In most jurisdictions, land and improvements are taxed at the same rate. The District
of Columbia is no exception to this general rule. Consider two homes in the District, each
valued at $100,000. Home A is a modest home on a large lot; suppose the land and
structures are each worth $50,000. Home B is a more substantial home on a smaller lot; in
this case, suppose the land is valued at $20,000 and the improvements at $80,000. Under
current District law, both homes would be taxed at a rate of .96 percent on the total
value and thus, as Table 1 shows, the owners of both homes would face property taxes of
$960. 1
But property can be taxed in many ways. Under a graded, or split-rate tax, land is
taxed more heavily than structures. Suppose the District were to decide to change its
property tax so that the tax rate on land were twice as high as the rate on structures and
rates were adjusted so that the two homes in this example continued to pay a total of
$1,920. As the second panel of Table 1 shows, this would require a tax rate on structures
of .71 percent and a tax on land of 1.42 percent. Taxes on Home A would rise to $1,067 and
taxes on Home B would fall to $853. A pure land tax, or site-value tax, takes this line of
reasoning to its logical limit. Under a pure land tax, the tax rate on structures is set
at zero and all revenues are raised by taxing land. An equal yield land tax in this
example would require a tax rate of 2.74 percent. The owner of Home A would pay $1,371 in
taxes under a pure land tax while the owner of B would pay $549. This report is an
analysis of graded property taxes and pure land taxes. It has the following organization.
Section II presents the history and theory of land taxes. Section III examines domestic
and international experience with graded taxes, with a particular focus on Pittsburgh.
Section IV looks at the distribution of the burden of graded property taxes and land taxes
and Section V examines questions of implementation. Section VI offers a brief summary and
conclusion.
II. The History and Theory of Land Taxes
Economists have had a longstanding interest in land taxation.2
The physiocrats, Adam Smith, David Ricardo, James Mill, and John Stuart Mill all wrote
extensively on the subject. But Henry George (1839-1897) is perhaps the person most
closely associated with land taxation. George was a sailor, prospector, printer, reporter,
San Francisco newspaper editor and publisher, political activist, and political economist
(Gaffney, 1987). He was immensely popular, so popular in fact that he nearly won the New
York mayoral election in 1886.
George is best known today for his 1879 book, Progress and Poverty. As Tideman
(1994) explains, George's primary concern was not to devise a better tax system, but
rather to explain why a growing incidence of poverty accompanied rapid economic growth and
to offer a remedy. From his experience in California in the 1850s and 1860s, he saw a
causal connection between the returns to owning land during economic booms and the
concurrent fall of wages. He viewed this as a result of an artificial scarcity of land
caused by speculators withholding land from production. His proposed solution: abolish all
taxes except for a tax on land values (and so people often talk about George's proposal as
a single tax). George argued that this tax would make more land accessible to those who
wanted to use it productively and make land speculation unprofitable. The greater
accessability of land and removal of other taxes would raise wages and lower prices,
thereby rasing workers' standard of living. George suggested that the tax equal virtually
all land rents; if the taxes collected were greater than government spending, the
remainder would be returned to voters.
Very few economists today would agree with all of Henry George's proposals or with his
analysis of the causes of poverty. Many would, however, support a call for substituting
taxes on land for other taxes. We now turn to a discussion of the basic appeal of a tax on
land.
Neutrality and Tax Incidence
It has often been argued that a tax on land is neutral, where neutrality is defined
along the following lines. All taxes impose a burden on taxpayers. It is tempting to view
this burden simply as the amount of money paid to the tax collector. In many cases,
however, there is a second element that requires attention. Most taxes will lead people to
distort their decisions. If the federal government taxes wages, some people are likely to
work less; if it taxes returns to capital (dividends, capital gains, interest) some people
will save and invest less; if local governments impose high taxes, some people will choose
to live in a different community. Economists call the economic cost of these distortions
the excess burden, or deadweight loss, of a tax. Thus the total burden of a tax is the sum
of the tax revenue collected (the direct burden of the tax) and the excess burden. Taxes
can be judged according to many criteria. Excess burden is one sensible criteria. A
"good" tax from this perspective is one that leads to only small distortions of
people's choices. The best tax would be one that is neutral, i.e., one that does not in
any way change anyone's behavior.
The standard property tax, like most taxes, imposes an excess burden and thus is not
neutral. Faced with the prospect of higher property taxes, landowners are likely to
develop their property less intensively than they otherwise would; a commercial developer
might, for example, decide to build a smaller office building or a homeowner might decide
not to renovate. In some cases the property tax will reduce the anticipated returns from
development so sharply that a landowner will choose to leave a property undeveloped.
There are some interesting historical examples of the deadweight loss from the property
tax. In 1696, England introduced a tax on the number of windows in a house, which was
believed to be a proxy for the size of the house and hence its value (Grinath, 1988).
Clearly, the number of windows may be correlated with house value, but such a tax would
place equal levies on homes with equal number of windows regardless of location or state
of repair. The tax may also lead to underconsumption of windows. When homeowners attempted
to replace windows with other methods of letting in air and light, the tax was expanded to
include all other forms of ventilation. A whole row of houses was built in Edinburgh,
without a single window in the bedroom story of the homes. Other homeowners would stop-up
windows and then reopen them after the assessor had been by their house. The Window Tax
remained in effect well into the 1850's.
A tax on land is an important exception to this general rule that taxes distort
decisions and thus generate deadweight losses. The analysis of a land tax in a static
setting where we ignore the use of the additional revenue is straightforward. The supply
of land is fixed. If land is taxed, people cannot reduce its supply in response; in the
jargon of public finance, the supply of land is perfectly inelastic. Landowners can do
nothing to escape the burden of the tax and thus a land tax does not distort economic
decisions. It is neutral; it neither encourages nor discourages development. And this
neutrality, of course, is part of its appeal.
If, on the other hand, we consider an increase in land taxes where we use the revenues
to reduce the tax on structures or other improvements, the economic consequences will be
very different. As we saw above, a tax on structures discourages development and thus is
not neutral; it reduces the intensity with which land is used. Thus arguments that a tax
on land can stimulate development turn in part on an implicit assumption that the revenues
will be used to reduce the tax on structures. But let us be clear. In that sense, any tax
will stimulate development as long as the negative effects of collecting the tax are less
than the positive effects of reducing the tax on structures. Land taxes have no special
claim as a tool to foster development. Almost certainly, for example, a nation wide head
tax would stimulate development in this sense.
Things become more complicated in a dynamic setting. Bentick (1979) and Mills (1981)
have argued that land value taxation need not be neutral with respect to the timing and
nature of land development. In particular, the taxing of land values may distort
the choice between earlier and later development of unused land parcels in favor of those
projects that promise an earlier stream of net receipts. The implication of their models
is that a movement in the direction of land taxation may hasten economic development,
perhaps to an extent that is excessive on purely efficiency grounds. This effect, however,
depends upon an important and controversial assumption concerning the way in which land is
valued for tax purposes.
