American think tank networks and expert debates around the Global Financial Crisis: Keynesian insurgents against austerity defenders

Abstract This article applies social network analysis and a Bourdieu-inspired understanding of think tanks to a particular historical period, the aftermath of the Global Financial Crisis of 2007–2009, when the neoliberal policy consensus was fractured and struggles to replace it by Keynesian policies ensued. The analysis finds that think tank and expert networks have played a key role in mobilising policy ideas, contesting or defending policy discourses. Experts advanced their proposals for policy changes in all those fields of power where they have gained a foothold. However, no consensus has been reached on monetary and fiscal policies, and the policy communities remain divided in these domains.


Introduction
After decades of mobilising neoliberal ideas through the media, universities, national and international organisations, right-wing think tanks in the US contributed to bring about in the 1980s a policy consensus around a set of policies broadly in line with a free-market, neoliberal ideology (Cockett, 1994;Marchak, 1991). In the 1990s, this shift triggered a wave KEYWORDS american think tanks; expert networks; fields of power of free trade agreements and substantially reduced the regulation of big business and financial institutions, privatised multiple activities formerly carried out by the state and increased tax cuts for large corporations (Burris, 2008;Wrenn, 2015). The 'common sense' underlying the neoliberal policy paradigm has also had an enduring impact on social institutions. For example, it initiated a trend to gradually diminish support for environmental protection, erect barriers to trade union organising and cut welfare spending. More importantly, the principles of solidarity and fairness, which until the 1970s guided welfare institutions, were gradually replaced by policies underpinned by the principle of individual responsibility and the notion of welfare dependency. These paradigmatic changes were intentionally pushed forward, and a particular set of think tanks eagerly contributed in naturalising them (see Harvey, 2007, pp. 40-44;Medvetz, 2012) so that neoliberalism increasingly permeated most US institutions (see Crouch, 2011).
The Global Financial Crisis from 2007/2009, which originated in the US (Krugman, 2013;Stiglitz, 2009), showed that this neoliberal consensus had reached a point at which it had become extremely unstable. Deregulation of financial markets encouraged, inter alia, by neoliberal think tanks in the years leading up to the crisis had led to speculation with increasingly risky financial instruments, especially in the housing markets. These then led to bank failures and subsequent tax-financed 'bailouts' to rescue 'too big to fail' banking and other financial institutions, and to a massive redistribution of wealth (Crotty, 2009).
In the wake of the crisis, a search for new policy ideas ensued. Among its consequences was a resurgence of Keynesianism among policy-makers and economists. In the US, new proposals in the closely related fiscal and monetary domains entailed a paradigm shift towards Keynesian, demand-side economics (e.g. reduction of interest rates, buying public and private assets by the Federal Reserve Bank and later expansionary fiscal policies). When near zero interest rates proved insufficient to stimulate growth, new expansionary policies such as increases in government spending and tax cuts to boost job creation and consumption began to be explored. While Keynesianism was moving towards the 'mainstream' again, it did so not unopposed. Experts affiliated with neoliberal think tanks argued against government spending because they saw deficit and debt reduction as the key to getting out of the Great Recession which had followed the financial collapse. They engaged in an ideological debate to challenge Keynesian policies (Lachman, 2016;Young, 2013), recommending a much greater commitment to fiscal austerity programmes.
Thus, even if the Keynesian revival did not last long in the US, no consensus has been reached in the policy community around austerity programmes and the struggle to define a new policy discourse continues.

Research questions and methodological framework
The struggle to define a new policy consensus in the US has entailed an intense mobilisation of ideas and knowledge by members of the policy community, notably think tanks. Networks are one of the main mechanisms through which policy experts from these organisations transmit ideas to influence state officials and public opinion in general, and transform dominant policy discourses.
But how do these networks emerge, mobilise ideas and shape the policy-making process? Complex forms of interlocking between think tanks, academic institutes and other organisations from inside and outside the state apparatus can emerge from the composition of the boards of directors or trustees, from how research teams are put together and from how research programmes are determined with the involvement of external stakeholders. The resultant web of interactions can be examined with the help of network analytical tools to answer different questions, but four are especially relevant for following the mobilisation of ideas and knowledge: (1) Do think tanks interlock according to their policy orientation?
