loss in rent that results directly from physical

depreciation. The building increases in value if the additional rent that results from the increase

in demand exceeds the loss of rent from physical depreciation. Profit maximizing developers

therefore have an incentive to erect new buildings of larger-than-currently-optimal size to take

advantage of the expected increase in value that arises from the building’s additional square

footage.12 Developers then have an incentive to charge prices that are apparently money-losing,

to be able to rent out the larger-than-currently-optimal building. Thus, the sight of discounts for

space in new buildings in a growing city is likely to reflect the developer’s anticipation of

continued urban growth, rather than being a sign that developers have overestimated demand.

7. Conclusion
We develop a model that divides the value of a property between the value of the land and the

value of any building, and we use it to analyze the effects of external shocks on the separate

values of land and buildings. A key result is that changes in land value affect the value of

existing buildings, decoupling the value of a building from its construction costs.

12
The actual building size that a developer will choose at time 0 depends on the expected duration of the economic
expansion as well as the discount rate.

22
 Our model offers a prediction for future empirical research on the economic effects of

transportation improvements. So far, empirical analyses have examined how transportation

improvements affect either the value of vacant land13 or total property values,14 providing

considerable empirical evidence that transportation improvements lead to higher property values.

However, several analyses in the second category refer to “housing values” rather than “property

values,” hence possibly suggesting to an inattentive reader that transportation improvements

raise the value of houses, considered separately from the value of the land on which the houses

sit. We are not aware of an empirical analysis that seeks to identify explicitly the separate effects

of transportation improvements on the values of land and existing buildings. Our model

provides a framework for undertaking such analyses, and we predict the finding that

transportation improvements lower the value of existing buildings.

13
See, for example, McDonald and Osuji, 1995, Craig et al., 1998, Coffman and Gregson, 1998, Knaap et al., 2001,
Cervero and Kang, 2011.
14
See, for example, Bowes and Ihlanfeldt, 2001, McMillen, 2004, McMillen and McDonald, 2004, Gibbons and
Machin, 2005, Yiu and Wong, 2005, Pagliara and Papa, 2011, Duncan, 2011, Billings, 2011, Mayor et al., 2012,
Yan et al., 2012, Wang, 2015, Kim and Lahr, 2014, Sagior, 2016.

23
References:
Arnott, R.J., MacKinnon, J.G., 1977. Measuring the costs of height restrictions with a general
equilibrium model. Regional Science and Urban Economics 7, 359–375.
Billings, S. B., 2011. Estimating the value of a new transit option. Regional Science and Urban
Economics 41, 525–536.
Bowes, D. R, Ihlanfeldt, K. R., 2001. Identifying the impacts of rail transit stations on
residential property values. Journal of Urban Economics 50, 1–25.
Bertaud, A., Brueckner, J., 2005. Analyzing building-height restrictions: Predicted impacts and
welfare costs. Regional Science and Urban Economics 35, 109–125.
Brueckner, J., 1980. A vintage model of urban growth. Journal of Urban Economics 8, 389–
402.
Brueckner, J., 1987. The structure of urban equilibria: A unified treatment of the Muth-Mills
model. In: Mills, E. (ed.). Handbook of Regional and Urban Economics. Vol. 2. North-
Holland, Amsterdam.
Brueckner, J., 2000. Urban growth models with durable housing: An overview. In: Huriot, J.
and Thisse, J. (eds.). Economics of Cities. Cambridge University Press, Cambridge.
Bruecker, J., Sridhar, K. S., 2012. Measuring welfare gains from relaxation of land-use
restrictions: The case of India’s building-height limits. Regional Science and Urban
Economics 42, 1061–1067.
Brueckner, J., Fu, S., Gu, Y., Zhang, J., 2017. Measuring the stringency of land use regulation:
the case of China's building height limits. The Review of Economics and Statistics 99(4),
663–677
Cervero, R., Kang, C. D., 2011. Bus rapid transit impacts on land uses and land values in Seoul,
Korea. Transport Policy 18, 102–116.
Coffman, C., Gregson, M. E., 1998. Railroad development and land value. Journal of Real
Estate Finance and Economics 16(2), 191–204
Colwell, P. F., Ramsland, M., 2003. Coping with Technological Change: The Case of Retail.
Journal of Real Estate Finance and Economics 26(1), 47–63.
Craig, L., Palmquist, R., Weiss, T., 1998. Transportation improvements and land values in the
antebellum United States: a hedonic approach. Journal of Real Estate Finance and
Economics 16(2), 173–189.
Duncan, M., 2011. The impact of transit-oriented development on housing prices in San Diego,
CA. Urban Studies 48(1), 101–127.
Enis, T.W., 1961. Empire State sold; price is 65 million. New York Times, August 23, 1961.
http://query.nytimes.com/mem/archive/pdf?res=9C01E5DC1430EE32A25750C2A96E9C94
6091D6CF (accessed on 12/11/2017).
Gibbons, S., Machin, S., 2005. Valuing rail access using transport innovations. Journal of
Urban Economics 57, 148–169.

