Publications
Countercyclical Capital Buffers: A Cautionary Tale
with Gary G. Richardson and Christoffer Koch.
Mentioned in the Wall Street Journal
Mentioned in Brookings Hutchins Roundup
Countercyclical capital buffers (CCyBs) are an old idea recently resurrected. CCyBs compel banks at the core of financial systems to accumulate capital during expansions so that they are better able to sustain operations during downturns. To gauge the potential impact of modern CCyBs, we compare the behavior of large and highly-connected com- mercial banks during booms before the Great Depression and Great Recession. Before the former, core banks did not expect bailouts and were subject to regulations that incentivized capital accumulation during booms. Before the later, core banks expected bailouts and kept capital levels close to regulatory minima. Our analysis indicates that the pre-Depression regulatory regime induced money-center banks to build capital buffers between 3% and 5% of total assets during economic expansions, which is up to double the maximum modern CCyB. These buffers enabled those banks to continue operations without government as- sistance during severe crises. This historical analogy indicates that modern countercyclical buffers may achieve their immediate goals of protecting core banks during crises but raises questions about whether they will contribute to overall financial stability.
In the Eye of a Storm: Manhattan's Money Center Banks During the International Financial Crisis of 1931
with Gary G. Richardson, Explorations in Economic History, Volume 68, Number 2, April 2018.
In 1931, a financial crisis began in Austria, struck numerous European nations, forced Britain to abandon the gold standard, and spread across the Atlantic. This article describes how banks in New York City, the central money market of the United States, reacted to events in Europe. An array of data sources - including memos detailing private conversations between leading bankers the governors of the New York Federal Reserve, articles written by prominent commentators, and financial data drawn from the balance sheets of commercial banks - tell a consistent tale. Banks in New York anticipated events in Europe, prepared for them by accumulating substantial reserves, and during the crisis, continued business as usual. Leading international bankers deliberately and collectively decided on the business-as-usual policy in order to minimize the impact of the panic in the United States and Europe.
Bank Leverage and Regulatory Regimes: Evidence from the Great Depression and Great Recession
with Gary G. Richardson and Christoffer Koch, American Economic Review, Papers and Proceedings, Volume 106, Number 5, May 2016.
The regulatory framework for commercial banks evolved over the 20th century. During the boom before the Great Depression, capital requirements for commercial banks were low and fixed. Bankers faced double liability. Failing banks were not bailed out. During the boom before the Great Recession, capital requirements were proportional to risk-weighted assets. Bankers faced limited liability. Banks deemed too big to fail received bailouts. Across these regimes, we compare banks’ capital choices using balance-sheet data. We document how the largest institutions’ choices changed over the business cycle. During the Roaring 20s, the largest banks increased capital holdings as asset prices rose to unprecedented levels. During the boom from 2002 to 2007, the largest institutions kept capital levels near regulatory minimums. Our results suggest that more market discipline would have induced the largest U.S. banks to hold greater capital buffers prior to the financial crisis of 2008.
Did the Reserve Requirement Increases of 1936-1937 Reduce Bank Lending?
with Haelim M. Park in Journal of Money, Credit, and Banking, Vol. 47, No. 5, August 2015.
Evidence from a Quasi-Experiment:We analyze the impact of contractionary monetary policy through increases in reserve requirements on bank lending. We compare the lending behavior of banks that were subject to the requirement increases in 1936–37, Federal Reserve member banks, to a group of banks that were not subject to the reserve increase, Federal Reserve nonmember banks. After implementing the difference-in-difference estimators, we find that the increases in reserve requirements did not create financing constraints for member banks and lead them to reduce lending. Therefore, the actions of the Federal Reserve concerning the required reserve ratios cannot be blamed for instigating the economic downturn of 1937–38.
Fetters of Debt, Deposit, or Gold during the Great Depression? The International Propagation of the Banking Crisis of 1931
in Jahrbuch für Wirtschaftsgeschichte / Economic History Yearbook, Vol. 52, No. 2, December 2011.