(2) What are the characteristics of the most central think tanks in the network, that is, those having greater capacity to connect, transfer ideas and shape policies? (3) What other public and private agencies are involved in the network producing connections between the think tank field and other fields of power? And lastly (4) What are the main ideological struggles in the network to define a new policy consensus?
The research strategy followed several steps to answer these questions. First, a sample was assembled that included only those think tanks from the US ranked by McGann's (2014) lists as the top economic policy think tanks. McGann's lists include the most important think tanks in the world, according to country and issue (see introduction to this thematic issue for a discussion of his rankings). A total of 35 US think tanks was found in these lists, although only 33 with sufficient information available to discover their connections (Table 1 registers these think tanks, their assets, the number of links they generate among themselves and their ideological orientation).
Second, a database was constructed with the names of all think tankers, i.e. directors of the boards, trustees, fellows and researchers affiliated in 2015 with these organisations. Affiliation means, in this study, that they work for the think tank in a full-or part-time, paid or unpaid research capacity. The ratio to include these individuals is based on the assumption that is they who generate connections, participate in media debates and define research programmes.
Third, formal social network analysis was carried out. This methodology measures relational data between groups of social actors and organisations, in the case of this study think tankers and think tanks. 1 With the help of network analysis software -Pajek -the database was converted into matrices which were processed to discover different patterns of interlocking, particularly: (a) centrality of think tanks and think tankers. Centrality is defined as the number of connections they produce and the speed, or distance, at which they can connect with the rest of think tanks and think tankers; and (b) the intensity of connections, and the degree to which it corresponds to the ideologies they espouse. However, only individuals with two or more positions effect links, having greater capacity to transfer ideas and information between organisations and generate consensus. Only these linkers were included in the network analysis.
Fourth, once the main members of the network with two or more positions were identified, the biographies of these experts were examined in order to establish how they have moved throughout their careers between different fields of power, in particular academic, government agencies (ministries, advisory committees and councils), national and global corporations and financial institutions, national and international policy-making organisations and the media. This circulation between fields of power (the so-called revolving door) entails mobilisation of (and competition between) policy ideas, as Crotty (2009, pp. 577, 578) documents in the case of financial institutions.

1
Social network analysis has been employed to examine inter-corporate relations (Mizruchi & Schwartz, 1992). though a static type of analysis, longitudinal perspectives add some dynamism. However, it is a powerful instrument to analyse relational data at particular moments in time, as in the case of this article. Source: ideological orientation according to inside Government http://think-tanks.insidegov.com/ and FaiR http://fair.org/extra-online-articles/fair%e2%80%88study-think-tank-spectrum-2012/. *degree: number of connections generated by each think tank; **think tankers: total number of directors and researchers. Last, but not least, after briefly mapping the ideological orientation of American think tanks following the characterisation of Inside Government and FAIR, an analysis of coincidences and rifts between think tanks' experts around monetary and fiscal policies was carried out. 2 Inside Government is a government website that offers information about campaign finance and elections, government spending, think tanks and other issues and FAIR is a media watch group. 3 With the help of these forms of social network analysis, this article argues that many of the neoliberal and Keynesian policy initiatives and ideas put forward during and after the Global Financial Crisis were conceived and developed within the think tank network; that think tanks were instrumental in forming policy communities that pushed policy agendas forward, triggering a 'war of ideas' (Stiglitz, 2009, p. 211); that some of these organisations have, moreover, hosted the very same public officials later made responsible for implementing these agendas by George W. Bush and Barack Obama; and that until today no consensus has been reached in the US in hardly any economic policy domain.

