Financial Reform after Silicon Valley Bank
This past March, the United States once again experienced a significant financial breakdown. Problems at a single institution, Silicon Valley Bank (SVB), produced such a strong threat of contagion that the entire deposit-insurance system faced a crisis and had to be swiftly overhauled.
To effectively cover all deposits at SVB, federal regulators extended deposit insurance beyond the $250,000 legal limit. The Federal Reserve, our nation's lender of last resort, also put in place a new facility called the Bank Term Funding Program (BTFP) to make additional funds available to financial institutions in case of a bank run. Whereas British economist Walter Bagehot famously said that in a crisis the central bank must lend without limit at a high interest rate against good collateral, the BTFP will lend at a low interest rate against long-term securities that will be treated as having retained book value, even though their market value may be less than that.
Since the founding of the Federal Reserve in 1913, the United States has dealt with the financial system by providing emergency backing for banks. This creates moral hazard — a lack of incentive to guard against risk that results when one is protected from potential costs — which in turn necessitates regulation. But the Houdini of bank risk has invariably escaped its regulatory straitjacket, and unfortunately, there is no simple solution. For all the mathematical modeling of economists and the formulas of practitioners, bank regulation is not an exact science. Phenomena like bank runs and credit cycles are driven not by economics, but by social psychology.
Roughly speaking, government guarantees — including deposit insurance — subsidize risk-taking, while safety and soundness regulations (such as capital requirements) amount to a tax on risk-taking. Ideally, policymakers would incentivize banks to take only the right risks at the right time. Given recent events, it's become clear that the current system does not foster such behavior.
This essay considers a way to involve the private sector in regulating bank safety. One of the challenges with government regulation arises from the fact that political incentives usually favor the pattern we have repeatedly observed: Excessive risk-taking builds up, regulators fail to discourage it soon enough, a crisis ensues, government undertakes emergency measures to contain the crisis, these measures increase moral hazard, policymakers enact new rules intended to control risk-taking, and, eventually, excessive risk-taking builds up once again. Public officials may say they want to break this pattern, but we must consider the possibility that it actually works for them. Escaping this destructive rut may ultimately require stabilizing the financial system by making it easier to fix, rather than trying in vain to make it harder to break.
WHAT BANKS DO
As households, we like to own riskless short-term assets like checking accounts. At the same time, we like to create risky long-term liabilities like mortgages and business loans. Banking allows us to do this. Banks hold our risky, long-term loans as assets while issuing the riskless, short-term deposits as liabilities. Thanks to banks, household balance sheets in the aggregate hold riskless short-term assets and issue risky long-term liabilities, while banks do the opposite.
But some households own the banks as shareholders, which means they also own the risks the banks undertake. So how does this work?
To invest directly in risky long-term assets, investors, whether they be households or banks, require a high rate of return; they would accept a lower rate of return on short-term, less risky deposits. The spread between the interest rate on long-term risky assets and the interest rate on short-term, less risky deposits gives the bank an opportunity to swoop in and earn a profit.
To see how this works, imagine the bank as a mutual association, like a credit union. We deposit money with the bank, thereby lending money to the institution. The bank pays us a low interest rate to hold these deposits. It then pools the deposits of all its customers and invests those funds by lending them to others. Banks hope to earn a higher rate of interest on these loans than they pay us for borrowing our deposits, which enables them to turn a profit.
Why don't we, as depositors, keep our deposits and invest them ourselves? The most obvious reason is that the bank employs dedicated experts to manage the risk of lending, meaning it does a better job at selecting and managing investments than we could on our own. But a more fundamental reason is that banking reduces the risks associated with lending. Our deposits, individually speaking, are short term, and are therefore risky for the bank to hold as liabilities. But when considered as a collective, they become less risky long-term liabilities, since we do not all withdraw our deposits at once. By pooling our deposits, banks reduce the risk they might otherwise incur by taking on just a handful of short-term liabilities.
Banks reduce risk further by pooling their loans. A diversified portfolio of loans is less risky than any single loan. And because different loans pay off on different dates, diversification across time ensures that interest-payment flows into the bank are steady.
Finally, as depositors, we are insulated from the lending risks the bank takes because the value of our deposits does not depend on how well the bank's loans perform. But for those of us who own shares in the bank, the value of our shares depends a great deal on such performance.
