Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
NORMATIVE CORPORATE INCOME TAX
DYNAMIC RENT, CORPORATE POLITICAL EXPENDITURE, AND NORMATIVE
CORPORATE INCOME TAX
Mihoko SHIMAMOTO
Abstract
Across developed countries, a "race to the bottom" has driven down corporate taxes
since the 1990s. Governments can improve their finances by collectively implementing
normative tax rates. The normative corporate tax rate is the sum of monopoly and
transfer rents, which firms take from other economic agents and which do not include
Schumpeter rents and the corporate tax rate. We calculate the U.S. rents of 234 S&P 500
companies using the Cobb-Douglas function and dynamic profit optimal conditions.
Examining the relationship between rent and lobbying using a Granger test, we find that
rents both cause and are caused by lobbying.
JEL codes: H25 D72 P16 H26 D63
I. INTRODUCTION
Production and industrial organization theory have developed various indicators to
evaluate corporate activities. For example, production efficiency and total factor
This work was supported by JSPS KAKENHI Grant Number16K00685. Faculty of Social Sciences, Hosei University, Aihara-cho, Machida-shi, Tokyo [email protected] Total Word Count is 8546 words.
1
NORMATIVE CORPORATE INCOME TAX
productivity measure the efficiency and the productivity of a company's production
activities, and the degree of monopoly and monopsony are indicators for measuring
their price formation in the markets. Corporate activities have been evaluated on a
practical level using these indicators.
I.A. Corporate Political Expenditure and Rent
However, should current studies of corporate activities focus only on their market
activity? In recent years, corporate political (i.e., nonmarket) activities have increased as
have studies of this behavior. Drutman (2015) pointed that the average lobbying
expenditure of the 1066 companies in the U.S. that have been listed in the S&P 500 has
more than doubled between 1981 and 2004. The passage of the Lobbying Disclosure
Act (LDA) in 1995 has made data on total lobbying spending by corporations more
readily available; this spending increased from $1.13 billion in 1998 to $2.09 billion in
2010. According to de Figueiredo and Richter (2014), the expenditures of corporations
and trade associations (i.e., business lobbying groups) comprise the vast majority of
total lobbying expenditures. Wrona and Sinzig (2018) refer to various literatures and
state that the more regulated industries (e.g., industries such as the steel, oil,
telecommunications, and pharmaceutical industries) tend to engage in political activity
more frequently. Political behavior by corporations is not limited to the United States.
2
NORMATIVE CORPORATE INCOME TAX
Gorostidi-Martinez and Zhao (2017) find that corporations’ attempts to shape
government policy favorably are commonly employed across many countries. For
example, firms in Sweden, Japan, and Germany formally participate in the public policy
process; and in the U.S., Canada, and Mexico, companies informally compete with a
variety of other interest groups to affect public policy.
In addition, it has been argued that political activity by corporations generates high
returns. For example, Alexander et al. (2009) examined the costs and returns of
lobbying in support of the American Jobs Creation Act. The law allowed U.S.
corporations with multinational operations a one-time opportunity to deduct 85 percent
of dividends received during a single year from a foreign subsidiary, such that only 15
percent of this repatriated income would be taxed. As a result, lobbying activity had a
22,000 percent rate of return. In another example, Etzioni (2018) noted that Whirlpool
spent $1.8 million over the course of two years on lobbyists and secured a renewal of an
energy tax credit for manufacturing high-efficiency appliances, which increased its
profits by $120 million, a return of approximately 6700% on the initial lobbying
expenses.
These relationships have also been tested using quantitative analysis. Kerr et al.
(2014) analyzed the relationship between lobbying and return using panel data from
3
NORMATIVE CORPORATE INCOME TAX
3,260 firms headquartered in the U.S. Among relatively large firms, the average firm
that engaged in lobbying sold roughly four times more than firms that did not lobby;
also, firms that did lobby were only slightly more likely to engage in research and
development than those that did not.
Most of this corporate unproductive rent is classified as transfer rent or monopolistic
rent in the public choice literature. These resources should have been distributed to
other economic agents but were diverted to the firm as a result of barriers to entry or
political actions, which reduces economic efficiency. Moreover, as lobbying expenditure
is regarded as unproductive itself, Bhagwati (1982) has labeled it Directly Unproductive
Profit-seeking (DUP) activity.
I.B. Taxation for Rents
For this reason, governments should prohibit these unproductive and unfair activities.
However, this is difficult to accomplish because corporate political expenditures profit
the politicians who have the authority to legislate political rules. Alternatively, they
could charge corporations for the unproductive rent. This would be also difficult
because the government would fear that companies would move overseas, reducing the
number of jobs and the corporate tax base. In fact, the average corporate tax rate in
OECD member countries has dropped considerably since the 1990s, indicating that a
4
NORMATIVE CORPORATE INCOME TAX
“race to the bottom” is occurring.
However, reducing tax rates has not necessarily been effective in encouraging
companies to keep their profits in their home countries. Cobham and Jansky (2017) and
Clausing (2016) provide evidence that profit-shifting has grown significantly even as
effective tax rates have fallen sharply.
In recent years, increasing tax evasion has eroded the financial base for public
services across countries. Alstadsæter et al. (2018) cite data collected by the Bank of
International Settlements that shows that wealth equivalent to about 10 percent of world
GDP is held in tax havens globally. While Scandinavian countries hold wealth
equivalent to only a few percent of GDP offshore, this figure rises to about 15% in
Continental Europe and to as much as 60% in Russia, Gulf countries, and a number of
Latin American countries. Crivelli et al. (2016) and Cobham and Jansky (2018)
estimated the global tax loss as a result to be between $650 billion and $5 trillion
annually.
In recent years, international organizations have finally started taking action. Zhu
(2016) reported that the “G20 + OECD” regime has taken the initiative since 2012 to
fight against tax evasion. The basic role of the G20 is to set the agenda and provide the
political consensus guiding and steering the whole governance process, while the OECD
5
NORMATIVE CORPORATE INCOME TAX
provides technical support and facilitates the implementation of the consensus.
Weinzierl (2018) argued that achieving substantial reform of the international tax
system would be made easier if policy-makers could point to an additional,
complementary, normative logic.
Although measures to prevent tax evasion have been studied internationally as above,
they have not stopped international competition for legal corporate tax cuts. The
solution for this issue is to calculate a normative corporate tax rate for each country,
especially for large corporations. All countries should then work together to impose
their prescriptive tax rates on corporations.
I.C. Normative Corporate Tax
Orthodox taxation theory in public economics has developed various arguments
against imposing corporate taxes. Stiglitz and Rosengard (2015: 713) define corporate
income tax as follows: “In a perfectly competitive economy, there would presumably be
no pure profits, so the tax is just a tax on the return to capital.” They also point out the
following essential problem: “Corporate earnings that are transferred to individuals in
the form of dividends are taxed twice – once in the form of the corporate tax and again
by way of the individual income tax” (729). Generally, neoclassical taxation theory
assumes an efficient state under perfect competition is assumed before taxation in order
6
NORMATIVE CORPORATE INCOME TAX
to identify the distortions caused by taxes.
