The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
MARCH/APRIL 1992 ECONOMIC PERSPECTIVES A review from the Federal Reserve Bank of Chicago D eterm ining m argin for futures co n tracts: the role of private interests and the relevance of e x ce ss volatility State and local governm ents' reaction to recession A FEDERAL RESERVE BANK OF CHICAGO Contents D eterm ining m argin for futures co n tracts: the role of private interests and the relevance of e x ce ss v o la tility........................................................2 Jam es T. M oser Do lower margin levels cause increased stock price volatility? The author argues against this view and suggests that a different theory may explain the relation between margin levels and price volatility: futures exchanges raise margin levels when volatility increases, in order to compensate for their increased risk. State and local governm ents' reaction to re ce ssio n ...................................................................................... 19 Richard H. M atto o n and W illiam A . Testa During economic contractions, state and local governments often manage to maintain spending in the face of decreased revenue growth and requirements to balance their budgets. The authors explain how they do it. ECONOMIC PERSPECTIVES March/April 1992 Volume XVI, Issue 2 Karl A. Scheld, Senior Vice President and Director of Research ECONOMIC PERSPECTIVES is published by the Research Department of the Federal Reserve Bank of Chicago. The views expressed are the authors’ and do not necessarily reflect the views of the management of the Federal Reserve Bank. Single-copy subscriptions are available free of charge. Please send requests for single- and multiple-copy subscriptions, back issues, and address changes to Public Information Center, Federal Reserve Bank of Chicago, P.O. Box 834, Chicago, Illinois 60690-0834, or telephone (312) 322-5111. Articles may be reprinted provided source is credited and The Public Information Center is provided with a copy of the published material. Editorial direction Carolyn McMullen, editor, David R. Allardice, regional studies, Herbert Baer, financial structure and regulation, Steven Strongin, monetary policy, Anne Weaver, administration Production Nancy Ahlstrom, typesetting coordinator, Rita Molloy, Yvonne Peeples, typesetters, Kathleen Solotroff, graphics coordinator Roger Thryselius, Thomas O’Connell, Lynn Busby-Ward, John Dixon, graphics Kathryn Moran, assistant editor ISSN 0164-0682 Determ ining m argin for futures contracts: the role of private interests and the relevance of excess volatility Jam es T. M oser Margins should be made con sistent to control speculation and financial leverage. —Brady Report On Monday, October 19, 1987, the Dow Jones Industrial Average declined 508 points. The marketplace on the following day is usual ly described as melting down. This analogy to a runaway nuclear reaction reflects the fear during the morning hours of October 20, 1987 that overheated trading activity had over whelmed trading systems. Studies were com missioned to investigate the events of these two days and to propose remedies. One of these studies, the Brady Report, recommends raising margins on stock index futures contracts in order to reduce the chances of a future financial meltdown.1 Support for the higher margins proposed by the Brady Report stems from the view that low margins result in greater speculation which, in turn, leads to greater volatility. According to this view, volatility produced by speculative trading can be controlled by regulating margin. I call this view the Excess Volatility Argument. Another explanation of the link between vola tility and margin levels is founded on the recog nition that stock and futures exchanges face increased risk when stock market volatility increases. According to this view, stock and futures exchanges raise margin levels when volatility increases in order to compensate for the increased risk. I call this view the Pruden tial Exchange Hypothesis. 2 This article examines the relation between volatility and margin levels in order to assess the plausibility of the Excess Volatility Argu ment and the Prudential Exchange Hypothesis. The next section discusses the private interests involved in setting margin levels and their relevance to the justification of the Prudential Exchange Hypothesis. The Excess Volatility Argument is critiqued in the following section. Analysis of the theory underlying the Excess Volatility Argument, a review of existing evi dence on the links between margin and volatili ty, and new tests of the theory all fail to support the proposition that raising margins leads to reductions in volatility. Evidence for the Pru dential Exchange Hypothesis is mixed. Tests relating margin changes to previous levels of volatility fail to confirm the hypothesis. A cross-sectional approach to test this hypothesis is introduced and some preliminary results are reported. Conclusions concerning the Pruden tial Hypothesis and the Excess Volatility Argu ment are summarized in the last section of the article. P riva te in tere sts in d e te rm in in g m arg in re q u irem e n ts According to the Prudential Exchange Hypothesis, stock and futures exchanges both have an interest in managing their exposure to James T. Moser is a senior econom ist at the Feder al Reserve Bank of Chicago. The author is indebted to Janet Napoli and Jeff Santelices fo r research assistance. Comments from Herbert Baer, Ramon P. DeGennaro, Douglas Evanoff, Virginia Grace France, Carolyn McMullen, Janet Napoli, and Steven Strongin