A banking crisis began in Austria in May 1931 and intensified in July, when runs struck banks throughout Germany. In September, the crisis compelled Britain to quit the gold standard. Newly discovered data shows that failure rates rose for banks in New York City, at the center of the United States money market, in July and August 1931, before Britain abandoned the gold standard and before financial outflows compelled the Federal Reserve to raise interest rates. Banks in New York City had large exposures to foreign deposits and German debt. This paper tests to see whether the foreign exposure of money center banks linked the financial crises on the two sides of the Atlantic.
The Small, Private Banker in New York and Regulatory Change, 1893-1933
in Essays in Economic and Business History, May 2010.
A movement began in New York State in the late nineteenth and early twentieth centuries to protect depositors in small, private banks. These depositors were often located in immigrant communities that did not possess the local level of capital to support a national or state-chartered bank. Small private banks, the only entities that often served certain poor areas, were less regulated than state-chartered or national banks, especially in regards to capital levels and quality of assets. The fact that bank managers possessed the ability and incentives to assume additional risk exposed depositors in these small private banks to an increased danger of moral hazard. Implicitly recognizing this risk after the Panic of 1893, state officials in New York began a movement that spanned three decades to reduce the incentives for moral hazard and increase the protection for depositors in these immigrant communities.
Intensified regulatory scrutiny and bank distress in New York City during the Great Depression
with Gary G. Richardson in Journal of Economic History, Vol. 69, No. 2, June 2009.
Bank distress peaked in New York City, at the center of the United States money market, in July and August 1931, when the banking crisis peaked in Germany and before Britain abandoned the gold standard. This article tests competing theories about the causes of New York's banking crisis. The cause appears to have been intensified regulatory scrutiny, which was a delayed reaction to the failure of the Bank of United States, rather than the exposure of money center banks to events overseas.
Working Papers
Excess Reserves during the Great Contraction: Evidence from the Central Money Market of New York City, 1929 to 1932
Recent turmoil in the financial sector has once again thrust bank liquidity and balance sheet management into the spotlight. History offers some insight into liquidity management in times of financial distress. The banking industry experienced a significant amount of turmoil during the Great Contraction of the 1930s. One aspect banks had control over was their reserves. This paper exploits a unique data set of state banks that were Federal Reserve members and state banks that were not members of the Federal Reserve System in New York City. Nonmember banks, without access to a lender of last resort, had a precautionary demand for reserves, increasing their ratio of excess reserves to assets after the first banking panic. Compared to state member banks, nonmember banks held a higher ratio of assets in excess reserves. The discount rate, representing the price of liquidity, had a positive relationship with reserves, suggesting that Federal Reserve policy could have affected excess reserve holdings during the Great Contraction. These results provide evidence that the actions of banks in the central money market support what banking theory suggests about the demand for liquidity during time of a financial crisis. They also provide evidence that banks were reacting to the banking panics in the pre-holiday shocks as they were after 1933.
Too-Big-to-Fail and Leverage of Systemically Important Commercial Banks: Evidence from the Booms before the Great Depression and Great Recession
with Gary G. Richardson and Christoffer Koch.
The regulatory framework for commercial banks evolved over the 20th century. So did the response of commercial banks to economic booms. During the boom before the Great Depression, capital requirements for commercial banks were low and fixed. Bankers faced substantial liability if banks failed. Failing banks were not bailed out by the government. During the boom before the Great Recession, capital requirements were proportional to risk-weighted assets. Bankers faced limited liability. Banks deemed too-big or too-connected to fail received bailouts. Across these regimes, we document differences in commercial banks' responses to asset-market booms. During the Roaring 20s, the largest and most connected banks in America's money center of Manhattan raised capital and lowered leverage in absolute terms and relative to smaller, less connected banks. During the boom at the beginning of the twenty-first century, America's largest and most connected institutions kept capital levels constant and near regulatory minimums. Our results suggest that more market discipline would have induced the largest U.S. banks to hold greater capital buffers prior to the financial crisis of 2008.
Works in Progress
Financial Fragility, Bank Liquidity and Leverage: Trends and Cycles during the National Banking Era
with Christoffer Koch.
