The history of the US banking system is in part the history of attempts to prevent bank runs. It's almost inconceivable now: you hear a rumor that your bank might fail and you run to the bank to withdraw all your cash, hoping that they'll even have it on hand to be able to give it to you. You find a long line of people waiting outside, or a scene of people demonstrating and even panicking. You might feel lucky even if you get back some of your deposits.

Bank runs occurred sporadically in the 1800's, and reached a climax during the Great Depression, which experienced repeated waves of bank runs over 4 years and reduced the trust of Americans towards holding bank accounts. Many stuffed cash in their mattresses. Bank runs still occur today (e.g. Wamu and IndyMac) but they tend to be quieter and dissipate quickly. Consumers are not as fearful because they believe that their money is insured by the FDIC, which is backed by the American government and ultimately the taxpayer.

In the past, many runs were false alarms based on rumors, and ultimately the banks returned to business as usual. But in many cases the banks were failing, and customers might suffer a significant loss on their deposits. In the banking panic of 1907, for example, customers of failed banks lost up to 20% of their deposits. However, overall the number of failed banks was small, and the losses due to them was very small, approximately 0.2% of total commercial bank deposits on a yearly basis. Nevertheless, bank failure was always a significant concern, and consumers were generally well versed in the issue of solvency of banks. Today we don't worry about this because we assume the government will protect us.

Unfortunately this protection comes at a great price. In the last 2 years, the Federal Government has spent over $2 trillion dollars to protect the banking system and prevent banking panics, as well as increasing insurance to bank accounts, up to $250,000 per account, as well as (temporarily) for money markets which were previously unprotected, and even accounts at investment banks. The taxpayer is now on the hook for any shortfalls if these banks fail, and certainly for propping up these banks, many of which are zombies. For example, Citigroup is by some accounts a zombie bank - surviving only through direct and indirect government support. (Citigroup received a grade of 'F' from Institutional Risk Analytics, the company hired by the FDIC to evaluate solvency of its member banks.)

The Federal Reserve and the Treasury justified their support of investment banks not because these organizations were fundamental in the economy. Indeed, as most people know, these institutions exist primarily for the purpose of making a profit, and unashamedly so. Instead, they justified their support of these institutions by the fear that if they fail, a banking panic would spread through the entire financial system, as it did in the Great Depression, and take down consumer banks as well. Federal Reserve Board Governor Ben Bernanke, a student of the Great Depression, knew very well that this must be avoided at all costs, because a severe impairment of a banking system deepens and lengthens a depression.

Whether the taxpayer receives back his $2 trillion investment to prop up the banking system remains to be seen. I suspect that much money will be lost because the government is still artificially inflating the housing market through the nationalization of Freddie and Fannie, and purchases by the Federal Reserve. Furthermore, banks are probably still not valuing their housing-market based assets fairly. It's likely that the housing bubble will pop again, and banks will need even more government support, which may or may not be given depending on the public's appetite for more bailouts. A new wave of bank runs may occur on a scale never before seen.

Anatomy of a Bank Run

There are two basic types of bank runs: runs on 'good' banks and runs on 'bad' banks. A good bank is one that is solvent and can meet customers demand for withdrawals, even if it does not have the cash immediately on hand. It can sell assets to raise cash, hopefully in a good market, restore confidence, and then resume normal operations. Runs may spread to other banks, but will soon dissipate.

However, a run on a 'bad' bank, by contrast, does not end so well. In this case, the bank runs out of available cash and does not liquidate more assets. The governing body, state or federal bank regulator, will examine the bank's holdings and declare if the bank is solvent (if its assets are greater than its liabilities). If it's not solvent, the bank will be suspended and customers will have to wait to retrieve their deposits. Depending on if the bank is insured (essentially all banks are insured these days) the customer will receive back some or all of their deposit. This type of bank run is more likely to spread to other banks, as others question the solvency of their own banks. Thus there is the potential domino effect of banks failing.

