Isaiah Champion

Runs, Repos and Liquidity in the Crisis


In the aftermath of the financial crisis of 2007-08, leading up to the passage of Dodd-Frank and extending to the present, regulators, academics and (some) bankers have voiced common refrains: banks are too large, too leveraged and too illiquid. Each of these claims has a history within the U.S. regulatory state, and unique tools have been deployed to solve each problem. The central roles of excessive leverage and liquidity freezes in precipitating the crisis are at this point well understood by academics and regulators, but unfortunately this understanding materialized after the passage of Dodd-Frank, and politicians invested in the law’s preservation or repeal remain deeply confused about the causes of the crisis, leading to a misinformed public.

The first section of this article explains the difference between “leverage” and “liquidity,” as well as their relationship to classical banking. The second section provides a short (framed) history of U.S. banking regulation, and the third section describes the central role of liquidity in the 2007-08 financial crisis.

Bankers’ Speak

“I am very aware . . . it is another drug we r on”1
Herbert McDade, an executive of Lehman Brothers, April 2008

To understand financial crises generally, it’s first necessary to understand the difference between “leverage” and “liquidity” (and how they’re related).

Leverage – Buying on Credit

“Leverage” is a measure of how heavily banks rely on borrowed money to fund their day-to-day operations, as opposed to paying expenses up front with cash (or “equity”). Historically, bank regulators have limited how leveraged banks may be by imposing capital controls. “Capital” is sometimes compared to a “cushion” that protects depositors (and other bank creditors) against losses if the bank becomes insolvent, but this metaphor is easily misunderstood. It’s likewise said, sometimes by ill-intentioned bankers, that capital is analogous to a rainy-day fund, or “money in the vault,” comparable to a survivalist’s underground inventory of canned goods to provide sustenance during the next nuclear winter. The implication is that while requiring banks to “hold” more capital might protect against catastrophic losses, every dollar in the “vault” is a dollar’s worth of productive lending drained from the economy. Thus, we hear the refrain: “capital is expensive” – costs and benefits must be balanced carefully (and capital controls implemented sparingly).

The metaphor of capital as a “rainy-day fund” reflects a confusion: it is liquidity requirements, not capital regulations, which require banks to “hold” assets. Capital requirements instead regulate how much of the money a bank uses to fund its activities, such as making loans or purchasing bonds, comes from its owners (shareholders), and how much comes from its depositors (creditors). The higher the capital requirements, the less debt a bank can use to extend loans and purchase bonds. Capital is analogous to a down payment – just as homeowners may be required to provide a 10% down payment when purchasing a home, banks may be required to fund investments with 10% “capital.” In each case, 10% of a purchase must be funded with equity. In a simplified example: if a bank purchases $100,000 of AAA-rated, high-quality securitized bonds backed by the full faith and credit of the U.S. government, regulators may require that they pay $10,000 of that with equity, and permit them to borrow the rest;2 likewise, if a homeowner buys $100,000 worth of house, they pay $10,000 in cash, and take out a mortgage to fund the rest. Capital requirements don’t require banks to save money, but rather limit how indebted they can become.

In banker-speak, using borrowed money to fund a purchase or loan is called “leverage,” since it allows the borrower to bet big on an investment without paying much cash up front. A bank that uses borrowed money to fund much of its purchases and loans is said to be “highly levered.” In good times leverage can magnify returns, as the bank profits as if it made a big bet on an investment, when in actuality it didn’t spend much cash. However, in bad times leverage makes things worse, as losses on a bad investment can far outstrip the meager amount of cash paid up front. While banks are generally far more levered than ordinary businesses, reasons put forth by bankers to justify this state of affairs are suspect. Critics argue that banks take on excessive leverage precisely because, when things are bad, it is the taxpayers who will suffer, not the banks. In this view, excessive leverage in the banking industry is a direct result of moral hazard created by "too big to fail" (TBTF), as banks would presumably not be so levered if they didn’t think the government would bail them out.

