A Better Way to Revive Glass-Steagall

John Crawford *


The financial crisis of 2007-2008 repeatedly forced regulators to face terrible choices between risking catastrophic contagion by letting particular firms or markets fail, and intervening to bail them out. 1 See generally Fin. Crisis Inquiry Comm’n, The Financial Crisis Inquiry Report, 280-386 (2011), (describing a series of such dilemmas during 2008). One explanation of why these dilemmas arose was that financial firms were “too big to fail.” Nearly a decade after the onset of the crisis, the major financial conglomerates in the United States are in many cases larger than they were in 2007. 2 For example, total assets for JPMorgan Chase & Co. came to $1.562 trillion at year end 2007, and to $2.49 trillion at the end of 2016. See JPMorgan Chase & Co., 2016 Annual Report (Form 10-K) 43 (2017),; JPMorgan Chase & Co., Annual Report (Form 10-K) 26 (2008), Figures as politically diverse as Donald Trump, Elizabeth Warren, John McCain, and Bernie Sanders all agree on a (partial) remedy: the revival of the Glass-Steagall Act, the Depression-era law that placed a firewall between the activities of commercial banks and investment banks. 3 “Glass-Steagall” refers to the Banking Act of 1933, §§ 16, 20, 21, 32, 12 U.S.C. §§ 78, 377 (1998) (repealed 1999) (current version at 12 U.S.C. §§ 24, 378(a) (2015)). Regarding popular support for a return to the law, see Donna Borak, Elizabeth Warren and Trump Agree on One Thing: Break Up the Banks, CNNMoney (Apr. 7, 2017, 2:46 PM ET), (reporting that Warren reintroduced a bill with McCain and two other senators to reinstate Glass-Steagall, and quoting an analyst’s view that “[a]t some point the market is going to have to accept that the Trump administration is serious about restoring the Glass-Steagall separation between commercial and investment banking”); Kevin Cirilli, Sanders Backs Reviving Glass-Steagall, TheHill (Jul. 7, 2015, 6:39 PM EDT), (reporting on Sanders’s support for a return to Glass-Steagall). By “revival,” supporters mean reinstating Glass-Steagall’s provisions prohibiting investment banks from affiliating with depository institutions—a prohibition that was eliminated in 1999. 4 Banking Act of 1933 §§ 20, 32, which prohibited affiliation between deposit-taking institutions and investment banks, were repealed by the Financial Services Modernization Act of 1999 (popularly known as the Gramm-Leach-Bliley Act). Pub. L. No. 106-102, §§ 101-103, 113 Stat. 1341-51 (1999). Splitting up behemoth conglomerates such as Citigroup and JP Morgan that engage in both investment and commercial banking would, they believe, help solve the too-big-to-fail problem and prevent future bailout dilemmas. 5 See Cirilli, supra note 3 (quoting Sanders’s public statement that “if we are truly serious about ending too big to fail, we have got to break up the largest financial institutions in this country”). A related reason sometimes proffered for reviving Glass-Steagall is the need to protect “ordinary” commercial banking from “risky” investment banking. See, e.g., Borak, supra note 3. This rationale is less compelling given the significant quantitative and qualitative constraints on transactions between banks and non-bank affiliates effected by Federal Reserve Act §§ 23A and 23B. 12 U.S.C. §§ 371c, 371c-1 (2015). But see Investment Co. Inst. v. Camp, 401 U.S. 617, 630 (1971) (describing more “subtle hazards” that may arise from the affiliation).