have been especially helpful. ECONOMIC PERSPECTIVES risks from trades routed through their exchange. Margins are an important means to this end. However, the nature of the risk in stock and futures markets differs hence, margin require ments play different roles in stock and futures markets. The next two subsections develop this distinction. The role of private interests in determining stock margin In stock markets, brokerage firms some times lend money to investors for the purchase of stock (see Box 1 for an explanation of how margin lessens the risk of stock brokers). Lend ing benefits brokerage firms because it increas es trading, thus increasing revenues from bro kerage fees. The risk inherent in lending is controlled by collateralizing these loans with stock. Brokerage firms further reduce risk by requiring investors to pay a portion of the pur chase price in cash. The amount of cash put up by the investor in a leveraged stock transaction is called margin. In particular, the amount of cash required when the position is initiated— the “down payment”—is referred to as the initial margin.2 Margin loans expose brokers to the risk that a stock price decline will produce losses in excess of the amount of posted margin. This risk increases both as the degree of leverage in the position increases and as the volatility of stock—the collateral—increases. Prudence motivates brokers to closely examine the ability of investors holding margined positions to cover their debt obligations. Increasing margin reduces the risk taken by the broker’s extension of credit. Thus, it is in the broker’s interests to require a prudential level of margin. The interests of the broker also include fees from trades executed on behalf of his or her customers. Lending facilitates trading by increasing the size of positions which can be held given the investor’s level of cash. Higher margins result in smaller loans, hence lower trading levels, other things equal. Thus, in creasing margins lessens brokerage fee income. Stock brokers set margin by considering both risk and profit, choosing the level of margin which is expected to yield a competitive return for the level of risk. Stock exchanges take the interests of bro kers into account when setting limits on margin lending. Exchanges consistently acting against the interests of their brokers lose business as FEDERAL RESERVE BANK OF CHICAGO brokers find more favorable routes for trades. Thus, the Prudential Exchange Hypothesis predicts that a stock exchange sets margin lev els which are consistent with the interests of stock brokers affiliated with the exchange. These interests, as previously identified, lead to levels of margin which balance revenues from trading activity with the risk of losses on credit extended to clients. The role of private interests in determining futures margin Determination of margin requirements for futures contracts raises concerns which are similar to those of the stock broker. Like stock brokers, futures exchanges, acting on behalf of their members, set futures margins to control their risks. However, the risks faced by the stock broker and the members of the futures exchange are not identical. In this section, I use a hypothetical futures contract on a stock index to develop the role of margin for futures positions. Futures contracts trade on a variety of assets. Examples are contracts on wheat, fro zen pork bellies, foreign exchange, Treasury bonds, and stock indexes. Contracts are distin guished by the price of the asset or commodity used to determine payments to the parties in the contract. As an example, consider the follow ing hypothetical contract. Over the next three months, for every point the Standard and Poor (S&P) 500 rises from its present level, Mr. Short will pay Ms. Long $1,000. For every point it falls from this level, Ms. Long will pay Mr. Short $1,000.3 Mr. Short and Ms. Long are referred to as counterparties in the futures contract. The counterparties are further identified as holding the long or short side of the contract. In this contract, Ms. Long holds the long side, which commits her to make payments when the fu tures price falls and entitles her to receive pay ments when the price rises. Conversely, Mr. Short holds the short side, which entitles him to receive payments when the futures price falls and commits him to make payments when the price rises. Payments between the counterpar ties are determined by marking the contract to the current price of other futures contracts on the same underlying basket of commodities or assets. This mark-to-market procedure is con ducted daily. Futures contracts feature terms serving two purposes. First, contract terms 3 determine the usefulness of contracts. Second, contract terms enable the exchange to manage customer insolvency problems. Futures are useful as low cost substitutes for transactions in the underlying asset. To see this, note that by carefully specifying a particu lar group of assets for determination of the final settlement price, the futures price will move closely with the price of the asset group. Thus, changes in the futures price for the S&P 500 are closely linked with changes in the prices of the 500 stocks used by Standard and Poor in con structing that index. The alignment of these prices is useful to individuals and firms seeking low cost means of altering the sensitivity of their portfolios to price changes. To see the usefulness of futures contracts, suppose Mr. Short owns a portfolio of stocks, many of which are included in the 500 stocks comprising the S&P Index. This portfolio is called