Leverage Effect (Financial Amplification)

Three usages

Three distinct concepts sit under the heading 'leverage effect'.

First, the general principle that leverage magnifies gains and losses. Investing with borrowed money enlarges available capital and brings either large profits or the risk of failing to repay. Leverage is defined as the ratio of asset value to the cash needed to buy it, and it expands and contracts with the business cycle in a procyclical manner. When asset values fall, collateral values shrink, margin calls follow and forced selling ensues, so prices can fall further than fundamentals warrant. Buying stock on 50% margin means losing 40% when the price falls 20%.

Second, the financial accelerator. In macroeconomics this is the process by which an adverse shock is amplified by deteriorating financial market conditions, with worsened conditions in the real economy and financial markets spreading financial and macroeconomic contraction. The core is asymmetric information: a firm's borrowing capacity depends fundamentally on the market value of its net worth, so falling asset prices weaken balance sheets and net worth, cut borrowing capacity and dampen investment. Reduced economic activity then drives asset prices down further.

Third, the statistical usage in quantitative finance. In the constant elasticity of variance (CEV) model (John Cox, 1975), when γ<1 a fall in the stock price and a rise in the leverage ratio increase equity volatility, a commonly observed feature of stock markets. In commodity markets γ>1 is common, so volatility rises with price increases instead.

Conditions and transmission channels

The financial accelerator framework rests on three conditions. Unless fully collateralized, external finance (debt) costs more than internal finance (equity); the external finance premium rises with the amount of funds needed and, for a fixed amount, falls as the borrower's net worth rises; and a fall in the borrower's net worth reduces the internal finance base while raising the need for and cost of external finance. Its theoretical microfoundation is the principal-agent problem in credit markets: lenders cannot costlessly obtain information about a borrower's investment opportunities, creditworthiness and risk-taking, and this agency cost generates the accelerator.

The framework is closely tied to the credit channel of monetary policy transmission. The credit channel is an indirect amplification mechanism in which policy changes affecting interest rates are amplified by endogenous changes in the external finance premium; it operates alongside the interest rate channel. The external finance premium is the difference between the cost of funds a firm raises internally and the cost of raising funds externally through equity and debt markets. The credit channel appears through two paths. The balance sheet channel holds that the size of the external finance premium is inversely related to the borrower's net worth; in the basic model it is expressed as a 'collateral-in-advance' constraint under which a firm's variable-input spending cannot exceed the sum of its total cash flow and the net present value of its assets. The bank lending channel holds that monetary policy changes affect banks' loanable funds and hence total lending.

The framework also applies to open economies. An adverse shock to a small open economy can be amplified by deteriorating international financial conditions, and initial shocks to productivity, world interest rates or country risk premia can cause a 'sudden stop' in capital inflows that accelerates the initial downturn. Agents in emerging economies often need external finance but face limited access to international capital markets because of information frictions or limited commitment.

History

The acceleration principle itself is older, in use from the early 1900s, with Aftalion's 1913 paper 『Les crises périodiques de surproduction』 regarded as its first appearance. Samuelson (1939) is a representative statement of the traditional view, arguing that a demand increase such as government spending raises national income and that this income induces consumption and investment, accelerating activity. The roots of the modern accelerator view go back to Fisher's (1933) debt-deflation theory. That theory, dealing with a downward spiral induced by over-indebtedness and reinforced by debt liquidation, deflation of asset and commodity prices, falling net worth and economic contraction, was neglected at the time behind Keynesian economics. Related earlier work includes Bernanke (1981, 1983), Bernanke & Gertler (1989, 'Agency Costs, Net Worth, and Business Fluctuations') and Kiyotaki & Moore (1997, 'Credit Cycles').

The term 'financial accelerator' was introduced into the macroeconomics literature by Bernanke, Gertler and Gilchrist's 1996 paper "The Financial Accelerator and the Flight to Quality" (Review of Economics and Statistics 78(1): 1–15). The paper's motivation was to explain the old puzzle that large swings in aggregate economic activity sometimes appear to originate in seemingly trivial shocks. As the view that financial market conditions matter for driving the business cycle gained ground, the framework revived as an analysis linking credit market imperfections to the propagation of downturns. Many economists today consider that it explains well the financial-macroeconomic linkages underlying the dynamics of the Great Depression and the subprime mortgage crisis.

