Content uploaded by Elise S. Brezis
Author content
All content in this area was uploaded by Elise S. Brezis on Apr 12, 2018
Content may be subject to copyright.
Journal of Macroeconomics 52 (2017) 175–188
Contents lists available at ScienceDirect
Journal of Macroeconomics
journal homepage: www.elsevier.com/locate/jmacro
Legal conflicts of interest of the revolving door
R
Elise S. Brezis
Department of Economics, Bar-Ilan University, Israel
a r t i c l e i n f o
Article history:
Received 6 October 2016
Revised 16 April 2017
Accepted 17 April 2017
Available online 22 April 2017
JEL classification:
H10
H70
O11
O43
Keywo rds:
Elites
Bureaucracy
Conflicts of interest
Bureaucratic capital
Revolving door
Economic growth
a b s t r a c t
This paper analyzes the conflicts of interest arising from the “revolving door”. The revolv-
ing door is a common phenomenon, and it is unlikely that most of it can be explained
by ‘regulatory capture’, a practice that is unlawful. Therefore, there is a need for a new
framework.
This paper proposes a framework wherein conflicts of interest arising from the revolv-
ing door are not unlawful , as is in the case of regulatory capture, but still lead to economic
distortions. The paper introduces a market for bureaucratic capital , which explains why in
equilibrium, the government allows this unethical, yet not unlawful, conflict of interest to
persist. Our first result is that the political elite finds it optimal to allow the existence of
the revolving door, as well as the creation of bureaucratic capital . The second result is that
in equilibrium, the revolving door leads to an excessive level of bureaucratic capital . As a
consequence, the interconnection of elites and the existence of the revolving door actually
lead to lower economic growth.
©2017 Published by Elsevier Inc.
1. Introduction
State capture, and the various conflicts of interest arising from the “revolving door”, have lately become subjects of
intensive economic research. These phenomena have become particularly relevant after the crisis of 2008; and the possibility
of a link between state capture and the financial crisis has even been raised in the literature.
1
The “revolving door” is a common phenomenon in the United States, and is defined as heads of state agencies, after
completing their bureaucratic terms, entering the very sector they have regulated. This phenomenon is also common in
many developed countries.
2 Over the years, three main empirical regularities have been identified vis-a-vis the revolving
door.
R I wish to thank Jean-Pascal Benassy, Francois Bourguignon, Joel Cariolle, Cecilia Garcia-Penalosa, Roger Guesnerie, Catherine Mann, Adam Przeworski,
Larry Samuelson, Thierry Verdier, seminar participants at the PSE, the European Public Choice meetings, the EEA meetings, the U4 workshop, as well as
the referees and the Editor, for their helpful comments.
E-mail address: elise.brezis@biu.ac.il
1 See OECD (2009) , and Carmen Segarra’s disclosure regarding Goldman Sachs. Previous financial crises were also traced to state capture (see Fisman,
2001 ).
2 See Charle (1987) and OECD (2009) . In France, it has been dubbed “pantouflage” [‘easing in’, as one does into their house slippers], and in Japan, it is
called amakudari [‘descent from heaven’].
http://dx.doi.org/10.1016/j.jmacro.2017.04.006
0164-0704/© 2017 Published by Elsevier Inc.
176 E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188
Fig. 1. Top revolving door employers in the US. Source : See website https://www.opensecrets.org/revolving/ .
The first is that the waltz-like tempo, according to which prominent figures move from public-sector positions into the
private sector, has become more intense over the last decade.
3 Moreover, the revolving door is not specific to the financial
sector; it is widespread over many sectors of the economy, but not in a uniform way.
4
Secondly, a firm’s use of the revolving door affects its profits and stock share prices. Indeed, Luechinger and Moser
(2014) showed that after the announcement of a former bureaucrat’s appointment to a board of governors, there were
positive abnormal returns to the firm. Goldman et al. (2009, 2013 ) found similar results.
Thirdly, over the last ten years, the revolving door phenomenon has been under scrutiny by organizations that combat
conflicts of interest, and documentation of the interconnection between the various power elites has increased.
5 However,
in most countries, the revolving door is not forbidden and is not declared unlawful, but rather is only mildly regulated via
a “cooling-off” period.
The theoretical literature on the revolving door has mainly emphasized conflict of interest taking the form of regulatory
capture , which focuses on fraud and unlawful behavior. Regulatory capture occurs when a former regulator is “captured” by
one specific firm, and while strict with the others, she is lenient with this firm in order to be hired by it after leaving
office. Note that this form of revolving door is unlawful in most Wester n countries and is perceived as pure corruption.
6
Nonetheless, it is difficult to believe that all the cases of revolving door as presented in Fig. 1 and Table 1 can be explained
by fraud and unlawful behavior.
This paper proposes a framework wherein conflicts of interest arising from the revolving door are not unlawful , as
is in the case of regulatory capture, but still lead to economic distortions. Broadly, this type of conflict of interest has
been coined ‘ abuse of power’, and encompasses the creation of excessive regulation while in public office. As defined by
3 See the list in OpenSecrets for the US and Corporate-Europe for the EU. The list of ‘revolvers’ cited in these websites is long, but to name just a few:
Alan Greenspan, Glenn Hubbard, Robert Zoellick, Dick Cheney, Larry Summers, and Madeleine Albright.
