The Revolving Door: Charitable Foundation Funding for Think Tanks ...

Jnchoi

Informations

• Publié par: elysonev
• Nombre de lectures: 122
• Langue: English

Extrait

The Revolving Door: Charitable Foundation Funding for Think Tanks

and Implications for Democratic Governance

Proposal Submitted to the

Stanford Center on Philanthropy and Civil Society Graduate Fellowships

Choi, Joon Nak

15 June 2007

Choi, Joon Nak

1

F

OUNDATIONS

, T

HINK

T

ANKS

AND

THE

R

EVOLVING

D

OOR

Executive

Branch

Think

Tanks

Campaign

Funding & Talent

F

u

n

d

i

n

g

P

o

l

i

c

y

Figure 1: The Revolving Door

Over the past eight years, philanthropic foundations gave over $230 million to non-profit policy

research institutes (“think tanks”). Such funding constitutes a substantial proportion of the

money entering American national politics. For example, the Harvard Law Review (2002)

estimated that the top twenty conservative think tanks spend more money annually than all soft

money contributions to the Republican Party combined. Yet, the financial link between think

tanks and foundations has escaped scholarly attention.

The research gap is particularly surprising considering

how partisan think tanks use foundation grants.

1

To

understand why this is important, start by considering a

party that has just lost the executive branch. Where do

its appointed officials go? Many go out the “revolving

door”, riding out hostile Presidencies in partisan think

tanks. Eventually, they come back in the revolving door

(Figure 1), resurfacing in the next friendly Presidency.

By funding the revolving door, foundations are

supporting “…governments in exile…where officials of

the party whose presidential candidate has been defeated can seek gainful employment while

they lick their wounds, wait for their party to come back to power and (hopefully) come up with

new ideas.” (Weaver 1989: 569) This government-in-exile refines a shared ideology, blueprints a

policy agenda, and recruits like-minded individuals amongst leading businessmen, lawyers and

academics. Consequently, the revolving door reinforces continuity in American politics.

A N

EED FOR

A

DDITIONAL

R

ESEARCH

Scholars have addressed the history of think tanks (Smith 1991), their role as sources of expert

advice (Ricci 1993) and made cross-national comparisons (McGann and Weaver 2000). Although

some scholars (e.g. Weaver 1989) are aware of the revolving door, they have yet to produce

research focused on this process and its consequences.

What accounts for this state of affairs? Part of the problem undoubtedly lies in the difficulties

associated with systematically collecting data on policy organizations. A second barrier may be

the division of labor in the social sciences. Social network analysis is useful for simplifying

complex inter-relationships between organizations like foundations, think tanks and presidential

administrations. However, few political scientists employ network analysis techniques, while

network-oriented sociologists consider the Beltway the exclusive realm of political science.

A research program designed to study the revolving door should proceed in three stages. In the

first stage, a basic descriptive analysis should chart how think tanks, foundations and presidential

administrations are tied together. This analysis will facilitate a second stage of research focusing

1

Although think tanks generally advertise themselves as unbiased analysts, many of them engage in highly partisan rhetoric. According to

Weaver (1989) and Ricci (1993), think tanks founded after 1968 have a stronger partisan flavor, while older think tanks (e.g. RAND) stay closer

to their traditional mission of providing independent policy analysis. Unfortunately, mo independent criteria for measuring think tank partisanship

currently exists. Thus, for the purpose of this proposal, I follow the mainstream media’s characterizations of various think tanks.

Choi, Joon Nak

2

on these networks’ meso-level impacts (e.g. presidential appointee selection, policy network

evolution, ideological reinforcement). Such research not only addresses pressing questions in

social network theory, but also facilitates a third stage of research focusing on the revolving

door’s impacts on democratic governance. I speculate that these effects include the circulation of

political elites and institutional inertia.

As a graduate fellowship recipient, I intend to complete a descriptive map of the think tank

industry (Stage I) and investigate how the industry’s social structure affects two policy outcomes

of consequence (Stage II).

In the remainder of the proposal, I describe my proposed research in greater detail. I start with a

discussion of the data I have collected. Next, I summarize the analyses I have already completed.