his cash position to distinguish it from the futures contract. When the prices of stocks BOX 1 Leverage, risk, and the role of margin The relation among leverage, risk, and the role of margin is most easily illustrated in the case of stocks. Borrowing to purchase stocks has leverage implications for both the borrowers (investors) and lenders (brokerage firms). This point can be illustrat ed with a simple T account. Market value of shares purchased $10,000 $6,000 Loan from broker $4,000 Equity placed by purchaser In the example, the initial margin requirement is 40 percent.1 Stock valued at $100 per share requires the purchaser of 100 shares to pay $4,000 of their purchase price. The broker lends the pur chaser $6,000. This combination of funds produces $10,000 paid to the seller of the stock. To see the consequences of leverage for borrow ers and lenders, we examine the effect of stock price changes. First, suppose the stock price rises to $110. After this price change, the T account looks as follows: Market value of shares purchased $11,000 $6,000 Loan from broker $4,000 Equity placed by purchaser $1,000 Gain on stock Thus, the $4,000 invested has gained $1,000 for a 25 percent return on invested funds. Had the inves tor not purchased the stock on margin; and paid the full $10,000 for the stock, the rate of return would 4 have been only 10 percent. The margined position earns 2.5 times the percentage change in stock prices (2.5 x 10 percent = 25 percent). These gains can be realized by selling the shares for $ 11,000, repay ing the loan balance of $6,000 from the proceeds, leaving $5,000. Examining the potential downside from a mar gined purchase explains why most stock purchases do not use margin. Suppose the stock price declines to $90. Now the T account looks like this: Market value of shares purchased $9,000 $6,000 Loan from broker $4,000 Equity placed by purchaser ($1,000) Loss on stock The $4,000 invested results in a loss of $1,000 for a 25 percent loss. Had the purchase price been paid in cash, the percentage loss would have been only 10 percent. The alternative way of seeing this is to recognize that the ability to hold 2.5 times more shares implies that any losses will be magnified by 2.5. Further, as shown later, equity balances must be restored when these balances fall below a preset level. Compliance with this rule may require inves tors to sell other asset holdings to meet the call for additional equity.2 Thus, from the investor’s per spective, margined stock purchases lever up risk. The leverage factor is 1 + Loan/Equity. For the initial position, this is 1 + 6,000/4,000 = 2.5. Now consider the above transaction from the lender’s point of view. The lender will also have an interest in this leverage factor. Suppose the stock price declines to $60, so that the T account is: ECONOMIC PERSPECTIVES in the portfolio decline, the value of Mr. Short’s cash position declines. However, his short futures contract position entitles him to receive payments from Ms. Long when stock prices decline. These payments lessen the extent of losses realized from the cash position. Thus, futures contracting can reduce an investor’s sensitivity to price changes. This use of futures contracts is called hedging. Ms. Long finds the contract useful for a different reason. Generally, her cash position Market value of shares purchased $6,000 $6,000 Loan from broker $4,000 Equity placed by purchaser ($4,000) Loss on stock The broker faces a problem. Liquidating the position at its current market value insures that the outstanding balance of the loan is paid off. Not liquidating the position puts the broker at risk that the stock price will decline further and that the inves tor will not be able to make up the difference from other sources. If the latter case occurs, the broker suffers a loss. The extent of this loss depends on the additional decline in stock price and the amount the broker can recover from the other resources of the investor. Thus, once the investor has lost the equity in the position, the broker relies on estimates of the extent of these other sources. To avoid the risk inherent in these estimates, the broker establishes a maintenance margin requirement. When the level of equity falls below the maintenance margin re quirement, a call for additional margin is made. Receipt of the called-for funds decreases the broker’s reliance on estimates of other sources of wealth. Once funds are received, the broker’s risk is reduced. An additional decline in stock price will, with cer tainty, be absorbed by the investor up to the new margin deposit. ’Currently initial margin requirements are 50 percent. The example uses 40 percent to clarify which portion is required from the investor (40 percent) and which is lent by the broker (60 percent). bankruptcy law prevents access to certain assets to meet financial obligations. FEDERAL RESERVE BANK OF CHICAGO consists mostly of low risk bonds. At times she has concluded that stocks are undervalued. Taking the long side of a futures contract al lows her to increase the sensitivity of her port folio to changes in stock prices. In particular, when her assessment that stocks are underval ued proves true, she realizes gains from her futures position. This use of futures contracts is called speculation. These uses of futures contracts are a cost effective means to the respective ends of Mr. Short and Ms. Long. Both results could be accomplished using transactions in the stocks themselves. Mr. Short could reduce his sensi tivity to stock price changes by selling stocks and investing the proceeds