24
 Glaeser, E. L., Gyourko, J., 2005. Urban decline and durable housing. Journal of Political
Economy 113(2), 345-375.
Glaeser, E. L., Gyourko, J., Saks, R., 2005. Why have housing prices gone up? NBER Working
Paper no. 11129.
Internal Revenue Service, 2016, How to depreciate property. IRS Publication 946, Cat. No.
13081F.
Kim, K., Lahr, M. L., 2014. The impact of Hudson-Bergen Light Rail on residential property
appreciation. Papers in Regional Science 93(S1), S79–S97.
Knaap, G. J., Ding, C., Hopkins, L. D., 2001. Do plans matter? The effects of light rail plans on
land values in station areas. Journal of Planning Education and Research 21, 32–39.
Mayor, K., Lyons, S., Duffy, D., Tol, R. S. J., 2012. A hedonic analysis of the value of rail
transport in the Greater Dublin Area. Journal of Transport Economics and Policy 46(2),
239–261
McDonald, J. F., Osuji, C. I., 1995. The effect of anticipated transportation improvement on
residential land values, Regional Science and Urban Economics 25, 261–278.
McMillen, D. P., 2004. Airport expansions and property values: the case of Chicago O’Hare
Airport. Journal of Urban Economics 55, 627–640.
McMillen, D. P., McDonald, J. P., 2004. Reaction of house prices to a new rapid transit line:
Chicago's Midway line, 1983-1999. Real Estate Economics 32(3), 463–486.
Pagliara, F., Papa, E., 2011. Urban rail systems investments: an analysis of the impacts on
property values and residents’ location. Journal of Transport Geography 19, 200–211
Quigley, J., Raphael, S., 2005. Regulation and the high cost of housing in California,” American
Economic Review 95(2), 323–328.
Rosenthal, S., Helsley, R., 1994. Redevelopment and the urban land price gradient. Journal of
Urban Economics 35: 182–200.
Sagior, J., 2016. Measuring the short-term, perceived impacts of high-speed rail on residential
property values in a Florida county. Journal of Property Tax Assessment & Administration
13(1), 49–83.
Tideman, T. N., 1999. Taxing land is better than neutral: land taxes, land speculation, and the
timing of development. In: Kenneth C. Wenzer (ed), Land-value taxation: the equitable and
efficient source of public finance, M.E. Sharpe: Armonk, NY.
Wang, X., 2015. Subway capitalization effect in Beijing: theory and evidence on the variation of
the subway proximity premium. Papers in Regional Science, DOI: 10.1111/pirs.12203.
Yan, S., Delmelle, E., Duncan, M., 2012. The impact of a new light rail system on single-family
property values in Charlotte, North Carolina. The Journal of Transport and Land Use 5(2),
60–67.
Yiu, C.Y., Wong, S. K., 2005. The effects of expected transport improvements on housing
prices. Urban Studies 42(1), 113–125.

25

View publication stats