4 Figure 1 documents the strong inclination of US companies to hire large cohorts of revolving door personnel, especially in the financial sector, but also
in the defense and pharmaceutical sectors. Moreover, Table 1 presents the data on the numbers of ”revolvers” in Goldman Sachs, Citigroup, and Fannie
Mae.
5 See for instance OpenSecrets, U4 Anti-Corruption Center, Transparency International , and OECD (2009) .
6 Except for Kaufmann and Vicente (2011) . Note that the wrongdoin arises while the worker is employed as a regulator, and any corruption might be
difficult to prove in court. The literature on regulatory capture is presented in the next section.
E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188 177
Tabl e 1
The revolving door in three major US financial firms.
Revolving door flow Number of revolved regulators by category
Tota l P NP
Goldman Sachs (GS)
1. Public to GS 19 ( 5 ) 5 ( 1 ) 14 ( 4 )
2. GS to Public 12 ( 3 ) 10 ( 2 ) 2 ( 1 )
3. GS to Public to GS 6 ( 1 ) 4 (0) 2 ( 1 )
Tota l 37 ( 9 ) 19 ( 3 ) 18 ( 6 )
Citigroup (CG)
1. Public to CG 20 ( 10 ) 3 (0) 17 ( 10 )
2. CG to Public 1 (0) 0 (0) 1 (0)
3. CG to Public to CG 5 ( 2 ) 4
( 2 ) 1 (0)
Tota l 26 ( 12 ) 7 ( 2 ) 19 ( 10 )
Fannie Mae (FM)
1. Public to FM 11 ( 6 ) 2 ( 1 ) 9 ( 5 )
2. FM to Public 3 ( 2 ) 0 (0) 3 ( 2 )
3. FM to Public to FM 12 ( 4 ) 6 ( 3 ) 6 ( 1 )
Tota l 25 ( 12 ) 8 ( 4 ) 17 ( 8 )
Source: Data collected by the author from official company websites, LexisNexis
Academic, OneSource (Avention), and business websites in 2014. I thank Gina Li
for excellent research assistance.
Notes : a. The revolving door flows are divided into three types: Type 1, public-
to-private: Former members of administration who currently hold an executive
position in a regulated company. Type 2, private-to-public: Former executives of
a regulated company who are currently members of a relevant administration.
Type 3, private-to-public-to-private (two-sided): Executives who have engaged in
both type 1 and type 2 movements. In parenthesis is the number of female re-
volved regulators.
b.
For the purposes of calculation, these regulators are sorted according to their
position in the private sector. They are classified into two categories. Powerful
revolved regulators ( category P ) are top-level government officials and legisla-
tors. Non-powerful revolved regulators ( category NP ) are individuals with lower-
level positions in the government or in a relevant administration.
Transparency International (2011) , this ‘abuse of power’ arises when “bureaucrats abuse their power to ingratiate them-
selves to potential future employers”.
The model developed in this paper focuses on the distortions due to the abuse of power that takes the general form of
over-regulation . In other words, the bureaucrat takes actions and makes decisions while in office enabling her to cash in
later when joining a firm she has regulated. While her actions can take various forms, they all incorporate what is termed
the creation of bureaucratic capital .
The most common manifestation of bureaucratic capital is investing in good relationships with the lower-level bureau-
cracy, ties which will help the senior bureaucrat in the future. Bureaucratic capital also takes the form of developing exces-
sive regulation. These are the negative effects of the revolving door stressed in this paper.
But not only has the literature cited the revolving door’s negative effects; there are also positive effects thereof: The
revolving door enables recruiting qualified bureaucrats. Hence, the model assumes that bureaucrats are heterogeneous in
their abilities, and more able bureaucrats do a better job of regulation, so that indeed allowing the revolving door enables
recruitment of high-quality bureaucratic elite.
In other words, by incorporating positive as well as negative effects of the revolving door, and those that are not unlawful ,
this model enables us to analyze the net effects of the conflicts of interest, arising from the revolving door, on economic
growth. This model enables deriving two main conclusions: The first is that it is not optimal to prevent this abuse of power,
and to stop the revolving door, even if it is unethical. The second is that in equilibrium, the revolving door leads to an
excessive level of bureaucratic capital , which in turn leads to over-regulation and to lower economic growth.
Moreover, these two conclusions explain the frequently observed empirical regularities noted above: why the revolving
door is in wide use; why the political elite does not try to prevent it; and why it affects firms’ market values, leading to
abnormal returns for corporate appointments.
The main novelty in this model is the existence of a market for bureaucratic capital wherein the various players in this
market are the three groups of elites which form the power elite: the bureaucratic elite, the business elite, and the political
elite.
7 Bureaucratic capital is supplied by the bureaucratic elite, while the demand comes from the business elite, and the
political elite appoints the bureaucrats.
Let us define more precisely the specific role of each elite, starting with the bureaucrats, who are appointed by the
political elite in order to regulate the economy efficiently. Yet the bureaucrats do not merely enact efficient regulation; they
7 The power elite was coined by Mills (1956) . See the literature on the various elites in the next section.
178 E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188
also create ‘ bureaucratic capital ’ by adding rules and regulations, and by investing in good relationships with the lower-level
bureaucrats. As the architect of these rules, regulations, and relationships, the regulator has insider knowledge of the system,
including any loopholes that might exist.