Oates and Schwab (1997) presented a simple example to illustrate the Bentick and Mills
argument and to highlight the key issues. In that example, landowners can either devote
their land to Use A and earn $1,000 per period forever or wait one period, develop their
land in Use B, and earn $1,100 per period. Land use decisions can never be changed once
they have been made. At a 10 percent interest rate and with no taxes, landowners would be
indifferent between these two options.
Now suppose the local government introduces a land tax. The impact of this land tax
turns crucially on how it is administered. There are two possibilities. Bentik and Mills
assume implicitly that the government would, in periods 2 and beyond, treat land developed
in Use A differently than land developed in Use B. In their view, since land use decisions
are irrevocable and the two land uses generate different land rents, land devoted to Use B
should be taxed more heavily than land devoted to Use A. In this setting, all land would
be developed in the first period. That is to say, the imposition of a tax on land would
stimulate development.
At first blush, this line of argument seems compelling. But as David Wildasin (1982)
and T. Nicolaus Tideman (1982) have pointed out, this form of a land tax is inconsistent
with most peoples' view, including Henry George's, of a land tax. They would argue that,
in the context of our example, the non-neutrality of land value taxation results from the
practice of taxing land on the value associated with its chosen use. If land were always
assessed at each point in time for tax purposes on the basis of its "highest and
best" use, irrespective of any commitments to a particular use, then land-value
taxation would indeed be neutral. Taxation at such a standard value (Vickrey 1970)
would be use-independent and, hence, neutral. In terms of the Oates and Schwab example,all
parcels would be taxed as if they produced a rental income of $1,000 in period one and
$1,100 in all subsequent periods. In this case all land is treated identically, and a land
tax must clearly be neutral. Taxes in this case are independent of a landowner's decision.
Thus any decision that maximizes the value of a parcel of land in the absence of the land
tax will continue to maximize value in the presence of the tax.
2See Tideman (1994) for an excellent discussion of
the history of thought on land taxation.
III. Domestic and International Experience with
Graded Taxes
While it is true that the District's property tax is similar to the property tax in
most cities, (that is to say, land and improvements are taxed at the same rate), there are
a number of U.S. examples of graded taxes and a handful of examples of pure land taxes.
Moreover, graded taxes and land taxes are used extensively outside the U.S. We now turn to
a discussion of the experience with land and graded taxes.
The Pittsburgh Experience3
Pittsburgh is the only major U.S. city that uses a graded property tax. In 1979 and
1980, Pittsburgh restructured its property tax system so that land was taxed at more than
five times the rate on structures. Oates and Schwab (1997) sift through the evidence from
this recent "natural experiment" in Pittsburgh, and we summarize their study
here.4
In order to understand the effects of land value taxation in Pittsburgh, it is
important to place this tax reform in the context of the ongoing economic evolution of the
city and metropolitan area.5 Since the end of World War
II Pittsburgh has gone through a largely successful though often painful economic
transition. The city was once a major manufacturing center. As the steel and other heavy
manufacturing sectors shrank, the Pittsburgh economy became more oriented towards
white-collar jobs. The banking and other service sectors, for example, grew very quickly
as Pittsburgh became the regional financial leader. In 1940, manufacturing employment in
the four-county Pittsburgh MSA accounted for almost half of the total work force; in 1985,
manufacturing employment constituted only 16 percent of total employment.
Pittsburgh has launched several major urban renewal projects. Renaissance I, initiated
in the 1940's, was a major effort to revitalize the central business district through a
public-private partnership. The project enjoyed a number of important successes. It was
followed by a second major renewal effort in the late 1970s: Renaissance II. As before,
the renewal effort involved an extensive partnership between public and private agents
with a major focus on continued development of the central business district. Several
major corporations decided to expand their headquarters in Pittsburgh and with public
assistance constructed a series of major office complexes. The result was a striking surge
in levels of commercial construction activity: there were commercial contract awards in
1980 for 9.5 million square feet of new space with (as discussed below) continued high
levels of building activity through most of the decade.6
Pittsburgh has had a graded property tax system since 1913, a system under which land
was taxed at a rate twice that of the structures on the land until 1979. As Table 2
indicates, Pittsburgh introduced a striking restructuring of the city's property tax in
1979 and 1980, raising the tax rate on land while leaving the rate on structures
unchanged, thus raising the tax rate on land to about five times the rate on structures.
This increased "tilt" of rates has been maintained and even increased slightly
during the decade following the restructuring.7
Table 2 should be interpreted cautiously, however. Properties in the city of Pittsburgh
are subject to taxation not only by the city government, but also by the county and the
overlying school district. These latter two jurisdictions do not participate in the graded
tax system but instead employ a conventional property tax that applies the same tax rate
to land and structures. As the last column of the table indicates, this results in total
tax rates on land in the city of Pittsburgh that are something more than twice the rate on
structures. Properties outside the city are, in contrast, subject to conventional property
taxation.
Oates and Schwab assembled time-series data on new building activity for a sample of 15
cities and metropolitan areas in the Rust Belt including Pittsburgh.8
Table 3 presents figures from their study for the real value of new building permits for
the cities in their sample based on the Dun and Bradstreet data. The trends in Table 3 are
striking. In 13 of their 15 sample cities, construction was higher during 1960-79 than in
1980-89. In some cases the trends are dramatic: for example, construction fell by
two-thirds in Youngstown and one-half in Erie. Columbus shows a slight rise. But
Pittsburgh is a remarkable outlier: the real value of building permits on an annual basis
rose by some 70 percent in the 1980's relative to the twenty-year period preceding the tax
reform!
Oates and Schwab point to two further interesting patterns. First, the Pittsburgh boom
was a central city phenomenon; construction in the Pittsburgh suburbs was actually lower
in the 1980's than in the mid and late 1970's. Second, virtually all of the increase in
Pittsburgh construction was in the nonresidential sector. Residential construction rose
only slightly while, in sharp contrast, commercial and industrial construction more than
tripled in annual value.
Oates and Schwab then turn to the key question: what role did Pittsburgh's decision to
increase the tax on land play in the city's construction boom? They argue that the
increase in the land tax was probably not an important direct explanation of
Pittsburgh's building boom. Some pieces of informal evidence support their conclusion.
First, in their interviews with "development experts," the Pennsylvania Economy
League (1985) found no evidence that the increase in rates of land taxation exerted a
noticeable impact on construction activity.9 Second,
the League found that several of the major projects that were begun in 1981 were well
along in the planning stages before the increase in the graded-tax ratio (though
admittedly, plans are often left fulfilled and it is possible that the increase in the tax
on land increased the number of proposed projects that were actually built).