However, with global companies taking over economic activity, it is necessary to
assume an oligopolistic market structure in the product market and the factor market in
the first place. If rents are created as a result, they should be attributed to other
economic entities and should be taxed and redistributed appropriately. It is often argued
that even in an oligopolistic market, competitive prices are formed in a contestable
market; however, this proposition should be verified empirically to determine whether
rent is being generated. In addition to the oligopoly in the product market and the factor
market, companies should be considered to have a monopsony for corporate taxes. In
principle, corporations conduct business activities utilizing human resources that have
been locally educated and trained, public safety resources, and various infrastructure in
the country where they are located. Corporate tax should be considered as payment for
public services, which is one of the production factors of a company. However, the
threat of capital flight may cause governments to receive too little in corporate taxes as
payments for public services. Based on these considerations, this paper defines the
normative corporate tax rate as the corporate tax rate that includes monopolistic and
transfer rents in the product, factor, and public service markets.
7
NORMATIVE CORPORATE INCOME TAX
I.D. The Issues Taken up in This Article
This paper will examine the following three issues. First, we construct a method to
divide the profits of companies into proper rewards obtained through market
competition and rents obtained through political activities and monopolistic means.
Second, we calculate rents for the past 33 years for 234 companies from the S&P 500
that have lobbying data in LDA reports1and estimate a normative corporate tax rate for
large US companies. Third, we analyze the causal relationships between rent and the
amount of lobbying reported in the LDA report and R&D investment.
II. DEVELOPMENT OF INDICATORS FOR MEASURING COMPANY
PERFORMANCE AND DATA LIMITATIONS
We assess corporate activities using the process discussed above. In order to develop
indicators that can be used widely in economic analysis and administrative procedures,
it must be easy to obtain data. We discuss this issue in this section.
II.A. Production Efficiency
First, we select a purely technical indicator of the efficiency of producers. A number
of prior studies address the measurement of this concept. Farrell (1957) indicated that
production efficiency could be measured as the distance between an observation and an
1The US Congress passed the Lobbying Disclosure Act in 1995, which requires registration and disclosure of certain lobbying activities. The LDA database is available on the website of US Senate: https://www.senate.gov/legislative/Public_Disclosure/LDA_reports.htm.
8
NORMATIVE CORPORATE INCOME TAX
estimated ideal referred to as an efficient frontier.2 Subsequent empirical studies of
productive efficiency have been divided into two streams: parametric and nonparametric
approaches. In parametric approaches, the production frontier is estimated using a
functional form such as a Cobb-Douglas function (see Aigner and Chu 1968). The use
of stochastic methods, such as in the survey of Bauer (1990) and panel data analysis by
Schmidt and Sickles (1984), has improved the accuracy of the estimations. The
nonparametric approach might be said be to have progressed more. Varian (1984)
consolidated the theoretical background by using revealed preference theory to show the
relationship between actual data and the production frontier behind it. Charnes, Cooper,
and Rhodes (1978) reduced the maximization of a ratio of weighted outputs to weighted
inputs (an index of a production efficiency) to a linear programming form. This
watershed discovery, which has come to be called Data Envelop Analysis (DEA), has
become the workhorse of efficiency studies today, and it is widely used for practical
analysis in the field of operational research as well as economics.
Data on the quantity of inputs and output are required to calculate this indicator.
However, in existing empirical analyses, only one industry is targeted, and the region
examined limits the quantity of data that can be obtained, as can be confirmed from the
2 Førsund, Lovell, and Schmidt (1980) provide an instructive review and outline of the initial research in this area.
9
NORMATIVE CORPORATE INCOME TAX
above empirical analyses and literature review. Although output and labor input data are
relatively available, data for other production factors are extremely difficult to obtain.
With regard to DEA, some analyses use accounting information on revenue and
expenses of as proxy variables for output and the input of production factors. However,
if the markets are not competitive, such an approach would not be theoretically
supported.
II.B. The Degree of Monopoly or Monopsony
The indicator that we imagine first as measuring the market performance of
corporations would be profit as defined in economics or accounting. However, markets
do not always function properly. Indicators that measure whether or not corporations are
operating under market competition are also necessary. The Lerner index (Lerner 1934)
captures the degree of monopoly in a supplier's market generally by dividing the
difference between the price of the product and its marginal cost by the price. In order to
obtain this value, calculation of the price elasticity of demand is necessary. A similar
index can be used to measure monopsony, and the calculation of the price elasticity of
supply is necessary to in order to obtain it. The empirical studies of Appelbaum (1982),
Schroeter (1988), and Azzam and Pagoulatos (1990) utilize standard models to obtain
these parameters by solving an equation system. Shimamoto (2018) reviews these
10
NORMATIVE CORPORATE INCOME TAX
studies in detail.
There are, however, limitations on the availability of data to calculate these
parameters. The estimation requires quantity and price data for both products and
production inputs; although most empirical studies employ cost or profit functions at the
firm level or even the industry level, it is difficult to obtain data for labor and energy
inputs, which are used in many different industries and production processes.
II.C. Rent
The indicators of corporate activity in orthodox economics have been related to
corporate production activities and market economic activities as described above. But
public choice has focused on the profits of corporate nonmarket behavior, that is,
political action. Rent is defined as additional income that is higher than that earned from
economic efficiency. The actions of people who lobby governments and bureaucrats to
achieve their advantageous business environment and secure rents are called rent-
seeking.
Tullock (1967) argued that both the dead weight loss caused by monopolies and tariffs
and the rents accruing to firms from these measures are social costs. Rents obtained
through monopoly price formation due to entry barriers are called monopoly rents, and
rents generated by tariffs, taxation, or subsidies set by government policies are called
11
NORMATIVE CORPORATE INCOME TAX
transfer rents. Krueger (1974) argued that rent-seeking activities are often competitive,
and rewards from rent-seeking activities can be expended on non-productive activities
for further rent acquisition.
However, since the 1990s, public choice theory seems to have shifted its focus to
game-theoretic clarification of political phenomena rather than exploring methods of
rent quantification. In order to analyze this relationship empirically, it is important to
measure rent with a standardized method that can be compared with other analyzes.
Nevertheless, as Shimamoto (2018) noted, rents have only been measured at the macro
and industrial levels using methods that are unsystematic, and a comprehensive and
versatile method for measuring rents has not been established. One of the objectives of
this paper is to establish a standard method of rent measurement using only value data
in financial statements.
III. CONCEPT OF DYNAMIC RENT
As noted above, the profits of a company can be divided into normal profits
generated by production efficiency and rents. Because monopoly rents or transfer rents
are surplus taken away from other economic entities and are social costs, according to
Tullock’s (1967) logic, they should be taxed and an original distribution should be
12
NORMATIVE CORPORATE INCOME TAX
restored.
Shimamoto (2018) explained in detail how to calculate dynamic monopoly rents for
each corporation using their accounting data and analyzed them for Japanese
companies. This is a method of calculating the time series of excess profits on the
assumption that companies are under monopolistic price formation. If, under the market
price given by the shadow price (i.e., the marginal profit at the optimal production
volume of the monopoly model = marginal cost at the same volume) in the monopoly
price formation model rather than the actual selling price (monopoly price), the
producer maximizes the dynamic profit, as in a competitive market, the actual time
series of input-output would be the result of long-term profit maximization in this model
in theory. We calculated dynamic monopoly rents as the difference between the actual
price and the shadow price multiplied by production amount.