This is actually a simplistic view, because banks are not so clearly 'good' or 'bad'. In fact, probably many good banks failed during the Great Depression. The reasons for this are complex, but one important component was that, because the overall economy was very poor, banks had to sell assets, particularly bonds, into a depressed market to raise cash in the event of a run.  This lowered the overall asset base and forced them into insolvency, which would then spread to the next bank. If they could have waited, then they would have survived. An influential view at the time was that the economic contraction was necessary to purge the economy of the weak institutions, so such bank failures were actually considered 'healthy'. Economists at the time did not appreciate the importance of a robust banking system in overcoming an economic downturn, and the damaging effects of an impaired banking system.

Banking Panic of 1907

The banking panic of 1907 started with the failure of several speculative copper industry banks, which spread to the Knickerbocker Trust in New York City a week later. This was before the FDIC and the Federal Reserve, and so depositors panicked in their attempt to withdraw money, which they feared could be lost forever, and the panic spread to other trusts as well as the stock market. To ease fears, New York State stepped in with $25 million to loan to the bank network, but that did not stop the panic. JP Morgan raised another $25 million, and this seemed to calm the nervous public.

However, the fear by this time had spread throughout the country. Banks responded by imposing a restriction on cash withdrawals, as they had done several times previously when threatened with runs. This was a painful limitation on a population so dependent on cash for every day needs as well as industrial requirements. However, the banks permitted other types of transactions, such as cashier's checks and bank transfers. The restriction lasted up to 3 months in some parts of the country. This prevented failure of many banks because they had the 'breathing room' to restore customer confidence. Although a painful experience, a severe bank run and a wave of bank failures was avoided and the economy recovered.

However, the population had enough of restrictions on withdrawals. The government responded with the Aldrich-Vreeland Banking Act of 1908, which provided for loans to be made to banks quickly in the event of a run. This act was used only once, in 1914 related to the start of WW I, and seemed to work well.

This financial mechanism was superseded by the Federal Reserve System in 1914. Although we think of the Federal Reserve today as governing monetary policy to optimize economic conditions (by setting important interest rates), back then the purpose was primarily to prevent bank runs to avoid restrictions on payments. This was done by offering secured loans through its 'discount window'. Adherence to the gold standard, as well as the governors' poor understanding of monetary policy coupled with a very cautionary approach, prevented the System from having much influence on the economy.

Unfortunately the cure was worse than the disease. In the attempt to prevent bank runs and restrictions on payments, the end result was a series of waves of the worst bank runs in history (from 1930 to 1933), along with the most severe restrictions on payments ever, the banking holiday of March 1933, which lasted for a week, during which banks were shut down completely, resulting in economic chaos and suspension of half of of them.

Thus the panics of 1907 and 1933 were greatly different in character. According to Milton Friedman's A Monetary History of the United States:

If the 1907 banking system had been in operation in 1929, restriction of payments might have come in October 1929 when the stock market crashed. If not, it would surely have come at the latest at the end of 1930 -- probably on the occasion of the failure of the Bank of United States on December 11, 1930. Had the restriction come then, it seems likely in retrospect that it would have produced an immediate shock and reaction much more severe than the unspectacular worsening of conditions that occurred in the fall of 1930, but would also have prevented the collapse of the banking system and the drastic fall in the stock of money that were destined to take place, and that certainly intensified the severity of the contraction, if they were not indeed the major factors converting it from a reasonably severe into a catastrophic contraction. (p. 168)

The reform measure finally enacted -- the Federal Reserve System -- with the aim of preventing and such panics or any such restriction of convertibility in the future did not in fact stem the worse panic in American economic history and the severest restrictions of convertibility, the collapse of the banking system from 1930 to 1933 terminating in the banking holiday of March 1933. (p. 698)

To be sure, Friedman is not advocating for the return of restrictions on payment. He believes that Fed policy was far too tight, and was in fact the main cause of the Great Depression. Relaxing the policy (i.e. 'easy money') would have been a far better solution, in his opinion. Bank runs and resulting suspensions would be reduced through greater liquidity, and perhaps more importantly, there would be a better financial environment in which banks could operate with greater stability.