This Essay advances several claims in opposition to this narrative. First, in addressing the bailout problem, the size of individual firms is far less important than the existence of vast amounts of short-term, “runnable” debt outside the traditional banking sector. 6 See generally Jack Bao et al., The Runnables, FEDS Notes (Sept. 3, 2015), (measuring runnable liabilities in the U.S. economy). Bailouts occur because regulators fear the knock-on effects of a firm’s failure. The severity of those knock-on effects is driven more by the type of claims that suffer losses due to a firm’s failure than by the magnitude of a firm’s first-order losses. Imposing losses on short-term creditors tends to generate vicious externalities that typically do not arise when long-term creditors and equity claimants absorb losses. 7 See infra notes 26-29 and accompanying text; see also Ben S. Bernanke, Chairman, Some Reflections on the Crisis and the Policy Response, Speech at the Russell Sage Foundation and the Century Foundation Conference on “Rethinking Finance” (Apr. 13, 2012) (noting that first-order stock market losses following the dot-com crash were as large as the peak-to-trough decline in all residential real estate from 2006-2011 and attributing the more severe impact of the housing crash to the fact that it triggered a banking panic). The phenomenon of non-depository institutions issuing oceans of short-term debt defines the “shadow banking” system, which lay at the heart of the crisis of 2007-2008. 8 See generally Morgan Ricks, The Money Problem: Rethinking Financial Regulation (2016) (providing an account of shadow banking and the crisis). Second, a different provision of Glass-Steagall, one that has never been repealed, attempted to address the shadow banking problem, with less and less success as the decades wore on. The provision, Section 21 of Glass-Steagall, prohibits investment banks from receiving deposits, 9 Banking Act of 1933 § 21, 12 U.S.C. § 378 (2015). the paradigmatic example of runnable debt. This prohibition was a key ingredient in preventing panics during the middle decades of the twentieth century. 10 See infra text accompanying notes 42-45. Unfortunately, the narrow way in which Section 21 defined deposits ultimately allowed investment banks, in the years leading up to the crisis, to comply with the letter of the law while violating its spirit—a loophole they ruthlessly exploited. 11 See infra text accompanying notes 46-49. They did so by issuing short-term debt instruments that did not fall within the technical definition of “deposit,” but that served, for institutional investors looking for a place to park their cash, as the functional equivalent of deposits. 12 Id.

If Section 21 had targeted not only deposits, but also any instruments designed to serve the same economic function as deposits, its enforcement would have gone a long way to eliminating the key vulnerabilities that spawned the crisis. From a stability perspective, it is much more important to effectuate the spirit of the prohibition on investment banks receiving deposits than it is to try to prevent commercial and investment banks from sharing a common corporate parent. A better way to revive Glass-Steagall, then, would be to bolster Section 21 rather than to reenact the anti-affiliation provisions.

I. Background on Glass-Steagall

The Glass-Steagall Act is the popular name for four provisions of the Banking Act of 1933, 13 See supra note 3. aimed at separating the activities of commercial banks 14 Commercial banks are the most prominent type of (and should be understood here as a synonym for) depository institution. For an account of the (complicated) history of legal entities that have been deemed a “bank” for purposes of the Bank Holding Company Act (BHCA) of 1956, see Saule T. Omarova & Margaret E. Tahyar, That Which We Call a Bank: Revisiting the History of Bank Holding Company Regulation in the United States, 31 Rev. Banking & Fin. L. 113, 138-88 (2011). and investment banks. 15 Investment banks are not depository institutions, but specialize in classic Wall Street activities such as underwriting securities offerings, acting as market makers, and advising on corporate mergers. See Investment Banking, Investopedia, (last visited May 17, 2017). Sections 20 and 32 prohibited commercial banks and investment banks from affiliating with each other. 16 See Banking Act of 1933, §§ 20, 32, 12 U.S.C. §§ 78, 377 (1998) (repealed 1999). The original rationale for the anti-affiliation provisions was not based on size, but on the possibility of conflicts of interest and a sort of “brand confusion” by retail customers. See, e.g., Investment Co. Inst. v. Camp, 401 U.S. 617, 630 (1971) (describing Congress’s focus on the “subtle hazards that arise when a commercial bank goes beyond the business of acting as fiduciary or managing agent and enters the investment banking business either directly or by establishing an affiliate to hold and sell particular investments”). These prohibitions were repealed by the Gramm-Leach-Bliley Act of 1999. 17 See supra note 4. Sections 16 and 21 of Glass-Steagall, on the other hand, relate to the activities investment banks and commercial banks themselves are allowed to engage in: Section 16 prohibits commercial banks from engaging in investment banking activities such as underwriting and dealing in securities, 18 Banking Act of 1933 § 16, 12 U.S.C. § 24 (2015). Note that the arbitrage of this provision may (or may not) cause problems for the deposit insurance fund, but is less important than Section 21 from a stability perspective, as the problem of panics is less likely to strike (insured) commercial banks. See, e.g., infra notes 33-34 and accompanying text. and Section 21 prohibits investment banks from receiving deposits. 19 Banking Act of 1933 § 21, 12 U.S.C. § 378 (2015). These provisions have not been repealed, and remain the law of the land.