In this paper, we consider this method to calculate both the monopoly rent and transfer
rent of a company, with special attention to companies’ bargaining power over public
services. For this purpose, we propose the following behavioral hypothesis. The
government provides various public services to companies and receives corporate
income tax from the companies in exchange. Multinational companies can move their
bases to foreign countries at any time, putting pressure on the government to reduce
13
NORMATIVE CORPORATE INCOME TAX
corporate taxes. This can be regarded as analogous to the markets for public services are
under monopsony. Governments are forced to discount the price of public services for
companies, so they have to reduce the amount of the other public services (such as
social security and education) provided to other actors in the country. Therefore, the part
of the corporate tax that the corporation has deducted can be considered as monopsony
rents in the markets for public services or as rents transferred from other members of
society. Any subsidy received by the company is included in the calculation of product
market rents because they are accounted for in revenue but are transfer rents in nature.
IV. MODEL
We apply the mechanism discussed in the previous section to the model constructed
in Shimamoto (2018) to calculate the company's rent.
IV.A. Short-term Equilibrium Conditions
We assume monopolistic and monopsonistic product markets. Producers maximize
short-term profits and also have a long-term equilibrium.
Assume a producer uses a general Cobb–Douglas production technology to produce
one product using four production factors. This function is denoted by
y=α5 v1α 1 v2
α 2 v3α3 K❑
α 4 (1)
where y is the output quantity, v1, v2 and v3 are the quantities of variable inputs, and K is
14
NORMATIVE CORPORATE INCOME TAX
the quantity of the fixed input, namely capital. In the short-term equilibrium, K is a
given value, and α 1+α2+α 3<1, which ensures that the marginal cost function is convex.
As the producer simultaneously faces both a monopolistic product market and
monopsonistic factor markets, the short-run profit maximization problem is given by
max πm=p ( y ) ∙ y−w1 ( v1 ) ∙ v1−w2 (v2 ) ∙ v2−w3 (v3 ) ∙ v3−rK
s.t. (1) (2)
The optimal conditions come from differentiating the Lagrange equation by the
variable factors v1 , v2 ,¿v3. These are given as
{ p' ( y ) ∙ y+ p ( y ) }∙ f v i
' ={wi' (v i ) ∙ v i+w i(v i)}, i =1, 2, 3. (3)
{ p' ( y ) ∙ y+ p ( y ) } can be expressed as (1+γ)∙p(y), where now γis assumed to have a
constant inverse demand elasticity and −1<γ ≤0. pm(y) can be defined by
pm(y)= { p' ( y ) ∙ y+ p ( y ) }=(1+γ)∙p(y) (4)
{wi' (v i ) ∙ v i+w i(v i)} can be expressed as (1+σ i¿ ∙wi, where now σ i is assumed to have a
constant inverse factor supply elasticity and σ i≥ 0. In the same way, wℑ(v i) can be
defined by
wℑ(v i) = {wi' (v i ) ∙ v i+w i(v i)}=(1+σ i)∙ wi(v i), i =1, 2, 3 (5)
The short-run optimization conditions in the monopoly and monopsony markets can be
expressed by arranging eqs. (1), (3), and (4) as follows.
y=α5 ∙( w1 m1−α2−α 3 ∙ w2m
α 2 ∙w 3mα 3
α 5 ∙ α 1❑1−α2−α 3 ∙α 2α2 ∙ α 3α 3∙ pm Kα 4 )
α 1α 1+α2+α3−1 ∙( w1m
α 1 ∙w2m1−α1−α 3 ∙w3 m
α3
α 5 ∙ α 1❑1 ∙ α 21−α 1−α3 ∙ α 3α 3∙ pm Kα 4 )
α 2α 1+α2+α3−1 ∙( w1 m
α 1 ∙ w2mα2 ∙ w3m
1−α 1−α 2
α 5 ∙ α 1❑1 ∙ α 2α 2 ∙ α 31−α1−α 2∙ pm Kα 4 )
α 3α 1+α 2+α3−1 ∙Kα 4
(6)
15
NORMATIVE CORPORATE INCOME TAX
vi=α 5∙( wℑ1−αj−αk ∙ w jm
αj ∙ wkmαk
α 5 ∙ αi❑1−αj−αk ∙ αjαj ∙ αkαk ∙ pm Kα 4 )
1α1+α 2+α 3−1, i, j, k=1,2,3 (i≠ j ≠ k ¿ (7)
It is important to note that these equations are not normal supply and factor demand
functions: pm(y) and wℑ ( v i ) are endogenous variables and differ from the exogenous
prices of a product and factors in the case of competitive markets. This interdependence
makes it difficult to find an optimal point and to formulate an empirical model.
To facilitate these calculations, we can utilize the relation between imperfect
competition models and perfect competition models. Now y t∗¿¿ indicates the short-term
optimum production level, and pt∗¿¿ is the equilibrium price in this imperfect
competition model, as described in Figure Ⅰ. A superscript t indicates a value in the t
period.
[Figure Ⅰ here]
Under the same production technology and the same given value of K t, pt∗¿(1+γ t)¿ is
defined as a given market price with constant pt∗¿¿and γt in a perfect competition. Then,
they bring the same production level y t∗¿¿as the short-term competitive equilibrium
value. γt is the degree of monopoly in this imperfect competition model at the optimal
point.
We can consider the production factor markets in the same manner. In this model, the
producer is a monopsonist in the factor markets. v1t∗¿ ,v2
t∗¿ ,v3t∗¿¿ ¿¿are the short-term optimal
factor quantities, and w1t∗¿ ,w2
t∗¿,w3t ∗¿¿¿¿ are the equilibrium prices in this monopsonic
equilibrium, as described in Figure Ⅱ. Under the same production technology and
the same given value of K t, w1t∗¿ (1+σ1
t ) , w2t∗¿ ( 1+σ2
t ) ,w3
t∗¿ (1+ σ3t ) ¿
¿ ¿ are defined as the market prices with
16
NORMATIVE CORPORATE INCOME TAX
constant w1t∗¿ ,w2
t∗¿,w3
t ∗¿ ,σ1t ,σ2
t ,∧σ3t ¿¿ ¿, which bring the short- term equilibrium factor quantities
v1t∗¿ ,v2
t∗¿ ,v3t∗¿¿ ¿¿in a perfect competition. Then, we can get mark-up rates of σ 1
t , σ2t , σ3
t , which
satisfy the condition where w1t∗¿ v1
t∗¿, w2
t∗¿ v 2t∗¿, w3
t∗¿ v3t∗¿¿¿
¿¿¿ ¿are the costs for each realized by profit
maximization in the imperfect competition in the t period.
[Figure Ⅱ here]
Therefore in consideration of the long-term equilibrium conditions, the supply
equation (6) and factor demand equation (7) can be regarded as the supply function and
factor demand function, respectively, under the given Kt, pt∗¿(1+γ t)∧w it∗¿ (1+σi
t ) ¿¿, which satisfy
the short-term equilibrium conditions in the perfect competitive model.