Thus, to understand how capital controls benefit depositors (if not by providing a “rainy-day fund”), consider the following: When you “deposit” money with a bank, you are giving the bank your money—making a loan, in a sense—but of course you expect to get it back. As the bank now owes you that money, you are a creditor of the bank. What distinguishes depositors from ordinary creditors (such as bond investors) is, first and foremost, that depositors may demand their money back immediately, without notice. Which is to say: bank deposits are on-demand loans. They are the opposite of Hotel California – you can leave at any time.

Why require banks to fund some minimum amount of their investments with capital? Simple: the more levered a bank is, the more likely it will blow up (if things are bad). Generally, creditors only lose money if the bank becomes insolvent – otherwise, even if things are bad and the owners of the bank lose money, all creditors (including depositors) will be paid back in full. This is because, under U.S. bankruptcy laws, creditors have priority over shareholders (or owners) of the bank. In exchange for standing to benefit when times are good, shareholders agree to take the first hit when times are bad. Capital requirements are intended to ensure that, if things do go bad, there will be a sufficient “cushion” of shareholder value (cash or equity) lying around to absorb the losses, leaving creditors (including depositors) largely unscathed. In other words, capital controls are intended to insure that when a bank makes bad investments, the bank doesn’t blow up as a consequence.

Liquidity – Floods and Storms

By contrast, liquidity (or “reserve”) requirements3 really do make banks “hold” cash in a “vault.” To understand the difference, consider the following: depositors loan cash to a bank in exchange for a promise to return the cash in the future on demand. The humble bank takes that money and dutifully extends low-interest commercial loans to virtuous small business owners across the country.4 This is how banks are supposed to make a profit: they engage in “maturity mismatch,” meaning that they transform short-term, low-interest liabilities (i.e., borrowed funds from depositors that are withdrawable on demand) into higher-interest, long-term commitments (i.e., extensions of credit that are “locked in” for the borrower). In a functioning capitalist economy, banks succeed or fail based on their ability to perform this service; pricing credit risk, knowing when it is (and isn’t) prudent to extend funds to a borrower, is the name of the game.

Of course, this business model wouldn’t work if banks were required to have enough cash on hand (or “in the vault”) to pay back all their depositors. You or I can walk into a bank and demand all our money at any time, but if you and I and all the other depositors walked in on the same day, the bank wouldn’t have nearly enough cash to pay us. The deep assumption of this business model, without which commercial banking wouldn’t exist,5 then, is that depositors won’t all demand their money back at the same time. To see this, imagine a simplified, non-FDIC insured bank (“Classic Bank”) that takes money from depositors and extends mortgages to homeowners. Classic Bank is sufficiently capitalized, meaning that it uses a good percentage of its own money to extend mortgages. However, if all of Classic Bank’s depositors demand their money back contemporaneously, Classic Bank will be unable to pay its depositors back, as their money is “out of the vault” and locked up in mortgages. The problem isn’t that Classic Bank doesn’t have the money, exactly; the bank may well have made smart, low-interest loans to virtuous homeowners that will be dutifully repaid someday. Rather, the problem is that Classic Bank, unlike its depositors, can’t replenish its coffers at will by recalling whatever loans it has extended. It can’t go to all the people to whom it has issued 30-year mortgages and simply demand that they pay up today. So Classic Bank inevitably faces the risk that its depositors will all demand to be repaid at the same time, and it will be unable to pay up. This is the classic bank run, which results from maturity mismatch, and dealing with this risk is at the heart of (and uniquely characterizes) the business of banking.

It’s also fundamentally a liquidity crisis. To see this, imagine a simplified, non-FDIC-insured bank (“Modern Bank”) that takes money from depositors and, instead of extending mortgage loans, uses that money to make long-term investments in securitized bonds (because, after all, it’s a modern bank). Don’t worry too much about “leverage.” One day, the NYT publishes an Op-Ed by Paulson-child, arguing that securitized bonds are in fact securitized junk (that is, bad loans which will never be repaid). This happens to be false — the securitized bonds consist of perfectly good loans — but Paulson-child knows that if he talks loudly enough the markets will hear his roar. As depositors rush to reclaim cash, Modern Bank rushes to sell (or “liquidate”) its securitized bond holdings in the secondary market (meaning that it resells the bonds second-hand to investors). As Modern Bank’s securitized bonds flood the market, the increase in supply (on top of Paulson-child’s roar) causes their prices to plummet in what’s known as a “fire sale” – and as a result, while Modern Bank previously had enough valuable assets to satisfy its creditors (including depositors), it doesn’t anymore, because prices have dropped.