By depositing money into banks, depositors are indirectly investing in risky long-term projects. This is socially beneficial in that it supports greater economic activity and productive investment than would take place without banking. But it also generates two problems: bank runs and Minsky cycles.
A bank is normally solvent but illiquid. It is solvent in the sense that its assets are worth more than its liabilities. It is illiquid in the sense that if it had to sell its assets in a hurry, it would receive less than what those assets would ultimately be worth, and probably less than its total liabilities, including deposits. This is because the bank's loans are best understood by its own experts. Other institutions, which understand those loans less well, will be reluctant to pay full price to purchase a failing bank's loans, meaning selling those loans to other banks in a hurry is difficult.
Was SVB insolvent or illiquid? There is some controversy about this, but we have reason to believe it was insolvent. Suppose I own a 10-year bond with a face value of $100 at an interest rate of 2%. I then borrow $100 in short-term funds at an interest rate of 1%. If interest rates remain steady, I would be solvent. However, interest rates have risen: The long-term interest rate is now 4%, and the short-term rate is now 5%. If I have to sell the bond today, I will only receive $85 of the original $100 I borrowed. (This is the market value of a bond with a 2% coupon discounted at a 4% rate.) Since I still owe $100 on my short-term loan, I am insolvent.
But would I be solvent if I did not have to sell the bond right away?
Some people say yes, because in 10 years I will receive the full $100 I borrowed initially. But this is misleading: I still owe $100, and for the next 10 years, I will have to pay more in interest on my short-term loan than I will receive in interest from the long-term bond. Even if my lenders carry me for 10 years, I am still insolvent. This conclusion would suggest that SVB was insolvent — as perhaps are other banks that claim to be solvent because they account for assets on a "hold-to-maturity" basis.
Because a bank holds illiquid assets, it is vulnerable if too many depositors try to withdraw funds at once. With large amounts of unexpected withdrawals, the bank must sell its illiquid loans at a loss. After too many forced sales, it will become insolvent.
Depositors, meanwhile, are not in a position to know whether the bank is solvent, nor can they know how other depositors are likely to behave. If I, as a depositor, suspect that my bank is insolvent, or if I suspect that other depositors are going to withdraw enough funds to make my bank insolvent, prudence would direct me to withdraw my own funds. When enough depositors arrive at that conclusion, a run on the bank occurs.
Because of a run, an otherwise solvent bank might become insolvent. A bank run can also lead to an inferior equilibrium for an insolvent bank. It would probably be better for the insolvent bank to gradually sell its assets, with depositors eventually getting back most, but not all, of their money.
Deposit insurance offers a solution to bank runs for ordinary banks (i.e., banks that are solvent but illiquid). If people try to run on an ordinary bank, the deposit insurance kicks in, and the bank does not have to sell its assets at a discount. Deposit insurance also eliminates the incentive to run on the bank in the first place, since depositors don't have to worry about losing their deposits.
The problem with deposit insurance is that it reduces a bank's incentive to take only prudent risks. If the insurance kicks in regardless of how reckless a bank's loans are, banks will be more willing to make risky loans — after all, the greater the risks they take, the larger their rewards may be. At the same time, taking greater risks increases the chance that a bank will become insolvent.
This problem is called "moral hazard." Private insurance companies address moral hazard by using a combination of incentives and rules. If you install a home security system, for instance, firms will often offer you a discount on your homeowners-insurance premiums.
The Federal Deposit Insurance Corporation (FDIC) does a poor job of dealing with moral hazard. The problem arises in part from the fact that its employees have no skin in the game; they get paid regardless of whether the FDIC manages the banking system well. SVB's failure did not lead to the firing of a single FDIC employee, even though the event was widely viewed as a regulatory debacle. Instead, the costs of the FDIC's missteps passed to bank shareholders in the form of higher assessments, which will be used to refill the FDIC's insurance fund.
In principle, the FDIC should take timely action to close a bank before its net worth falls too close to zero. But instead, as a government agency, the FDIC has an incentive to "extend and pretend" — to allow the bank to keep operating longer than it should, and to overvalue the assets on the bank's balance sheet to make it appear that there is no reason to close it. Compared with leaving a shaky bank open, closing the bank takes more effort and exposes the agency to more political flak. Extend-and-pretend is an attractive alternative, especially for high-level political appointees: If the strategy backfires, the problem will blow up on someone else's watch.