IV.B. Long-term Equilibrium Conditions
Long-term equilibrium conditions are derived by maximizing the time-series total of
the discounted present value of profits defined by the short-term pseudo-competitive
equilibrium model minus capital costs. By deriving these long-term pseudo-competitive
profit maximization conditions, we can find the optimal γt , σ1t , σ2
t ,∧σ3t that make the
time series data of Kt, pt∗¿ y t∗¿ ,w1
t∗¿ v1
t∗¿ ,w2
t∗¿ v2t∗¿ ,¿w3
t∗¿ v 3t∗¿ ¿¿
¿¿¿¿¿ ¿ for each of the four periods the optimal dynamic
solution. These past data can be regarded as results that have satisfied both the short-
term and long-term equilibrium conditions. Then we can determine the rent ratios,
γt∧σ it, which are the proportions of rent in the product price and factor prices,
respectively.
Now let us formulate the long-term pseudo-competitive profit maximization
conditions. It is maximized for discrete time periods from 1 to T. The pseudo-
17
NORMATIVE CORPORATE INCOME TAX
competitive profit function in period t is defined as follows.
π t=p t∗¿ ¿¿)∙ y t ¿
−w1t∗¿ (1+σ1
t ) ∙ v1t ¿ ¿
−w2t∗¿ (1+σ2
t ) ∙ v2t ¿ ¿
−w3t∗¿ (1+σ3
t ) ∙ v3t ¿ ¿
−Qt ∙ I t (K t−1 , K t , δ t)❑
(8)
wherey t ( ∙ )∧v it ( ∙ )are defined by Equations 6 and 7, and γt , σ 1
t , σ2t ,¿σ3
t are assumed to
change over time. Investment in period t, I t ( K t−1 , K t , δt ) is defined as follows.
I t(K t−1 , K t , δ t)❑= K t−(1−δ t)∙ K t−1 (9)
where δ t is the depreciation rate in period t, and Qt is the exogenous unit price of
investment in period t.
The long-term equilibrium condition arises from maximizing the sum of the
discounted present value of π t from period 1 to T based onK t, as follows.
maxK t
Π=π1+∑t=2
T
∏s=2
t 1(1+rs )
π t (10)
Thus, the necessary condition for optimization, is given as follows,
∂ Π∂ K t =∏
s=2
t 1(1+r s )
∙¿¿
−w3
t∗¿ (1+σ3t ) ∙ ∂v3
t
∂ K t −Qt ∙∂ I t
∂ K t ¿+∏
s=2
t+1 1(1+rs )
[−Qt+1∙ ∂ I t+1
∂ K t ]=0 (11)
By modifying the partial differentiation using logarithmic differentiation,3 this condition
3 See Shimamoto (2018) for the detailed derivation process. Generally, the following equation holds for partial differentiation.
18
NORMATIVE CORPORATE INCOME TAX
can be finally arranged into the following simple equation,
(1+γt )∙ p t yt
K t −(1+σ1t ) ∙ w1
t v1t
K t −(1+σ2t ) ∙ w2
t v2t
K t −(1+σ3t ) ∙ w3
t v3t
K t
−α 1+α2+α 3−1−α 4
∙[Qt− 1(1+r t+1 )
∙ Qt+ 1∙ (1−δ t+1 )]=0
(12)4
V. APPLICATION TO RENT CALCULATION FROM CORPORATE ACCOUNTING
DATA
As mentioned in Section 3, the profit of a company can be separated into the profit
gained by market competition and rent. In following sections, we will calculate the rent
(monopolistic rent + transfer rent) by applying the model in the previous section using
corporate financial data.
Each variable of the model in Section 4 is operationalized using the following financial
data. The output value (py) is the total sales, and the four production factors
∂ B∂ A
= ∂ B
∂ ln B ∙ ∂ ln B
∂ ln A ∙ ∂ ln A
∂ A
Referring to the short-term optimizing conditions, eqs. (6) and (7):∂ ln y∂ ln K
=−α1 α 4
α1+α 2+α3−1+
−α 2α 4
α 1+α2+α 3−1+
−α 3 α 4
α1+α 2+α3−1+α 4=
−α 4
α1+α2+α 3−1
∂ ln v i
∂ ln K=
−α4
α1+α 2+α3−1 , i = 1, 2, 3
Using ∂ ln x /∂ x=1/ x, ∂ y∂ K =
−α 4
α1+α2+α 3−1 ∙ y
K
∂ v i
∂ K =
−α 4
α1+α2+α 3−1 ∙
v i
K, i = 1, 2, 3
4 In this empirical study, the data are collected from the financial statements of each company. The depreciation amount for a year is included as an item among the costs. Therefore, we setδ t+1 as zero in order to avoid double counting.
19
NORMATIVE CORPORATE INCOME TAX
w1t v1
t ,w2t v2
t , w3t v3
t , Kare operating expenses, non-operating expenses plus extraordinary
losses, corporate taxes, and total assets.
Among the four parameters γt , σ1t , σ2
t ,∧σ3t of the model, one should be removed when
R(= α1+α2+α 3−1
−α 4) is regarded as fourth variable. At this time, σ 2 will be set to 0
because of the nature of the production factor; γ represents the degree of monopoly in
the product market; and σ 1 represents the degree of monopsony for production factors
(labor, raw materials, etc.). As described in Section 3, σ 3 represents the degree of
monopsony for public services, assuming that taxes are discounted due to monopsony in
markets for public services.
When we regard the variable R relating to scale as the fourth parameter,γ , σ1 , σ3 and
R can be obtained by quadratic programming using data for at least four years, as
calculated in Shimamoto (2018). However, when the parameters are determined in this
way, the time series values of R fluctuate drastically from year to year, which makes the
sequential values of rent unstable. Therefore, we define a scale variable, S, as follows:
S=α1 +α 2+α3+α 4 (13)
The calculation is performed by setting S to 1 (i.e., R=1) for the time being. Sensitivity
analysis will be performed with S = 1.2, which is the upper bound of induced returns to
scale in the U.S. in recent years according to Boussemart et al. (2018).
VI. CAPITAL REWARD AND RENT DISTRIBUTION
In the case of constant returns to scale (i.e., S=1), total rent is regarded as −γt pt yt+
σ 1t w1
t v1t +σ2
t w2t v2
t +σ3t w3
t v3t . However, if the yield shows increasing returns to scale, it is
20
NORMATIVE CORPORATE INCOME TAX
necessary to devise profit and rent distribution. In order to explain this point, we must
define the distribution of capital reward and rent specifically.
Transforming Equation 12,
p t yt
K t −w1
t v1t
K t −w2
t v2t
K t −w3
t v3t
K t
¿ R ∙[Qt− 1(1+r t+1 )
∙Qt+1∙ (1−δt+1 )]+−γt pt y t
K t +σ1
t w1t v1
t
K t +σ3
t w3t v3
t
K t (14)
In the accounting data, the capital depletion δ is recorded as an item in w 1t v1
t as
depreciation expense, so δ = 0. Also, regardingI t as the investment amount, both Qt and
Qt+1 are equal to 1. Taking these things into account, we multiply both sides of Equation
14 by K.
pt y t−w1t v1
t −w2t v2
t −w3t v3
t
¿ R ∙ rt+1
1+r t+1 ∙ K+¿+σ 1t w1
t v1t +σ2
t w2t v2
t +σ3t w3
t v3t ¿ (15)
Since rt+1 is the interest rate, it can be generally regarded as (discounted present value
of) the marginal efficiency of capital. When R = 1, R ∙ r t+1
1+rt+1 ∙ K means a competitive
and normative capital reward. Thus, this equation can be interpreted as,
Net Income = Capital Reward + Rent (16)
In other words, in the case of constant returns to scale, rewards above the marginal
efficiency of capital are rents.