Third, it is clear that the dramatic changes in the Pittsburgh economy significantly
increased the demand for office space but that the supply of office space was slow to
respond. This excess demand for office space is apparent in data on office vacancy rates
for Pittsburgh and some of the other cities in our sample. Vacancy rates in Pittsburgh
ranged from less than one percent to three and one-half percent during 1978-1982. The data
suggest, moreover, that the construction of several massive new office buildings in the
early 1980s effectively brought the market back towards equilibrium, as office vacancy
rates rose sharply by the middle of the decade.
How then do Oates and Schwab account for the Pittsburgh building boom? They assign a
major role to a fundamental imbalance between the supply and demand for office space
generated by the growing importance of the financial and service sectors in the Pittsburgh
economy and the commitment from a number of major corporations to the Pittsburgh CBD.
Some proponents of land taxation have been disappointed by the Oates and Schwab
conclusion, arguing that they have understated the importance of the change in tax policy.
Oates and Schwab argue that this criticism is a serious misinterpretation of their
position. Economic theory, they argue, tells us that a major increase in land value
taxation in Pittsburgh should have been (roughly) neutral. The critics of land value
taxation have suggested that the Pittsburgh tax reform was unimportant because it had
little effect on development. The point here is that if land taxation is neutral, we would
expect it to have no effects on any decisions. This is its very appeal: it does not
distort economic choices! Thus, the responses of those interviewed are fully consistent
with the traditional view of the neutrality of land taxation. Land taxation should not,
and apparently did not, in itself hasten development.
Does this mean that Pittsburgh's land taxes are an ineffective policy tool and that the
decision to increase the tax on land was irrelevant? The answer, Oates and Schwab argue,
is "no". What would have happened if the city had not decided to tax land more
heavily? Pittsburgh was under severe fiscal pressure in the late 1970's, and some type of
tax increase was virtually unavoidable. Had an increase in land value taxation not been
introduced, city officials would have turned to another form of taxation: perhaps higher
taxes on structures, or (more likely according to contemporary reports) the introduction
of a significant increase in the city's wage tax. Both theory and empirical evidence
strongly suggest that such tax increases would have had a serious negative effect on the
Pittsburgh economy. It is against the backdrop of such alternatives that the tax on land
values needs to be considered. The role of land value taxation in Pittsburgh should be
understood in a setting of differential taxation. The relevant issue here is how
the effects of the land value tax compared with those of the available alternative
sources of tax revenues. It appears that a land tax did not cause a building boom in
Pittsburgh, but it did allow the city government to avoid policies that might have
undercut that boom.
Graded Taxes and Land Taxes in Other U.S. Cities
A number of Pennsylvania cities other than Pittsburgh have used a graded tax at various
times, and their experience offers a further opportunity to study the effect of split-rate
taxation. Tideman and Johnson (1995) looked at construction in 55 Pennsylvania cities and
towns over the 1980-1992 period. Seventeen of their cities used a split-rate tax over at
least part of the sample period (Aliquippa, Carbondale, Clarion, Coatsville, Du Bois,
Duquesne, Harrisburg, Hazleton, Lock Haven, McKeesport, New Castle, Oil City, Pittsburgh,
Scranton, Titusville, Uniontown, and Washington).10
They found no evidence to support the argument that the difference between the tax rate on
land and the tax rate on structures led to higher construction.
Stephen Cord (1983) interprets the Pennsylvania experience quite differently. Cord
compares Scranton with a similar, neighboring city Wilkes-Barre. Although these cities had
nearly equal revenue per capita, as well as similar ethnic characteristics, the former had
a history of using the graded tax. In 1979, Scranton nearly doubled the tax rate on land
and removed the property tax from new construction while Wilkes-Barre kept the standard
flat-rate property tax. Cord shows that in the two years following the tax change, average
annual building permits increased 22 percent in Scranton and decreased 44 percent in
Wilkes-Barre from the three previous years. His analysis of McKeesport came to similar
conclusions. In 1980, McKeesport increased the tax rate on land, decreased the tax rate on
buildings, and offered three year tax abatements for new construction. Construction in
McKeesport rose in 1980-81 relative to the preceding three years but fell in two
neighboring cities that maintained the standard property tax.
The Center for the Study of Economics also presents some interesting information
regarding economic development in Pennsylvania cities using the graded tax. Washington,
which adopted a graded tax in 1985 and expanded it throughout the next decade, compares
favorably with the similar, neighboring Uniontown, which also adopted a graded tax but
quickly rescinded the tax after one year. Average annual construction per person over the
1987-1995 period was 23 percent higher in Washington. Similarly, they found that New
Castle experienced a 70 percent increase in the number of building permits issued within a
three year period following its change to a graded tax. Neighboring towns retaining their
flat rate experienced a 66 and 90 percent decrease.
Amsterdam, New York is one of the few cities outside Pennsylvania to experiment with a
graded tax. Amsterdam hoped that a split-rate tax would increase economic development
(Reeb, 1993). The state of New York passed the required enabling legislation in 1993 and
the graded tax went into effect July 1, 1995, the beginning of a new fiscal year. The tax,
however, was repealed June 30, 1996, a victim of a mayoral election loss. Under the short
lived graded tax, 25 percent of the property tax burden fell on land as compared to12
percent under Amsterdam's standard property tax. For the year it was in effect, very few
people noticed the change. In fact, Don Reeb, an economist heavily involved in drafting
the law, discovered that no one attending a local Chamber of Commerce meeting was aware of
the tax change. Reeb surmised that these business owners were not alerted by an increase
in their property tax bill since property taxes are so small relative to the businesses'
value of sales. The Amsterdam assessor claimed that only three of Amsterdam's 20,000
residents contacted the assessor's office to ask questions about the tax change.
There are at least two U.S. towns that were founded on Henry George's principles. As
Woolery (1982) explains, Fairhope Alabama was founded in 1894 on the premise that all
public revenues would be derived from land taxes. The original founders formed the
Fairhope Single Tax Corporation which owned and managed the land through leases rather
than issuing deeds. As the need for public services increased, so did the necessary rent
charged the lessees. This led to disputes which more than once ended in court. The
corporation has faced many lawsuits, including a case it appealed to the Alabama Supreme
Court over the legality of the Fairhope Single Tax Corporation. Surviving these, it struck
oil in the late 1970's. With the prospects of oil income from the land, the corporation
may no longer need rental income to support public outlays, thus diverting the experiment
away from a test of the George idea.
Arden, Delaware was also founded on Henry George's principles (Wiencek, 1992). Frank
Stephens, a sculptor and businessman, spent seven years as a lieutenant for George,
working to bring the single tax movement into the American political realm. Following a
failed attempt to elect Georgist politicians in 1896 Congressional races across the state
of Delaware and George's death in 1897, Stephens returned to Delaware in 1900 to buy a
162-acre farm north of Wilmington. On this land he formed Arden as a single tax community.