This means that prior to 1999, an investment bank could neither receive deposits nor affiliate with a firm that could—for example, by sharing a corporate parent. After 1999, investment banks still cannot receive deposits, but they can affiliate with firms that do. Thus, for example, JPMorgan Chase & Co., a holding company, can own as subsidiaries both JPMorgan Chase Bank, N.A., a commercial bank, and J.P. Morgan Securities, LLC, an investment bank. 20 Fixed Income Information, JPMorgan Chase & Co, (last visited May 17, 2017). Only the commercial bank, however, can receive deposits, and only the commercial bank is covered by deposit insurance. 21 What’s Covered, Fed. Deposit Ins. Corp., (last visited May 17, 2017). Furthermore, the commercial bank’s interactions with affiliated entities are strictly circumscribed by Sections 23A and 23B of the Federal Reserve Act. 22 12 U.S.C. §§ 371c, 371c-1 (2015).

Those calling for a return to Glass-Steagall seek to reenact the anti-affiliation provisions that were repealed by Gramm-Leach-Bliley, so that JPMorgan Chase & Co. will be forced to divest itself of either its commercial bank or its investment bank. There are respectable arguments for such an approach, 23 See, e.g., Luigi Zingales, Why I Was Won over by Glass-Steagall, Fin. Times (Jun. 10, 2012), (arguing, among other points, that “Glass-Steagall helped restrain the political power of banks”). but as an anti-bailout measure, it would be markedly inferior to putting bite into (the still operative) Section 21.

II. Shadow Banking and the Crisis

The reason Section 21 is important is that it has the potential to address the problem at the heart of the recent crisis: shadow banking. 24 See generally Gary B. Gorton, Slapped by the Invisible Hand: The Panic of 2007 (2010) (explaining the crisis as a classic bank run in a new institutional setting—that of shadow banks rather than banks). As noted, shadow banking refers to non-depository institutions funding long-term investments with lots of short-term debt. 25 See Ricks, supra note 8, at ix (noting that while “shadow banking” is used in diverse ways, to the crisis response team at the U.S. Treasury Department, it meant “the financial sector’s use of vast amounts of short-term debt to fund portfolios of financial assets”). This funding structure is a defining feature of traditional (commercial) banks, and is sometimes referred to as “maturity transformation.” The maturities of a bank’s liabilities are overwhelmingly short, whereas the bank’s assets typically have long maturities. Thus, the bank must repay depositors on demand, but must wait months or years to be repaid by those who have borrowed from it. This makes the bank vulnerable to “runs”—that is, en masse withdrawals of depositors’ funds. The bank does not keep depositors’ cash in vaults, but lends it out; these loans cannot be called in and it is often difficult for the bank to sell such loans to another institution on short notice for full value. 26 This is true to the degree that banks’ assets are not perfectly liquid. See generally John Crawford, Capital Accounts: Bank Capital, Crises, and the Determinants of an Optimal Regulatory Approach, 66 Hastings L. J. 1161, 1167 (2015) (book review) (explaining liquid assets). Runs tend to generate terrible externalities. 27 See, e.g., Ricks, supra note 8, at 109-10 (describing the panic crunch in credit availability that widespread runs cause). Widespread runs constitute a panic; panics are the essence of a financial crisis. 28 Id. at 102-03.