Theoretical framework
Although some scholars (Medvetz, 2012;Pautz, 2014) acknowledge think tanks networks can play a key role in the process of policy-making, except for Burris (2008), little work has been done to apply a more systematic network analysis to the understanding of how they shape that process in the US. This article attempts to fill this void and apply network tools to make sense of the initiatives American think tanks' experts undertook to construct a new policy consensus after the Global Financial Crisis. Like Medvetz (2012), I draw on Bourdieu's (2005) concept of 'field of power' , but I aim to show that networks within and between different fields can facilitate the flow of policy ideas and knowledge throughout the structure of power, as well as the development of advocacy coalitions. Bourdieu (2005) defines field of power as the network of relations between organisations and agents competing to control resources in differentiated spaces (economic, political, cultural and symbolic). The structure in each field -that is, the pattern of relations predominating therein -guides the strategies followed by occupants (agents and organisations, in the case of this study experts and think tanks) to maximise social, economic, political and cultural capital. Each field represents an arena of struggle in which the structure of power and the predominant groups are reconstituted. But, as Bourdieu acknowledges, top elites easily move from one field to another and, occasionally, operate simultaneously in different fields, imbricating networks and fields.
Medvetz, in turn, focuses on the space think tanks occupy in the US structure of power. In his view, such space has become a boundary space, 'a hybrid subspace of knowledge production' (Medvetz, 2012, p. 42): a space where experts affiliated with think tanks -who have academic, political, corporate, media and scientific backgrounds -meet, interact and struggle for different objectives. Among these, two stand out: redefining the institutional rules that certify and legitimise the knowledge they produce and disseminate; and constructing dominant policy discourses on the basis of this knowledge (Medvetz, 2012).
While the Advocacy Coalition Framework (ACF) put forward by Weible and Sabatier (2006, p. 123) can help 'understand and explain belief and policy change when there is goal disagreement and technical disputes involving multiple actors' , Bourdieu's field theory can in addition account for the struggles within policy networks to define what knowledge is acceptable and how this knowledge flows in different directions.
Following Medvetz' application of Bourdieu's field theory, this article analyses, first, the characteristics of the field of think tanks in the US, using network analysis tools. Second, it examines the interaction of the field of think tanks with other fields. Third, the struggles between think tanks and experts to construct a consensus around a new policy discourse are studied. In the next section, I describe the main characteristics of the network, in terms of concentration of interlocks, centrality of think tanks and experts, intensity of connections and the degree to which these indicators evince similarities and differences in political orientation.

The American economic policy network
The group of 33 think tanks included in the sample has produced a complex set of connections among themselves and, as shown in the next section, with other public and private agencies. To analyse these connections, all the members of the governing structures -boards of directors/trustees -and the research staff, fellows and researchers listed on their web sites were included. The main criterion to build the database was not to leave out anybody with the potential of generating interlocks since a member of the board in one think tank may be an expert affiliated with another. Given the focus of the article, in the case of think tanks pursuing research in many policy fields (e.g. Brookings, Heritage, CATO, NBER), only experts dealing with economic issues were included (see Table 1 for the acronyms of the 33 think tanks). This means that in the case of the National Bureau of Economic Research (NBER), a total of 876 members occupying positions as research associates and faculty research fellows were included, while for RAND and the American Enterprise, Institutes 530 and 436 were included, respectively. However, as demonstrated below, while these groups seem large, the dynamics of the network depends on only a few individuals. These are the experts effecting most links and, consequently have greater capacity than those generating fewer links to broker interests and ideas, transmit knowledge relevant to the policy-making process and accumulate the social capital associated with the positions they hold and the links they produce. However, greater capacity by these individuals does not necessarily translate into coordinated efforts to define policy agendas.
According to this network analysis, interlocking is highly concentrated -a few think tanks are extremely well connected whilst the majority are not. A total of 5075 positions are held by 4836 individuals in the 33 think tanks of the database. Out of these, only 209 (4.3%) simultaneously hold two or more positions, 187 hold only two, 15 hold three, 6 individuals hold four and only 1 (Lawrence Summers) holds five. Out of the 33 think tanks in the database, three -Economic Policy Institute (EPI), Global Policy Forum (GPF) and Linktank -are not linked to the network. The 209 think tankers (directors and experts) who have two or more positions generate 13,400 interlocks; only 20 think tankers (less than 10%) generate 2248 (almost 17%). Going beyond the level of the individual and onto the level of the think tank as an institution, Table 1 shows that, out the 448 interlocks generated by the 30 think tanks connected in the network, five (less than 20%) concentrate more than 50% of the interlocking activity.