in low risk assets such as Treasury bonds. Ms. Long could in crease her sensitivity to stock price changes by selling some of her bonds and buying stocks. Each prefers to accomplish his or her respective end at the lowest possible cost. Futures con tracts often provide the least costly route to adjusting portfolio sensitivity. However, contracts which are not depend able will not be useful. In the stock index fu tures contract described above, both Mr. Short and Ms. Long find the contract advantageous in the sense that it represents a low cost means of altering their sensitivities to changes in a broad measure of the stock market. However, Mr. Short might regard such a contract as worthless if he had reason to believe that, should prices fall, Ms. Long would be unable to make the required payment.4 Similarly, Ms. Long’s concerns about Mr. Short’s ability to pay lower her assessment of the value of such a contract. Except for this insolvency issue, both find the contract useful. Thus, each party has an inter est in resolving the insolvency problem at rea sonable cost. Resolution of the insolvency problem is the role of the exchange. Exchanges fulfill this role by requiring that all contracts clear through members of the clearing association affiliated with the respective exchange. In this process, the clearing association becomes counterparty to each side of all contracts traded on the ex change. Should either the long or short side fail to perform its obligations, the loss is realized by the clearing association rather than the original counterparty. Continuing the above example and introducing the role of the exchange, sup pose the stock market rises ten points. Mr. Short owes Ms. Long $10,000. If he has be 5 come insolvent, the contract guarantee assures that Ms. Long is paid the $10,000.5 This per formance guarantee removes the respective credit risk concerns and focuses the attentions of the counterparties on contract price. Neither party finds it necessary to expend resources to evaluate the credit risk of the other party. This resolution of the insolvency problem increases the value of futures contracting for both parties. Performance guarantees provided to the coun terparties are clearly costly. The exchange, acting to maintain the solvency of its clearing association, attempts to manage its potential for loss. This is accomplished by managing the exchange’s exposure to the credit risk stem ming from each participant in the contract. Management of the exchange’s credit risk uses an overlapping system of solvency require ments, mark-to-market arrangements, and mar gin requirements. To see the role of the com ponents of this system, I begin with an ideal characterization of the marketplace, then relax various assumptions in order to explain how each of these components is used to manage the credit risk of a futures exchange. Evidence of solvency is the first level of protection. We can see the role of solvency requirements by imagining an ideal market place where monitoring of the wealth of each party is perfect and continuous. With the addi tional assumptions of immediate access to the wealth of these parties and unlimited liquidity in markets where assets can be immediately and costlessly sold off; no counterparty would be exposed to risk. Under these conditions, at the instant when a party is determined to be insolvent, that party’s assets would be immedi ately attached, their futures positions closed out, and assets sold with the proceeds used to cover shortfalls arising from the futures posi tion. Thus, with this characterization of the marketplace, the exchange avoids all risk of loss by relying on its legal authority to close out futures positions as counterparties become insolvent. Relaxing the assumption of costless asset liquidation, the exchange incurs transactions costs in liquidating positions. This is readily resolved by applying “haircuts” to asset values when computing net worth for solvency purpos es. That is, the value of each asset in the inves tor’s portfolio is reduced—haircut—by the amount of transaction cost incurred on sale. 6 Thus, solvency requirements are sufficient for the exchange to manage its exposure with this characterization of the marketplace. If the assumption that assets can be liqui dated immediately is dropped, exchanges prefer asset holdings which can be used to settle pay ment obligations. On determining that a coun terparty has become insolvent, the exchange seeks to avoid risk by closing positions and disbursing payments quickly. Delays encoun tered in the liquidation of assets increase the exchange’s risk of realizing further losses. Since futures contracts require that positions realizing gains be paid in cash, exchanges have a strong preference for asset holdings in cash or readily convertible to cash. This enables the exchange to attach assets which can be immedi ately applied to fulfill its required payments of gains. Thus, margin requirements amend the solvency requirement by stipulating that futures positions be supported by liquid asset holdings. The requirement that margin balances be de posited with the exchange further enhances this liquidity requirement: funds are immediately available to the exchange. Mark-to-market arrangements augment the arsenal of exchange protections against credit risk by substituting for perfect monitoring of wealth. Frequent marking to market creates a flow of information to the exchange on the solvency of counterparties. To see this, recall that mark-to-market rules require positions incurring losses to cover these losses with cash