Abstract
Scholarship on the Federal Reserve argues that banks do not lobby on interest rates, holding that banks would not waste scarce resources advocating for a position that central banks already know they embrace. I challenge this assumption theoretically by arguing that banks have different interests and influence on monetary policy, with the largest banks having the most influence. I run a firm-level panel regression on a novel dataset drawn from the balance sheets of banks that have sat on the Federal Advisory Council (FAC) between 2003 and 2019. The two main contributions are that for all sizes of banks, only the four largest banks that sit on the FAC influence the Federal Reserve on monetary policy; this influence can lead to the Federal Funds rate falling as well as rising. However, for influence to occur, banks must be sitting on the FAC, and the salience of financial regulation must be low. This shows how the lobbying of very largest banks has an outsize role in setting policy in times of quiet politics, regardless of outside options.
Does a central bank’s interest rate policy respond to banks’ interest rate preferences? In this paper, I challenge two common ideas regarding bank lobbying and the political power of finance: first, banks do not have homogenous interest rate preferences. Instead, banks’ interest rate preferences for higher or lower rates depend upon their precise business models—for example, being geared more towards investment or retail banking—and their specific balance sheet compositions. Second, that banks lobby on monetary policy, and that they can use the unique access that is provided to them by their membership of the Federal Advisory Council (FAC) to lobby the Federal Reserve (hereafter the Fed).
Crucially, an important factor within the multi-causal story behind how a bank can successfully exert influence upon a central bank seems to depend on having a particular constellation of structural prominence and instrumental power. Very large banks exhibit more structural power due to their size (their “prominence”), not only because of the gravitational effect their balance sheets can exert upon the financial system, but also because a central bank recognizes the consequences of its policies upon these very large banks, which it depends upon to transmit monetary policies into the economy. Banks, in the most benign view of their policy-advice role, have highly specialized knowledge, given that much of the impact of central bank policy is frontloaded onto them. Therefore, their unique insight is highly valuable to central banks for the efficient operation of financial markets, as banks can provide a central bank with specific expertise beyond the economic advice that their own advisory board can easily get from its own economic research teams.
I empirically explore these arguments by analyzing the influence of American banks on the interest rates policy of the Fed. That is, I estimate the effect of banks who are members of a leading banking industry advisory group, the Federal Advisory Council (hereafter the FAC), during the period 2003 to 2019, and who in this position make policy recommendations to the Fed. I infer their preferences regarding interest rates based on five selected indicators which are most affected by interest rate fluctuations and which I draw from the banks’ balance sheets. Accounting for the size of the banks (small, large, and prominent, that is, one of the “Big-Four”), and their membership of the FAC over time, the selected panel of US banks is used to predict the probability of these banks’ interest rate preferences influencing the Federal Funds (FF) rate, a benchmark rate through which the Fed communicates its interest rate sentiments, via lobbying at the FAC.
Culpepper and Reinke (2014) argue that though large banks in the US system enjoy a privileged position, their lack of exit options means that their structural power was inadequate to fend off state bailouts following the 2008 financial crisis (hereafter referred to as “the crisis”). This paper shifts the lens to test empirically the political power of finance over monetary policy, before and after the crisis. These timeframes (2001-2007, and 2013-2019) were periods of “quiet politics” where the salience of financial regulation was low, and there was a greater opportunity for regulators to be amenable to the interest of banks (Baker, 2010; Culpepper, 2011).
A bank’s preference towards interest rates and how it might lobby for those preferences, is a gap in the literature that has only appeared because of structural changes in modern economies due to financialization. Though the process of financialization can be traced back to the 1970s, the period where change in the banking industry was most intense has been since about 1990 and especially after 2000 (Krippner, 2011). Prior to 1990, banks’ monetary policy preferences were posited as consistently preferring higher interest rates (Woolley, 1984: 76). This was when the “3-6-3 rule”—borrow at 3%, lend at 6%, and be on the golf course for 3 p.m.—was a well-worn joke in banking circles, though the industry by the 1980s was much more competitive and quickly deregulating (Walter, 2006). If one defines financialization as the growing importance of financial activities and profits for national economies, the effects in the form of a preponderance of the financial industry over all other economic activities began in the 1990s (Krippner, 2011).
Crucially, this growing importance has led banks to become more politically powerful corporate actors (Hacker and Pierson, 2011; Johnson and Kwak, 2010). This is not to argue that banks necessarily act in concert—rather, and as I argue in this paper, lobbying at the FAC hews more closely to the “disorganized combat” analyses of the political power of banks (Woll, 2016). The largest banks in the US can have preferences that conflict with each other, related to their specializations in either retail or investment banking, and given the political power of large banks, it is therefore important to scrutinize fora in which banks make policy recommendations and the decisions taken there in order to measure their political influence.
One such forum is the FAC, comprising twelve members of the financial industry, whose role is to consult and advise the Fed “on all matters within the Board's jurisdiction” (Federal Reserve, 2016). The FAC’s quarterly advice paper to the Fed usually incorporates a mix of recommendations regarding current loan and financial market conditions, regulations, trade policy, and topical issues that affect the industry, closing always with monetary policy advice.
One might argue that, given the size and importance of the “Big-Four,” attendance at the FAC would not be necessary to have their opinions heard. The meetings could be demoted to a brisk videoconference with junior Fed staff and still adhere to the spirit of the statutory requirements, if not the letter (Federal Reserve Interview 1, 2017). Yet, they take place in Washington D.C. with the most senior banking officials and Fed staff in attendance. As emphasized by a senior FAC representative, the FAC has not been relegated in importance over time but rather continues to rank at the Fed just below the meetings of the Federal Open Market Committee (FOMC) in terms of importance, with the full Board of Directors in attendance at all meetings.