‘Bureaucratic capital’ therefore enables the bureaucrat to cash in later thereon, after exiting the public sector and joining
a firm in the sector she previously regulated. Thus, the bureaucrat can abuse her previous position to increase her own
returns in a perfectly legal way.
The second player in this framework is the business elite, which finds the knowledge accumulated by the bureaucrat
valuable. Thus, once the latter has left the civil service, she is offered a position such as joining a firm’s board of directors,
thereby enabling her to cash in on the bureaucratic capital she accumulated. In turn, this bureaucratic capital under the
business elite’s control enables increasing the firm’s revenue.
The third player is the political elite, which appoints the bureaucrats and care about enabling the economy the highest
economic growth possible. The question raised by the existence of a market for bureaucratic capital is: Why does the political
elite, for which economic growth is a priority, not find a way to prevent the bureaucratic elite from creating bureaucratic
capital ?
This paper shows that the political elite does not act to eliminate the revolving door and to prevent the accumulation
of bureaucratic capital . On the contrary, they find it optimal to allow the bureaucratic elite to create a certain amount of
bureaucratic capital , even if doing so has some negative effects on economic growth.
The explanation behind this result is the following: Bureaucrats are heterogeneous in their abilities, and more able bu-
reaucrats enable higher productivity and higher economic growth. In order to recruit high-quality bureaucratic elite, govern-
ments must pay them well. However, relative salaries in the public sector are not very high. An easy way to let regulators
have high income, so that they will be of high quality, is by legislators’ closing their eyes to the fact that the bureaucrats
can cash in on the bureaucratic capital they create while serving as heads of agencies.
Thus, the political elite faces a tradeoff between on the one hand, having high-quality bureaucrats, and on the other
hand, letting the high-quality bureaucrats create bureaucratic capital . The optimal solution is a non-corner one, wherein
bureaucratic capital is created. This is the first result of the model.
The second result of this paper is that the market equilibrium will not bring the economy to the highest rate of growth,
and the level of bureaucratic capital is higher than the optimal one. The reason is that while this bureaucratic capital is
valuable to the firm, it is in fact a social waste.
In conclusion, by analyzing the interconnection between the three elites in power –the political, the bureaucratic, and
the business elites – this paper explains the emergence of bureaucratic capital and its effects on economic growth. The
interconnection of elites and the existence of the revolving door actually lead to lower economic growth.
The paper is divided in five parts. In Part II, we present the related literature. In part III, we present the model. In Part
IV, we present the equilibrium, and Part V concludes.
2. Related literature
The literature related to this paper is mostly about the power elite and the revolving door. The economic literature on the
role of the power elite –the political, business, and bureaucratic elites –is not very extensive, and mainly centers around
three topics: (i) the elite structure, (ii) the interrelations between the elite and the economy, and (iii) the interconnections
within elites, which is the topic of this paper.
The elite structure refers primarily to the elite’s social background, its recruitment and promotion patterns, and its ge-
ographic and ethnic origins. The recruitment analysis investigates the transparency of selection and the channels whereby
such selection takes place.
8
The economic literature on the interrelations between the elite and the rest of society devotes special attention to dis-
tributive conflicts and political institutions. The research shows that members of the elite, who have power and wealth,
establish institutions that serve their own interests and exclude the masses from benefits.
9 For example, one avenue of
research argues that wealthy elites with enough political power to block changes will not accept adopting practices that
would enhance growth, as the latter might hurt them.
10
The third topic is the interconnections between elites, which have mainly been the focus of sociologists. This literature
is divided into two main lines of thought: On the one hand are the sociologists who believe that in democracy, there
is competition among the numerous types of elites, and their interconnection has no strong effect on the economy. This
“pluralistic-democratic” position was espoused by sociologists such as Dahl (1957) and Parsons (1963) , who argued that in
Western democracy, the existence of elite groups within the power structure is not an empty fiction. Western social order is
characterized by a dissociation from and distribution of power, or a ‘polyarchy,’ in contrast to the social order in communist
countries, where all such groups are unified in the single-party system. This plurality of elites ensures competition, and that
they do not form a ‘power elite’ separated from the ‘masses.’
8 See Turner (1960) . Some of the sociologists also explore the elite’s attitude formation and behavior, coined as “positional and decision-making”.
9 See Bourguignon and Verdier (20 0 0) and Easterly (2001) .
10 See Acemoglu et al. (2011) and Rajan and Zingales (1998, 2004 ).
E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188 179
This school of thought further espouses the idea that democracy should a priori impose some control on the power of
the ruling elite. Indeed, Schumpeter (1954) claimed that the democratic process permits “free competition among would-be
leaders for the vote of the electorate” and that the masses can choose between various elites. Ye t they all agree that in
non-democratic polities, there is collusion between elites who have political and economic power, and who typically act on
behalf of their own interests.
In contrast to the “pluralist-democratic” view, classical elite theorists such as Mosca (1939), Pareto (1935), Michels (1915) ,
and Mills (1956) espoused the idea that despite the democratic character of a given regime –where power is meant to reside
in the demos (the people) –power is in fact concentrated in the hands of a few –the oligarchy – which Mills (1956) called
the “power elite” .