payments. Cash paid by customers to brokers is forwarded to the clearing member and then to the clearing association. Brokers observing the payments made by their customers can infer their ability to continue to cover losses. Like wise, by observing delays in payments made by clearing members, the clearinghouse can infer their members’ abilities to continue to cover losses. Delays in making mark-to-market pay ments reveal liquidity problems which may develop into solvency problems. The cost of obtaining this information is decidedly less than the cost of direct monitoring systems which might be regarded as nearly ideal. As the frequency of marking contracts to market increases, the exchange approximates the ideal case of continuous monitoring of counterparty wealth. However, this approach is costly. Reducing the mark-to-market frequency places the exchange at risk that the counter ECONOMIC PERSPECTIVES party has become insolvent since the position was previously marked to market. Thus, fu tures margin balances are used to collateralize the completion of the obligation to make markto-market payments. Margin balances bond the performance of contract holders to make the cash payments required when contracts are marked to market.5 Failure to complete this obligation creates an exercisable claim on the margin account. By exercising this claim while simultaneously closing out the futures contract, the maximum loss of the exchange is the loss on closing out the futures position netted against the margin balances for the account.7 Thus, futures exchanges rely on solvency, mark-to-market arrangements, and margin to control the credit risk inherent in futures con tracting. Margin provides the clearinghouse with liquid assets which lowers the cost of making payments to contract holders. Mark-tomarket arrangements provide a signal of the level of liquidity available. The combination of mark-to-market arrangements and margin limits the credit risk exposure of the exchange. This combination of lower credit risk, lower costs of transacting, and the presence of an information generating process for customer liquidity low ers the cost of providing guarantees against counterparty risk. This increases the usefulness of futures contracting by increasing its depend ability. Distinctions in margin assessments provide additional support for the idea that futures ex changes rely on multiple avenues to manage their exposure to credit risk. For example, qualified hedgers have long or short cash posi tions in the asset underlying the futures con tract. Because losses and gains on futures posi tions are offset by changes in the value of the underlying asset, hedgers expose the exchange to less credit risk exposure than do speculative positions. Recognizing their exposure is less, futures exchanges specify lower margin re quirements for qualified hedgers than for more speculative positions.8 Clearing members of the exchange are another category of partici pants having reduced margin requirements. Clearing associations closely monitor the risk of clearing member insolvency. Having in curred the cost of this additional monitoring activity, the clearing association increases its reliance on these solvency assessments and, consequently, reduces the level of margin re quired for clearing member positions. FEDERAL RESERVE BANK OF CHICAGO Private interests and the Prudential Exchange Hypothesis The above discussion shows that private interests motivate both the stock broker and the futures exchange to require margin. Use of margin facilitates trading of stocks and futures contracts, thereby increasing revenues from fees paid to stock brokerage firms and to mem bers of futures exchanges. However, inade quate margin levels for stock positions increase the riskiness of loans made by brokerage firms. Inadequate margin levels for futures contracts increase the cost of contract-performance guar antees. In both cases, the risk of loss encourag es the affected parties to reduce these risks by increasing margin levels. Both stock and fu tures exchanges have incentives to keep mar gins at an optimal level at which fees from increased trading provide an adequate return for the risks they bear. Clearly, an increase in stock price volatility increases the potential losses of investors and hence increases the risk of insolvency. Thus, it would make sense for exchanges to respond to increased volatility by increasing margin re quirements. This might reduce revenues from trading activity, but will clearly decrease the risk of losses from insolvency. Conversely, a decrease in volatility lessens the threat of insol vency. So, it would make sense for exchanges to respond to decreased volatility by lowering margin requirements in order to increase reve nues from trading activity. The Prudential Exchange Hypothesis is the hypothesis that exchanges do indeed act in the way just de scribed, raising margins in response to in creased volatility and lowering margins in response to decreased volatility. A positive association between observed changes in vola tility and subsequent changes in margin levels would be evidence in favor of the Prudential Exchange Hypothesis. Below I describe the results of research investigating the relation between changes in volatility and changes in margin levels and discuss the implications for the Prudential Exchange Hypothesis. M arg in d e te rm in a tio n and th e Excess V o la tility A rg u m e n t While the Prudential Exchange Hypothesis suggests that increases in volatility should lead to increases in margin, the Excess Volatility Argument suggests that