Three think tanks have the highest centrality, meaning they can interact more rapidly not only among themselves but with the rest of the network. By far the most central is NBER, producing a total of 85 links. This organisation, which has been characterised as 'non-partisan' (Dolny, 2013), brings together prestigious left-, centre-and right-leaning experts. It has a far-reaching web of connections with think tanks of different ideological preferences, but mainly with those classified as centrist and conservative. On the right, it shares 18 experts with AEI -i.e. these experts appear simultaneously as researchers for both think tanks -14 with HOOVER and 2 with CATO; in the centre, it shares 25 experts with the Peterson Institute of International Economics (PIIE), 11 with Brookings, 10 with the Center for Global Development (CGD), 5 with the Council of Foreign Relations (CFR), 2 with the New American Foundation (NAF) and the Levy Economics Institute (LEVY); in the centre-left, it shares only three experts with the Center for American Progress (CAP) and URBAN. The number of interlocks exhibits a more or less intense relationship between think tanks demonstrated by the thickness of the lines in Graph 1.
AEI, widely considered a conservative (Dolny, 2013;McGann, 2007) or right-wing (Burris, 2008) think tank, has 43 connections in the network. Besides those effected with NBER (18), AEI has generated seven interlocks with HOOVER, six with the Manhattan Institute (MI), three with CATO, Heartland and WINEP and two with ATLANTIC, ATLAS and CSIS. This means the most intense connections of AEI are effected with other conservative think tanks. A similar picture emerges with PIIE, coming next with a total of 41 interlocks. This think tank, classified as both centrist and non-partisan, tends to interlock with similar organisations. It shares eight links with CGD, seven with CFR, four with Brookings and three with Atlantic, all of them centrist or non-partisan. Both NBER and BROOKINGS function as policy brokers, linking think tanks on the right, centre and left and, as shown below, facilitating advocacy coalition building, as Weible and Sabatier (2006) would anticipate.
Thus, close inspection of interlocks reveals the weight of ideological resemblance, particularly among conservative and centrist think tanks, a trend which can be traced at least to the 1980s (Burris, 2008). Left-wing think tanks (liberal and progressive) do not generate that many interlocks among themselves or with think tanks of a different political orientation. This more marginal position of left-wing think tankers (see Graph 1) might be due to their preference for other mechanisms to influence policy, such as the circulation of think tank elites between different fields, which is examined next.

Circulation of experts between different fields of power
Network building is among the most important mechanisms to set a policy agenda in the economic policy domain and transfer ideas within policy communities. But networks are compounded by other mechanisms that play an equally important role in mobilising ideas, connecting knowledge with power and building consensus. These mechanisms include the revolving door, ad hoc discussions, hearings before Congress Committees, joint projects bringing together experts from different think tanks with policy-makers and journalists, among others (Abelson, 2004, pp. 228, 229). This article shall concentrate on the revolving door -that is, on the circulation of experts between different positions of power in other fields.
To discover how experts of the think tank network travel between different fields, movingand in practice implementing -policy ideas and to gauge the scope of this mechanism, I examined the biographies of all members of the 30 think tanks having two or more positions in the network. The aim was to identify the positions they have held throughout their careers in other fields, particularly in government, universities, corporations and financial institutions, national and international organisations and the media.
As Table 2  There is an intense circulation of elites throughout these fields, although certain positions cannot be held simultaneously. Usually, public officials have to leave, at least temporarily, the posts they hold in think tanks, universities, international organisations or corporations. It is common, however, to be simultaneously members of universities and think tanks, a practice that is encouraged by think tanks, which offer prestigious scholars grants or fellowships to carry out research projects in the fields these organisations are most interested in. For some individuals, think tanks are springboards for high positions in the public service, as well as places to return to upon leaving public office. Thus, at least 28 of the 209 members with two or more positions in the network have held a senior status in the President's Council of Economic Advisors, 13 of them presided over this Council and in 2015 were affiliated with NBER as research fellows. This is perhaps the most obvious way to link knowledge with power, but also to look at policy ideas from different standpoints.