What about the case of increasing returns on scale? When S>1 ,R<1 and R
converges to 0 as S increases. Regarding R ∙ r t+1
1+rt+1 ∙ K as capital reward according to
Equation 15 in this case, as the scale harvest increases, the proportion of capital rewards
21
NORMATIVE CORPORATE INCOME TAX
among net income decreases and the proportion of rents increases.
[Figure Ⅲ here]
However as shown in Figure Ⅲ, as the yield increases, MC decreases (from MC1 to
MC2), and the ratio of rent to revenue or profit increases, covering the grid area rather
than just the grey area. Therefore, in the case of increasing returns to scale, the
distribution of net income needs to be revised. It may be appropriate to consider the
capital reward under the constant return to scale multiplied by the scale factor S as the
capital share. In this case, the capital reward and the rent distribution can be expressed
as the first and second terms on the right side of the following equation.
Net Income =¿ S ∙ r t+1
1+rt+1 ∙ K
+{(−γ t p t y t+σ 1t w1
t v1t +σ2
t w2t v2
t +σ3t w3
t v3t )++R ∙ rt+1
1+rt+1 ∙ K−S ∙ rt+1
1+rt+1 ∙ K } (17)
In this case, all of the scale effects are attributed to the capital, so this distribution
method can be regarded as the upper limit of the capital reward.
Some people may be dissatisfied with the absence of the Schumpeter rent gained by
operations that exceed those of other companies, including R&D investments. They
would suggest counting Schumpeter rent separately. In this study, as described in the
next chapter, the rent could not be divided into γ and σ i segments. But, if it could be
split, the Schumpeter rent would be considered part of γ segment.
VII. DATA AND CALCULATION
The system of equations was solved using MATLAB software.5 Using matrix
5 MATLAB ver. R2016a. These equations systems were solved using the “lsqlin” command in MATLAB, which is a solution for quadratic programming (i.e., constrained linear least-squares problems).
22
NORMATIVE CORPORATE INCOME TAX
expressions, simultaneous equations that solve for γ, σ 1, ¿σ 3 in Equation 12 can be
specified and then estimated with three years of financial data as follows:
[ R Qt+V 2t
RQ t+1+V 2t+1
RQ t+2+V 1t+2]=[ Y t −V 1
t −V 3t
Y t+1 −V 1t+1 −V 3
t+1
Y t+2 −V 1t+2 −V 3
t+2] [ (1+γt )(1+σ1
t )(1+σ3
t ) ] (18)
where Qt ≡Qt− 1
1+rt+1 ∙ Qt+ 1, Y t ≡ pt yt
K t , V it ≡
wit v i
t
K t . γt, and σ it (i=1,3) are the mark-up
rates for output and inputs in periods t to t+2. If the coefficient matrix has an inverse
matrix, equation 19 holds.
[ ( 1+γ t )(1+σ 1
t )(1+σ 3
t ) ]=[ R Qt+ V 2t
RQ t+1+V 2t+1
RQ t+2+V 1t+2][ Y t −V 1
t −V 3t
Y t+1 −V 1t+1 −V 3
t+1
Y t+2 −V 1t+2 −V 3
t+2]−1
(19)
By calculation of these equations, we obtain γ, σ 1, and σ 3. When the actual
calculations were made, we found a considerable difference between the value on the
left side and the value on the right side of Equation 18. Therefore, the parameters were
modified by an algorithm for absorbing the difference between the values on both sides
in the order of γ , σ1 ,¿ σ3. Therefore, the value of the resulting total rent (the second term
of Equation 15) is appropriate, but the respective values of the parameters γ , σ1 ,¿ σ3 are
considerably distorted.
Using financial data covering 36 years, it is possible to set t from the first year to the
34th year. For each t, the solutions ofγ, σ 1 and σ 3 are given; however, the equation for
period t contains the discount rate of the t+1 period, rt+1, and so the maximum length of
the time series of solutions is 33.
We collected the following data. The accounting data of 234 corporations registered
23
NORMATIVE CORPORATE INCOME TAX
in the US S&P 500 in 2018 were downloaded from Mergent online; these were the
corporations for which there was lobbying data. The lobbying data was taken from LDA
Reports available on the US Senate website. These companies were classified into 28
industries according to their Standard Industrial Classification (SIC) codes.6
Data on discount rates are U.S. interest rates for each year, as found in the
International Financial Statistics of IMF. The unit price of investment (Q) was set as one
in every year.
VIII. RESULTS
VIII.A. Characteristics of Rent in Industrial Sectors and in Time Series
First, we compare industries based on the ratio of rents to corporate net income. Table I
shows that the time series average rent rate for each industry in cases of scale variable S
was set at 1 and 1.2. Comparing the cases, the rent rate is slightly lower when assuming
economies of scale (i.e., S=1.2). The relative rankings among industries show that the
top industries are almost the same, although their ranks changed slightly.7
The industries with the highest rent rates are Business Services, Measuring Analyzing
and Controlling Instruments, Chemicals, Health Services Products, and Petroleum
Refining.
This is almost consistent with the results of previous studies introduced at the
6 While the numbering followed SIC codes wherever possible, several sectors that had few firms were consolidated. For this reason, there are numbers up to 33, but there are only 28 industries.7 For the results in Tables I and , Equation 18 was solved for the case of S = 1, and the rent Ⅱcalculated from the value of the parameter after error correction was used. However, in the case of S = 1.2, it is necessary to adjust the capital reward by the scale factor, so the rent is the value obtained by subtracting the capital reward from net profit. As a result, it seems that there is a slight difference in the ranking between industries when S = 1 and when S = 1.2.
24
NORMATIVE CORPORATE INCOME TAX
beginning. Since the size of the rent for one company is not known from the rent rate,
the average rent amount for a single corporation in the case of S = 1 is also shown in
Table I. The largest rent amounts are for Petroleum Refining, Communications, Mining,
and Chemicals sectors.
[Table I here]
How would the corporate tax rate change if the rent calculated in this study were
added to the corporate tax? Table shows a comparison between the ratio of corporateⅡ
income tax to profit before tax and the ratio of corporate income tax plus rent to profit
before tax. From this result, even considering scale economy (i.e., S=1.2), it can be seen
that the corporate tax rate including rent is around 20% higher than the current corporate
tax rate.
[Table here]Ⅱ
Figure shows the time series of rent and corporate income taxes derived from theⅣ
accounting data. We find in the last 30 years that the corporate tax rate has fallen
gradually while the rent rate has risen monotonically. A rough analysis of this trend
shows that the rent rate rose at roughly the same pace that the corporate tax rate fell
until around 2000. Since then, the rent rate has increased more rapidly than the tax rate
has declined.