The fragility of banks’ funding structure caused periodic panics throughout the nineteenth and early twentieth centuries in the United States. 29 See, e.g., Theodore Gilman, Federal Clearing Houses 183 (1899) (making the claim in 1899 that “[s]ince 1793 panics have occurred [in the United States] in the following years: 1797, 1811, 1813, 1816, 1819, 1825, 1837, 1847, 1857, 1866, 1873, 1884, 1890, and 1893”); Ben Bernanke, The Federal Reserve and the Financial Crisis 9-10 (2013) (identifying six banking panics between 1873 and 1914). The problem of panics in the traditional banking sector was largely solved by a combination of (i) deposit insurance to prevent runs and (ii) regulatory constraints and supervision to address moral hazard concerns arising from insurance, 30 Ricks, supra note 8, at 160-63. yielding a long, panic-free period in U.S. history. 31 See Gorton, supra note 24, at 11. The crisis of 2007-2008 marked the end of this “Quiet Period.” 32 Id. The end came because banks’ fragile funding model had migrated to non-depository institutions such as investment banks, which lay in the regulatory “shadows,” outside the safety net and without the prudential supervision applied to banks. 33 See generally id. at 1-12 (describing the rise of shadow banking and the role it played in the crisis).

It is important to note that shadow banks were not created by the repeal of Sections 20 and 32 of Glass-Steagall 34 Banking Act of 1933 §§ 20, 32, 12 U.S.C. § 78, 377 (1998) (repealed 1999) (prohibiting commercial banks and investment banks from affiliating with each other). : affiliation with a commercial bank is irrelevant to the question of whether an investment bank or other non-depository institution is itself functioning as a shadow bank by funding its activities with large amounts of short-term debt. Thus, the problem with Lehman Brothers and Bear Stearns was not that these classic Wall Street investment banks were affiliated with depository institutions—they were not—but rather that they served functionally as depository institutions themselves. 35 See generally Ricks, supra note 8, at 32 (describing the functional equivalence of deposits and short-term funding used by other financial firms); Fin. Crisis Inquiry Comm’n, supra note 1, at 280-91, 324-43 (describing the failures of Lehman Brothers and Bear Stearns). For example, in the years leading up to the crisis, large institutional investors who needed a place to park their cash for short periods of time began to eschew traditional bank accounts and instead make short-term collateralized loans to these investment banks and others like them. 36 This type of lending arrangement was called “repo,” short for “repurchase agreement.” Tobias Adrian & Hyun Song Shin, Fed. Reserve Bank of N.Y., Staff Rep. No. 382, The Shadow Banking System: Implications for Financial Regulation 8 (2009), (“In a repo, the borrower sells a security today for a price below the current market price on the understanding that it will buy it back in the future at a pre-agreed price.”); see also Fin. Crisis Inquiry Comm’n, supra note 1, at 103 (“Unlike banks and thrifts with access to deposits, investment banks relied more on money market funds and other investors for cash; commercial paper and repo loans were the main sources.”). The loans might pay a slightly higher interest rate and the collateral provided security to the lender. The loans were functionally just like a deposit—they were short-term, often overnight, and routinely rolled over until the funds were needed for a particular transaction, at which point the investment bank could seek out other lenders to replace the withdrawn funding. 37 Adrian & Shin, supra note 36, at 6. Just like traditional banking, this system served a valuable purpose for those engaging in it, but created vulnerability to en masse, panicked withdrawals and all the pernicious knock-on effects for the real economy that such panics generate. 38 See, e.g., Ricks, supra note 8, at 106-09.