All of this does not necessarily mean that the FDIC is fatally flawed. The optimal number of bank failures is not zero: Not all banks are healthy, and in a competitive system, some of these banks will — and probably should — fail. SVB's failure was a disaster primarily because it was mishandled.
Ordinarily, when a bank fails, the FDIC negotiates with another bank to take over the failed bank. The FDIC may have to pay out of its insurance fund in order to induce the other bank to risk such a takeover, but as a result, the bank failure proceeds gracefully. Yet in SVB's case, the rapid run on deposits and the slow pace of negotiations meant that the bank's status remained uncertain long enough for a wider panic to ensue.
Even if we should not mind occasional bank failures, the periodic systemic crises we've witnessed are a sign that something is wrong with our current approach. We should therefore consider alternatives to government-provided deposit insurance.
Deposit insurance is neither necessary nor sufficient for dealing with normal banks. For banks that are solvent but illiquid, Bagehot's lender-of-last-resort function, properly executed, can stem bank runs. That is, if the Fed is willing to lend money to failing banks against their good collateral (collateral that is relatively stable in value) at a high interest rate, banks will be able to meet demands for withdrawals.
Banks themselves can also do a better job of preventing runs. For example, they can offer term deposits that require advanced notice for withdrawal without penalty. If you give the bank a month's notice that you plan to withdraw $5,000, you pay no penalty. If you give only a week's notice, you pay a penalty of $100. If you withdraw immediately, you pay $500. This would induce some depositors to wait to withdraw, giving the bank time to arrange short-term loans — including loans from the Fed.
If we are going to have a system that includes deposit insurance as well as a lender of last resort, private insurance providers might be more agile and effective than the FDIC. Much the way the FDIC is supposed to operate today, a private insurer would be responsible for setting standards for distinguishing an illiquid bank from an insolvent bank. It would use these standards to close a bank before it becomes insolvent, and then sell the failed bank to another bank.
A private insurer would be responsible for determining the solvency of its insured banks. Since the firm's insurance fund would diminish if it failed to shut down a bank on a timely basis, it would have less of an incentive to extend and pretend, as the FDIC does now. The private insurer would monitor each bank's portfolio performance and risk-management practices on a quarterly basis. If the insurer believes a bank is insolvent, it will be able to meet withdrawals at the bank by borrowing from the Fed against the bank's good collateral at a high interest rate.
This can be done through an insurance contract, which would specify the conditions for imposing sanctions ahead of time. An insurance contract should be legally required to have a finite term of, say, four years. If the bank cannot obtain a new insurance provider at that time, it should be closed, with the incumbent insurance provider responsible for any losses involved in selling that bank to another institution. These rules would signal that the insurance company cannot get away with delays in addressing a bank's insolvency (unless a new insurance company is foolish enough to take responsibility for insuring the insolvent bank).
Like other private insurers, competing deposit-insurance companies, each with skin in the game, would come up with incentives and rules that reduce the moral hazard that deposit insurance generates. To help ensure that private firms can handle bank failures, the Fed would have to require them to maintain a minimum amount of capital on hand — this capital would be available to induce another bank to take over a closed bank. Minimum capital requirements would have to be low enough to attract capital to the industry but high enough to cover the cost of paying a new bank to purchase a failed bank.
The insurer would charge different premiums to different banks based on their portfolios and risk-management practices. In cases of bank failure, a private insurer would impose sanctions on the bank or force it to merge with another bank, depending on the process spelled out in the contract. Banks would thus face incentives that are more finely tuned than those that accompany FDIC insurance.
The other problem with banking comes in the form of the Minsky cycle, named for economist Hyman Minsky.
Minsky warned that financial stability can lead to instability. As time passes without a financial crisis, firms become more optimistic, and lenders loosen credit standards. This encourages firms to take on riskier projects. Eventually, something triggers a crisis. Firms become pessimistic, and they have a harder time obtaining credit. Economic activity falls off, and firms forgo profitable projects. Over time, confidence gradually recovers, and firms begin financing profitable projects once again. Then time passes, optimism becomes excessive, and the cycle resumes.