The connections of American think tanks with top university professors not only meanas Medvetz (2012) contends -academic prestige and credentials. They also nurture a common language that validates the knowledge produced by these organisations and attests argumentative authority. Thus, at least 30 of the experts in the network have held academic positions at Harvard, 10 at Columbia and 7 at MIT, among other universities.
Not fewer than 35 of the experts have an active presence in print and electronic media, where they write columns and broadcast their views. Often, the 209 experts mix academic and journalist activities, as in the cases of Larry Summers, Kenneth Rogoff, Joseph Stiglitz, Paul Krugman, Dani Rodrik and Martin Feldstein, all of whom hold a high position in the ranking of Ideas, which is assembled considering the number of citations in RePec. 4 And although number of citations does not necessarily mean recognition, it shows they are part of the main debates.
Overlaps of American think tanks with international organisations producing and disseminating policy knowledge are frequent too, as can be testified in the positions held by at least 18 experts in the World Bank and 16 in the IMF. Experts also move to and from private national and international organisations, as is the case for 12 affiliated with the Trilateral Commission, 5 belonging to the Group of Thirty, 6 attending the Bilderberg Group Conferences and 2 the World Economic Forum.
Some experts build bridges between various fields. Several cases are particularly interesting in this regard, not only for the number of positions accumulated in the field of think tanks but also for their presence in other fields. One of the most visible members of the network is Joseph Stiglitz (third in the 2016 Ideas-REPEC list of most cited economists), affiliated with four think tanks (three centrist and one progressive), and generating 105 connections with other members of the network (see Table 3). Stiglitz, Nobel Prize in Economics (2001), has circulated through all fields but the corporate field. His positions have included membership on the advisory board of the Center for Economic and Policy Research (CEPR), a progressive think tank whose board includes Walden Bello and Robert Solow, a professorship at Columbia University, the office of chief economist at the World Bank and Chairman of the Council of Economic Advisers. In the last 15 years, Stiglitz has adopted a very critical view of IMF, World Bank and laissez-faire economists in general (Stiglitz, 2015).
Kenneth Rogoff (eighth in the 2016 list of most cited economists) is affiliated with four think tanks (all centrist and/or non-partisan), generating 118 interlocks in the network. He has circulated through the IMF, the academic and media fields (see Table 3). He is a professor at Harvard and writes a monthly column on global economic issues reproduced in over 50 countries with Project Syndicate. He defends free trade (including the TPP), a flat tax and anti-deficit policies.
Paul Krugman (24th in the 2016 list of most cited economists) is affiliated with NBER and PIIE. He was awarded in 2008 the Nobel Prize in Economics and has been a professor at Stanford, Yale, London School of Economics and Princeton. Since the beginning of the Global Financial Crisis, he has been among the main proponents of a Keynesian approach and has criticised the advocates of 'spending cuts and tax hikes even in the face of high unemployment' (Krugman, 2013, p. xi). In his column for the New York Times, he has argued for more financial regulation and public investment, particularly in education and infrastructure.
Lawrence Summers, one of the most influential economists in the US in the last two decades (29th in the 2016 list of most cited economists), has been affiliated with five of the most central think tanks, generating 137 connections with other members of the network and circulating across all surveyed fields, including the corporate field (see Table 3). In 1999, he became the Secretary of the Treasury during the Clinton administration, playing a leading role in the deregulation of the financial system, particularly the repeal of the Glass-Steagall Act and the approval of the Gramm-Leach-Bliley Act, allowing the incorporation of investment and commercial banks. During the Obama administration, he became influential in the response to the Great Recession while presiding over the National Economic Council. In the discussions for the American Recovery and Reinvestment Act (ARRA) -approved by the House and the Senate in early 2009 -he advocated tax cuts, while other economists (in particular De Fazio, Stiglitz and Krugman) argued in favour of more spending in infrastructure, education, health and policies to create jobs. In the academic field, Summers was a professor and president of Harvard University. In the field of international organisations, he was the Chief Economist of the World Bank (1991)(1992)(1993) and member of the Bilderberg Group. In the corporate field, he was the director of Citigroup and in the media, a columnist for the Financial Times and the New York Times. Martin Feldstein (13th in the 2016 list of most cited economists) produces 126 interlocks in the network and has revolved through all of the studied fields. In academia, he has been a professor at Harvard University. In the public administration, he presided over the Council of Economic Advisers during the Reagan Administration, where he was widely considered a 'deficit hawk' , calling for painful measures to reduce the deficit. He has been active in the corporate field too, as director of AIG, JP Morgan, Eli Lilly and TRW. Internationally, he is affiliated with the Trilateral Commission and has attended several of the Conferences organised by the Bilderberg Group.