The literature on the FAC’s impact on US monetary policy extends to three authors: First, Havrilesky (1990) argues that Fed monetary policy responds to the signals from the banking industry. Second was Weise (2008) who concurred, holding that monetary policy matches the preferred stance of the banking industry (but only after 1979, not before). Both authors were interested in the periods preceding and following the “Volcker shock,” with Havrilesky studying the years between 1973 and 1985, and Weise taking the data up to 2000. However, what was not examined by these authors was the fundamental structural changes brought about by intense financialization in the 1990s onwards. Instead, they treated the banking industry as a relatively homogenous entity.
Furthermore, even though these articles assessed the impact of the FAC on the Fed’s monetary policy activities and found there to be a positive effect, the wider literature on bank lobbying argues that this type of channel is ignored by the financial sector, focusing instead on the banking committees and the tax-writing process in Congress (Lavelle, 2013: 157). Therefore, the paradigm shift that is financialization, with its implications of heterogeneous policy preferences for finance, ensures that this paper provides a timely update to not only the literature on the FAC by not treating the banking industry as a unitary actor, but also on bank lobbying and the creation of monetary policy more generally.
Employing a novel dataset and innovative measure of interest rate preferences, I contest the position that banks sitting on the FAC do not exploit their privileged position on this advisory board to lobby in favor of their interest rate preferences, (Woolley, 1984: 76-77).
My findings suggest that only the four largest banks in the US, the structurally prominent banks (Young, 2015) that are even larger than those globally systemically important banks deemed “too big to fail,” are more likely to successfully further their interests through lobbying at the FAC. In other words, even the four largest banks’ (but only those four) structural power—their prominence—must be augmented by their instrumental power (here activated by membership of the FAC) if they want to be heard. My results also suggest that though the structural power of prominent banks waned during the financial crisis, there is evidence that it returned in earnest for those same banks in the post-crisis period. Furthermore, my analysis implies that for prominent banks that are highly financialized, there are clear situations where those banks prefer lower interest rates, sit on the FAC, and have their preferences met the by the Fed, even when the Taylor Rule would predict that interest rates would be set higher.
The following section expands on how the changing material interests of banks brought about the structural changes to modern economies that is financialization. The next section shows how this shift in material interests feeds into how banks’ interest rate preferences are diverging. A section delineating economic orthodoxy, exemplified by the Taylor Rule follows, before outlining the mechanism through which finance generally exerts it political power. Hypotheses follow, before details about the data and methodology are provided. Following that are the results, then conclusion. The study opens new avenues for the study of bank lobbying and the political power of finance.
Banks’ political power, their interest rate preferences, and the Taylor Rule
Financialization can be understood as both an analysis of the current stage of consistently evolving forms of capitalism, or as the quantitative and qualitative evolution of the centrality of finance and the financial sector (Sawyer, 2013). These viewpoints are not mutually exclusive, but this paper focuses on the latter. In other words, this paper will focus on financialization as the shift from industrial to finance capitalism in developed countries, that began in the 1970s, which highlights the “pattern of accumulation in which profit making occurs increasingly through financial channels rather than through trade and commodity production” (Krippner, 2005: 14). The result of this pattern of accumulation was the financial industry becoming not only the largest and most important sector of the global economy (Krippner, 2011), but itself becoming a truly global phenomenon (for an overview, see Van der Zwan, 2014).
The process of financialization shifted the US banking sector from a “patient finance” model to one—driven by deregulation, financial innovation and greater competition—with shareholder value as the dominant corporate objective which emphasized short-term gains, and prioritized high-risk, high-return activities (Crotty, 2005). The process of deregulation culminated in the repeal of Glass-Steagall, a law passed in the US in 1933 that forced investment and commercial banks to split into separate entities, thereby greatly reducing the risk of losses to “Main Street” savers. The repeal of Glass-Steagall, formally known as the Gramm-Leach-Bliley Act of 1999, allowed banks to develop further financial services activities distinct from traditional banking. Large banks capitalized quickly on this repeal to expand into non-traditional activities such as securities underwriting, insurance sales, and retail brokerage, all of which produce non-interest income. As the non-interest income of large banks in the US doubled (Yin and Yang, 2013), the banks have moved into areas where, in some cases, it made financial sense to act as borrowers rather than solely as lenders. And thus, the stereotype of the conservative banker making prudent profits from sensible loans, as envisaged by Woolley (1984), looked increasingly like a distant memory.
Large banks’ changing material interests and motives, as financialization has entrenched, has created a division between large banks (including prominent banks) and small banks. The move from relationship-based, patient banking (Culpepper, 2011) to the shareholder-value model, is illustrative of this division. For example, two divergent non-interest income strategies can be observed in US banking activity (DeYoung and Rice, 2004). First, for large banks, the most profitable strategy is to take advantage of their economies of scale in the creation, servicing, and securitization of consumer loans. Competition is fierce in this market, so profits are not derived in the main from interest income, but rather the fees that can be charged for providing the services. Since the repeal of Glass-Steagall, non-interest income is increasingly derived from non-traditional sources, such as securities underwriting, insurance sales, and retail brokerage (DeYoung and Rice, 2004). Second, and conversely, smaller banks operate in local markets where relationship-based banking is important, but also where the costs for depositors to switch to rival banks are high. Interest income continues to be the main profit driver for them. Thus, non-interest income is either less important for small banks, or unattainable because of their lack of economies of scale.
Furthermore, Esposito et al. (2015) and English et al. (2012) find that bank stock prices decline significantly following an unexpected increase in the level of interest rates. Declining share prices are taken very seriously, in the dominant shareholder-value model where short-term profits driving share-price rises and dividend pay-outs are the sine qua non of corporate goals (Lazonick and O'Sullivan, 2000). Therefore, a firm might lobby against a short-term interest rate hike if it could potentially cause a fall in their share price. In other words, the shareholder-value model that dominates large banks is the antithesis of the relationship-based model that drives smaller banks, leading to the conclusion that pronouncements that banks always want higher interest rates (Woolley, 1984) should thus be taken with more than a grain of salt.