increases in margin should lead to decreases in volatility. The 7 Excess Volatility Argument originated as an argument to justify the regulation of margin on stocks.9 The argument is frequently extended to margins for futures contracts. This section explains the Excess Volatility Argument as it is applied to stocks. I then demonstrate problems with the argument. Federal regulation of margin requirements on stocks began with the Glass-Steagall Act of 1934. The act empowered the Federal Reserve to specify margin requirements for stock.10 This portion of the act was motivated by con cern that margins prior to the 1929 stock mar ket crash had been too low. Following the 1929 crash, proponents of the Excess Volatility Argument felt that low stock margin require ments encouraged speculation which exacerbat ed price swings. The claim that there is a direct relationship between speculation and volatility is based on the view that trends in market prices can be identified as they occur and that specula tors respond to these trends by taking positions which profit from near term anticipated price changes. This combination produces a band wagon effect or speculative bubble. For exam ple, according to this view, if speculators per ceive markets as rising, they think that easy profits can be had by buying into the market quickly to take advantage of the next round of price increases. The added pressure of these orders to buy elevates prices further. Each round of profits increases interest in “jumping on the bandwagon.”11 Proponents of the Excess Volatility Argu ment believe that private brokerage firms can not be relied on to limit speculation by requir ing high margins on stocks because high mar gins would decrease trading volume and the profits from brokerage fees. The solution to the problem of excessive volatility, according to the Excess Volatility Argument, is to move control of margin from the securities industry to government. By raising the cost of speculative positions, episodes of excessive speculation could be managed by officials who do not ben efit from increased trading activity. Further, these officials are answerable to the public for their decisions, making them sensitive to the concerns of the public. Problems with the Excess Volatility Argument The Excess Volatility Argument as applied to stocks depends on a number of implicit as sumptions. First, investors are assumed to 8 ignore the risk of participating in speculative excesses. Second, brokerage firms are assumed to ignore their risks in facilitating the trades of these investors. Third, investors are assumed to lack opportunities to avoid margin require ments. If any of these implicit assumptions are not plausible, then the argument is less credible. First, consider the assumption that most investors ignore the risk involved in specula tion. According to the above scenario, inves tors buy in response to price increases produced in previous rounds of buying. They ignore fundamentals, such as the ability of the firm to make expected dividend payments, which de termine the fundamental value of stocks. For the scenario to work, investors must ignore the fact that as stock prices rise they become fur ther removed from fundamental values.12 In vestment motivated by this reliance is risky. The larger the distance from the stock price to its fundamental value, the greater the necessary correction. Buy orders which increase upward pressure face the risk of increasingly large losses. Thus, investors placing these orders are ignoring the risk that the price correction will produce a loss. As risk averse investors raise their assessments of risk, they require higher returns. However, in this case, expected returns must decline as the size of the necessary correc tion increases. It is not plausible to claim that in general, investors ignore the risks of specula tion in this way. The Excess Volatility Argument also ne glects the incentive of brokerage firms to set margins prudentially. As previously demon strated, individual brokerage firms face the risk that margin loans will not be repaid if customer losses exceed available funds. To control this risk, brokerage firms have incentives to raise margin levels. These incentives mitigate the higher revenues from increased trading activity. The exchanges recognize that brokerage firms near bankruptcy may compete for broker age fees by lowering margins. These firms will be more willing to require lower margin be cause lower margin increases the number of orders placed through these firms and increases revenues from brokerage fees. This additional business prevents bankruptcy provided the realized losses from insolvent customers are small relative to the additional revenue from fees. The incentive to take this chance is great est for firms which have the least to lose; that is, brokerage firms which are nearly bankrupt. ECONOMIC PERSPECTIVES However, this form of competition harms via ble brokerage houses in three ways. First, com petition for business reduces the immediate revenues from brokerage fees for viable firms. Second, bankruptcy of a brokerage firm lessens industry good will.13 This intangible asset is the capitalized value of trading activity which stems from confidence that brokers properly represent customer interests. Evidence that brokerage firms are aggressively pursuing their own interests damages this confidence, reduc ing the value of their good will. Third, brokers must be confident that commitments made with other brokers will be honored. Insufficient margining by individual brokerage firms less ens confidence