These five economists have held central positions in the network (see Table 3) and have revolved through different fields, although only Summers and Feldstein (both austerians) have held positions in the corporate field. Besides, Stiglitz and Krugman have advocated Keynesian demand-side economic policies, challenging on several occasions the supply side policy proposals put forward by Feldstein, Summers and Rogoff, and even questioning the quality of their research. Stiglitz and Summers, for example, have engaged in policy fights since the beginning of Bill Clinton's administration. In 1993, Summers -then Assistant Treasury Secretary -wanted to accelerate financial liberalisation in South Korea, while Stiglitz -who chaired the Committee of Economic Advisors -favoured a more gradual approach. The austerity package forced by the IMF to respond to the crisis increased instability, and despite Stiglitz recommendations the IMF was pressured by the executive board (where Summers represented US interests) to pursue this course of action. Just before he was ousted from the World Bank, Stiglitz accused Summers of supporting Wall Street interests (Engdahl, 2000) and, more recently, for his emphasis on fiscal policies to create an attractive investment climate by controlling inflation, debt and deficits. In this way, he is part of what Krugman calls the 'confidence fairy' approach (that is, the idea that if fiscal deficit and debt are contained, confidence will stimulate investment) even if, as Krugman (2015b) contends 'the benefits of improved confidence failed to make their promised appearance' .
Both Stiglitz and Krugman have also been involved in heated debates with Rogoff. After Stiglitz criticised the IMF for its response to the Asian financial crisis (in his 2002 book Globalization and its discontents), Rogoff, who was then serving as economist in this organisation, addressed in 2002 an open letter to Stiglitz where he justifies the strategies followed in the region (Reinhart & Rogoff, 2013). Later, Krugman challenged austerians over the debt threshold thesis, according to which economies falter when debt surpasses 90% of GDP. Krugman accused them of withholding data and never dispelling the confusion provoked by a paper that seemed to justify austerity policies in the US and Europe. Moreover, he argued in The Guardian […] all of the economic research that allegedly supported the austerity push has been discredited. Widely touted statistical results were, it turned out, based on highly dubious assumptions and procedures -plus a few outright mistakes -and evaporated under closer scrutiny. (Krugman, 2015b) In his blog, the Conscience of a Liberal, 5 Krugman has pointed out time and again to the inconsistencies of Feldstein's policy proposals. He strongly opposed expansionary monetary policies, as they would produce runaway inflation. Feldstein's arguments against quantitative easing -Krugman said -were built on the grounds of inflation fears, even when inflation remained at its lowest.
The circulation of policy knowledge and ideas moves in different directions. Sometimes think tanks initiate the process with the introduction of key ideas, sometimes they elaborate on existing ones, but they can also play a more important role in propagating, legitimising or challenging ideas introduced by policy-makers. Thus, neoliberal experts (Young, 2013) confronted the set of ideas revolving around the Keynesian resurgence. Keynesian experts, in turn, challenged and disqualified austerity programmes. According to Krugman, All of the economic research that allegedly supported the austerity push has been discredited [and] Since the global turn to austerity in 2010, every country that introduced significant austerity has seen its economy suffer, with the depth of the suffering closely related to the harshness of the austerity. (Krugman, 2015b) In the next section, I map the ideological cleavages in the think tank network and discuss where the most important experts stand on specific policies. Differences and similarities will be sketched by examining the arguments of think tanks' most visible experts.

Ideological cleavages in the American think tank network
According to the classification of Inside Government and/or FAIR of the 33 think tanks in the database, 12 are predominantly conservative, 12 predominantly left-wing (liberal or progressive) and by far the largest number (25) are centrist (non-partisan or centrist) (see Tables 1 and 4). Although not strictly aligned with this classification, Keynesian economic policies are found among left-wing and centrist think tanks, while neoliberal policies are found among conservative, libertarian and centrist ones. Conservative and libertarian think tanks are predominantly neoliberal in the economic domain but in the social and political spheres they support traditional values.