[Figure here]Ⅳ
25
NORMATIVE CORPORATE INCOME TAX
In the next step, we estimate the normative tax rate. While one possibility would be to
consider the normative tax rate to be the sum of the tax rate and the rent rate, this
overstates the optimal rate as Schumpeter rents would not be deducted. Schumpeter rent
should be considered as a part of the γ segment of rent. However, as explained in
Section XII, the rent could not be extracted in this analysis. We attempt to estimate the
normative tax rate under the assumption that Schumpeter rent results from R&D
expenditures. As shown in Figure , the ratio of R&D expenditure to total costs,Ⅴ
including taxes, increased from 2% to 6.8% between 1982 and 2014. Assuming that the
entire rent rate in 1982, that is, 3.0% of profit before tax in S=1.2, resulted from R&D
expenditures, we could set the 2014 Schumpeter rent rate at 3.4 times the rent rate, or
10.2%. In 2013, the Schumpeter rent rate was 3.3 times the rent rate, or 9.9%. The sum
of tax rate and rent rate is 56.7% in 2013 and 77.4% in 2014. Therefore, the normative
tax rate would be 46.8% in 2013 and 67.2% in 2014. Even if Schumpeter rent were
deducted in a similar manner from annual rents, the normative tax rate in recent years
would be at least around 40%.
[Figure here]Ⅴ
VIII.B. Relationship of Rent with Political Action by Panel Data Analysis
26
NORMATIVE CORPORATE INCOME TAX
Do the political actions of corporations generate high returns? To assess this, panel
data analysis was performed using rent as the explained variable and lobbying amount
as the explanatory variable. Since rent should be heavily influenced by the size of the
company, we added annual sales to the explanatory variables.
First, we estimated a regression without lagging lobbying, and we used a Hausman test
to determine whether fixed or random effects models were appropriate. Then, the
coefficients of determination were compared for models of the lobbying data with
various time lags. When moving the lag of the lobbying data forward and backward,
peak values of the coefficient of determination appeared in each direction. The models
in Table use the lag structures that generated the two peak values. Based on theⅢ
results, when S = 1, a $1 increase in lobbying expenditure would increase the rent by
about $52; and at S = 1.2, it would increase the rent by about $43. However, we note
that this result is based on a correlation and not a causal relationship between lobbying
and rent, an issue we address below.
[Table here]Ⅲ
Another concern is that each of these variables is likely to have an autoregressive
process. Therefore, we also tried Population-Averaged estimation (PA) for setting the
autoregressive process for the error term of each panel. However, it is difficult to find
27
NORMATIVE CORPORATE INCOME TAX
significant results, and only the successful case is included in Table .Ⅲ
A more difficult econometric problem occurs when both of the independent variable
and the dependent variable are unit root processes. If this is the case, observed
relationships between the variables may be spurious. Therefore, a unit root test was
performed on rent and lobbying data. In recent years, unit root tests for panel datasets
has been developed; we used a method developed by Harris and Tzavalis (1999) (HT).8
Since missing values are not allowed, the tests were conducted using a balanced dataset
of 65 companies during the years from 1999 to 2014. For both variables, the null
hypothesis was rejected in the test. However, since the null hypothesis states that all
panels contain unit root in HT, this result still allows the possibility that some of the
panels have unit root processes.
Next, using this balanced data, the causal relationship between the rent and the
lobbying expenditure described above was tested. Although Granger causality testing
(Granger 1969) has long been used for single time series, methods for panel data have
developed more recently. We performed a panel Granger test developed by Dumitrescu
and Hurlin (2012).9 The lag period was set to one and we used a bootstrap procedure. As
a result, both of the null hypotheses that rent was not the cause of lobbying expenditure
8 The xtunitroot package in Stata version 15 was used for the test.9 It was calculated using a library called xtgcause of stata ver15. See Lopez and Weber (2017) for a specific explanation.
28
NORMATIVE CORPORATE INCOME TAX
and that lobbying expenditure was not the cause of rent were rejected. That is, it was
concluded that rent was the cause of lobbying expenditure for at least one panel and that
lobbying expenditure was the cause of rent for at least one panel.
From the above results, it can be said that while some companies have obtained high
rents by lobbying, other companies have invested in lobbying because they have
obtained high rents in the past.
VIII.C. Which Generates Rents: R&D or Lobbying?
It may be argued that rent is the outcome of investment in corporate R&D rather than
the outcome of looting from other economic agents. We tried to analyze this issue with
panel data for 73 companies whose accounting data include R&D expenditures. We
considered adding R&D expenditures to the explanatory variables, but unfortunately the
correlation coefficient between R&D and lobbying was 0.7214, creating a
multicollinearity problem. We performed separate panel data analyses using lobbying
expenditure and sales and then R&D expenditure and sales as explanatory variables.
Table shows the results. The coefficients of determination for the two models areⅣ
very similar, and both models have almost the same explanatory power. However, in
this context it is not possible to separate the relationships between these two variables
29
NORMATIVE CORPORATE INCOME TAX
and rent.
[Table here]Ⅳ
IX. CONCLUSIONS AND REMARKS
As mentioned in Section I.D., this article makes three contributions. The conclusions
and the remarks for each are as follows.
The first contribution was the construction of a method to separate corporate profits
as appropriate compensation from market competition from the rents derived from
political expenditure, monopoly, and monopsony. Applying the method of Shimamoto
(2018), we solved the pseudo-competitive dynamic profit maximization model and
found the optimal conditions when the markets of products, production factors, and
public services were all monopolistic. The results showed that profits (net income after
tax in accounting) can be divided into capital compensation and rents (Equation 15). In
the case of increasing rate of return, adjustments were made with the scale factor.
The second contribution was the estimation of the normative corporate tax rate for large
US companies using the S&P 500 accounting data. Orthodox taxation theory has
basically assumed competitive markets and considered corporate tax as a tax on capital
rewards. Therefore, sometimes it has been argued that corporate tax is a double taxation.
However, this paper considered corporate tax to be a payment for public services and
30
NORMATIVE CORPORATE INCOME TAX
calculated the corporate tax rate including the monopsony rent associated with public
service. Figure shows the time series of this rent.Ⅳ However, Schumpeter rent was
deducted assuming that all such rents in 1982 came from R&D and taking into account
the recent increase in R&D expenditure as a percentage of total costs. As a result, the
normative tax rate in recent years has been at least around 40%.
If accurate estimates ofγ, σ 1 and σ 3 could be made, a clearer conclusion on this point
would have been obtained. This is because Schumpeter rent occurs in the product
market; it is, a part of γ. However, after calculating γ, σ 1 and σ 3 by the quadratic
programming method, considerable errors were identified by checking Equation 17.
Recalculation was performed using an algorithm that absorbs the error sequentially into
γ, σ 1 and σ 3. Therefore, we were able to estimate total rent, but an effective result could
not be shown this time for each value of γ, σ 1 and σ 3. Finding or developing software to
calculate more accurate parameters is a task for future research.
If γ, σ 1 and σ 3 can be estimated more accurately, it will be possible to calculate the
normative corporate tax rate for each country from corporations’ accounting data. By
building international cooperation based on these results, each country will be able to
achieve an appropriate corporate tax rate. Also, if the OECD+G20 regime takes more
actions on tax havens, it will be possible for each government to achieve greater
31
NORMATIVE CORPORATE INCOME TAX
financial soundness.
The third contribution was the identification of a causal relationship between rent and
lobbying, which was assessed with a panel Granger test on rent and lobbying data
obtained from LDA reports. It was concluded that rent was the cause of lobbying
expenditure for at least one panel and that lobbying was the cause of rent for at least one
panel.