Arden's deed established that no property would be owned privately; instead, the land
would be held in trust by three trustees and leased to residents in 99-year leases.
Residents would pay an annual a land rent based on the value of they parcel they occupy,
as determined by elected assessors. This land rent would be the single tax necessary to
fund the community. An architect and single-tax apostle Will Price designed the community,
allocating nearly half the land to forests, paths and roads. The community, once founded,
attracted a unique group of residents who were not required to pledge faith in the
teachings of George. Artisans, woodworkers, authors and poets of all political persuasions
moved into the community. They found the cost of residency low since they were not allowed
to buy land, but only needed to build a modest dwelling and pay the annual land tax.
International Experience
In Denmark, as early as 1844, the national property tax was assessed on the value of
the land but not the improvements on the land. The site-value property tax was abolished
in 1903 and replaced with a flat-rate on the total value of land and improvements.
However, this tax change hurt the many small farmers in Denmark who had invested heavily
in improvements but had limited holdings of land. These farmers joined together to lobby
for a return to the site-value tax. In 1922, legislation was introduced to split the
national land tax into a composite-rate or graded tax. Today, all cities in Denmark use a
graded property tax (Center for the Study of Economics, pamphlet). The effects of a graded
tax on development have not been studied extensively in Denmark despite the widespread use
of the graded tax because the proportion of overall revenues raised by the property tax
has decreased with time (Silagi, 1994).
Australia has experimented extensively with the graded tax (Edwards, 1984). A federal
land tax was initiated in 1911 and then revoked in 1952. Prior to 1976, all of the states
in Australia taxed property based on the value of the unimproved land. Today, all but the
state of Tasmania continue with the land tax, although the revenue raised by the states is
relatively small.
Many of the local governments in Australia also use variants of the graded tax. The
Sydney City Council, for example, began levying a tax on the unimproved value of land in
1916 (Archer, 1972). The tax is proportional and the rate is determined annually to meet
the budgetary needs of the city. Valuation of land is based on fair market value, which
implies that it is based on the property's value in the highest potential use and not
current use. This is in contrast to methods employed in the United States, where typically
site value is assessed as the "land value proportion of the current market value of
an improved property," (Archer, 1972, p.23). Assessments rely on market data, which
includes sales evidence. Not only sales of vacant land, but also sales of parcels for
redevelopment serve to reveal unimproved capital values of land.
Lusht (1992) and Flaherty and Lusht (1996) studied the effects of land value taxation
in Australia looking at economic development among the municipalities in the Melbourne
statistical district. In 1919, municipalities were granted the choice among property tax
bases, including land values, a variation of capital value, or a combination of the two.
The freedom to choose a property tax base did lead to some early switches, but few changes
have happened since the 1960's. By 1991, 27 of the 56 municipalities in the district were
using land taxation, 28 were using capital value taxation, and one used a combination of
the two tax bases. Both studies focused on 28 of the municipalities which were classified
as containing "substantial land available for residential development," fifteen
with land value taxation and the other thirteen with a flat-rate tax on both land and
improvements. Lusht looked at economic development among the municipalities in the
Melbourne statistical district. Comparing the municipalities, he found those using a land
tax were located closer to the center of the city and had residents with slightly higher
income. Controlling for these differences, his time series analysis showed that those
areas with land value taxation had higher rates of residential development over the
1983-1987 period. Flaherty and Lusht compared the selling prices of residential lots and
the stocks and flows of residential development for this cross-section of municipalities.
The authors could not clearly identify the effect of a land value tax on the selling price
of lots. The inability to show a relationship between the tax used and lot values can be
explained as the combination of capitalization of the land tax decreasing prices and the
favorable tax treatment of improvements increasing prices. Their study showed that those
areas with land value taxation had more dense residential development and a larger stock,
but they found no association between land value taxation and residential flows for
1990-1993.
A number of countries in Africa use some form of land value taxation, including South
Africa. The new South African constitution, passed in 1996, restricts the means by which
local governments can raise revenue to property taxes, excise taxes, and utility charges.
South African law allows local governments to choose between a flat-rating approach, a
composite-rating approach, or a site-rating approach. Flat-rating entails levying an equal
tax rate on both the value of the land and the improvements built on the land (and is thus
equivalent to the standard property tax in the U.S.). Composite-rating entails levying
different tax rates on the land and improvements (i.e., it is a graded tax), where the
improvements are usually taxed at a lower rate. Site-rating entails levying a tax only on
land.
A study carried out by Michael Bell and John Bowman (1997) presents an interesting
analysis of property taxes in many jurisdictions within South Africa. Johannesburg
currently uses a site-rating approach. The Greater Johannesburg Transitional Metropolitan
Council expected nearly 23 percent of its revenue for 1995-96 to come from the property
rate. Land value assessments are based on the status of the land as either non-income
producing or income producing. For non-income producing property, i.e., residential, both
improved value and land value are calculated. The latter is calculated by "relating a
number of sales in a particular township to a hypothetical stand of 1,000 square meters
using land conversion factors and then adjusting this value by the percentage positive and
negative land attributes unique to that particular stand," (Bell and Bowman 1997,
p.51). For income producing property, assessment of land value is based on "fair
market value", where comparable sales at comparable times are used to capture the
market value of the land in question.
The City of Cape Town currently uses a flat-rating approach but has plans to change to
a site-rating approach. In 1991, the city council appointed a Committee of Enquiry to
study the benefits of site-rating. The committee recommended a goal "to attain a
ratio of 10: 1 in favor of the rating of land as against improvements," (Bell and
Bowman 1997, p.98). Despite the recommendation for composite-rating, for political reasons
the city council decided to first value sites and implement the tax on land. As local
jurisdictions are integrated with the end of apartheid, new areas are being brought under
Cape Town's jurisdiction. Many of these areas are former black local authorities, where
prior to the mid-1990's blacks were forbidden to own much of the land. Without ownership,
no local government in the past could levy a property tax, so the properties had never
been assessed. Now that ownership is allowed, the city must assess values before it can
tax the properties, a process which requires time and money. Leaders believe that the time
necessary to assess the values of improvements and land may undermine the legitimacy of
existing property taxes if existing assessments, which were done in 1979, become grossly
out-of-date while the new areas are considered. Clearly the city is in a transition stage
in its development.
From 1903-1913, Canada introduced site value taxation in the western provinces in an
effort to encourage the breakup of large areas of land held by absentee owners, to prevent
land speculation, and to spur construction (Kitchen, 1995). Assessments reflected the
highest-and-best-use of the land in question. Provinces have since turned over the
property tax to the municipalities. Today, cities in the four western provinces either
exempt improvements from the property tax base, or record the improvements on the
assessment roll at a percentage of their full value (Perry, 1990). The most common levels
at which improvements are assessed is three-fifths to two-thirds of the full value. Both
Quebec and Newfoundland require separate assessments of land and improvements, even though
the provinces levy a flat-rate property tax.