This understanding of the US-think tanks ideological spectrum, guided by a rather simple left-right view of economic theories and policies, can be refined by fleshing out where think tanks stand on more specific policies. Considering the purposes of this article and space constraints, only the views of the most visible and vocal experts in the monetary and policy domains will be analysed. Their views are representative of the neoliberal and Keynesian policy communities, advocating austerity and expansionist policies, respectively, as the way out of the financial crisis. These policy communities struggled to redefine the dominant policy discourse in the think tank network.
In the monetary domain, the main divisions roughly align with patterns of interlocking. Advocates of hard currency policies, including those wanting a return to the gold standard, have opposed advocates of different kinds of stimulus packages, such as the Troubled Asset Relief Program (TARP) approved at the end of George W. Bush second term, the Recovery Act passed at the beginning of Barack Obama's Administration and most notably quantitative easing (QE). Adam S. Posen (president of PIIE and member of CFR and CGD) provided de intellectual foundation for QE and for adapting monetary policy-making and debt management to low interest guidelines throughout many countries. The QE idea quickly spread to most centrist think tanks, which -as shown in Graph 1 -are tightly interlocked. Joseph Stiglitz, Paul Krugman, Dani Rodrik, David Wessel, Joseph Gagnon and other representative experts of centrist think tanks (like PIIE, CGD, CFR, BROOKINGS, all with a high degree of centrality) actively defended QE policies that in the context of the financial crisis would promote growth. On the right, AEI and CATO strongly opposed QE and favoured hard currency policies. Desdmond Lachman (2016) -affiliated with conservative AEI -believed QE would lead to asset price bubbles, and several experts from MI (conservative) warned it would excessively enlarge FED's balance sheet, increase inflation and weaken the dollar. Lawrence White (2013) -senior fellow at CATO and policy expert at Heartland, strongly interlocked libertarian think tanks -contends that an 'automatic' gold standard would help tie the government's hands against printing money to cover its expenses. In contrast, Krugman (2015a) has argued against this standard, which in his view would further depress the economy.
In the fiscal domain, think tanks' position on tax revenue policies, debt, budget and deficit differs greatly, and ideological alignments are more complex, though they correspond roughly with the structure of the network. In this domain, the 'fiscal responsibility ideology' is very old and has spread throughout the political spectrum (Grosser, 2010). Right-wing think tanks (both conservative and libertarian) understand fiscal responsibility as anti-tax, anti-deficit, anti-debt and anti-government spending which translates into austerity and deflationary policies. The most serious problem in the federal budget, according to conservative experts -see Boccia (2016) and Bromund (2011), affiliated with HERITAGE and Capretta (2016) of the AEI -is the growth of entitlement programmes, particularly Social Security, Medicare and Medicaid.
Centrist think tanks and experts are divided between those endorsing anti-debt policies (see debate over Reinhart and Rogoff 's debt threshold thesis above) and those preferring government investment, especially when interest rates are low; and liberal and progressive think tanks, which argue that to reduce the deficit and increase spending in education, health and infrastructure, the budget should be balanced, increasing tax revenue, particularly from corporations and the wealthy. However, according to Krugman, fear-mongering about the size of the deficit has become a widespread strategy to slur left-wing proposals on government spending and raising corporate taxes. 'Unfortunately, policy-makers won't do the right thing, largely because they keep listening to fiscal scolds -people who insist that public debt is a terrible thing even when borrowing costs almost nothing' . Jared Bernstein (2016), senior fellow at left-wing CBPP and former Chief Economist, contends the deficit was aggravated by tax cuts during George W. Bush's administration and Obama's first term. In a testimony before the Congressional Budget Office, he insisted 'there is room for significant improvements in our tax code. Especially the business side of the code [… is] riddled with subsidies and loopholes' (Bernstein, 2016, p. 10). Thus, defining fiscal responsibility has become the main struggle in this domain.