By including panel data for R&D expenditure, we tried to test the proposition that rent
was due to R&D rather than corporate political spending. However, the correlation
between lobbying expenditure and R&D expenditure was high and no clear results were
obtained.
References
Journal. Aigner, D.J., Chu, S.-F. (1968) On estimating the industry production function.
American Economic Review 58: 826.-839.
Journal. Alexander, R., Mazza, S., Scholz, S. (2009) Measuring Rates of Return on
Lobbying Expenditures: An Empirical Case Study of Tax Breaks for Multinational
Corporations. The Journal of Law & Politics 25: 401-457.
Journal. Alstadsæter, A., Johannesen, N., Zucman, G. (2018) Who Owns the Wealth in
32
NORMATIVE CORPORATE INCOME TAX
Tax Havens? Macro Evidence and Implications for Global Inequality. Journal of
Public Economics 162: 89-100.
Journal. Appelbaum, E. (1982) The Estimation of the Degree of Oligopoly Power.
Journal of Econometrics 19: 287–299.
Journal. Azzam, A., and Pagoulatos, E. (1990) Testing Oligopolistic and Oligopsonistic
Behavior: An Application to the United States Meat-Packing Industry. Journal of
Agricultural Economics 41:362–370.
Journal. Bhagwati, J.N. (1982) Directly Unproductive, Profit-seeking (DUP) Activities,
Journal of Political Economy 90: 988-1002.
Journal. Bauer, P.W. (1990) Recent Developments in the Econometric Estimation of
Frontiers. Journal of Econometrics 46: 39-56.
Journal. Charnes, A., Cooper, W.W., Rhodes, E. (1978) Measuring the Efficiency of
Decision Making Units. European Journal of Operational Research 2: 429-444.
Journal. Clausing, K.A. (2016) The Effect of Profit Shifting on the Corporate Tax Base
in the United States and Beyond. National Tax Journal 69: 905-934.
Journal. Cobham, A., Jansky, P. (2018) Global Distribution of Revenue Loss from
Corporate Tax Avoidance: Re-estimation and Country Results. Journal of
International Development 30: 206-232.
Journal. Cobham, A., Jansky, P. (2019) Measuring misalignment: The location of US
multinationals’ economic activity versus the location of their profits. Development
Policy Review 37: 91-110.
33
NORMATIVE CORPORATE INCOME TAX
Journal. Crivelli, E., de Mooij, R., Keen, M. (2016) Base Erosion, Profit Shifting and
Developing Countries. FinanzArchiv 72: 68-301
Journal. De Figueiredo, J.M., and Richter, B.K. (2014) Advancing the Empirical
Research on Lobbying. Annual Reviews of Political Science 17: 163-185.
Book. Drutman, L. (2015) The Business of America is Lobbying: How Corporations
Became Politicized and Politics Became More Corporate. Oxford: Oxford
University Press.
Journal. Dumitrescu, E.-I., Hurlin, C. (2012) Testing for Granger Non-Causality in
Heterogeneous Panels. Economic Modelling 29: 1450-1460.
Journal. Etzioni, A. (2018) Theoretical Implications of ‘High Rent.’ Sociological
Forum 33: 529-538.
Journal. Farrell, M.J. (1957) The Measurement of Productive Efficiency. Journal of the
Royal Statistical Society 120: 253-281.
Journal. Førsund F.R., Lovell, C.A.K. (1980) Schmidt, P. A Survey of Frontier
Production functions and their Relationship to efficiency Measurement. Journal of
Econometrics 13: 5-25.
Journal. Gorostidi-Martinez, H., Zhao, X. (2017) Corporate Political Strategies: A
Contemporary Literature Review. Journal of Advances in Management Research
14: 375-404.
Journal. Granger, C. (1969) Causal Relations by Econometric Models and Cross-
Spectral Methods. Econometrica 37: 424-438.
34
NORMATIVE CORPORATE INCOME TAX
Journal. Harris, R.D., Tzavalis, E. (1999) Inference for Unit Root in Dynamic Panels
Where the Time Dimension is Fixed. Journal of Econometrics 91: 201-226.
Journal. Kerr, W.R., Lincoln, W., Mishra, P. (2014) The Dynamics of Firm Lobbying.
American Economic Journal: Economic Policy 6: 343-379.
Journal. Krueger, A.O. (1974) The Political Economy of the Rent Seeking Society.
American Economic Review 64: 291-303.
Journal. Lerner, A.P. (1934) The Concept of Monopoly and the Measurement of
Monopoly Power. The Review of Economic Studies 1: 157-175.
Theses, reports, and other unpublished materials. Lopez, L., Weber, S. (2017) Testing
for Granger Causality in Panel Data, University of Neuchatel Institute of Economic
Research IRENE Working Paper 17-03, Neuchatel, Switzerland.
Journal. Schmidt, P., Sickles, R. (1984) Production frontiers and panel data. Journal of
Business and Economic Statistics 2: 367-374.
Journal. Schroeter, J.R. (1988) Estimating the Degree of Market Power in the Beef
Packing Industry. Review of Economics and Statistics 70: 158–162.
Journal. Shimamoto, M. (2018) Examining the Relationship between Rent and Political
Expenditure: Using Rent Information Obtained from Financial Statements. Journal
of Economics and Political Economy 5: 194-208.
Book. Stiglitz, J.E., Rosengard, J.K. (2015) Economics of the Public Sector, 4th ed. New
35
NORMATIVE CORPORATE INCOME TAX
York: W. W. Norton &Company, Inc.
Journal. Tullock, G. (1967) The Welfare Costs of Tariffs, Monopolies, and Theft.
Western Economic Journal 5: 224-232.
Journal. Varian, H.R. (1984) The Nonparametric Approach to Production Analysis.
Econometrica 52: 579-597.
Journal. Weinzierl, M. (2018) Review of Global Tax Fairness, Thomas Pogge and
Krishen Mehta, Editors. Journal of Economic Literature 56: 673-684.
Journal. Wrona, T., Sinzig, C. (2018) Nonmarket Strategy Research: Systematic
Literature Review and Future Directions. Journal of Business Economics 88: 253-
317.
Journal. Zhu, J. (2016) G20 Institutional Transition and Global Tax Governance. The
Pacific Review 29: 465-471.