in the completion of these com mitments. This leads to increased costs as brokers replace the surety afforded by adequate margin balances with increased monitoring of the financial well being of the other brokers. To reduce these costs, exchanges, acting in the interests of the industry, set minimum margin requirements. These minimums prevent nearly bankrupt firms from increasing their risks to attract additional brokerage fees at the expense of the remainder of the industry. Third, for margin regulation to work as proponents of the Excess Volatility Argument suggest, investors must lack alternative sources of funds. Margin requirements specify the amount of collateral which must be deposited for loans which are collateralized by stocks purchased with the funds provided. These requirements can be understood as restrictions on leverage which can be avoided. For exam ple, individuals can avoid margin restrictions by seeking loans on their other sources of wealth, such as funds from a second mortgage or borrowing against the cash value of insur ance contracts. These sources can be used to create “homemade” leverage at higher levels than those allowed using credit collateralized with stock holdings. In addition, Fishe and Goldberg (1986) point out that if leverage pref erences exceed those available under margin regulations, firms can increase their debt to provide any desired level of leverage. The ability to avoid restrictive margin requirements suggests that the regulation will be relatively ineffective.14 The above objections show that the Excess Volatility Argument as applied to stock markets has a number of weaknesses. Consequently, it FEDERAL RESERVE BANK OF CHICAGO does not present a strong case for the claim that controlling margin will influence the volatility of stock prices. Proponents extend the Excess Volatility Argument to futures markets.15 This extension ignores the differing roles of margin in the respective markets. The objections de scribed above also hold for the Excess Volatili ty Argument as it applies to futures markets. Furthermore, there may be additional difficul ties for the case of futures markets since margin plays a different role in futures contracts than in stock transactions. Analysis of the terms of futures contracts reveals no compelling reason to expect margins to control volatility in futures markets. In this section, I have described some of the conceptual difficulties for the Excess Vola tility Argument. In the next section, I consider the empirical evidence concerning the effects of margin changes on the volatility of prices. Evidence o f th e e ffe c ts o f m arg in changes on v o la tility A number of empirical findings do not support the claim that margin levels affect volatility. First, in order to have an effect on stock price volatility, equity positions funded by margin loans would have to constitute a sizable portion of investments in the stock market. Figure 1 graphs the dollar value of securities margin loans on equities as a percent age of the market value of corporate equity over the period 1968 to 1988. The dollar value of margined securities positions are a small portion of total stock holdings. Thus, attemptFIGURE 1 Extent of margin positions: NYSE margin as a percent of market value 9 ing to decrease the number of margined securi ties positions by raising the cost of holding these positions would influence stock prices only if speculative activity affecting a small portion of stock holdings could have a signifi cant impact on prices for both margined and unmargined stocks. With such a large percent age of equity holdings unaffected by the level of required margin, policies influencing the level of margin required to purchase equities are unlikely to significantly affect volatility.16 Considerable empirical research examines the links between margins on securities and the volatility of security prices. This literature is extensive and is not reviewed here.17 A repeat ed finding is that changes in equity margins are not related to subsequent changes in stock price volatility.18 Similar research for a wide array of futures contracts shows that margins on futures con tracts are an ineffective tool for reducing vola tility. Previous work by Furbush (1988) com pares S&P 500 volatility before and after mar gin changes on the S&P 500 futures contract, and finds no significant change in volatility. On the other hand, Kupiec (1990) finds a posi tive association between daily volatility esti mates for the S&P 500 index and previous initial margin rates (the amount of margin di vided by contract value) for that contract. Both results contradict the negative association pre dicted by the Excess Volatility Argument. In this section, I present additional evi dence that raising futures margins does not lower the volatility of the futures contract price. As with any literature testing for a nonzero effect, econometric difficulties can bias the test toward finding no effect. Recognition of this problem encourages careful researchers to try alternative approaches and repeated testing of a nonzero effect. My evidence improves on the existing literature in several ways. First, I use a new econometric technique to obtain volatility estimates. The procedure uses a method which improves the measurement of volatility and isolates changes in margin from changes in the level of futures price. Second, I test both the Prudential Exchange Hypothesis and the Excess Volatility Argument. My procedure consists of testing the hy pothesis that margin changes are associated with the volatility of two financial