11
Michels has coined this view as the “iron law of oligarchy”, claiming that an oligarchy is inevitable within any organiza-
tion as part of the "tactical and technical necessities" of organization ( Michels, 1915 ).
12
Consequently, there can be collusion
even in democracies. Various elites may not be mutually competitive and may not control and balance each other; instead,
they may be intertwined as a unanimous, cohesive power elite.
13
A strong interconnection among elites leads to all sectors of the economy being ruled by a single group that thinks
monolithically. Two schools of thought have related a monolithic group to negative economic growth: The first claims that
a monolithic group leads to the stagnation of ideas and attitudes, which in turn may prevent the adoption of technological
breakthroughs ( Bourdieu, 197 7 ).
14 The second school focuses on the lack of competition in a monolithic powerful group,
generating corruption, which has harmful consequences for growth.
This paper focuses on bureaucrats’ quality. The economic literature on bureaucracy is voluminous and diverse. For exam-
ple, there are descriptive studies on the evolution of bureaucratic institutions (see Johnson and Libecap, 199 4 ), and on the
typical profile of bureaucrats, including political affiliations, the degree of professionalism and the bureaucrats’ efficiency
( Rauch 1995 ). Numerous articles have been written on the quality of bureaucracy and on its effects on economic perfor-
mance, and more specifically on economic growth (see Krueger, 1993; Shleifer and Vishny, 1993 ; and Mauro, 1995 ). La Porta
et al. (1999) presented a synthesis on the various views.
15
Regarding the revolving door, the mainstream of the economic literature has mainly emphasized conflict of interest in
the form of regulatory capture, in which regulators can be induced to act in the firm’s interest.
16
This literature started with
the works of Stigler (1971) and Peltzman (1976) followed by Eckert (1981) . Most researchers in this field have focused on
the potentially undesirable effects of corruption and regulatory capture, and solutions thereto that could be implemented
( Spiller, 1990; Brezis and Weiss, 1997 ). However, there are also studies that show that there may be positive aspects of the
revolving door that should not be overlooked ( Salant, 1995; Che, 1995 ).
An empirical literature has recently arisen that investigates the revolving door’s effects on firms’ performance.
17 As an
indication of the strong link between political connections, the revolving door process, and corrupt practices, cross-country
analyses ( Faccio, 2006; Faccio and Parsley, 2009 ) showed that the gap in economic returns between firms with and without
political connections increases in highly corrupt environments. Moreover, Faccio (2010) showed that politically connected
firms pay lower taxes than do other firms.
There are also differences between firms with and without state connections in the finance realm. Khwaja and Mian
(2005) and Boubakri et al., (2012) showed that firms with state connections are more likely to be bailed out of financial
distress.
18
Political connections created by the revolving door increase firms’ value by re-directing consumer demand in their fa-
vor. Goldman et al., (2013) showed that, following the 1994 House and Senate elections, the presence of former politicians
affiliated with the winning political party on the boards of U.S companies increases the total value of awarded public pro-
curement contracts.
In a study on Italy, Cingano and Pinotti (2013) showed that corporate appointments of local politicians do not lead to
higher productivity; and Kramarz and Thesmar (2013) showed that French firms hiring directors and CEOs with former ca-
reers in the civil service, underperform. Moreover, they found that connected CEOs are better paid than their non-connected
counterparts, are less likely to be fired, and are associated with lower performance.
Finally, in Russia, Slinko et al., (2005) find that politically powerful Russian firms adversely affect the performance of
small or politically-powerless firms, by lobbying regulators to excessively regulate the latter, and also by diverting govern-
11 A coherent synthesis of these differing views can be found in Dye et al. (2011) .
12 This view was adopted by Hunter (1959) and Domhoff (1970, 2010) for the US, and Aaronovitch (1961) and Miliband (1969) for England.
13 For a summary on the economic consequences of the absence of elite competition, see Brezis and Temi n, (2008) .
14 Brezis and Crouzet (2006) incorporated a bias àla Bourdieu, and analyzed its effect on economic growth.
15 There is also an extensive literature on the quality of politicians. See Besley (2005), Besley et al. (2011), Mattozzi and Merlo (2008) , Galasso and
Nannicini (2011, 2015 ), de Paola and Scoppa (2011) , and Kotakorpi and Poutvaara, (2011) .
16 For a review on regulatory capture see Dal Bo (2006) . Lately, de Haan and Veltrop (2014) developed a model of regulatory capture based on social
identity. There is also another type of state capture –the lobbying capture , wherein after leaving office, the bureaucrat is hired by a lobbying
firm. But this
capture is not directly linked to the revolving door focusing on individuals entering a firm in the sector they have regulated (see Blanes et al., 2012 ).
17 The empirical research on the revolving door focuses on both directions, i.e., from the public sector to the private sector and vice versa, while this
paper focuses only on the transition from the bureaucratic elite to the business elite.
18 In contrast, Luechinger and Moser (2014) related the value of the firm on the stock market to the timing of the revolving door. See also Fisman
(2001) and Faccio et al. (2013) .