3 This section of the report draws heavily on Oates
and Schwab (1997) and Oates and Schwab (forthcoming).
4 There were at least three earlier studies of the
effects of land value taxation in Pittsburgh. Henry Pollakowski (1982) was unable to find
much in the way of "adjustment effects" as measured by the number of property
transactions. However, his data extended only from 1976 through 1980. Steven Bourassa
(1987) explored the effects of Pittsburgh's tax system on housing development. Using
monthly data on the value of new residential building permits as his dependent variable,
Bourassa found that the tax rate on improvements, but not the rate on land, was a
statistically significant determinant of the level of residential building activity.
Bourassa's findings, while of some interest, are limited in scope, for, as we shall see,
the major impetus to development in Pittsburgh has been in the non-residential sector. Of
more relevance to our concerns is an interesting study undertaken in the mid-1980s by the
Pennsylvania Economy League (1985). At the request of Mayor Richard Caliguiri, the League
examined the effects of the graded tax on both the development of the city and the equity
of the tax system. Drawing both on extensive interviews with "local development
experts" and some quantitative analysis of the graded-tax ratio and development of
different properties, the League concluded that "The graded tax has very little
effect on development" (p. ii).
5 For a useful description of the historical
evolution of Pittsburgh with a focus on the renewal efforts under Renaissance I and II,
see Shelby Stewman and Joel Tarr (1982).
6 The commercial building boom in Pittsburgh under
Renaissance II has encompassed several major projects: PPG Place (six buildings, including
a forty-story office tower), One Oxford Center (a forty-six story office tower and retail
complex), The Steel Plaza/One Mellon Bank Center (a fifty-three story office tower and
retail complex that includes the main station of the Light Rail Transit [LRT] system),
Allegheny International's headquarters, Liberty Center, the Hillman Complex, and several
others.
7 The assessment-sales ratio in Pittsburgh is .25 so
that the nominal tax rates appearing in Table 2 must be divided by 4 to obtain measures of
effective tax rates.
8 See Oates and Schwab (1997) for a more detailed
description of the sources and nature of the data.
9 Clearly, not everyone would agree with this
assessment. Walter Rybeck (1991), for example, quotes the Pittsburgh City Council
President as follows: "I'm not going to say the land tax is the only reason a second
renaissance occurred, but it's been a big help" (pp.4-5).
10 Allentown adopted a split-rate tax in 1997. Two
of the cities cited (Hazleton and Uniontown) rescinded the graded tax within a year after
adopting it.
IV. The Incidence of a Graded Property Tax in the
District of Columbia
Our report to this point has focused largely on some of the efficiency issues
surrounding the graded property tax. But efficiency is clearly not the only criterion
anyone should use in the evaluation of tax policy. The incidence of a tax is also an
important issue. Here we offer some evidence on a fundamental question: who would bear the
burden of a split-rate property tax in the District? Two earlier studies have looked at
this problem: The Department of Finance and Revenue's (DFR) 1994 study "The Impact of
a Split-Rate Property Tax in the District of Columbia" and the Pro-Housing Property
Tax Coalition's 1991 study "Real Property Tax Rates for Tax Year 1992". They
offer strikingly different views on the incidence of a graded property tax. The DFR report
concludes that a move to a split-rate tax would "shift property taxes burdens
generally onto residential property owners and away from other property uses" (p. 1);
the Coalition finds that a split-rate tax would reduce taxes on owner occupied housing by
12 percent. Our hope here is to shed some light on the key issues that lead to such
different conclusions.
Our analysis is based on DFR data on real property assessments by neighborhood and
class.11 These data include separate assessments of
the value of land and improvements and allow us to evaluate the impact of a split-rate
tax. There are at least two important limitations to our analysis. First, while we can
have confidence in the assessed valuation of the sum of improvements and land, it is
unclear how much faith we can place in the accuracy of the separate assessment of land and
improvements. Under current District tax policy, land and structures are taxed at the same
rate and thus it would make little sense for the District to put a great deal of effort
into developing accurate measures of land values; for all practical purposes it makes no
difference if we think of a $100,000 property as $25,000 of land and $75,000 of structures
or $75,000 of land and $25,000 of structures. If the District did adopt a graded tax it
would need to determine land values much more carefully and it is quite possible that, as
a consequence, our view of the distribution of the burden of the tax could change
significantly. Nonetheless, we proceed as though the valuations of land and structures are
accurate. Second, we focus on the statutory incidence of a graded tax but do not consider
the economic incidence of the tax. That is, we have not tried to estimate the extent to
which the burden of a graded tax would be shifted.
Our analysis proceeds along the following lines. We have data on the number of parcels,
the assessed value of land, and the assessed value of improvements for each of the
District's five classes of property and each of the District's neighborhoods. We begin by
examining current tax policy. To do so, we calculate property taxes for each class and
neighborhood given current tax rates. As shown in the first panel of Table 5, current tax
rates in the District range from .96 percent for owner occupied housing (Class 1) to 5.00
percent for vacant property (Class 5). Since the District has a traditional property tax,
the tax rate on improvements and land in the first panel of Table 5 is the same. Our
calculations of tax revenues reflect the $30,000 homestead exemption but do not reflect
special provisions for senior citizens. Throughout this report we refer to these estimates
of tax liabilities under current tax policy as the BASELINE.
We then examine five alternative policies. All five alternatives raise the same revenue
as the BASELINE, and thus our analysis is best thought of as an example of differential
tax analysis where we compare the effects of a possible tax change relative to
those of current policy. We summarize each alternative in Table 4 and for each alternative
we present the required equal yield tax rates in Table 5, relative tax revenues for all of
the District for each of five classes of property in Table 6, and relative tax revenues
for each neighborhood and each property class in Tables 7 through 11. We also present a
set of maps (Figures 1 through 3) which summarize the information on owner occupied from
Tables 7.
As Table 4 shows, ALTERNATIVE ONE is a pure land tax: structures are not taxed, all
revenue is raised by taxing land, and all land is taxed at the same rate regardless of its
current use. Taxing all land at the same rate requires the elimination of classification
and the homestead exemption. ALTERNATIVE ONE is very similar to DFR's "complete
split-rate tax" (Department of Finance and Revenue 1994, p. 16). The DFR analysis is
based on tax rate of 5.00 percent, chosen so as to avoid cutting taxes on vacant (i.e.,
Class 5) property. As DFR explains, a 5.00 percent land tax raises significantly more
revenue than does the District's current property tax system. Our analysis is based on a
land tax of 3.67 percent (see Table 5) which is the equal yield tax rate.