Again, the dual “solution” to banks’ fragile funding structure—the combination of a safety net and prudential regulation—does not extend to shadow banking, which metastasized in the decades leading up to the crisis and which remains robust today. 39 See, e.g., Gary Gorton & Andrew Metrick, Regulating the Shadow Banking System, Brookings Papers on Econ. Activity, Fall 2010, at 261, 265 fig.3; Volcker All., Unfinished Business: Banking in the Shadows 14 fig.1 (2016), History has taught that widespread maturity transformation without deposit insurance or its equivalent tends to end in tears. 40 Ricks, supra note 8, at 160-63. Avoiding bailouts and achieving stability would therefore be much better served by prohibiting non-depository institutions such as investment banks from functioning as de facto depository institutions than it would be by allowing them to function as such, but prohibiting them from sharing a corporate parent with de jure depositories.

III. Looking Back and Looking Forward

As noted, Section 21 attempted to address the precise problem adumbrated above by forbidding investment banks like Lehman Brothers from receiving deposits. 41 Banking Act of 1933 § 21, 12 U.S.C. § 378 (2015). Section 21 is one of several laws prohibiting non-depository institutions from receiving deposits. 42 See, e.g., N.Y. Banking Law § 131 (McKinney 2017). The fact that this prohibition was not heavily arbitraged during the middle decades of the twentieth century 43 See, e.g., Gorton & Metric, supra note 39, at 261, 265 fig.3. played no small role in producing the long, panic-free Quiet Period. 44 See Gorton, supra note 24, at 11.

By the 1990s, however, investment banks had begun ruthlessly to exploit a flaw in Section 21 (and similar prohibitions). The flaw is that the provision relies on a very narrow, formalistic definition of “deposit.” 45 12 U.S.C. § 378(a) (2015); see also 12 U.S.C. § 1813(l) (2015). Section 21 prohibits investment banks 46 More precisely, the prohibition applies to “any person, firm, corporation, association, business trust, or other similar organization, engaged in the business of issuing, underwriting, selling, or distributing, at wholesale or retail, or through syndicate participation, stocks, bonds, debentures, notes, or other securities.” 12 U.S.C. § 378(a)(1). from receiving “deposits subject to check or to repayment upon presentation of a passbook, certificate of deposit, or other evidence of debt, or upon request of the depositor.” 47 Id. The types of short-term credit with which investment banks were funding themselves prior to the crisis—and with which they continue to fund themselves 48 Volcker All., supra note 39, at 14 fig.1. —did not fall within the literal prohibition of Section 21, but served precisely the same economic function, and created the same risks, as the instruments that were covered by Section 21.

Protecting the financial system is first and foremost about panic prevention. 49 See, e.g., Ricks, supra note 8, at 102. It is virtually impossible to have vast amounts of uninsured short-term debt funding long-term investments without creating the risk of panic. 50 See, e.g., id. at 142, 215 (describing panics as a “pathology of short-term debt,” and noting that the advent of deposit insurance brought unprecedented stability to the U.S. financial system until “emergence on a huge scale of shadow banking”). The first step to recapturing the conditions that led to the Quiet Period, then, is to limit widespread maturity transformation to insured depositories. An important move in this direction would be to amend Section 21 of Glass-Steagall to prohibit investment banks from receiving deposits or the economic equivalent of deposits, defined functionally as a short-term debt claim. 51 Id. at 230-37.


Shadow banking caused the crisis of 2007-2008: the run on the shadow banking system transformed what would otherwise have been a large but hardly unusual crash in asset prices into a macroeconomic catastrophe. Bailout decisions were driven largely by a desperate desire to stop the panic. The best way to try to prevent a recurrence is to tackle shadow banking, not to prohibit shadow banks from affiliating with commercial banks. If Section 21’s prohibition were enforced in broadly functional terms, it would not matter much from a stability perspective whether investment banks, even giant ones, affiliate with commercial banks, as their failure would be unlikely to spark a panic. As long as investment banks are engaged in shadow banking, on the other hand, prohibiting them from having the same corporate parent as commercial banks will do little to promote stability.

A revival of Glass-Steagall holds promise for stabilizing the system and ending the implicit government guarantees enjoyed by large non-depository financial institutions. But the revival should focus on bolstering Section 21, which remains on the books, rather than reenacting its repealed provisions.

* Professor of Law, University of California Hastings College of the Law.