After all’s said and done, there’s a shortfall between the amount raised by Modern Bank in liquidating its bonds (which sold well below what it initially paid for them due to the fire sale), and the amount depositors are demanding that Modern Bank repay (equal to what it initially paid). There are two ways this scenario can end. If Modern Bank’s “cushion” of capital is less than the shortfall, shareholders are wiped out and depositors take a hit. Or, if the “cushion” covers the shortfall, shareholders are hit but depositors take their money and run. In the latter case, Modern Bank was adequately capitalized; in the former, it wasn’t. However, in either case, Modern Bank loses virtually all of its assets, and thus, its value as a going concern. Modern Bank will be liquidated, with the only distinction between the two outcomes being whether depositors, in addition to shareholders, suffer losses. When it is revealed the next day that Paulson-child was mistaken and securitized bonds rise gloriously in value, Modern Bank remains in the ditch.

This is the sense in which bank runs are liquidity panics. Classic Bank invested in rights to receive payments under mortgage loans, and got in trouble because it couldn’t convert its mortgages into ready cash at will, due to the term of the mortgages. Likewise, Modern Bank invested in rights to receive payments under securitized bonds, and got in trouble because it couldn’t convert its bonds into ready cash at will, due to the effect of fire sales. Just as Classic Bank couldn’t recall its loans (even though they’re high-quality), Modern Bank couldn’t resell its bonds (even though the price drop was an anomaly).6 In each case, a storm of depositors rushed the bank and, due to an inability to convert assets to cash (without causing a fire sale), the bank had to close up shop. This dynamic has repeated itself throughout history, and after an unusual seven decades of relative quiet in the U.S., it resurfaced in the 2008 crisis.

The Calm Before the Storm

In addition to distinguishing between leverage and liquidity, it’s important to understand that post-crisis concerns over “systemic risk” and “too big to fail” (TBTF), as well as the ongoing trend towards consolidation in the banking industry, are recent phenomena. Historically, U.S. banks were constrained by a unique charter system that limited their ability to conduct business across state lines (and thus their size). TBTF was effectively prohibited in the U.S. prior to the civil war by state law, as banks were required be “chartered” by, and thus subject to the laws of, their home state. The National Bank Act of 1963 created a system of “federal” charters for banks, but even then federally-chartered banks remained subject to branching restrictions limiting their ability to expand. It wasn’t until the Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994 dramatically reduced branching restrictions for federally chartered banks that TBTF became a regulatory possibility. By contrast, while the “repeal” of Glass-Steagall in 1999 has captured the ire and imagination of U.S. politicians, the effect of this repeal was simply to enable broker-dealers (institutions that trade in securities, and are regulated by the SEC) and “banks” (institutions that accept FDIC-insured deposits, and are regulated by the Fed, FDIC and OCC) to merge under common bank holding companies. While this development may have contributed to TBTF, the deep issue surrounding Glass-Steagall’s repeal concerns the scope of activities banks should be permitted to engage in, not their size (as discussed in footnote 4). If TBTF can be attributed to any (misguided) legislative action, it’s Riegle-Neal, not Glass-Steagall.

By contrast, leverage and liquidity are timeless concerns for bank regulators, in the U.S. and internationally. (After all, bank runs are essentially liquidity crises exacerbated by leverage.) Prior to 1933, bank runs were a common phenomenon in the U.S. While regulation of both federal and state chartered banks was less extensive prior to the Great Depression, governments in the 19th century commonly employed both capital and liquidity controls to limit bank runs within their jurisdiction. However, these regulations were crude and ineffective: capital controls failed to adequately adjust for the riskiness of assets held by banks, while “reserve” requirements stood in inexorable tension with the nature of banking by constricting maturity mismatch. The greater the reserve requirement, the less ‘banking’ banks could do – and when storms hit, depositors ran in such high concentrations that “reserves” would be exhausted. As a result, the regulations failed and bank runs proliferated.