A close look at Graph 1 shows that, except for SPN, all think tanks classified as libertarian (see Table 1) are a tightly knit group located in the far right of Graph 1; conservative think tanks (mainly HOOVER, MI, AEI and HERITAGE) also make up a group in the middle right; centrist think tanks are located towards the left side, showing less tightly knit patterns of interlocking. However, centrist think tanks focusing primarily on economic policies (PIIE, CGD, NBER and BROOKINGS) are also tightly interlocked. And last, liberal and progressive think tanks are more loosely interlocked among themselves and with the network as a whole. These patterns of interlocking facilitated ideological alignments in both the monetary and fiscal domains, which played a key role in the aftermath of the Global Financial Crisis. But the presence of policy brokers in the network also allowed cooperative efforts and the formation of advocacy coalitions around specific fiscal policies, as it is expected in the ACF (Weible & Sabatier, 2006). Brookings organised several discussion groups that included representatives of conservative, centrist, liberal and progressive think tanks to reach minimum consensus around problems such as the fiscal deficits, debt and QE (Brookings, 2014(Brookings, , 2015. This think tank also led several cooperative efforts such as the Joint Center for Regulatory Studies, the Fiscal Wake-Up Tour (Campbell & Pederson, 2014, pp.76, 77) and, in particular, two coalition of experts confronted around fiscal policies. First, one coalition bringing together experts affiliated with HERITAGE (Stuart Butler and Allison Fraser), BROOKINGS (Ron Haskins and Paul Cullinan), AEI (Joseph Antos) among other think tanks advanced the proposal 'Taking Back Our Fiscal Future, ' which recommended cutting back spending in Medicare, Medicaid and Social Security (Heritage, 2008). Second, a coalition including experts affiliated with CBPP, CEPR, BROOKINGS and AEI opposed this proposal because it '[…] could jeopardize the health and economic security of the poor, the elderly, and people with serious disabilities' (Aaron & Schultze, 2008).

Discussion
American think tanks and experts have constructed a tightly knit network that encompasses the most important think tanks at the left, centre and right of the ideological spectrum. Interlocking activity is highly concentrated in a few organisations and experts, denoting higher concentrations of social and cultural capital and greater capacity to mobilise policy knowledge.
Ideological differences between think tanks affect the patterns of interlocking, and vice versa. Conservative and libertarian think tanks connect more intensely among themselves; so do centrist and non-partisan ones, but they also connect with conservative think tanks and only marginally with libertarian, liberal and progressive ones. The intensity of connections means sharing common spaces within think tanks where ideas and knowledge can be transmitted, validated or revised.
These patterns of connections reveal a heterogeneous integration into the field of think tanks, where right and centrist organisations have become intensely intertwined, whereas left and progressive think tanks occupy a more marginal position. This is not only because their assets are scarcer -and, consequently, the number of researchers and connections lower (see Table 1) -but also because experts affiliated with these think tanks prefer other mechanisms to influence policy, such as the revolving door and active presence in universities, the media and at congressional hearings. These findings could be expanded in future research with selective interviewing of think tankers.
The mechanisms to connect knowledge with power are manifold, but one of the most important is the revolving door, whereby experts affiliated with think tanks circulate through universities, making academic knowledge available to the policy-making community and giving policy proposals a scientific outlook. Besides, many of these experts circulate through national and international policy organisations, bringing their ideas and ideological convictions with them; they disseminate policy knowledge in the media to persuade public opinion and challenge their adversaries; and they imbricate all these different fields, as Bourdieu suggests.
Disagreements between top economists in the US reveal the struggle of Keynesian insurgents against austerity advocates to gain credibility and to convince policy communities that their proposals are better suited to overcome the Great Recession. But they also reveal the efforts by Keynesian experts to redefine the rules that legitimise policy-making knowledge, one of the main motivations driving think tanks. These disputes are often accompanied or followed by cooperative efforts and advocacy coalitions, whereby experts try to reach a consensus. Patterns of competition and cooperation taking place not only in the media, but in workshops, discussion groups and congressional hearings could be further analysed with the tools offered by the ACF to understand more thoroughly the dynamics of the network and the strategies preferred to influence policy-making.