In the context of banking, the yield spread — the difference between revenues on debt instruments of varying maturities, credit ratings, issuers, and risk levels — varies with the Minsky cycle as well. In the optimistic phase, banks are eager to try to earn a profit using short-term deposits to fund riskier long-term loans — a practice called "earning the spread." Competition for deposits and loans intensifies, and the spread narrows. But banks gradually become overconfident. When they take on too much risk and a crisis hits, they pull back, and the spread widens. Households and businesses that are still willing to obtain risky loans find them very costly, if they are available at all. Then as time passes, banks become more willing to take risks to earn the spread, and the spread narrows again.
The Minsky cycle poses a systemic risk. It's not a matter of a single bank failing for idiosyncratic reasons; it's a phenomenon that affects the psychology of the entire market. Consider the financial crisis of 2007-2008. Leading up to the crisis, the housing market had experienced years of price appreciation and declining interest rates. As a result, the mood was optimistic. But the run-up of home prices eventually reversed, resulting in rising mortgage defaults and rating downgrades of mortgage-backed securities (investment products issued by banks that consist of bundles of real-estate loans). The downgrades caused such securities to be much less attractive to banks and other loan-holders, and made those securities unusable as collateral for repurchase agreements. These developments produced widespread financial distress. Among other measures, policymakers responded with the $700 billion Troubled Asset Relief Program to stabilize the financial system, while the Fed massively increased its purchases of long-term bonds and mortgage-backed securities through a tool called "quantitative easing" in order to reduce interest rates and increase the money supply.
The 2007-2008 financial crisis illustrates another important problem: Policymakers did not anticipate it. When a Minsky cycle occurs, regulators are caught up in the optimism along with everyone else. In 2006, top policymakers, including Fed chairman Ben Bernanke, expressed confidence that banks were using reliable techniques to manage risk and that any problems with sub-prime mortgages would be contained. Bernanke and other top officials were not prepared for what ensued — but they could have been. According to a recent research paper by Harvard Business School's Robin Greenwood and others:
[R]apid credit and asset price growth over the prior three years, whether in the nonfinancial business or the household sector, is associated with a 40% probability of entering a financial crisis within the next three years. This compares with a roughly 7% probability in normal times, when neither credit nor asset price growth is elevated. Our evidence challenges the view that financial crises are unpredictable "bolts from the sky" and supports the Kindleberger-Minsky view that crises are the byproduct of predictable, boom-bust credit cycles. This predictability favors policies that lean against incipient credit-market booms.
It's also worth noting that, prior to 2007, policymakers did all they could to encourage mortgage lenders to extend credit to what they termed the "underserved" market. Afterward, they berated banks for "predatory lending" and passed the Dodd-Frank Act, which imposed strong credit requirements for underwriting loans when the market was already weak. If anything, the policy community amplified the housing cycle.
We must hope that policymakers will explore ways in which the government could dampen such credit cycles, or at least avoid reinforcing them, in the future. Perhaps an independent commission could advise public officials about systemic risk factors, including the credit cycle. Unlike the current Financial Stability Oversight Council, which is primarily comprised of incumbent regulators, an independent commission would include members from academia and the business world with more diverse opinions and contrarian inclinations.
EASY TO FIX
Another viable policy approach would involve backing away from making the financial system difficult to break and focusing instead on making it easy to fix. The first method lulls households, businesses, and policymakers into a false sense of security, amplifying the severity of each new crisis. Instead, we should accept the fact that the financial system will experience shocks. We should also realize that the key to real stability is encouraging institutions to fail gracefully and ensuring that no one institution's ruin triggers havoc across the financial sector.
The economic consequences of the Dot-com Crash of 2000 were mild in part because the crash primarily affected equity markets, not credit markets. In light of this fact, policymakers should encourage greater use of equity and less use of credit. In the housing market, instead of subsidizing mortgage lending, we could focus on down-payment assistance. In the tax system, instead of punishing firms that rely on equity by raising corporate tax rates (which encourages borrowing by raising the value of the ability to deduct interest payments from corporate income), we could incentivize firms to use equity by reducing those tax rates as well as taxes on dividends and capital gains.
In the financial sector, the resolution of each crisis reinforces the tendency to bail out large financial institutions. This gives an advantage to the largest banks. The result has been a dramatic increase in financial concentration: We have put most of our country's financial eggs in the basket of about half a dozen gigantic banks.