This changed in 1933 with the establishment of FDIC deposit insurance, which reinvented the dynamics of runs. Under the Glass-Steagall reform (which has not been repealed), the Federal government would insure that depositors received their money back if a bank became insolvent. Thus, depositors no longer needed to worry about credit risk, as their deposits were insured by the Federal government in the event the bank went bust. This resulted in a 70+ year “quiet period” in American banking, where the frequency and severity of bank runs fell dramatically. It also resulted in “moral hazard,” where banks became able to borrow at lower and lower rates due to their perceived government guarantees.

In response to this newfound quiet, banking regulators began focusing exclusively on capital, under the implicit assumption that illiquidity had been “fixed” by deposit insurance. “Safety and soundness” guidelines were established by the OCC for federally-chartered banks, which standardized the amount and quality of capital required to be held against assets by federally-chartered banks. “Margin” regulations were promulgated by the Fed, which likewise limited the degree of leverage that customers could assume with respect to securities transactions. In 1975, the SEC passed Rule 15c3-1, or the “Net Capital Rule,” which extended uniform capital requirements to broker-dealers. And in 1988, when the Basel Committee on Banking Supervision (“BCBS”), an international committee of central bankers, convened to establish a uniform global framework for banking regulation, the committee emerged with “Basel I,” a regulation that revolutionized capital across the world but did relatively little to affect liquidity.

The “Run on Repo”

“There’s this whole area called ‘shadow banking.’ That’s where the experts tell me the next potential problem could come from.”7

“Understanding that the shadow banking system is, in fact, real banking… is the only reasonable basis for new policies.”8

In addition to the elimination of branching restrictions, a number of other developments took place over the last three decades that transformed the banking industry. Interest rate caps on deposit accounts were gradually eliminated as limitations on the scope of activities banks could historically engage in were expanded (primarily by regulatory interpretations, not by the “repeal” of Glass-Steagall). “Derivatives,” historically limited to use by Midwestern farmers and Chicago commodities traders, expanded dramatically to become an essential risk management tool for modern corporations (and to dominate Wall Street trading profits). More importantly for understanding the crisis, “securitization” developed as a new technique for bolstering liquidity. Instead of making individual loans to individual borrowers and holding those loans to maturity, banks began “pooling” loans into investment vehicles and selling interests in such vehicles to investors. As returns on these vehicles were easier to predict and standardize than on individual loans (due to averaging effects across multiple borrowers in different circumstances), securitization led to the creation of a new secondary market in commercial and residential loans. As a result, instead of being forced to (illegally) terminate long-term loan commitments to raise money to stem a depositor flood, Modern Bank could now sell interests backed by its loans (“asset-backed securities” or ABS) to secondary investors to raise funds. Fueled by an array of regulatory sticks and carrots, the market in securitization soon became the engine through which the majority of residential and commercial credit in America was extended.

In addition, as the banking industry consolidated, so did the asset-management industry. Institutional investors, such as mutual funds, grew to hold an increasing proportion of Americans’ savings. As a result of this gradual buildup of cash outside the traditional banking system, a new market developed to find productive uses for such cash – namely, the repo market. On paper, a repo is an agreement to buy and sell assets, but economically it is a lending transaction. In a repo, Party A agrees to sell assets to Party B at an agreed upon price, and Party B agrees to sell those assets back to Party A at some point in the near future at a slightly higher price. The difference in the initial purchase price (paid by Party B) and the ‘re-purchase’ price (paid by Party A) represents interest paid (to Party A) on the assets loaned and held (to and by Party B) through the duration of the transaction. Importantly, while repos are often drafted to contemplate such purchases and re-purchases occurring within as short a period as a single day, in practice, parties to a repo ‘roll over’ their daily obligations, with the economic result that the borrower continues holding assets and paying interest to the lender until assets are called.