Banks’ interest rate preferences
Three types of non-interest income are significant regarding interest rates: valuation effects on securities; hedging through derivatives; and fees and commissions, which represent more than 60% of total non-interest income for banks (Borio et al., 2015). Types of fees include those directly linked to lending and deposit activity (e.g., credit lines, transaction services), and those related to more investment-banking-type activities (e.g., trading, mergers and acquisitions, and market making). In times of lower interest rates, professional services are more in demand because portfolio management is more necessary when the option for sufficient interest rate income is not available to clients (Albertazzi and Gambacorta, 2009). Thus, I argue that banks who provide services, and who profit more from non-interest areas of banking, may lobby for lower interest rates.
Banks must finance their investments via some form of debt, traditionally done in the form of deposits, but nowadays also done through various securities sold on the open market. When market interest rates fluctuate, so do the costs for banks to fund their activities; whereas retail banking prefers low inflation, investment banking conversely prefers low interest rates (Wheelock, 2016). Though interest income is an important driver of small banks’ earnings, the largest investment banks fund their commercial activities through loans, and therefore seek low interest rates on those loans.
Higher interest rates are consistent with higher rates of loan defaults. Banks with a high percentage of non-performing loans should be expected to be particularly sensitive to interest rates being raised. In the aftermath of the crisis, FAC members’ balance sheets would have included many non-performing loans; some banks may have preferred to “extend and pretend” (Barseghyan, 2010; Borio et al., 2015), delaying the raising of interest rates until non-performing loans could be turned around, or worked off their balance sheets.
As interest rates fall, bank profits swell due to short-term gains in the value of collateral and lower default risk on loans repriced in response to interest rate decreases (IMF, 2017). Rising interest rates can also affect profitability through two mechanisms: via liquidity spirals, where weakening net worth increases bank risk aversion, which in turn depresses the market value of assets and lending volumes; and via disinflationary spirals, whereby the value of cash increases due to its “safe asset” reputation (IMF, 2017, citing Brunnermeier et al. (2016)).
Bank numbers in the US have been dropping steadily, from 14,500 in the mid-1980s to 5600 in 2015, and bank mergers have been driving the decline (Kowalik et al., 2015). Larger banks benefit most from the scale economies resulting from efficient management of wholesale funding costs (IMF, 2017), and low interest rates make mergers and acquisitions cheaper. Additionally, the possibility for economies of scale derived from more efficient deposit management would point towards consolidation as an attractive proposition. Both these scale efficiencies are advantageous for larger banks, and one might wonder whether larger institutions might have used low interest rates to squeeze their competitors, or even acquire them completely. It is precisely this type of strategic interest rate advocacy—calling for rates to be lower in a market where only the fittest (or most diversified) survive—that this paper is interested in. But to show that lobbying is effective, it is necessary to demonstrate where advocacy diverges from economic orthodoxy. This orthodoxy is exemplified by the Taylor Rule.
The Taylor rule
To achieve its interest rate goals, the Fed’s Federal Open Market Committee seemed to follow the so-called Taylor Rule under Chairman Alan Greenspan’s tenure (1987–2006) (Meltzer, 2014); his successor Chairman Ben Bernanke (2006-2012) argued that he hewed closely to an only slightly modified Taylor Rule (Bernanke, 2015). The Taylor Rule is a formula designed to provide recommendations for how short-term interest rates should be set by a central bank, to attain its statutory objectives. It is a composite measure that includes standard economic indicators, while also including an economic performance logic.
If there is diversity in banks’ interest rate preferences and some banks have privileged access to the Fed through the FAC at specific time periods, it is worthwhile to assess even small adjustments to Fed policy and to examine whether these can be linked empirically to bank interest rate preferences. The Taylor Rule tracks the Federal Funds rate closely, as shown in Figure 1, and therefore these small adjustments may be marginal, that is, merely ten or twenty basis-point movements away from the projected Taylor prescription. But it is here that we can uncover the influence of banks. If we can identify such a discrepancy that cannot be attributed to the economic indicators that are available to the Fed already, after careful exploration, we may infer that this is, at least partially, due to the unique preferences of those banks that sit on the FAC. To help explain how preference attainment comes about, the next section will detail the sources and consequences of banks wielding their political power. The Taylor Rule as a predictor of the Federal Funds Rate. Source: Federal Reserve Bank of St. Louis
The political power of finance
Young (2015) shows how the distribution of firms in modern economies track a “power law,” with a few, highly prominent firms, at the top. The US case shows that there are four such banks operating at this level: Citigroup, Bank of America, Wells Fargo, and JPMorganChase. As such, I argue that it is these prominent banks that should be most successful in lobbying the Fed on monetary policy. The mechanisms through which members of industry generally shape outcomes to meet their preferences are their instrumental power (Miliband, 2009; Mills, 1956) and their structural power (Block, 1977; Lindblom, 1977). Instrumental power encompasses the numerous means, distinct from the core functions of the firm, through which business influences politics (Culpepper and Reinke, 2014). It is normally implemented via lobbying activities or campaign donations. Structural power, in turn, refers to the privileged role of those private-sector firms, because the states in which they are located rely upon these firms to provide employment and create economic growth. Governments may privilege these industries or firms by favoring them with key roles in policy deliberations, ensuring that their interests are safeguarded. This is particularly the case for financial firms, given how financial activities have now displaced manufacturing and other forms of industry and business as the dominant force in the global economy (Bell and Hindmoor, 2016; Krippner, 2011).