36
NORMATIVE CORPORATE INCOME TAX
37
Figure Ⅰ Imperfect Competition and Perfect Competition in the Products Market
NORMATIVE CORPORATE INCOME TAX
38
Figure Ⅱ Imperfect Competition and Perfect Competition in the Factor Market
Marginal Factor Cost
w it (v i
t ) vit +w i
t(v it)
Marginal Revenue Products
pt∗¿(1+γ t)f vi' ¿
0
Factor Supply Curve
w it(v i
t)
vivit*
wit*
wit*(1+σit)
wi
NORMATIVE CORPORATE INCOME TAX
Table . Average Rate of Rent Occupying NetIncome in Each Industry from 1982-2014Ⅰ
S=1 S=1.2
39
pMC1
MC2
q0
D
MR
Figure Ⅲ Increasing Returns to Scale and Capital Reward
rent2
rent1
NORMATIVE CORPORATE INCOME TAX
Industry SamplesRent
RateRanking rent (US$)
Rent
RateRanking
2 Mining 15 0.261 1,144,003,230 0.210
3 Construction 4 0.232 112,732,040 0.172
4 Food, Tabacco 6 0.296 480,769,673 0.246
5 Textile and Apparel 4 0.303 112,477,315 0.235
6 Wood,Furniture and Paper 5 0.242 195,833,288 0.186
7 Printing and Publishing 1 0.216 140,561,234 0.159
8 Chemicals and Allied Products 17 0.411 4 1,087,798,371 0.363 3
9Petroleum Refining and Related
Industries2 0.394 6 1,623,837,565 0.362 4
1
0
Rubber and Miscellaneous Plastics
Products2 0.369 10 324,167,006 0.298
1
2Stone, Clay Glass and Concrete Products 1 0.247 571,116,385 0.208
1
3Primary Metal Industries 1 0.284 246,663,068 0.231
1
4
Fabricated Metal Products, Except
Machinery and Transportation
Equipment
2 0.247 133,539,785 0.184
1
5
Industrial and Commercial Machinery
and Computer Equipment8 0.396 5 305,172,979 0.343 6
1
6
Electronic and Other Electrical
Equipment and Components, Except
Computer Equipment
10 0.313 359,995,846 0.264
1
7Transportation Equipment 6 0.252 536,212,084 0.204
1
8
Measuring Analyzing and Controlling
Instruments; Photographic, Medical and
Optical Goods;
15 0.423 2 318,938,582 0.363 2
1
9Miscellaneous Manufacturing Industries 1 0.394 7 192,234,039 0.325 9
2
0Transportation and Postal Service 8 0.178 478,307,429 0.140
2 Electric, Gas and Sanitary Services 22 0.089 104,824,650 0.051
40
NORMATIVE CORPORATE INCOME TAX
1
2
2Communications 4 0.286 1,146,942,982 0.239
2
3Wholesales Trade 6 0.385 9 237,060,851 0.327 8
2
4Retail Trade 18 0.392 8 536,566,790 0.329 7
2
5
Finance, Security and Commodity
Brokers, and Holding and Other
Investment Offices
22 0.077 51,612,337 0.038
2
6Insurance 17 0.048 110,421,389 0.032
2
8Real Estate 6 0.299 145,190,501 0.231
2
9Health Services 5 0.421 3 394,481,552 0.358 5
3
0Business Service 25 0.470 1 636,786,584 0.415 1
3
3Public Administration 1 0.365 66,261,437 0.309 10
Note: 1) The average for each industry is the average of the average from 1982 to 2014 for each
sample.
2) Rent Rate = Rent / Net Income
41
NORMATIVE CORPORATE INCOME TAX
Table Corporate Income Tax Including RentⅡ
S=1 S=1.2
Industry W3/TI(a) (W3+rent)/TI (W3+rent)/TI(b) (b)-(a)
2 0.382 0.571 0.534 0.152
3 0.294 0.502 0.459 0.166
4 0.321 0.545 0.505 0.184
5 0.282 0.508 0.460 0.178
6 0.356 0.599 0.555 0.199
7 0.067 0.402 0.274 0.207
8 0.189 0.520 0.484 0.295
9 0.391 0.669 0.647 0.256
10 0.419 0.670 0.624 0.204
12 0.261 0.580 0.542 0.281
13 0.331 0.521 0.485 0.153
14 0.313 0.502 0.457 0.144
15 0.182 0.520 0.476 0.294
16 0.230 0.446 0.417 0.188
17 0.301 0.491 0.456 0.155
18 0.297 0.638 0.592 0.295
19 0.267 0.564 0.516 0.249
20 0.308 0.424 0.399 0.092
21 0.352 0.419 0.390 0.038
22 0.415 0.625 0.593 0.178
23 0.304 0.480 0.462 0.158
24 0.359 0.610 0.570 0.211
25 0.272 0.292 0.300 0.028
26 0.265 0.310 0.297 0.033
28 0.186 0.439 0.384 0.197
29 0.386 0.655 0.616 0.230
30 0.289 0.650 0.610 0.321
33 0.278 0.509 0.473 0.196
Average 0.296 0.524 0.485 0.189
42
NORMATIVE CORPORATE INCOME TAX
Note: W3=Income Tax TI=Profit before Tax
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
S=1, Rent Rate
S=1.2,Rent Rate
Income Tax / Profit before Tax
S=1, (Income Tax+Rent) / Profit be-
fore TaxS=1.2, (Income Tax+Rent) / Profit be-
fore Tax
%
43
Figure Ⅳ Time Series of Rent and Income Tax
NORMATIVE CORPORATE INCOME TAX
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
012345678
%
44
Figure Ⅴ R&D Expenditure Occupying Total Costs
NORMATIVE CORPORATE INCOME TAX
Table Rent and lobbying of corporationsⅢn=234 Independent variables
Scale
parameterType
lag or
lead
operator
lobby py R2 method
S=1 l1 b) 52.21511 0.043241 0.4644 re
(0.002) a) (0.000)
f1 34.60893 0.043641 0.4692 re
(0.029) (0.000)
log-
logl2 0.08999 0.965632 0.4595 re
(0.002) (0.000)
f3 0.060189 1.048933 0.4865 re
(0.029) (0.000)
f3 0.122773 0.928
(0.005) (0.000) pa ar(4)
(d
S=1.2 l1 42.79533 0.041408 0.4445 re
(0.009) (0.000)
f3 29.26516 0.042817 0.4451 re
(0.049) (0.000)
log-
logl2 0.089021 0.940016 0.4766 re
(0.003) (0.000)
f6 0.128032 1.014605 0.5195 re
(0.000) (0.000)
45
NORMATIVE CORPORATE INCOME TAX
Note a) () is p value
Note b) fn is lead operator of n periods, ln is lag operator of n periods.
Note c) fe: fixed-effects estimator, re: random-effects estimator, pa:
population-averaged estimator
Note d) ar(4) is a fourth order autoregressive process.
46
NORMATIVE CORPORATE INCOME TAX
Table Comparison of lobbying and R&D as explanatory variablesⅣ
n=73dependent
variable
lag or
lead
operator
Independent variables
S=1 rd py R2metho
d
rent l2 b)0.507037
7
0.087868
1
0.739
4fe c)
(0.000) a) (0.000)
rent f20.451807
5
0.087436
8
0.741
6fe
(0.000) (0.000)
lobby py R2metho
d
rent75.86993
00.075207
0.744
8re
(0.013) (0.000)
rent f1 101.68440.075180
6
0.751
3re
(0.001) (0.000)
S=1.2 rd py R2metho
d
rent l2 0.4755820.083766
8
0.720
6fe
(0.000) (0.000)
rent f20.419551
2
0.083239
7
0.722
7fe
(0.001) (0.000)
lobby py R2metho
d
rent 64.141240.071694
4
0.731
9re
(0.034) (0.000)
rent f 89.576070.071685
5
0.737
9re
47
NORMATIVE CORPORATE INCOME TAX
(0.002) (0.000)
Note a) () is p value
Note b) fn is lead operator of n periods, ln is lag operator of n periods.
Note c) fe: fixed-effects estimator, re: random-effects estimator,
48