How to prevent a bank run, Part II: FDIC

After the panics of March 1933, the public became even more distrustful of banks and continued hoarding cash. The economy continued to contract. The Banking Act of 1933 established the FDIC, which created government backed insurance for all bank accounts in member banks, up to a limit of $2,500. This policy has been greatly successful in restoring confidence in the banking system, and has prevented severe bank runs up to the present day. (However, the ultimate cost of the policy I believe will be very high -- far greater than the loss of assets it was intended to protect.)

Before the FDIC, deposits were not guaranteed, and everyone knew that they could lose their money in a failed bank (although as stated before this was rare). Now, bank accounts were protected by the Federal government. This was part of FDR's New Deal (though not actually requested by FDR), which included other social safety nets such as unemployment insurance and Social Security. People no longer had to worry nearly as much. Security of deposit became a right.

Most economists agree that government insurance of deposits was a great idea, including Milton Friedman and JK Galbraith. According to Galbraith, on the establishment of FDIC: "As a result one grievous defect of the old system, by which failure begot failure, was cured. Rarely has so much been accomplished by a single law." It seemed to be the panacea that everyone was searching for.

Moral hazard

As with all insurance, however, came the problem of 'moral hazard'. Because bank accounts were guaranteed, bankers and depositors didn't need to worry about the health of their banks as much, and so banks were subject to decreased market discipline. This presented a risk to the taxpayer, who was on the hook in the event of failure. Therefore, it was necessary for the government to regulate banks to make sure they acted responsibly. A regulatory framework was therefore developed, and the public put its trust in that framework.

The Savings and Loan crisis of the 1980's resulted from the breakdown of the regulatory framework of the FDIC. The purpose of Savings and Loans, when they were created, was to provide funding for mortgages, and S&L's were limited to home mortgages only. In the early 80's the banks were brought in under the FDIC umbrella. However, in the mid 80's, as interest rates rose, the banks became insolvent because the value of their fixed-rate mortgage assets decreased, while the interest they were required to pay on deposit accounts (in order to remain competitive) increased. They were allowed to continue, and were permitted to branch into higher yielding, riskier loans in commercial real estate, in the hopes that they would regain solvency. This strategy was a tremendous failure. They were finally shut down in the late 80's, and resulted in a cost to the taxpayer of $140 billion. The FDIC insurance fund was depleted and covered only a fraction of total losses.

Because these banks were insured and regulated, the consumer had little reason to question the solvency of their bank or the soundness of its investments, which were in some cases complex real estate transactions (e.g. Whitewater). Instead it was left up to regulators, who for various reasons did not take action until it was too late. This was a prime example of the 'moral hazard' of insurance.

The cost to the taxpayer of the S&L crisis was far greater even than the total losses during the Great Depression, which amounted to about $7 billion total, which is about $50 billion in 1987 comparable dollars. This provides strong evidence that the real cost of insuring accounts is far greater than the potential loss of assets. When the depositor is monitoring his bank, problems with the bank are likely to be detected much earlier on.

Too Big to Fail

The "Too Big to Fail" doctrine is not new. In fact it was first enunciated in 1984 in response to problems at Continental Illinois Bank, which was the seventh largest bank at the time. Large losses on various assets resulted in a bank run by its uninsured depositors. The FDIC, in an attempt to prevent the panic from spreading to other banks, decided to offer full coverage for all accounts. While this worked, and allowed for an orderly wind down of the bank's assets (over several years), it had two unanticipated results. First of all, the FDIC had to extend coverage nearly universally, as small banks demanded equal treatment. Secondly, it introduced the doctrine of "Too Big to Fail". As Comptroller of the Currency Todd Conover testified, if the bank was allowed to fail:

We could very well have seen a national, if not an international, financial crisis the dimensions of which were difficult to imagine. None of us wanted to find out.