Table 6 shows clearly that this land tax would shift a significant portion of the
burden of the property tax from the owners of rental residential and non-residential
property to homeowners. All of the entries in Table 6 are index numbers that show tax
revenues from an alternative relative to a second tax policy. The first panel of Table 6
shows tax burdens for ALTERNATIVE ONE for each class of property relative to the BASELINE.
The first entry in that table of 151.7 implies that owner occupied housing (Class 1) would
pay 51.7 percent higher taxes under ALTERNATIVE ONE than under the BASELINE. Taxes for all
other class of property would fall.
At first blush, it might seem that these results mean that shifting taxes from
structures to land would always harm homeowners. We would argue that a somewhat different
interpretation would be more helpful. The key issue here is the District's classified
property tax. Under current policy, homeowners face a tax rate that is less than half the
rate on commercial property (see Table 5). A pure land tax does two things: it shifts the
burden from structures to land, and it taxes all properties at the same rate and thus
requires the end of classification. The uniform taxation of all property does harm
homeowners, but shifting the tax from structures to land does not.
ALTERNATIVE TWO is designed to make this point clear. As Table 4 explains, ALTERNATIVE
TWO is a standard property tax that taxes land and improvements at the same rate. It
eliminates classification and the homestead exemption, and thus treats property in all
classes equally. ALTERNATIVE TWO would require an equal yield tax rate of 1.53 percent
(see Table 5). As Table 6 shows, a standard property tax without classification would
impose a much higher burden on homeowners than would a pure land tax. The third panel of
that table shows tax burdens under ALTERNATIVE ONE (the pure land tax) relative to
ALTERNATIVE TWO (the standard property tax without classification). As Table 6 shows,
homeowners would pay 20.5 percent lower taxes under the land tax. Thus while it is true
that District homeowners would pay higher taxes under a pure land tax than under current
tax policy, it is the end of classification and not the shift of taxes from structures to
land that leads to this result. Homeowners would in fact be better off under a pure land
tax than they would be under a traditional property tax with uniform tax rates.
Would it be possible to design a tax system that shifts more of the burden from
structures to land but avoids raising the tax burden on homeowners? The answer is
"yes", and we present several below. But before doing so, it is important to
emphasize one point. As we argued above, a key feature of a pure land tax is neutrality;
it does not distort anyone's economic decisions. A hybrid graded tax coupled with
classification will clearly not be neutral; if we tax some uses at higher rates than
others then some land will be devoted to uses that are taxed relatively lightly. There
might very well be excellent reasons on equity grounds to maintain classification, but the
marriage of a land tax with property tax rates that vary across uses takes tax policy in
two very different directions at the same time.
There are at least two ways to implement a split-rate tax while trying to treat all
classes of property equally. ALTERNATIVE THREE is a graded tax where land is taxed at
twice the rate as structures.12 Tax rates in all
classes are changed by the same proportion; in particular, as the fourth panel of Table 5
shows, the tax rate on improvements is 31 percent lower under ALTERNATIVE THREE than under
the BASELINE (i.e., current tax policy in the District) and the tax rate on land is 38
percent higher. ALTERNATIVE THREE is very similar to the proposal in the Pro-Housing
Property Tax Coalition's 1991 study and DFR's "modified classified split-rate
tax".13
As the fourth panel of Table 6 shows, taxes on Class 1 and Class 2 property under
ALTERNATIVE THREE would be 9 percent lower than under the BASELINE and taxes on commercial
property would be five percent higher. The second column of Table 7 and Figure 1 present
tax burdens on owner occupied homes for each neighborhood. Here again, all of the data are
index numbers and show tax liabilities under ALTERNATIVE THREE relative to the BASELINE.
Thus, for example, Table 7 shows that homes in American University Park would face taxes
under ALTERNATIVE THREE that are 97.3 of the taxes they would face under current tax
policy. Figure 1 and Table 7 suggest that the benefits from this shift to a graded tax
would be widespread. Taxes would fall by 15 percent or more in 12 neighborhoods, fall by 5
to 10 percent in 36, and remain roughly unchanged in the remaining 8. Tax decreases would
be large in Anacostia and surrounding neighborhoods (see Figure 1), and thus there is at
least some evidence that low income homeowners would enjoy a significant part of the
benefits from a split-rate tax.
Alternatively, it would be possible to design a land or graded tax with different tax
rates for each class of property that generated the same revenue from each class as under
the BASELINE. ALTERNATIVE FOUR is a pure land tax which, as the fifth panel of Table 5
shows, incorporates tax rates of 2.91 percent (Class 1) to 5.20 percent (Class 5). By
design, tax revenues for each class of property under ALTERNATIVE FOUR are the same as
under the BASELINE and thus all of the entries in the fifth panel of Table 6 are exactly
100.0. While in the aggregate homeowners face the same tax liabilities, the impact of this
hybrid land tax would have quite different effects on different neighborhoods. The third
column of Table 7 shows that taxes under ALTERNATIVE FOUR would rise by more than 15
percent in 9 neighborhoods, rise 5 to 10 percent in 10, remain roughly unchanged in 9,
fall by 10 to 15 to percent in 8, and fall by more than 15 percent in the remaining 20.
Figure 2 suggests that low income homeowners would stand to gain the most under this
version of the land tax. Taxes would rise or remain unchanged in most of the neighborhoods
west of 16th Street and fall in most of the neighborhoods east of 16th. Here again, many
of the sharpest declines would be concentrated in Anacostia and surrounding neighborhoods.
ALTERNATIVE FIVE is a graded tax version that is similar in many ways to ALTERNATIVE
FOUR and DFR's "classified split-rate tax". Tax revenues are the same in each
class as under the BASELINE and thus all of the entries in the last panel of Table 5 equal
100.0. The tax rate on land is set at twice the rate on structures; the last panel of
Table 6 shows that the required tax rates on land range from 1.44 percent for Class 1 to
5.10 percent for Class 5. The last column of Table 7 and Figure 3 show that ALTERNATIVE
FIVE would have only a minimal impact on homeowners in most neighborhoods. Taxes would
remain essentially unchanged in 33 neighborhoods, rise by 5 to 10 percent in 7, and fall
by more than 10 percent in 16.
11 As discussed in the text, there are five classes
of property in the District and each has a different property tax rate. Some parcels have
mixed uses (e.g., they are part residential and part commercial). For this analysis, the
mixed use parcels were split into two or three smaller parcels. For example, a parcel with
a land assessment of $100,000 that was sixty percent commercial and forty percent
owner-occupied would have been divided into a $60,000 Class 4 parcel and a $40,000 Class 1
parcel.
12 Land represents roughly one-third of the total
assessed valuation of Class 1 property. We therefore incorporate a homestead exemption of
$10,000 for land and $20,000 for improvements in ALTERNATIVE THREE, ALTERNATIVE FOUR, and
ALTERNATIVE FIVE.