Firms have more consistent and systematic influence when they display both types of power (Fairfield, 2010). Yet, both instrumental and structural power can provide them with the means to thwart, or cultivate, the type of policies they choose. Fairfield argues that possessing instrumental power involves having the capability for deliberate and often collective action in policy-making fora, whereas structural power entails apolitical, market-coordinated decisions in the economic arena. One might hence conclude that the financial sector can only yield structural power if firms coordinate. However, I argue that the structural power wielded by Big-Four firms does not depend on market-coordination—Big-Four firms not coordinating is a tactic that is employed by US banks even during a financial crisis (Woll, 2016). What sets prominent banks apart is that their structural power—their unique position within finance and therefore the economy at large—makes them not only “too big to fail” but also gives their singular instrumental actions the weight of a collective action by a sector of smaller firms. In other words, their structural power enhances their instrumental lobbying power to an enormous degree.
This paper’s focus within the structural power literature is to build upon Kevin Young’s (2015) structural prominence theory, which argued that policymakers treat prominent firms differently and are more attentive to their needs in the policy-creation process. Yet he found that there was less evidence that structurally prominent firms are successful in preference attainment if in opposition to the stance taken by the body that is being lobbied. I will argue that when banks exhibit a preference for interest rates that contrasts with that suggested by the Taylor Rule, and if there is a corresponding predicted preference attainment, then this suggests that prominent banks can overcome the status quo bias that Young surmises may be at play when prominent firms are successful in preference attainment.
Turning to instrumental power, much has been written on the political power of banks, specifically their lobbying activities (Mahoney, 2007; Posen, 1995; Wright, 1996). Such instrumental power is derived from providing expertise on complex financial instruments, through lobbying using the substantial funds available to Wall Street firms (Johnson and Kwak, 2010), onto more subtle forms such as the concept of the “revolving door” from Wall Street to regulator, and back to Wall Street (Bó, 2006; Cornaggia et al., 2013). Instrumental power can further encompass firms’ informal social ties to policy-makers and regulators (Seabrooke and Tsingou, 2015), or firms providing policy advice, such as what happens quarterly at the FAC.
I argue that banks exercise their instrumental power at the FAC via their quarterly meetings with the Fed, where they provide policy advice—although those who sit on the council claim that their advice is given with no ulterior motive that might benefit themselves, even going so far as to issue a motto of “not talking your own book” when giving advice at council meetings (Federal Reserve Interview 1, 2017). But it is not enough to merely point to their giving of policy advice at Council meetings as evidence for the wielding of instrumental power. As outlined above, size is the single most important factor in determining the effect of banks political power. If it were just instrumental power that would suffice, then just being in the room at an FAC meeting, being any of the 12 banks that advise the Fed on Council meetings days would be enough. I argue that instrumental power needs to be augmented by structural power (arising from firm size) to be relevant.
The core argument of this paper is that banks have interest rate preferences, that these preferences can be for lower rates, that they will advocate strategically for these preferences at the FAC, and that their size should matter with respect to their success in lobbying. Though all banks have some structural power—derived from their privileged position in the economy—larger banks should have more power, commensurate with their structural prominence.
Turning first to structural prominence, and building on the power of finance literature, I take the position that the big-four US megabanks, those that are structurally prominent in the US banking industry, but crucially not those that are merely globally systemically important banks, can be successful in having their interest preferences met. This is even when they have an interest rate preference that is different from what the Taylor Rule would suggest the Fed do. I refine structural prominence to include only the four largest, most prominent, banks in the US, highlighting quite how contingent structural power is on relative size. I also suggest that the structural power of the biggest banks in the US is entrenching since the financial crisis ended, proposing that this period of quiet politics post-crisis is presented with more powerful banks than before the crisis started.
Thus, the Fed, interested as it is in the health of its central monetary policy transmission mechanism, is open to the interest rate preferences of large banks being accommodated. Therefore, it would seem plausible that the desires of prominent banks would be taken more seriously, and perhaps, given primacy over other pressing sectoral concerns.
Against this, it is reasonable to expect that banks use the privileged position they occupy in fora such as the FAC to convey these preferences and employ their instrumental power to augment their structural power. I hold that the FAC is a prime location for banks to exert their instrumental power through lobbying and their policy advice. I also expect that it is the combination of structural power (prominence) and instrumental power (FAC membership) that will lead to the largest predicted effect in the model. Therefore, the first hypothesis reads that:
Banks that are prominent and sit on the FAC will have a larger predicted effect on the Federal Funds rate. Additionally, it is useful to see how this influence of banks at the FAC might vary over time. According to Baker (2010), there is a lifecycle of regulatory capture, whereby in boom periods regulators feel unable, or are unwilling (Pagliari and Young, 2015; Seabrooke and Tsingou, 2009), to produce a narrative of events that would run contradictory to feel-good stories that emanate during high-growth periods. Counter-cyclical narrative issues also have little public salience, given that the distributional issues that build before a crash happens are little understood by the public. This leads to an overall more permissive environment, which I expect banks would use to lobby more successfully for their preferences. After the crash, however, regulation becomes high in public salience, and banks should find lobbying more difficult. Yet, this period of high salience is limited, so we should see salience waning after the financial crisis period of 2008 to 2012 has ended. This decreasing salience should in turn allow for an increased predicted effect post-crisis. Accordingly, the second hypothesis reads:
Banks that are prominent and sit on the FAC will have a larger predicted effect during non-crisis periods on the Federal Funds rate. Turning to how the interest rate preferences of banks may be sometimes for lower rates, I argue that though interest income remains a primary driver of profits for the banking industry as a whole, it is also true that the largest banks have diversified such that non-interest income is potentially as important for them, if not more so. Also, given the perilous position that some financial institutions found themselves in the lead up to and following the financial crisis, banks who need to borrow to close funding gaps, or who wish to “extend and pretend” until bad loan books can be deleveraged or until they come back above water, may lobby for lower interest rates, as depicted in the third and last hypothesis:
When comparing before the financial crisis to after, the negative predicted effects on the Federal Funds rate for banks that sit on the FAC will be larger after the crisis than before.