Thus a bank run was prevented, but at great cost. First of all, the taxpayer was now on the hook for the entire commercial banking sector. Secondly, the government had established a precedent for saving a 'Too Big to Fail' institution. If Continental was allowed to fail, certainly it would have been a shock to the banking system and might have taken down other banks. But the economy would have recovered. The government learned an important lesson, which was that this could be done, and in fact the Treasury under Hank Paulson used a similar strategy in 2008 to effect the $1 trillion dollar financial bailout of financial firms, Fannie and Freddie, and AIG. Certainly the failure of these institutions would have created a great shock to the system. However, as the failure of Lehman proves, the market is capable of withstanding such a shock.

The problem today

And as time went on, Federal insurance of bank accounts greatly expanded. Now, most bank accounts are insured up to $250,000, but individuals can divide their funds into multiple accounts, making the protection effectively unlimited. This increase of deposit insurance was not intended or anticipated, but was a necessary result of the system: as long as there was unprotected money in the banking system, a run was possible.

The total liability of the FDIC now is huge. There is no way that the FDIC could cover losses in a significant banking crisis, which means that the taxpayer must cover the shortfall as they did in the S&L crisis. Unfortunately the banking system now is not in a healthy state, with at least one huge zombie bank - Citigroup. If other large commercial banks aren't zombies, it is only because of the taxpayer support they've received, and further downturns may force them into insolvency again.

This is the inevitable result of FDIC policy. In order to prevent bank runs, it must eventually insure the entire market. Otherwise there will be a run on uninsured deposits. But this creates a huge and untenable liability for the taxpayer. Furthermore, the government gets sucked into supporting zombie banks to avoid FDIC insurance fund losses, further increasing the taxpayer liability. Once a zombie, the bank will have great difficulty returning to solvency, and will be subsidized by the taxpayer indefinitely. This increases the Federal debt, which in turn reduces the solvency of the country as a whole.

Big investment banks like Goldman Sachs can now borrow directly from the Fed discount window (they were given access as a result of the financial crisis).  They can borrow at low interest rates and use the money to invest in risky assets, repay their TARP, and even pay themselves.  All this was done in order to maintain stability of the financial system, since the FDIC alone couldn't do the job.

And lest one think that at least the problem of 'restriction of cash withdrawals' has been solved by the FDIC -- it certainly has not.  The Credit Crunch of 2008 was simply a back door restriction.  The problem is back.  It simply something that we need to learn to live with.

Return to basics

The Federal Reserve and FDIC were both attempts to prevent bank runs and suspension of withdrawals. While it succeeded for 80 years in that regard, the limits of the policy have been reached. In its attempt to save the 'good' banks, the policy has resulted in zombie banks, and if history is any guide, the cost to resolve these banks will ultimately be far greater than the potential losses of assets of a bank failure that was caught early on.

As shown above, bank runs can be prevented by temporary restriction of cash withdrawals. While this was a big problem in a cash-based economy in the 1800's, today the results would not be nearly as severe, since people can use other forms of payment, such as checks and credit and debit cards. Perhaps there would be limits on those methods of payments as well, but they would be temporary and the crisis would pass once confidence has been restored, or the bank is suspended. (This is a form of what happened in the Credit Crunch of 2008.)  Of course there would be a significant loss to the depositor if the bank is suspended (perhaps 20% - 50%), but this is a very rare occurrence and so the overall loss would not be great.

The consumer would have to learn to protect themselves against a bank failure, as they did in the 1800's. If one wants to ensure complete liquidity and safety, one can invest in short term US Government Treasury bills, and in fact banks can create such accounts, as well as other types of high-reserve/low-yield accounts. Beyond this safe harbor, customers can deposit into more standard types of bank accounts with higher yield, accepting the risk of a fall in value just as they do when investing in the stock market. Currently US Treasury bills are the safest investment of all, although even they are threatened by the continuing government support of failed institutions.

It is an ironic but unavoidable fact that the attempt to make everything safe puts everything at risk.