Financial crises are a defining feature of the U.S. National Banking Era of the late 19th century. Previously, aggregate and disaggregate investigations of this imporant era relied on a measurement of bank capital as a gross concept and found banks had substantially higher capital ratios than present-day financial intermediaries. We incorporate asset-side liquidity, lending to the banks’ own directors, to refine our understanding of the capital dynamics during the numerous crises of the national banking era. We develop a new, adjusted net measure of bank capital. To enable meaningful comparisons with modern day concepts of bank leverage, we net out lending of commercial banks to their own directors, who were usually large capital holders during that time period. We find that – depending on the time period – the levels and the dynamics of capitalization in the National Banking Era vary substantially using our new net measure of leverage. For example, during the crisis of 1884 our data reveals substantial inter-crisis repayments of director loans, which apparently served as liquidity buffers reminiscent of the present day High-Quality Liquid Asset (HQLA), to bloster the net capital while leaving gross capital little changed. Furthermore, at the end of the sample our new measure of net capitalization is substantially lower than the levels of capitalization banks possessed in the U.S. in the 2000s preceding the Global Financial Crisis. Our findings point to private precautionary off-balance sheet liquidity insurance in the absence of a government-sponsored lender of last resort provided by directors of small banks.
The Federal Reserve as a Start-Up: New Evidence from the Daily Discount Ledger from 1914 – 1917 (with Christoffer Koch)
How did the U.S. central bank, the Federal Reserve, operate in its founding years? Previous research primarily focused on aggregate consequences of the Fed's founding, yet to date little empirical work has been done to reveal details of the underlying operational mechanism---discount lending, along policy relevant dimensions such as seasonality, volumes, and maturity. We analyze a unique daily discount ledger from the Federal Reserve Bank of Dallas containing all discount transactions from 1914 - 1917. Details identify the transacting commercial bank, underlying collateral, location, size, and maturity of the discounted loan. Our initial findings reveal, first, that the operationalized discount rate was persistently above the floor published for the district. Second, we document an acceleration of the number and total volume of transactions. At the same time the average amount of transactions is little changed. However, we do find two distinct episodes that saw a clustered rise in the average volume of discounting. Third, maturity and the amount of the discounted collateral exhibit a negative relationship---the smallest size decile has nearly twice the remaining time to maturity compared to largest size decile. Fourth and finally, detailed transaction-by-transaction ledger data allows us to reveal important, otherwise hidden intra-week patterns . The average volume of transactions is markedly larger on Fridays, suggesting a role for the Fed to ensure liquidity over weekends. Saturday transaction volumes and maturities are substantially lower than during regular weekdays. These noticeable intra-week volume and maturity patterns suggests that the Fed helped the liquidity of banks balance sheets and permitted banks to carry less liquid assets.
The Economics of Bank Moratoria: The Milford Plan, Michigan, and Bank Survival during the Great Depression
Theoretical literature addresses the optimal arrangement between depositors and managers in the event of financial distress. Yet there are few instances where actual arrangements during a crisis have been observed. Archival evidence previously unused from the 1930s provides a rare example of how bank managers and depositors negotiated to inject capital into the bank while reducing the threat of mass withdrawals. These plans, known as “the Milford Plan,” were eventually enacted by more than 130 banks in the state of Michigan. Under the plan, every stockholder of the bank provided $100 for each share owned. These funds were paid into the surplus funds of the bank to strengthen its capital structure. In return for this recapitalization effort, the depositors would agree not to withdraw a certain percentage of their deposits as set by the board of directors of the bank. This arrangement would continue for one year after the moratorium was declared and approved of by the depositors. During this time, no dividends were to be declared and the cash contribution by the shareholders would not serve as a credit against their 100 percent liability in the event of insolvency. In essence, the Milford plan provided an injection of capital by the owners of the bank, under the condition that depositors retained enough funds in the bank to keep it solvent. This episode provides evidence that depositors and bank owners can negotiate agreements to keep the bank solvent and liquid in the absence of a lender of last resort.