180 E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188
ment spending. In contrast, they found that politically powerless firms, wherein which the firms’ concentrations of political
power are lower, invest more and are more productive.
In conclusion, the recent literature has shown that the revolving door erodes the tax base, reduces productivity, and
lowers aggregate growth. In the next section, we present a model analyzing the revolving door’s net effects on economic
growth.
3. The model
3.1. Introduction
This paper develops a model of the revolving door that includes both its negative and positive effects, and analyzes
the interconnection between the three power elites (political, business and bureaucratic), by introducing a new market:
the market for bureaucratic capital . In this market, the supply of bureaucratic capital is determined by the bureaucratic elite,
while the demand is determined by the business elite; and the equilibrium between supply and demand determines the level
of bureaucratic capital . Moreover, the bureaucrats are appointed by the political elite. This new market enables explaining
(i) why firms want to pay rents for hiring a previous regulator to sit on their boards, and (ii) why the political elite permits
the revolving door phenomenon.
The model analyzes whether this equilibrium leads to a high rate of economic growth, which is the political elite’s
goal. This paper shows that from the point of view of the political elite, it is optimal to allow the bureaucrat to create
bureaucratic capital . However, the level of bureaucratic capital determined by the bureaucratic and business elites is higher
than the optimal one for the political elite. This is so because, bureaucratic capital, ceteris paribus , generates higher income
for the firm, yet ultimately in equilibrium, it is a social waste.
The model will be presented in the following way. We begin by presenting the production functions, we then address
the behavior of the various elites; and then we display the rate of economic growth as a function of the elites’ behavior.
3.2. Production functions and the labor market
The economy has one main sector: the good sector. The model of this sector follows exactly the Romer (1990) model.
The good sector is growing over time due to R&D.
The workers work in two sub-sectors: either in the production of the final good, or in the R&D sector, and the workers
in these sub-sectors are homogeneous. There is mobility between these two sub-sectors, so that wages will be equal, and
we have that:
L
r + L
y =
¯
L (1)
where L
r is the size of the labor force in the R&D sector, and L
y the labor force in the output sector. The technical labor
force, ¯
L is constant.
In this economy, there are also a small amount of individuals who have specific abilities and skills such as physicians,
economists, or attorneys, and they provide services in their fields of specialization. Without any loss of generality, we assume
that the measure of these individuals is zero.
19 Moreover, these ‘specialists’ are heterogeneous in their ability, and each of
them will get a salary as a function of her ability, so that their salaries are not uniform (see Weiss, 1980 ). It is among this
group of individuals that the political elite will choose the bureaucrats who will regulate the economy.
In the labor and good markets, following Romer (1990) , we assume that the final good, Y is produced with labor and
intermediate goods, and the production exhibits constant returns to scale.
20
The intermediate goods x
j
are produced by mo-
nopolist firms, and the factor that leads to growth is the increase in the number of new technologies, which are developed
in the R&D sector, and which are embedded in new intermediate goods available on the market. There is no growth of
population, and capital is increasing through savings of this population, as in Romer.
The workers can work in two sub-sectors: either in the production sector, or in the R&D sector. The workers are ho-
mogenous in their ability and get wages determined endogenously in the model.
The production function of the final good is:
Y = L
1 −α
y
A
0
x
α
j
dj (2)
where Y is the output at each period; L
y
- the number of workers in the production sector; x
j
the number of intermediate
goods/machines from type j; and A, the level of technology, measured by the range of capital goods available. While the
19 Since this paper does not focus on que stions related to relative productivity of good and service sectors, we assume that the measure of the individuals
in each of these fields of skills is zero in the whole population. The number of fields of specialization is finite, therefore the whole group
has a measure of
zero and its aggregate income and consumption can be neglected in the market clearing conditions.
20 Since the production functions follow Romer (1990) , we therefore succinctly present the model, and more explanations on the assumptions of the
model can be found in his paper.
E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188 181
final good is produced in a perfect competitive environment, the intermediate-goods sector consists of monopolistic firms,
which each produce a specific intermediate good, x
j
.
The firms involved in the production sector, Y are maximizing profits:
Max L
1 −α
y
A
0
x
α
j
dj −w
y
L
y
−
A
0
p
j
x
j
dj (3)
w
y are the wages paid for labor in sector Y, and p
j
is the price of the intermediate good x
j
.
From the profit maximization in the production sector, we get:
w
y
= (1 −α)
Y
L
y
(4)
and
p
j
= αL
1 −α
y
x
α−1
j
(5)
3.3. The bureaucratic capital
In this section, let us present the supply of bureaucratic capital by the regulators, the demand of bureaucratic capital by
the intermediate-good firms, and the equilibrium in bureaucratic capital .
3.3.1. The bureaucratic elite and the supply of bureaucratic capital
The bureaucratic elite is comprised of individuals who are appointed to head government agencies. This paper focuses
only on top regulators, such as the governors of banks, heads of the Federal Trade Commission, the anti-trust division, the
FDA, and so on. While these agency heads are of measure zero in the total population, their effect on the economy is
important, and this paper will analyze their positive as well as negative effects thereon. Indeed, what does a bureaucrat do?