Data and methodology
To explore banks’ influence on the Fed rate, I employ a panel generalized least squares (P-GLS) regression, with clustered fixed effects. The dependent variable is the FF rate, the main instrument of the Fed to implement monetary policy (St. Louis Fed, 2024). This variable is differenced following a Dickie-Fuller test for stationarity. Given that the dependent variable (the FF rate) is a monthly series, it was necessary to interpolate the independent variables, which were quarterly series, using cubic spline interpolation. Cubic spline interpolation allows for a good cross-sectional fit in a wide range of settings, especially when approximating economic functions (Bowsher and Meeks, 2008).
I measure banks’ interest rate preferences through three latent factors derived from five balance sheet items: Tier-1 leverage ratio, Loan-to-Deposit Ratio, Texas Ratio, Net-Interest Margin, and Non-Interest Income Ratio (see Factor Analysis below and Appendix Tables A6 and A7 for detailed descriptions and summary statistics). Data about the balance sheet items comes from the Quarterly Call Report that each FAC member submits to the Federal Deposit Insurance Corporation (FDIC). These reports were collated by a private financial services information company, BankRegData (2024), who produce a searchable database. To ensure replicability, a random sample of reports from BankRegData were double-checked against the FDIC original records, and BankRegData’s database was consistently accurate. Since reliable banking data is available from the FDIC from 2003 onwards and variation in interest rate policy was minimal during the crisis, I distinguish between three periods: pre-Crisis (January 2003 to December 2007); Crisis (January 2008 to December 2012); and post-Crisis (January 2013 to December 2019). Giving a precise endpoint to the Crisis is not possible, but by the end of 2012, quantitative easing tapering was being announced in the US, and in the Eurozone, European Central Bank President Mario Draghi’s famous “Whatever it takes” speech in July of 2012 was bearing fruit.
The moderating variables are membership of the FAC, and a bank’s size (small, large, and prominent). As regards membership, 56 banks have sat on the FAC between 2003 and 2019. Some sat on the council for four quarters, some for up to 32 quarters. Membership data of the board of the FAC was obtained using the Fed’s annual reports, which lists the FAC’s members’ names and bank.
The “Large” bank indicator is comprised of those banks which are classified by US and international financial regulatory authorities as “Globally Systemically Important Banks” (GSIBs), but who are not one of the Big-Four. Big-Four banks, though also classified by regulatory authorities as GSIBs, are of an order of magnitude larger than non-Big-Four GSIBs. There are four banks in the US that fit this concept of prominence, namely, Citigroup, Bank of America, Wells Fargo, and JPMorganChase. In 2014, these “Big Four” had approximately $4.6 trillion in deposits between them—almost 45% of the total deposits for all US commercial banks (Nasdaq.com, 2015). These Big-Four banks comprise the “Prominent” bank indicator. The “Small” bank indicator are those banks who fall into neither of these two categories. The three indicators are operationalized as a period when one of these banks sits on the FAC board.
Following Brambor et al. (2006), I include all constitutive terms in the panel interaction models, compute robust clustered standard errors, and rather than rely on the regression table outputs, I focus on the predictive margins, also known as marginal effects graphs in interpreting the results.
A predictive margin is a statistic used to estimate the average outcome or response variable for a specific group or scenario, while considering the influence of other variables in a statistical model. I use predictive margins to estimate the average predicted response for certain interest rate preferences across different groups or scenarios (small/prominent bank; sitting/not sitting on the FAC). The use of margins has several benefits that particularly apply to their use in panel data with interactions. I employ three-way interactions, which means the effect of one variable changes depending on the level of the other two variables, rendering the simple coefficient of the interaction (as reported in the regression tables) very difficult to interpret correctly.
To create the graphs, I calculate marginal effects at representative cases (MERs), which show how the effects of variables (in this case interest-preference factors) vary by other attributes (in this case the size of the bank and its membership of the FAC). The representative cases are thus a range plotted using the minimum and maximum values of the interest-preference factors. The minimum and maximum values which bookend each marginal effects graph are the minimum and maximum values that the factor scores can take, in turn derived combining the original variables (Net-Interest Margin, Tier-1 Leverage Ratio, etc.) that make up the principal component.
Following this, by using the Stata commands margins and marginsplot, it is possible to see how the marginal effects of these interest-preference factors on the FF rate differs across that range of values (Williams, 2015). That is, it is possible to predict what the effect on the FF rate is at absolute maximum value, minimum or any value in between that has presented in the time series for these banks. Therefore, for a given factor, if the score is very low (indicating a preference for lower interest rates), we can hold the size of the bank and its membership on the FAC constant, and say that if a large bank is sitting on the FAC and presents with this low score for a particular interest-preference factor, we can look at the graph to see the predicted effect this would have on the FF rate.
For reasons of parsimony, two controls were chosen, namely a modified Taylor Rule and the unemployment rate. The unemployment rate is a seasonally adjusted monthly index. The Taylor Rule is a modified one that is suggested by former Fed Chairperson Ben Bernanke as one approximating the Taylor-type rules he himself, and his successor Janet Yellen, followed. The equation for calculating the Taylor Rule is
Factor analysis
Using a factor analysis, I determine a latent factor or factors that allow for approximating the interest rate preferences of banks based on specific balance sheet components. Given that all five indicators should measure one concept (interest rate preference), factor loadings were estimated using principal-component factor analysis (which is a distinct method, not to be confused with principal-component analysis). The factor loading threshold chosen is 0.4 as per Acock (2018), and the oblique rotation method oblimin was selected to allow the indicators to correlate.
Factors Loadings.
Note: (+) indicates a reversed scale. Positive values indicate a preference for higher interest rates, and vice- versa. The rotation method is Oblimin (Orthogonal). The loading threshold is +/− 0.4 (indicated in bold).