Leading up to the crisis, banks became increasingly reliant on repos with institutional investors (such as mutual funds and asset managers) to fund their operations. Institutional investors, which had amassed record volumes of assets under management, would lend their securities (as depositors lend money) to banks in exchange for interest payments. Just as depositors could (historically) earn interest by storing passive cash in deposit accounts, so institutional investors could earn interest by repoing out passively held securities. As these loans were short-term, banks were able to achieve better funding rates than on long-term lines of credit, as institutional investors (and depositors) tended to lend more cheaply knowing that their securities (or cash) were callable on demand. As with securitization, regulatory sticks and carrots also played a role in fueling the market, as did the growth of short selling (the practice of betting against a stock, which is accomplished by borrowing securities in a loan transaction, and thus presumes a liquid market for borrowing the stock). By 2000, the repo market had expanded to become an essential ingredient of bank financing.

However, unlike deposit accounts, repo facilities were not (and are not) entitled to FDIC insurance. In fact, as the volume of assets held by institutional investors grew to dramatically exceed $250,000 (the federal cap on deposit insurance), institutional investors increasingly looked for other, safer facilities in which to store assets. As obligations under repo facilities were generally fully collateralized, meaning that borrowers were required to post collateral with lenders to cover losses in the event of a default, repos were perceived as quite safe from a capital (or credit risk) perspective. Thus, with the carrot of deposit insurance unavailable, institutional investors increasingly parked assets in repo facilities, knowing that they would receive a steady return and, if disaster struck, could both get out of Dodge fast and seize collateral to cover losses.

Enter Postmodern Bank. Instead of accepting cash from depositors (to fund lending activities or securitized bond investments), Postmodern Bank accepts repoed securities from institutional investors (to, let's say, use as collateral for its derivatives trading). Postmodern Bank thus takes “deposits” that are callable on demand, but since these repoed securities are ineligible for deposit insurance, instead posts collateral to secure the "deposits." One day, the NYT publishes an Op-Ed by Paulson-child, arguing that securitized bonds are in fact securitized junk. (This happens to be false, but the Paulson-child knows if he talks loudly enough the markets will hear his roar.) As it turns out, the collateral posted under the repo facilities happens to consist of securitized bonds! Institutional investors begin to worry: the only thing protecting their assets if the bank goes bust is the collateral (which is junk!), as the FDIC has left them out in the cold. After consulting many financial models in light of Paulson-child's thesis, investors conclude that securitized bonds are, in fact, worth only 50% of what they initially thought. In finance, this is called a ‘haircut’ – the value of the collateral has fallen due to a ‘haircut’ that accounts for increased risk. As a result of their collective revelation, the investors demand that Postmodern Bank increase its collateral under the repo facilities by 50%. If the bank doesn’t or can’t post sufficient collateral, since the repos are (legally speaking) short-term loans, the investors can decline to ‘roll over’ their repos and demand the return of their repoed assets (or “deposits”).

After all’s said and done, there’s a shortfall between the amount of sufficiently liquid assets the bank can amass in a day, given the difficulty of converting illiquid assets to cash without triggering fire sales, and the 50% haircut imposed on its repo collateral. As with Classic Bank and Modern Bank, even if Postmodern Bank is otherwise sufficiently capitalized, the withdrawal of repo funding by institutional investors leaves the bank with insufficient funding to continue its daily operations. Unable to avoid defaulting on other contracts due to the loss of funds, Postmodern Bank is forced into bankruptcy. While it is revealed the next day that Paulson-child was mistaken, and institutional investors revise their models and recalculate haircuts accordingly, Postmodern Bank remains in the ditch.

While the crisis was exacerbated by many factors other than the repo market (and while Paulson-child's thesis turned out to be valid), this dynamic destroyed Lehman and was at the heart of the crisis. To the extent that Dodd-Frank was based on a theory of the crisis that didn’t appreciate the role of repo in causing liquidity freezes, certain reforms under Dodd-Frank may be suspect. The disease must be diagnosed before it can be cured, and despite everything we’ve heard about the dangers of “financial innovation,” the story remains alarmingly familiar.