On one hand, she maximizes her utility, given that she was appointed to be head of a regulating agency. On the other
hand, she has positive effects on the economy while regulating, which we discuss later. So, we focus now on her utility
function at the time she is appointed to this position.
The novelty of our paper is that during her time in office, the regulator regulates, but at the same time, she creates
bureaucratic capital . The bureaucratic capital are all the unnecessary regulations she is developing. Indeed, on one hand,
some regulations have an important effect on the economy as a whole, for instance, changing the reserve ratio in time of
recession, or blocking a medicine dangerous for human beings. On the other hand, she develops regulations and personal
connections, which do not have an effect on the economy. She develops them, for only one reason: she has the knowledge
on these regulations, and she has these connections. This knowledge and connections are valuable to the firms in the in-
dustry, and thus, once she has left the public service, the regulator can cash-in on this bureaucratic capital. This is the basic
idea of the paper.
One unit of bureaucratic capital can be understood as one piece of regulation. The regulator decides the optimal amount
of over-regulation she wants to develop, that will cost her effort, but that will permit her to get a higher income in the
future. As all individuals, a regulator lives on period, but which is divided into two sub-periods. During the first sub-period,
she works as a regulator, and during the second one, she works in the sector she has regulated.
During her term as a regulator, she acquires bureaucratic capital of size H , which costs her effort of size E . Since her
utility function is a function of her total consumption, C and effort, E, we get that:
V = U(C, E) and ∂ U/∂ C ≥0 ∂ U/∂ E ≤0 (6)
For matters of simplicity, we may assume that the utility function is linear:
V = U(C, E) = C −λE;λ> 0 (7)
The production function of bureaucratic capital is a function of effort, and we assume that the level of bureaucratic
capital is a concave function of the amount of effort invested, the same for all bureaucrats, which takes the specific
form:
21
H(E) = [(1 + γ) E]
1 / 1+ γγ> 0 (8)
Her lifetime salary when appointed as regulator is . After leaving her job as regulator, the bureaucrat works in the in-
dustry that she regulated. She receives in top of her lifetime salary, a rent which is a function of the amount of "bureaucratic
capital", H she has accumulated at price q.
22
21 We are aware that for some bureaucrats, who are either more social, or with less “ethical values”, it is easier to either create connection with other
people, or create redundant regulations. For purpose of simplicity, we assume that bureaucrats have the same “production” function of bureaucratic capital,
and that these social
factors are not linked to ability, since removing this assumption does not affect the results.
22 In other words,
is the lifetime salary of the regulator, which includes also what she gets after retiring from being a bureaucrat. qH is the ‘bonus’ she
gets by the company which will hire her after leaving office, for her over-regulation.
182 E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188
Fig. 2. Supply and demand of bureaucratic capital, and the trade-off between quality and bureaucratic capital.
In consequence, the utility function she maximizes is:
V = + qH −λE (9)
Eq. (9) can be rewritten as a function of the level of bureaucratic capital , by substituting E from equation ( 8 ). We get:
V = + qH −λH
1+ γ
1 + γ(10)
Taking the FOC of Eq. (10) with respect to H, we get the optimal level of bureaucratic power, H she wants to accumulate:
H = (q/λ)
1 /γ(11)
Eq. (11) describes the “supply” function of bureaucratic capital by the bureaucratic elite as an increasing function of the
price q , and which is displayed as the S function in Fig. 2 , part (I).
We later discuss the appointment of the bureaucrats by the political elite. We now turn to the demand for bureaucratic
capital, which is determined by the firms, and the entrepreneurs who stands at the head of these firms.
3.3.2. The intermediate-goods firms and the demand for bureaucratic capital
The business elite is composed of entrepreneurs, who are at the head of intermediate-goods firms, who own a patent
developed by the R&D sector, and who produce goods, x
j
, in a monopolistic competitive environment. In other words, the
intermediate-goods firms consist of monopolistic firms each with its own intermediate good, and in consequence, they are
regulated by regulators nominated by the political elite. The entrepreneurs maximizes profits of the regulated firms. This
will permit us to develop the demand of these firms for the “knowledge” of the “revolver” regulators, i.e., the bureaucratic
capital.
The output is a function of two factors of production. The first is capital, k
j
. Following the Romer model, we assume that
the production function takes the simple form:
x
j
= k
j
.
However, in our model, the output x
j
is also function of a second factor of production, which is the level of bureaucratic
capital the firm acquires. When a firm j hires a bureaucrat with a bureaucratic capital H
j
, the production of output j becomes
more efficient. This is so, because the regulator has knowledge and connections on the system. More specifically, it depends
on the relative level of bureaucratic capital by the different regulators of the different firms, since it is only the relative
knowledge which matters.
So the production function in sector j takes the form:
x
j
= k
j
H
j
H
φ
φ> 0 (12)
where H
j
is the level of bureaucratic capital produced by the regulator entering firm j, and H is the average level of bureau-
cratic capital owned by the other firms.
23
23 This formulation is quite in use in models with monopolistic competition, as for instance the Neo -Keynesians models with price setting and monopo-
listic competition (see Blanchard and Kyota ki , 1987 ).
E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188 183
Note that if H
j
= H , then the output is just: x
j
= k
j
, no matter the average level of bureaucratic capital. Although having
hired a bureaucrat to increase the productivity of the firm may bring advantage from an individual point of view, it is pure
waste from a social point of view.