Tests on the indicators in the pre- and post-crisis time periods passed Bartlett’s test of sphericity (p-value = .00), which tests the hypothesis that the chosen indicators are unrelated and unsuitable for detecting latent structures. Additionally, in tests on both time periods, a high Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy (pre-crisis KMO = 0.8; post-crisis KMO = 0.7) was calculated and supports suitability for factor analysis with values greater than 0.6.
The factor that is returned for the pre-crisis period should be interpreted as a scale, running from the theoretical minimum and maximum points of the factor, based upon the data provided from the balance sheet variables. In this case, the factor scale should be interpreted as having banks with problematic loan-to-deposit and Texas ratios ranking lower and towards the negative end. At the positive end are banks where conventional forms of profit earning that tend heavily towards interest-based forms of income (non-interest income scale being reversed), and with stable funding (based on a positive tier-one leverage loading), are represented. This factor may also be interpreted alternatively as a scale running from risk-inclined banks on the negative end, to risk-averse banks on the positive end. That there is so much diversity contained within one factor can be explained by the fact that, given the repeal of Glass-Steagall in 1999, by the beginning of our sample in 2003, the process whereby multiple banking functions were brought under one roof had already begun. This factor is labeled traditional because though it leans heavily towards interest-earning as the dominant profit center, and thus would indicate a preference for higher interest rates, there is conflicting preference for lower interest rates at the negative end of the scale, indicated by the negative values for loan-to-deposit ratio, and Texas ratio. Thus, we can see the factor as telling the story of the pre-crisis period, with banks having conventional interest preferences based on their profit-centers, but some leaning towards dragging rates down, because of their worrying balance sheets.
For the post-crisis period, two factors are returned, labeled retail and investment to indicate the dominant profit centers of the type of banks they represent (interest income in retail, versus non-interest income in investment). These labels should not be taken as being overly deterministic of precisely what a bank’s corporate approach might be, but rather it provides a useful heuristic of the macro preferences between types of banks.
The retail factor has some similarities to the traditional of the pre-crisis period, and the two have a relatively high correlation of 0.6 (see Table A8 in the appendix). This is because this factor also loads highly on positive values for the net-interest margin indicator and non-interest income (scale reversed) indicators, and the tier-one leverage ratio indicator. Both suggest traditional interest rate preferences, and conservative and prudent balance sheets. At the negative end of the scale, loan-to-deposit ratio dominates, suggesting that this factor also includes banks who have balance sheets that are not as healthy as perhaps would be desired by investors or regulators.
The second factor, investment, makes for an interesting contrast to the retail factor. Texas ratio loads positively on this factor, which indicates, when positive, that a bank has a majority of good-performing assets, and that it has a low probability of suffering a balance sheet shock from assets failing. On the negative end of the scale, we have net-interest margin and non-interest income. Both suggest that traditional forms of bank profitability, that is, interest income, are not so influential in this factor. Though non-interest income only loads at the bottom end of the inclusion threshold at 0.4, that net-interest income loads at −0.7 corroborates the analysis that non-interest income streams are dominant for this factor—signaling that this is related to how investment banks have successfully deleveraged loans made before the financial crisis, and now seek non-traditional forms of revenue. In sum, for the pre-crisis period, we have an expected result, yet for the post-crisis period, the results are rather counter-intuitive. Pre-crisis, the repeal of Glass-Steagall meant that investment banks and retail banks could merge, and we see this play out in the “broad church” that is the traditional factor. Yet, after the crisis, two factors emerged, whose structures indicate interest rate preferences that align with the dichotomous understanding of banking—retail and investment—that appeared obsolete after Glass-Steagall was rescinded.
The financial crisis period of 2008 to 2012 returns two factors when analyzed, which I have labeled Underwater and Overwater. As the name suggests, these represent banks who may be overburdened with financial assets that are worth less than their notional value (underwater), and banks who may be more diversified, or who may have been more prudent with their risk appetite (overwater). As these relate to interest rate preferences, underwater banks would prefer interest rates to remain low, such that they can “extend and pretend,” and then deleverage bad assets over time, and overwater banks have inverse preferences to those. Finally, as a robustness check, a factor covering the whole period from 2003 to 2019 was computed. In the next section, the results of these interest-preference factors will be analyzed.
Results
The traditional factor in the pre-crisis period, and investment and retail factors in the post-crisis period, serve as the main independent variables in the model, with the model predicting their effect on the dependent variable (the FF rate) when interacted with the bank’s size (small, large, and prominent) and FAC membership at the time. Figures 2, 3, and 4 show the main results (for full results, see Appendix, Tables A1-A4). Figure 2 below shows the three-way interaction between the factor traditional, sitting on the FAC, and a prominent bank. It shows that, moving along the scale of the traditional factor (plotted as the marginal effects at representative cases, bookended by the factor’s minimum and maximum values), in the pre-crisis period, prominent banks have a non-zero predicted effect on the FF rate. Yet, the graph highlights that the predicted capacity of prominent banks to affect the FF rate is conditioned by them sitting on the FAC. Pre-crisis prominent bank. Crisis-period prominent banks. Post-crisis prominent bank—retail factor.
This is shown by contrasting the confidence intervals for banks not sitting on the FAC, which track the zero line, meaning we cannot exclude that the average marginal effect is zero. In contrast, the line for banks sitting on the FAC is steeper, and with a non-zero marginal effect across the range of traditional values. That is, a prominent bank in this period sitting on the FAC produces a larger average predicted effect on the FF rate than a bank not sitting on the FAC. Small and large banks sitting on the FAC do not produce a significant effect (Appendix, Table A1, models 2 and 3).
Turning to how to interpret the traditional factor, we can see that on the negative end of the scale, where loan-to-deposit ratio and Texas ratio dominate (both indicators of troubling balance sheet issues, and suggesting a preference for lower interest rates), there is a smaller predicted effect on the dependent variable. Given the time frame, this makes intuitive sense. Before the financial crisis broke out, banks certainly were developing problems on their balance sheets, but did not recognize them as potentially disastrous problems until later.