Finally, I discuss the research I plan to complete through the end of my fellowship year.

D

ATA

The greatest obstacle in social network research tends to be the difficulty of systematically

collecting appropriate data. Fortunately, this hurdle has already been overcome in a collaborative

effort with Professor Gi-wook Shin. Last summer, I researched American opinion leaders on

Korean issues. Considering the influence of resident scholars at think tanks, we decided to

examine think tanks in greater detail. With the assistance of several undergraduate research

assistants, we collected network data on a sample including (1) the 25 most prominent foreign

policy think tanks, (2) 14 think tanks specializing in Asian affairs and (3) the 20 foundations

most actively funding think tanks. We collected four types of network data:

-

Funding streams linking philanthropic foundations with think tanks

. By federal law,

foundations must annually report their grants to the Internal Revenue Service. We

obtained these reports (1997-2005) and aggregated all of a particular foundation’s grants

to a particular think tank. The resulting data represents a valued 2-mode network

(henceforth

FUNDNET

) showing the relationship between foundations and think tanks.

-

Interlocking directorates for think tanks

. Most think tanks are governed by a board of

trustees, publicly disclosed on their websites and/or annual reports. Using archive.org,

we examined archived think tank websites from three years (1999, 2003, 2006) to obtain

historical board membership data. The resulting data represents a 2-mode network

(

DIRNET

) showing how shared governing board members link various think tanks.

-

Interlocking academic advisory boards for think tanks

. Many think tanks feature an

academic advisory board. We collected historical advisory board membership data the

same way we collected governing board memberships. The resulting data represents a 2-

mode network (

AANET

) showing how shared academic advisors link think tanks.

-

Scholar movement between think tanks

. Think tanks often invite visiting scholars from

other think tanks. Also, resident scholars may move between think tanks for a variety of

reasons. Again, we collected this data the same way we collected governing board

memberships. The resulting data represents a 2-mode network (

SCHOLNET

) showing

how scholar movements link think tanks.

Choi, Joon Nak

3

Although the data was originally collected to research American views on Korea, it includes all

significant foreign policy think tanks (according to longtime members of several foreign policy

think tanks). Thus, the data is suitable for examining a broader range of foreign policy outcomes.

S

TAGE

I: M

APPING THE

S

OCIAL

S

TRUCTURE

OF THE

T

HINK

T

ANK

I

NDUSTRY

This dataset describes the think tank industry’s social structure in some detail. With it, I have

made substantial progress towards a social map of the think tank industry. This analysis is best

represented as network visualizations (using multidimensional scaling).

Figure 2 visualizes

FUNDNET

. Foundations are marked as light-colored vertices; think tanks are

darker. Vertex size indicates the amount of funding disbursed (by foundations) or received (by

think tanks). The thickness of a line indicates the amount of funding flowing between connected

vertices. Vertices located closer together should be considered closer in social space.

Figure 2:

FUNDNET

Visualization

Figure 2 highlights two dimensions of the industry’s social structure. Ideologically similar think

tanks

2

are located near each other, indicating that they receive funding from similar sources.

Thus,

FUNDNET

demonstrates strong homophily. Second,

FUNDNET

shows a pronounced left-

right dimension resembling the ideological polarization of American politics.

Figure 2 also reveals how foundations disburse their funds. The center-left and conservative

think tanks receive funding from distinct groups of foundations. The center-left think tanks

receive substantial funding from several large foundations (e.g. the Carnegie Endowment). In

2

Two clusters represent useful reference points. The Hoover Institution (marked hoover), the American Enterprise Institute (aei), the Cato

Institute (cato) and the Heritage Foundation (heritage) represent prominent conservative think tanks. The Brookings Institution (brookings), the

Council for Foreign Relations (cfr) and the Carnegie Endowment for International Peace (ceip) represent their center-left counterparts.

Choi, Joon Nak

4

contrast, the conservative think tanks receive the bulk of their funding from just three

foundations (the Bradley, Olin and Sarah Scaife Foundations).