A more diversified financial sector would likely be more robust, and might provide credit to a wider variety of business enterprises. Policymakers could encourage more varied financial business models to develop by offsetting the subsidy to large banks provided by their "too big to fail" designation with a tax on the largest banks, encouraging them to spin off smaller units.
Private equity, including venture capital, usually requires investors to make long-term commitments. This serves to match risky, long-term assets with long-term liabilities. Perhaps instead of reviling private equity, policymakers ought to encourage it to grow.
THE POLITICAL-ECONOMY PROBLEM
When it comes to manufacturing, health care, and other sectors, economists can reasonably think of the government as a referee. They may consider the incentives that operate in the private sector and propose ways for policymakers to change those incentives and fill gaps in order to secure better outcomes. The fact that rent-seeking behavior on the part of special interests and selfish behavior by public officials will inevitably affect policy does not change government's role as referee — the state must never become a player in the game.
Finance is different. In this sector, government is a critical participant, because financial reform requires changing the incentives of public officials.
It is only natural for governments and banks to become intertwined, for they share an important characteristic: Perceived permanence is vital, while perceived instability can be fatal. If a government appears to be vulnerable to a revolution, it often loses its de facto power: Viewing their regime as doomed, citizens will see no need to obey its laws, government employees will see no point in going to work, and soldiers will not risk their lives to defend the nation. Likewise, if the public believes a bank is about to go under, the bank will lose customers, depositors will run, and the bank will find itself unable to tap other sources of credit. To maintain a public perception of stability, governments and banks lean on one another.
It is unsurprising, then, that wherever we behold a reliable government, we also observe a well-functioning financial system. And wherever a government's position is precarious, banks are unreliable. One does not see strong banks alongside weak governments or vice versa. When solid banks exist, a government can borrow in order to deal with a crisis or stave off military attacks. When a stable, effective government presides, banks can rest assured that their contracts with borrowers will be enforced. Government can also protect financial institutions against bank runs. By the same token, it can protect banks against risks they should have avoided.
This co-dependent relationship makes banking different from other sectors. In the banking industry, government is not a detached overseer or referee; public officials have ample motivation to support banks and to use them to channel credit to their preferred uses — primarily government's own spending. This state of affairs creates a dilemma.
The United States' large outstanding government debt stands as one of this country's biggest sources of systemic risk: The last time the debt-to-GDP ratio reached 100% of GDP was at the end of World War II. At that time, the large deficits generated by the war were behind us. Today, they are in front of us, due to the projected needs of the Social Security and Medicare programs.
The debt-to-GDP ratio cannot keep rising forever. How will it be contained? Neither political party has an appetite for large spending reductions. Large tax increases are unlikely, and even if enacted, they may not do much to close the deficit. Explicit default is unthinkable. That leaves only implicit default, via inflation. But inflation will not reduce the debt burden unless interest rates remain below the rate of inflation. That may require financial repression. In other words, the government would have to induce households and firms to hold their assets in government bonds at low interest rates by making other investments unattractive or illegal.
Americans experienced some financial repression in the 1940s, when interest rates were held artificially low prior to the Treasury-Federal Reserve Accord of 1951 and Regulation Q required banks to offer low interest rates on savings deposits and to pay no interest on checking accounts. Without money-market funds or other financial instruments, households and firms were forced to make do with bank deposits that did not keep up with inflation.
In today's computer-driven, globalized financial markets, it might be more difficult for the government to pull off repression. But it has the incentive to try. What risks that might pose for the financial system at large have yet to be seen.
In the end, the March banking crisis can be traced to the debt problem. The large deficits, and the quantitative easing used to finance them, triggered inflation. This in turn led to higher interest rates and large losses at banks holding long-term securities. Such developments underlined the reality that governments sometimes have goals that differ from, or even conflict with, the aim of maintaining a robust financial system. The government's incentive to channel credit to its own needs may override the goal of encouraging banks to take the right risks at the right time.
Periodic crises are bad for the public. A responsible approach to finance would seek to insulate Americans from the mistakes and cycles that occur in this sector. But for policymakers, each crisis represents an opportunity to take on more authority over the financial system. So long as this is the case, reform efforts face long odds, however necessary they may be.