“The greater the scientific advance, the more primitive the fear.”
—Don Dellillo, White Noise

Since regulators have begun to understand the nature of the crisis, attention has shifted back to liquidity. BCBS and the Fed have passed a Liquidity Coverage Ratio and Net Stable Funding Ratio, each of which are aimed squarely at liquidity risk, and meaningful structural reforms have been implemented in the short-term repo market. However, this newfound understanding by regulators and academics has yet to trickle into the public sphere. Many reforms of Dodd-Frank that are marketed as efforts to prevent another global meltdown, such as the Volcker Rule, are not designed to address the true causes of the 2008 financial meltdown, even if they’re good policy in other ways. Other reforms, such as the central-clearing of derivatives, have created grave risks of their own that are in danger of being ignored by ideologically driven politicians. While Elizabeth Warren has emerged as the leading leftist crusader for financial reform, her rage on certain issues is deeply misguided. And while Hilary Clinton has ingeniously marketed reforms that would actually help banks as “tough on Wall Street,” Sanders thinks Glass-Steagall would have prevented the crisis. This article is a step towards correcting these misunderstandings.

1Herbert McDade, an executive of Lehman Brothers, describing Lehman’s accounting treatment of sale and repurchase transactions (“repos”) in an April 2008 email (Examiner’s Report at 742). A recent report by the Office of Financial Research suggests that non-U.S. banks are still on the drug.

2The example is simplified as it assumes the bonds have a 100% “haircut.” Haircuts, and their importance for capital and liquidity regulation, are discussed in the “Run on Repo” section.

3Historically, liquidity requirements were referred to as “reserve” requirements in the U.S. – hence the “vault” metaphor.

4Or alternatively, the bank uses the cash to trade derivatives on aluminum for proprietary accounts, simultaneously investing in aluminum processing facilities in a scheme to manipulate prices to enhance trading profits (but only if the bank is Morgan Stanley or Goldman Sachs). Caveats aside, delimiting what kinds of activities banks should, and should not, engage in is important work, given that banks enjoy anti-competitive funding advantages over non-bank companies (due to TBTF) and play a uniquely important role in our economy. However, while post-crisis reforms like the “Volcker Rule” and the growing political momentum surrounding Glass-Steagall are central to this debate, the gradual expansion of banks into non-banking activities had little to do with the crisis.

5In theory, you could require that banks keep all of depositors cash “in the vault.” However, banks would then be constrained from extending loans and would have to earn profits from transaction fees. This would essentially turn banks into Paypal (with custodial services), and thus, they would cease being “banks” in any meaningful sense.

6It’s worth noting that Classic Bank could also (in theory) have resold its rights to receive payment under the mortgages to secondary investors, and thus, have avoided the fact that it can’t legally terminate its mortgages. However, as mortgage loans are highly customized to specific borrowers, meaning they are “bespoke” and require buyers to conduct substantial diligence to determine the quality of the investment, mortgages have historically not enjoyed a liquid secondary market. This means, if Classic Bank attempted to sell its rights under the mortgages to secondary buyers, it would encounter the same fire sale dynamic that plagued Modern Bank. Remedying this problem is, in fact, the central purpose of securitization – to (in theory) make the secondary market for loans more liquid.

7Hilary Rodham Clinton, speaking during the first 2015 Democratic presidential debate.

8Gary Gorton, an expert, whose work popularized the “run on repo” theory of the crisis (italics added).

The Hypocrite Reader is free, but we publish some of the most fascinating writing on the internet. Our editors are volunteers and, until recently, so were our writers. During the 2020 coronavirus pandemic, we decided we needed to find a way to pay contributors for their work.

Help us pay writers (and our server bills) so we can keep this stuff coming. At that link, you can become a recurring backer on Patreon, where we offer thrilling rewards to our supporters. If you can't swing a monthly donation, you can also make a 1-time donation through our Ko-fi; even a few dollars helps!

The Hypocrite Reader operates without any kind of institutional support, and for the foreseeable future we plan to keep it that way. Your contributions are the only way we are able to keep doing what we do!

And if you'd like to read more of our useful, unexpected content, you can join our mailing list so that you'll hear from us when we publish.