Figure 3 visualizes a combination of

DIRNET

,

AANET

and

SCHOLNET

. This 2-mode network,

however, has been reduced to a 1-mode network by assigning an edge between all think tanks

sharing individuals. Vertices located closer together should be considered closer in social space.

Figure 3: 1-mode Visualization of

DIRNET

,

AANET

and

SCHOLNET

Like Figure 2, this analysis highlights homophily and a left-right ideological spectrum.

Ideologically similar think tanks are located near each other, indicating that they not only receive

funding from similar sources, but also have the same kinds of members.

By 12/2007, I plan to identify think tanks that support Clinton Administration officials leaving

the revolving door, and others that provide the Bush Administration key personnel.

I intend to

publish these analyses as a descriptive overview of the think tank industry’s social structure.

S

TAGE

II: M

ESO

-L

EVEL

I

MPACTS

OF

T

HINK

T

ANK

S

OCIAL

S

TRUCTURE

After exploring the architecture of the think tank industry, I will investigate how its social

structure directly affects important meso-level phenomena. Although the data supports several

lines of inquiry, I will focus on two well-defined research subjects during my fellowship year.

In the first project, I will investigate how homophilous networks influenced policymaker

opinions on North Korea during the 2002-2003 nuclear crisis, when they publicly debated

whether to engage North Korea diplomatically, or to apply forceful sanctions. Although the

above analyses show think tanks’ homophily, they do it solely on the basis of popular perceptions

about think tanks’ ideological tendencies. Scholars have not determined whether the industry’s

social structure has a causal effect on policymaker opinions. I intend to address this gap.

Using

regression analysis, I will model how social networks (DIRNET, AANET and SCHOLNET)

Choi, Joon Nak

5

predict individuals’ expressed opinions on a key foreign policy debate

.

3

I plan to produce a

chapter for a conference volume, and then turn this work into a journal submission.

The second research project deals directly with the revolving door. The social networks literature

has long proposed that people find jobs through interpersonal relationships. Most observers of

American politics would agree that incoming administrations fill key positions using the

revolving door. Less obvious is the way governments-in-exile actively recruit new members into

their social networks. Anecdotal evidence suggests that policymakers build relationships outside

the revolving door, bring think tank colleagues with them into a new administration, and place

them in key positions.

I will test this proposition by modeling how social network analyses (on

DIRNET, AANET and SCHOLNET) predict key foreign policy appointees at the beginning of the

Bush Administration

. This study would not only bring academic rigor to a popular discussion

topic, but also engage the emerging academic literature on social network evolution.

C

ONCLUSION

By linking American politics and philanthropy using social network analysis, this project charts

previously unexplored territory in two ways. It is among the first studies on network evolution,

an increasingly urgent issue in the networks literature. At the same time, it will build the

foundations for a study of the revolving door’s broader implications for democratic governance.

Finally, the findings will hopefully renew interest in policy organizations among economic

sociologists and organizational scholars.

I am well-qualified to conduct this research. As a fourth-year (TGR) student, I have completed

advanced courses on network analysis and political organizations. Also, I understand the dataset I

will be using, having played a lead role in designing and implementing the data collection

process. Most importantly, my recent work on Korean political parties has demonstrated my

ability to conceptualize, execute and report social network analyses of political organizations.

R

EFERENCES

McGann, James G. and R. Kent Weaver. 2000. Think-Tanks and Civil Societies. New Brunswick,

NJ, Transaction Publishers.

Ricci, David M. 1993. The Transformation of American Politics: The New Washington and the

Rise of Think Tanks. New Haven: Yale University Press.

Smith, James A. 1991. The Idea Brokers: Think Tanks and the Rise of the New Policy Elite. New

York: Free Press.

“The Political Activity of Think Tanks: The Case for Mandatory Contributor Disclosure” 2002.

Harvard Law Review 115(5), 1502-1524.

Weaver, R. Kent. 1989. “The Changing World of Think Tanks.” PS: Political Science and Politics

22(9), 563-578.

3

In addition to the data described above, I will also be using Professor Shin’s data on individual opinion leaders’ newspaper quotations.