So the profit maximization for an intermediate good firm is:
Max πj
= p
j
( x
j
) x
j
−r k
j
−q H
j (13)
The two costs of factors of productions are (i) capital, k
j
where r is the interest rate, and (ii) the bureaucratic capital at
cost q . The last term in Eq. (13) is the amount paid to the regulator for her bureaucratic capital. Indeed, the bureaucrat who
owns bureaucratic capital of level H will “sell” it to the firm, at price q , which is endogenously determined by supply and
demand.
Each firm maximizes profits by finding the optimal amount of output, x
j
and bureaucratic capital, H
j
. Note that Eq.
(13) can be rewritten in the following way:
Max πj
= p
j
( x
j
) x
j
−r x
j
H
j
H
−φ
−q H
j (14)
where p
j
is given by Eq. (5) . Since the business elite is composed of monopolists who see the price of their good as nega-
tively related to the demand, the two first-order conditions for maximizing profits are:
p
j
( x
j
) x
j
+ p
j
( x
j
) −r
H
j
H
−φ
= 0 (15)
q H = ϕ r x
j
H
j
H
−φ−1
(16)
From Eq. (5) , we note that the demand elasticity of p
j
( x
j
) is equal to α−1. Substituting into Eq. (15) , and in a symmetric
equilibrium all H
j
are the same. Thus we get that:
p
j
= p =
1
αr (17)
H
j
= H =
φrK
qA
(18)
where the total amount of capital in the economy K is:
K =
A
0
k
j
dj .
Moreover, since all intermediate-goods firms sell for the same price, p, we get that: x
j
= x , and k
j
= k .
Eq. (18) represents the demand for bureaucratic capital, as a decreasing function of q , which is displayed as the D function
in Fig. 2 , part (I).
3.3.3. The equilibrium of bureaucratic capital
From the side of the bureaucratic elite, described in Section 3.3.1 , we get the supply equation of bureaucratic capital ( Eq.
11 ), and from the side of the business elite, described in Section 3.3.2 , we get the demand for bureaucratic capital ( Eq. 18 ).
By equating demand with supply we get the equilibrium stock of bureaucratic capital:
H∗= (φrK/λA )
1 / 1+ γ
and
q ∗= [ λ(φrK/A )
γ]
1 / 1+ γ
(19)
This equilibrium is presented in Fig. 2 , part(I). First, we note that when the parameter which represents the effect of
regulation on the firms, φincreases, then the level of bureaucratic capital increases. Moreover, if bureaucrats are less effi-
cient in producing bureaucratic capital, then γincreases, and the level of bureaucratic capital decreases. The last interesting
variable is K/A ; we discuss it below, when we show that, in a balanced growth path, it is constant.
Summarizing this section, the intertwining of the bureaucratic and business elites has led to the formation of bureau-
cratic capital of level H
∗. The creation of this capital has allowed the bureaucratic elite to cash in on their bureaucratic
capital stock after leaving their public-sector regulator positions, and enter the very businesses they once regulated. The
intertwining of the bureaucratic and business elites is thus the consequence of the supply of bureaucratic capital by the bu-
reaucratic elite, and the demand by the business elite. In other words, the revolving door allows the creation of bureaucratic
capital, that is legal and not unlawful.
Is this level of bureaucratic capital optimal from the standpoint of the political elite? In the following sections we show
that it is not, but we also show that the optimal bureaucratic capital level from the standpoint of the political elite is not
zero. In order to do so, we define the rate of growth of the economy that is determined by the R& D sector, and we describe
the political elite’s behavior.
184 E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188
3.4. The rate of growth in the economy
3.4.1. The R&D sector
Following Romer (1990) , the R&D sector develops new designs for new intermediate goods. The only factor that leads
to growth is the increase in the number of new technologies existing. We assume that the number of new inventions is a
function of the size of the labor force in the R&D sector, and also of the amount of machines already in existence, A. This
assumption is the usual externality of spillover effects which leads to a “size effect” in economic growth.
Moreover, based on Mauro (1995) and La Porta et al. (1999) , who have shown that the quality of government affects the
performance of firms, we assume that the ability of the regulator affects the productivity of the workers in the R&D sector
and we get that the number of new inventions is:
˙
A = δ(Q) L
r
A (20)
where δis a positive parameter function of the quality of the bureaucrat, Q , δ ≥0 and δ < 0. L
r is the size of the labor
force in the R&D sector, and A the amount of machines already in existence. In consequence we get that, in steady state,
the rate of growth of the inventions, g , which is also the rate of growth of the economy, is constant since we focus only on
the balanced growth path:
g =
˙
A
A
= δ(Q) L
r (21)
Indeed, as in Romer, and in many of the basic models of economic growth, the focus is on the balance growth path, in
which, capital, output, innovations and consumption all grow at a constant rate, g defined by ( 21 ).
24 So we have that K/A
and Y/A are constant.
The two elements affecting economic growth are the size of the labor force in the R&D sector, and the ability and quality
of the bureaucratic elite. In the next section, we explain how the political elite appoints the bureaucratic elite.