Therefore, we can expect that this end of the scale would produce less of a predicted effect for a bank lobbying the Fed at the FAC. Moving towards the positive end of the scale for the traditional factor, bank profitability via interest earnings, that is, the traditional means of bank profitability, shows a clear predicted effect. Specifically, the more a prominent bank sitting on the FAC earns its profits via interest income, the more likely it is that the FF rate will rise, in the pre-crisis period. To put this in theoretical context, structural power (indicated by bank size) matters, but to be truly effective, it must be activated by instrumental power. Instrumental power, indicated by FAC membership, is not enough in the pre-crisis period, as can be seen in the absence of an effect for small or large non-prominent sized banks.
The main takeaway from the pre-crisis period is that prominent banks dominated in terms of attaining their preferences, with not even large non-prominent GSIBs seeming to be able to influence the Fed. The looming Texas ratio problems of some banks also can be seen playing out in prominent banks exhibiting a preference for lower interest rates.
Turning to the period of the financial crisis, which as noted above returns two factors when analyzed, labeled Underwater and Overwater, no clear pattern emerges for bank size or membership influencing the FF rate (see Figure 3 above). This is to be expected, because though the structural forces that led to the financial crisis, and how banking has changed in its aftermath, is central to the analysis, the period between 2008 and 2012 provides less analytical leverage in understanding how banks might influence monetary policy. This is because in this period, interest rates dropped to what is known as the zero-lower bound (ZLB), which is functionally a zero percent interest rate. Because there was such little variation in interest rates during this period, and since the mechanisms by which central banks were forced to drop their rates to zero and in some case below are well understood and explained elsewhere (Gambacorta et al., 2014), the crisis period is best used as just a comparator for predicted effects of banks in non-crisis periods. However, both factors behave as theoretically expected, with overwater banks increasing as the scale moves from negative to positive, and underwater banks doing the opposite.
For the post-crisis period, there are two factors, retail, and investment. As expected, given the similarity between the retail and traditional factors, Figure 4 is similar to Figure 2 in some respects, with the largest predicted effect reserved for banks at the positive end of the scale. A positive score on the retail scale indicates traditional banking earning profits from interest income, with negative scores on the retail scale indicating worrying loan-to-deposit ratios. Both for banks that do and do not sit on the FAC, a non-zero effect on those banks who score on the negative range cannot be excluded. Yet, on the positive end of the scale, an interesting story is playing out, with clear indication that retail factor banks who sit on the FAC have a clear predicted effect on the FF rate. Thus, this shows the instrumental power of banks displaying the retail factor is only significant for prominent banks who want interest rates to remain high. This contrasts with the pre-crisis traditional factor, where those banks had a significant effect for both raising and lowering rates. Yet, though retail banks have a predicted effect on just the positive end of the scale, the magnitude of that predicted effect is potentially much larger than traditional banks exhibited.
The second post-crisis factor, investment, presents a strong and clear predicted marginal effect along both the positive and negative ends of the scale, seen in Figure 5 below, and this figure produces by far the strongest indication for a bank to prefer lower interest rates. For this factor, where the scale is negative, net-interest margin dominates. Banks that have low net-interest margins are those who do not derive as much profit from interest-earning activities as those banks with a higher net-interest margin. These banks with low net-interest margins are usually larger, and have diversified into alternative revenue streams. Post-Crisis Prominent Bank—Investment Factor.
Yet, this is not to say that conservative banking is absent from this factor. As with all the factors that have been significant for large banks, variables that indicate conservatism in banking are still important, and this can be seen in the clear effect that the Texas ratio, in the positive part of the investment scale, plays on the dependent variable. A high Texas ratio (note: reversed) indicates that a bank has been prudent with the loans it has made and may have successfully deleveraged any underperforming loans on its books after the crisis.
Turning to how these results relate to the hypotheses, for hypothesis 1, it is evident that only prominent banks have conclusively non-zero predicted effects across the scales of the three factors. The effects of either small or large-sized banks are statistically insignificant (see tables A1-A3 in the appendix). Therefore, it seems apparent that the size of a bank (its structural power) is an important predictor of its ability to affect the FF rate. (Appendix Figures 6-21)
However, there is also the additional membership component to hypothesis 1. Evidence for how membership of the FAC is crucial can be seen playing out across Figures 2, 3, and 4 with even prominent non-members of the FAC having little capacity to concretely have a predicted effect on the FF rate. This is mirrored by small and large banks, as can be seen in Appendix Tables A1-A3. Thus, the evidence supports hypothesis 1, that size and membership matter. That is, the models predict that structural power needs to be augmented by instrumental power to be successful in banks achieving monetary policy goals.
Moving on to hypothesis 2 and models 1, 4, and 7 (detailed in Appendix), we can see that the coefficients of the three interest preference factors (traditional, retail, and investment) are roughly similar (Beta = 0.025; 0.022; and 0.025, respectively, all with p < 0.01). The coefficient for the crisis period factors, overwater and underwater, however, are not significant (models 10 and 13 in the Appendix, and Figure 3 above), which means we cannot rule out that the effect is zero. Therefore, this supports hypothesis 2 that there is a larger predicted effect in non-crisis periods than during the financial crisis.
Comparing the pre-crisis period to the post-crisis period (Figures 2, 3, and 4), the predicted effect of large prominent banks sitting on the council is larger over the range of the investment factor than for the traditional factor. In addition, the retail factor at the positive end of the scale also has a larger predicted effect than the pre-crisis traditional factor. What can be inferred from this is that the period of “quiet politics” (Culpepper, 2011) that was evident in the pre-crisis period seems to have returned in earnest, and for prominent banks sitting on the FAC that exhibit the investment factor, the predicted effect is even more consistent than for prominent banks before the crisis.