13 In DFR's modified classified split-rate tax, the
tax rate on vacant property (i.e., Class 5) is left unchanged at 5.00 percent.
V. The Assessment of Land Values
From the standpoint of neutrality, accurate assessments of land values under a
split-rate or pure land tax are not an absolute necessity. All that is required is that
taxes be independent of the way land is used so that taxes cannot influence a landowner's
decisions. But from a legal and administrative standpoint, accurate assessments are
essential. And the accurate assessment of land is difficult in an urban setting, such as
the District, where developed vacant land is sold only rarely. In this section we look at
some alternative approaches to estimating the value of land.14
According to the Appraisal Institute, there are a number of textbook techniques for
assessing land values. When vacant site sales information is available, the assessor would
use this information to assess land values of like properties. However, in the District
such sales are rare, particularly for residential properties. Subdivision analysis is not
applicable in the District where very little residential land exists which can be
subdivided for sale. The income capitalization approach requires information on the cash
flow for the property. The District currently uses the allocation method for vertical
properties such as condominiums, where a percentage of the sales price is allocated to be
the value of the land, since such properties do not have distinct lots. The allocation
percentage is usually 30 percent, but can vary depending on the property's location within
the city.
The District plans to rely heavily in the future on the extraction method for the
valuation of commercial land. Under the extraction approach, the assessor measures the
sales value of the improved property and estimates the depreciated value of the buildings;
land value is then the difference between the former and the latter, i.e., land value is
measured as a residual. Residential land in the District is valued on a neighborhood
section basis according to one of four classes: single family, row house end, row house,
or flat (multi-family or apartments). Adjustments are made for the size of the lot,
location or presence of alleys, corner influence, presence of extra deep or wide lot,
location of property, use, etc. These adjustments are based on the experience and
judgement of the assessor. Since many neighborhoods are far from homogeneous, the District
has created sections within neighborhoods of approximately 600 to 800 properties. The
District is developing a land chart as a base for assessments which contains the median
values for the four categories within each section.
While the extraction method might in the end prove to be the best of the available
approaches, there is an important conceptual problem that deserves attention. A sensible
definition of the value of a parcel of a land is the amount someone would be willing to
pay for the land if it were undeveloped. Thus in an important sense, land should not be
assigned the residual value but instead the structures should be considered as the
residual. If we knew what someone would be willing to pay for the land alone and what
someone would pay for the sum of the land and structures, then the structures are worth
the difference between the two. The original cost of the structures in this view is
irrelevant.
Under some circumstances, this distinction is simply an academic issue. To see this,
suppose someone has developed a parcel of land for Use A. In this example, the interest
rate is 10 percent, structures cost $10,000, structures never depreciate, the project
generates an annual return of $1,500 forever, and there will never be an alternative use
for the land. Under these assumptions, almost any reasonable approach would lead to the
conclusion that the land is worth $5,000. Someone would be willing to pay $5,000 for the
land if it were undeveloped, since they could develop it in Use A at a total cost (i.e.,
land plus structures) of $15,000 and then earn returns with a present value of $15,000.
Similarly , someone would be willing to pay $15,000 for the land and structures. Under the
extraction method, the assessor would assign a value of $10,000 to the structures and
assign a residual value to the land of $5,000. Thus the extraction method and the direct
valuation of the land offer the same, correct answer.
But now suppose that the market changes and suddenly it became possible to develop this
parcel in Use B. In Use B, structures again cost $10,000 but the annual return is $1,600.
Someone would now be willing to pay $6,000 for the land since they would immediately
develop it in Use B and realize returns with a present value of $16,000. The extraction
method, however, would continue to value the land at $5,000. More generally, the
extraction method can offer misleading estimates of land values whenever the highest and
best future use of land is different from the current use.
The District could also explore the possibility of using regression techniques and its
computer assisted mass appraisal system to develop better estimates of land values. The
International Association of Assessing Officers defines computer assisted mass appraisal
(CAMA) as "the systematic appraisal of groups of properties as of a given date using
standardised procedures and statistical testing," (McCluskey 1997, p.2). Using one of
the many different variations of CAMA, assessors attempt to replicate a property market
which allows them to project property values onto parcels which have not been sold within
the relevant time period. Mass appraisal techniques have two goals: "to attain
acceptable standards of predictive accuracy" and "to facilitate explainability
of the assessed values," (McCluskey 1997, p.5). CAMA evolved from econometric
modeling and multiple regression techniques, where there exist well-established
statistical tests for accuracy and where individual property characteristics can be
valued, aiding in the explanation of total appraisal values.
The typical mass appraisal system has five stages: data collection, verification and
analysis; model development to reflect property type and location; examination of
statistical validity and integrity of the model; incorporation of geographic information
systems; and determination of final assessment values. Numerous mass appraisal techniques
exist, but the most commonly used method internationally is multiple regression (MRA)
Multiple regression analysis seeks to develop a predictive relationship between
property characteristics and property value such that the assessor can determine value
only from knowledge of characteristics (Mark and Goldberg, 1988). The models which are
developed usually represent this relationship as linear, where the total property value is
a sum of the values of the property characteristics. Property characteristics often
include internal factors such as square feet of living space or number of bathrooms;
external factors such as public services or economic conditions; and location factors such
as neighborhood or accessibility. The chosen relationship, or regression equation, is
estimated for properties with sales information. MRA techniques allow the assessor to
decompose the total value of a property into components based on the property
characteristics. Using these estimated implicit prices and knowledge of property
characteristics, the assessor can estimate values for properties with no recent sales
information and values for individual characteristics of interest, such as land value.15
14 We thank Minetta Coles, Philip Applebaum, and
Randy Vinson of the District government for very helpful discussions about the District's
assessment practices.
15 A simple analogy might make this clearer. Suppose
we did not know the prices of individual items in a grocery store but we knew how much
each person spent at the grocery and how many of each item they purchased. MRA would
relate expenditures to the quantities of goods in an effort to develop estimates of the
prices of the individual items.
VI. Summary and Conclusions
Like most U.S. cities, the District taxes improvements and land at the same rate. But
as we argued at the beginning of this report, property can be taxed in many ways. Under a
graded, or split-rate tax, land is taxed more heavily than structures. A pure land tax is
a special case of a graded tax; under a pure land tax, the tax rate on improvements is
zero and all revenue is raised by taxing land.
Our goal in this report has been to set out some of the important issues the District
would need to consider before shifting to a graded tax. Pittsburgh is the only major city
in the U.S. that uses a graded tax and we have summarized the impact of Pittsburgh's
decision to sharply raise the tax on land in the late 1970s and early 1980s. We have also
reviewed other U.S. examples of a graded tax and we have looked at some of the
international experience on this question. Our report includes estimates of the incidence
of land taxes and graded taxes in the District and a discussion of some important
implementation problems.