3.4.2. The political elite and the quality of the bureaucratic elite
For matter of simplicity, this paper assumes that the goal of the political elite is to maximize the rate of growth of the
economy given by Eq. (21) , and this in order to be reelected.
25
One instrument in the hand of the political elite is to appoint the bureaucratic elite, i.e., the regulators who regulate
the monopolistic firms. As emphasized above, regulators are chosen among the individuals with specific abilities and skills.
These individuals are heterogeneous in their abilities, and as emphasized by Weiss (1980) , when ability affects the produc-
tivity of a person, then wages are not equal for all: “workers’ wage is an increasing function of his ability” . Individuals with
high ability and quality earn more than ones with less ability.
In consequence, without loss of generality, we assume the following form :
26
W
s
= ξQ
s (22)
where W
s is the income of a specific individual, and Q
s is the ability of this individual.
Since quality of the regulator affects economic growth ( Eq. 21 ), the political elite will choose a regulator with the highest
ability possible and who gets an income related to quality given by Eq. (22) . We assume that the legislator possesses perfect
knowledge of each candidate’s ability.
In consequence, the political elite knows that the reservation income of the potential bureaucrat is given by ( 22 ) and
therefore the choice faced by the political elite is to hire a bureaucrat with ability such that:
Q
i
= Max
Q
s
|
ξQ
s
≤V
s
(23)
and the solution is:
Q
i
=
1
ξV
i (24)
where V
i
is the lifetime income of the bureaucrat i. Substituting Eq. (10) into Eq. (24) , we get the relationship between
ability and level of bureaucratic capital faced by the political elite:
Q
i
=
1
ξ[−λH
1+ γ
i
1 + γ+ q H
i
] (
The QH curve
) (25)
This QH equation describes the trade-off faced by the political elite while choosing the bureaucratic elite: Appointing a
regulator with higher ability means letting her accumulate a higher level of bureaucratic capital. This equation is therefore
24 We do not discuss growth rates in transition path. Moreover since this paper does not put the focus on savings and consumption, we follow Romer’s
equation (2) for capital accumulation.
25 Another alternative is to assume that the political elite is benevolent. If we would have included some “political economy externalities”, this would
reinforce our results, and therefore, we adopt a simple formalization of the political elite.
26 We are aware that there are models in which this relationship is not linear. For instance in the theory of “winners take it all”. But except for the very
top (which then, will not take a post in the public sector), the assumption of linearity seems reasonable. See Greenwald, (1979)
.
E.S. Brezis / Journal of Macroeconomics 52 (2017) 175–188 185
the production possibility frontier between bureaucratic capital and quality faced by the political elite. This QH equation
(which is described for the equilibrium price q
∗) is depicted in Fig. 2 , quarter (II).
27 The maximum amount of quality is
reached at H
=
H
∗.
It should be added that this result is quite intuitive: If the political elite wants to hire a high-quality bureaucrat, her
income has to be high. Since wages in the public sector are lower than the reservation wage of the bureaucrat, they have to
let her invest and accumulate bureaucratic capital so that she will be indifferent between being a bureaucrat or providing
private services as lawyer or economist. The political elite will get a high quality bureaucrat subject to Eq. (25) : a higher
quality means a higher amount of bureaucratic capital.
We are left with the question, which bundle of quality and bureaucratic capital will the political elite chose. In the next
section, we develop the equilibrium rate of growth chosen by the political elite.
4. Determination of the equilibrium and of the rate of economic growth
As in Romer, the equilibrium is obtained by equating wages earned by workers in the output and the R&D sectors.
28 So,
we have:
w
r = w
y (26)
where w
r and w
y are wages in the R&D and production sectors respectively. As previously mentioned, the total labor force
working in the production and the research sectors is constant and denoted by ¯
L .
L
r + L
y =
¯
L (1
)
where L
r is the size of the labor force in the R&D sector, and L
y the labor force in the output sector.
Since the salary earned by workers in the R&D sector is the value of the patent of their invention, we have that:
w
r =
˙
A
L
r
P
r (27)
where P
r is the price of a new-design patent, and
˙
A is the number of new inventions developed.
Moreover, remember that:
w
y
= (1 −α)
Y
L
y
(4
)
In order to solve Eq. (26) , following Romer, we apply the asset pricing arbitrage equation, so that :
29
r P
r = π+
•
P
r (28)
where
•
P
r is the change in the price of patents and πare profits. Since there is no increase in population, along the
balanced growth path, output Y , and inventions, A grow at the same rate, so that patent prices also are constant, and we
get:
P
r
=
π
r
(29)
Moreover, from Eqs. (17) , ( 2 ) and ( 5 ) we obtain that the profit for each of the firms is:
π= α(1 −α)
Y
A
−qH (30)
Equating Eqs. (4) and ( 27 ) and substituting from Eq. (30) , output can be written in the following way (see the online
appendix):
Y = AτL
y where τ= (
α2
r
)
α/ 1 −α(31)
Since from Eqs. (20) , ( 26 ), ( 27 ), ( 29 ) and ( 30 ), we get that:
(1 −α) Y
L
y
=
δA
r
[
α(1 −α) Y
A
−q
∗H]
Then we obtain:
30
L
y
=
r
αδ(Q)
+
q
∗H
α(1 −α) τ(32)
27