Turning to hypothesis 3, the expectation is that there will be a larger predicted effect for factors post-crisis, as compared to a pre-crisis factor. The results are mixed, but very interesting. The pre-crisis traditional factor (Figure 2) shows that there is a strong predicted negative effect, which indicates that there was some predicted preference for lower interest rates by traditional banking even before the crisis. When the traditional factor is compared to the post-crisis retail and investment factors, interestingly the large negative predicted effect disappears for the retail factor, but increases for the investment factor (when a prominent bank sits on the FAC).
It is of course indisputable that many banks prefer higher interest rates, given that most banks earn most of their profits from interest income. Yet, what the results above seem to show is that for prominent banks who exhibit the investment factor, these banks appear to be engaging their instrumental power to lobby for not only higher interest rates, but also lower rates, supporting hypothesis 3. This is additionally important, because whereas hypotheses 1 and 2 related to the magnitude of the predicted effects, hypothesis 3 adds to this by making inferences on a second level, that of preferences for both higher and lower interest rates.
Conclusion
The aim of this paper was to seek a re-interrogation of a financial forum that for too long has been dismissed as irrelevant. In doing so, the intention was to uncover how banks’ instrumental and structural power affects how successful they are at lobbying, how this power ebbs and flows over time, and whether, in fact, banks have heterogenous preferences for monetary policy, and not merely always a preference for higher interest rates.
By delving into the balance sheets of a wide spectrum of financial institutions, from small local banks to their globally significant brethren, some interesting results were uncovered. The first result is that the Fed listens to the very largest banks more than smaller banks. This is unsurprising, but disheartening, because the needs of Wall Street and Main Street rarely go hand in hand. What is interesting is that the results show that merely being classed as “globally systemically important” is not enough to bend the ear of the Fed. A bank must be structurally prominent to influence policy.
There is support for the waxing and waning of instrumental power regarding success in lobbying at the Fed. Though instrumental power was high before the crisis, it diminished considerably during the financial crisis before rebounding again. Post-crisis, from 2013 onwards, instrumental power again seems necessary, in that its combination is required to augment banks’ structural power. As noted by Fairfield (2010), structural and instrumental power need not covary, yet the results presented in this paper show that though we are moving back into a period of quiet politics where structural power should not be magnified (Block, 1977), the evidence demonstrates that prominence plus membership after the crisis produces an even stronger predicted effect, across a range of the investment interest-preference factor that indicates highly financialized banks. This is borne out by what we know regarding the resurgence of the largest banks in the US after the crisis (Ioannou et al., 2019) which only seemed to make those remaining (those now truly too-big-to-fail) even stronger. As noted by Woolley (1984: 86), “Bankers are usually not able to define a policy preference that is precise and clear and that can receive widespread support in the industry. . . . [In normal times] the Federal Reserve will receive correspondingly more ambiguous messages from bankers, and the field of influence may then be dominated by a relatively small group of powerful multinational banks that can speak with a clear, coherent voice.”
This entrenchment of the power of the very largest banks is important, because very large banks have interest rate preferences that do not necessarily align with small banks. This is not to say that they always prefer lower rates, as they also may use their influence to create “drag” on rates as they rise, slowing the rate of increase. This is particularly relevant for the literature on financial lobbying, as it opens a new area of analysis: monetary policy. But this also highlights how non-prominent banks simply do not have the political weight to counter prominent banks. The FAC should be the ideal forum to invoke instrumental power in the form of close-contact lobbying, but smaller firms simply do not seem to have the structural importance necessary to bolster their instrumental power.
This is not to say that the interest rate preferences of prominent banks are the dominant causal factor in determining the Federal Funds rate. This is a multi-causal story of which the preference of very large banks is just one small part (especially when compared to factors such as the unemployment rate). Yet, I argue that even though the effect size is much smaller than already well-established causal elements such the economy, it is no less an important one to focus on, because it highlights the disparity of power even when a multitude of banks have a seat at the (FAC) table.
Thus, because this paper shows that it is only the preferences of the largest banks that win out, it points to a difficult possible future for smaller banks, those not yet already acquired by larger banks. How this lack of interest rate preference attainment affects smaller banks is an important avenue of future study because the health of smaller banks rather than Wall Street is closely tied to the health of Main Street and the real economy. Given that the FAC appears to be dominated as a forum for the use of instrumental power by prominent banks, an avenue of further study with considerable merit would be a network analysis of the key policy-makers, and of bankers from institutions of all sizes, to see how the preference attainment of banks is conditioned by network relationships between bankers and policymakers (Ban et al., 2016; Winecoff, 2020) .
Turning to how these findings might be replicable to banks outside of the US, it should be noted that the FAC is unique, not only because of the importance and centrality of US banks to the global economy, but also that it is a stakeholder group made up solely of the banking industry, with consultations with the Fed mandated by law. In contrast, while the UK’s Bank of England Base Rate is comparable to the FF rate, there are no formalized bank meetings with the UK’s Monetary Policy Committee, comparable with the FAC. Similarly, the European Central Bank sets its Deposit Facility Rate every six weeks, yet the Banking Industry Dialogue forum meets only semi-annually at most. Yet, this should not dissuade further study of whether the results shown here can be replicated, because UK and EU banks have only gotten larger since the financial crisis (Howarth & James, 2022), and therefore structural prominence analyses of how banks lobby central banks (un)successfully must be taken seriously.
Supplemental Material
Supplemental Material - Too big to ignore: The federal advisory council, monetary policy, and the interest rate preferences of banks
Supplemental Material for Too big to ignore: The federal advisory council, monetary policy, and the interest rate preferences of banks by Ciarán O’Flynn in Competition & Change
Footnotes
Acknowledgements
The author would like to thank Marielena Ivory, Tina Freyburg, Mat Watson, and Manolis Kalaitzake for their support and helpful comments. The author would also like to thank the two anonymous reviewers for their time and effort.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