Cases

The 1929 stock market crash is an example of deleveraging spreading into the real economy. In the 1920s margin requirements were loose and leveraging up to 90% with debt was not uncommon; when the market contracted many individuals received margin calls and had to add funds or have their shares forcibly sold, and that selling drove prices down further and triggered more margin calls. According to Rappoport & White (1994), however, margin requirements had already risen to record levels from late 1928 or early 1929, so borrowing ratios were falling.

During the 1998 Russian financial crisis many hedge funds pursuing arbitrage strategies suffered large losses and had to cut positions, and the resulting price moves deepened the losses and triggered further liquidation. Financial contagion also appeared, with price moves in one market causing price moves in another, and the episode heightened concerns about market disruption and systemic risk, leading the Federal Reserve to orchestrate a rescue of Long-Term Capital Management (LTCM). In the August 2007 quant hedge fund crisis, hedge funds again hit capital constraints and had to cut positions, and at that point prices moved more with liquidity factors than with movements in fundamentals.

In the 2007-2009 crisis, subprime losses in 2007-2008 ran to hundreds of billions of dollars but amounted to only about 5% of total stock market capitalization; because leveraged financial institutions bore the losses, spiral effects amplified the crisis and stock market losses exceeded 8 trillion dollars. Krishnamurthy (2010) estimates direct losses from subprime mortgage defaults at up to 500 billion dollars. Credit default swaps (CDS) also played an important role in the run-up: because a CDS buyer could buy the instrument without holding the bond, pessimists could drive asset prices very low, and standardization of CDS made large volumes of trading possible.

Lehman Brothers showed accounting leverage of 31.4 times in its last annual financial statements (assets of 691 billion dollars divided by shareholders' equity of 22 billion dollars), and the bankruptcy examiner Anton R. Valukas determined that actual accounting leverage was higher because of contested accounting treatments such as so-called repo 105. Notional leverage was more than twice that because of over-the-counter dealings (Lehman's notional derivatives stood at 738 billion dollars at the end of 2007).

Policy and welfare

One way to break the accelerator mechanism is to reverse the fall in asset prices. When asset prices fall below a certain level, the government can buy assets at that price to raise demand and push prices back, and the Federal Reserve did in fact buy savings institution mortgage-backed securities at abnormally low market prices in 2008 and 2009. Another instrument is to sever the link between a borrower's net worth and borrowing capacity. In a liquidity crisis the central bank should expand money supply to increase liquidity provision, with the strongest effect when investors are close to financial constraints.

In an economy with very high leverage, a small number of investors have borrowed heavily from all lenders in the economy, so asset prices are determined by a small group. When the failure of an extremely leveraged actor raises the likelihood of simultaneous failure by other leveraged actors, that actor can become indispensable to the economy, leading to the 'too big to fail' problem. Leveraged arbitrageurs consider only the maximization of their own profits and not the effect of their decisions on asset prices; because asset prices determine other arbitrageurs' investment capacity through their wealth and financial constraints, externalities arise that may not be socially optimal.

Regulation

Before the 1980s quantitative limits on bank leverage were rare. Banks in most countries had reserve requirements, but these do not limit leverage and are conceptually different from capital requirements. National supervisors began imposing formal capital requirements in the 1980s, and by 1988 Basel I standards applied to most large multinational banks. Basel I is credited with improving bank risk management but had two flaws: it did not require capital for all off-balance-sheet risk, and it gave banks an incentive to select the riskiest assets within each risk bucket. Work on Basel II began in the early 1990s and was phased in from 2005, and the poor performance of many banks in the subprime crisis prompted calls to reintroduce a leverage limit. Under Basel III banks are expected to maintain a leverage ratio (tier 1 capital divided by total exposure) above 3%, where exposure includes off-balance-sheet items and derivative add-ons.