The decision to shift to a graded property tax is a difficult question. Theory tells us
that higher taxes on land are non-distorting but that taxes on structures are. Therefore,
a graded tax could be expected to offer significant benefits. Given the District's tax
rates that now vary significantly across classes of property, a pure land tax would shift
the burden of the property tax onto homeowners. It would be possible to design a graded
tax that addressed this issue, but the marriage of a graded tax and classification would
not be easy. Administering a graded tax is far from straightforward since it would require
the District to develop accurate estimates of land values that it quite likely would need
to be able to defend in court. Moving to a graded tax would produce clear winners and
losers; some property owners will gain and some will lose from such a change in tax
policy. Thus a move to a graded tax could offer important advantages to the District, but
it is a decision that needs to be considered carefully.
Archer, R.W., Site Value Taxation in Central Business District Redevelopment
(Sydney, Australia). (Urban Land Institute Research Report 19, 1972).
Bell, Michael E. and John H. Bowman, eds., Property Rates in South Africa: A
Preliminary Assessment of Current Practices, (The Johns Hopkins University Institute
for Policy Studies, June 1997).
Bentick, Brian L., "The Impact of Taxation and Valuation Practices on the Timing
and Efficiency of Land Use," Journal of Political Economy, Vol.87 (August
1979), pp. 859-68.
Bourassa, Steven C., "Land Value Taxation and New Housing Development in
Pittsburgh," Growth and Change (Fall 1987), pp. 44-55.
Center for the Study of Economics, "C.S.E. What It Is and What It Does,"
(pamphlet).
Cord, Steven B., "Taxing Land More Than Buildings: The Record In
Pennsylvania," The Property Tax and Local Finance, C. Lowell Harriss, ed.,
(The Academy of Political Science, 1983), pp.172-179.
Department of Finance and Revenue, Government of the District of Columbia, The
Impact of a Split-Rate Property Tax in the District of Columbia, (September 1994).
Edwards, Mary E., "Site Value Taxation in Australia," American Journal of
Economics and Sociology. Vol. 43, (October 1984), pp. 481-495.
Flaherty, John and Kenneth M. Lusht., Site Value Taxation, Land Values, and
Development Patterns, (The Mary Jean and Frank P. Smeal College of Business
Administration, Pennsylvania State University, October 1996).
Gaffney, Mason, "Henry George," The New Palgrave, (MacMillan Press
Limited, 1987), pp.514-515.
Grinath III, Arthur C., "Why the British Did Windows: The Window Tax and Its
Historical Context," (mimeo, December 13, 1988).
Kitchen, Harry M. Property Taxation in Canada. (Canadian Tax Paper, No. 92,
Canadian Tax Foundation, 1992).
Lusht, Kenneth M., "The Site Value Tax and Residential Development," (The
Lincoln Institute of Land Policy Working Paper, 1992).
Mark, Jonathan and Michael A. Goldberg., "Multiple Regression Analysis and Mass
Assessment: A Review of the Issues," The Appraisal Journal, Vol. 56 (1988),
pp.89-109.
McCluskey, William, "Mass Appraisal Modelling Techniques and the Application of
Geographic Information Systems," Designing Local Government for South Africa:
Structures, Functions & Fiscal Options, (Session 6, Policy-Oriented Conference hosted
by the Center for Business Law, University of South Africa and the Financial & Fiscal
Commission, July 23-25, 1997)
Mills, David E., "The Non-Neutrality of Land Value Taxation," National Tax
Journal, Vol. 34, (March 1981), pp. 125-129.
Oates, Wallace E. and Robert M. Schwab, "The Impact of Urban Land Taxation: The
Pittsburgh Experience," National Tax Journal, Vol. 50 (March 1997), pp.1-21.
Oates, Wallace E. and Robert M. Schwab, "The Impact of Urban Land Taxation: The
Pittsburgh Experience," in Helen F. Ladd, ed. Local Government Land and Tax Policy,
(Cheltenham, UK: Edward Elgar Publishing Company, forthcoming).
Pennsylvania Economy League, Development, Equity, and the Graded Tax in the City of
Pittsburgh (Pittsburgh, Pa.: March 1985).
Perry, J. Harvey, Taxation in Canada, fifth edition. (Canadian Tax Paper, No.
89, Canadian Tax Foundation, 1990).
Pollakowski, Henry O., Adjustment Effects of a Tax on Land: The Pittsburgh Case
(Cambridge, Mass.: Lincoln Institute of Land Policy, Lincoln Institute Monograph #82-8,
1982).
Pro-Housing Property Tax Coalition, "Real Property Tax Rates for Tax Year
1992," (June 1991).
Reeb, Donald J., "Political Economy of the Two Rate Property Tax Amsterdam, New
York," (mimeo, October 12, 1993).
Rybeck, Walter, "Pennsylvania's Experiments in Property Tax Modernization," NTA
Forum (Spring 1991), pp. 1-5.
Silagi, Michael, "Henry George and Europe: In Denmark the Big Landowners Scuttled
the Age-old Land Tax But the Smallholders, Moved by George, Restored It," American
Journal of Economics and Sociology, Vol. 53, (October 1994), pp.491-501.
Stewman, Shelby, and Joel A. Tarr, "Four Decades of Public-Private Partnerships in
Pittsburgh," in R.S. Fosler and R. Berger, eds., Public-Private Partnership in
American Cities: Seven Case Studies (Lexington, Mass.: Heath, 1972), pp. 59-127.
Tideman, T. Nicolaus, "A Tax on Land Value is Neutral," National
Tax Journal, Vol. 35 (March 1982), pp. 109-111.
Tideman, T. Nicolaus, "The Economics of Efficient Taxes on Land," in Nicolaus
Tideman, ed., Land and Taxation, (London: Shepheard-Walwyn in association with
Centre for Incentive Taxation, 1994), pp. 103-104.
Tideman, T. Nicolaus and Cathleen Johnson, "A Statistical Analysis of Graded
Property Taxes in Pennsylvania," (Lincoln Institute of Land Policy Working Paper,
1995).
Vickrey, William, "Defining Land Value for Tax Purposes," in D.M. Holland,
ed., The Assessment of Land Value (Madison: University of Wisconsin Press, 1970),
pp. 25-36.
Wiencek, Henry, "Laying Out the Idyllic Life in a Latter-Day Arden," Smithsonian
Magazine. (May 1992), pp.124-142.
Wildasin, David E., "More on the Neutrality of Land Taxation," National
Tax Journal, Vol. 35 (March 1982), pp. 105-108.
Woolery, Arlo. "The Fairhope, Alabama, Land Tax Experiment," Land Value
Taxation: The Progress and Poverty Centenary, Richard W. Lindholm and Arthur D. Lynn,
Jr., eds., (University of Wisconsin Press, 1982), pp.197-204.