futures con tracts (see Box 2 for details of this procedure). Using leads and lags of the margin change 10 variables allows a determination of the time ordering of the relationship between margin changes and volatility. That is, using Equation 2 (see Box 2), we can determine whether changes in volatility come before or after changes in margin. The approach utilizes the persistence of volatility to associate margin changes occurring around a volatility shock.19 The test employs rates of margin changes which occur before the date of observed volatil ity (margin change “lags”) and rates of margin changes which occur after the date of observed volatility (margin change “leads”) in a regres sion having volatility as the dependent variable. The coefficients on these before and after mar gin changes are relevant to two quite different hypotheses about the relationship between margin changes and volatility. According to the Prudential Exchange Hypothesis, futures exchanges respond prudentially to higher vola tility by increasing margin requirements. If this hypothesis is correct, then margin changes should occur after shocks to volatility. For example, if volatility of a futures price rises due to an oil crisis, margins on affected contracts should rise in response. Thus, there should be positive coefficients on margin changes occur ring after observed volatility. That is, positive coefficients on margin changes occurring after observed volatility indicate that futures ex changes, acting to protect their interests, raise margin when exchange officials observe in creases in volatility. Thus, positive coefficients on margin changes occurring after observed volatility can be taken as evidence affirming the Prudential Exchange Hypothesis. Proponents of the Excess Volatility Argu ment expect margin increases to reduce volatili ty. Evidence that volatility is persistent implies that volatility will not change unless a subse quent shock produces a change. Proponents of the Excess Volatility Argument argue that margin changes shock volatility by raising the cost of holding speculative positions. Thus, increases in margin lower volatility and de creases in margin raise it, according to propo nents of the Excess Volatility Argument. A finding of negative coefficients on margin changes occurring before observed volatility is consistent with this expectation. A related question concerns the length of time separating futures margin and volatility changes. The low cost of futures trading sug gests that responses to a change in margin are ECONOMIC PERSPECTIVES likely to be quickly observed. This suggests the time between margin changes and observed volatility need not be long. Alternatively, if margin changes produce purely transitory ef fects, they would not be a particularly useful policy tool.20 This motivates examining a long er interval. In order to test the Excess Volatili ty Argument, I looked at margin changes that occurred up to twelve trading days before ob served volatility. Twelve trading days are more than one-half month, so it seems reasonable to expect that any effects from a margin change would be observed during this interval. Also, if margin changes produce effects which persist BOX 2 Procedure to test association of margin changes and volatility Davidian and Carroll (1987) introduce a meth od later extended by Schwert (1989) to calculate daily volatility estimates. Schwert and Seguin (1990) show that, assuming normality, this proce dure gives unbiased estimates of daily return stan dard deviations. The procedure iterates between a specification for mean returns and a separate speci fication for volatility. Equation 1 gives the specifi cation for the mean return from a futures contract as follows: (1) r =X\ 6 + 8,; where rt is the continuously compounded return for a futures contract at time t. This return is condition al on information available at t such as the month of the year and previous returns. This information set is represented by X. The residual, £f, captures the effects on returns from unanticipated events occur ring at time t. The parameter B summarizes the contribution of information items in the determina tion of returns. The variance of Ef summarizes the volatility due to unanticipated events over the sam ple period. Under certain conditions e is an effi cient estimator of the true volatility.1 One of these conditions is that volatility is unchanging or homoskedastic. If the error terms are heteroskedastic, then we need to identify the source of heteroskedasticity in order to correct for it in Equation 1. That is, we need a theory which can be tested about the deter minants of volatility in futures returns. The Excess Volatility Argument is a testable theory that margin affects volatility. Equation 2 expresses the relevant theory as follows: (2) le I = Y\ a + k Z x dmi+i + p ,; i = -k, i*0 where l£(l is the absolute value of the residual from Equation 1, F are information-set variables which might affect the volatility of returns, and dmt are percentage changes in margin requirements at time t. The parameters a and y. summarize the impact of FEDERAL RESERVE BANK OF CHICAGO these variables on volatility. Nonzero values for these parameters imply that volatility is affected by the associated variable. Of primary interest here are the y. which summarize the effect of margin changes. A negative coefficient implies that margin increases are related to lower volatility, a positive coefficient implies that margin increases are related to higher volatility. Variables included in the information set, X for Equation 1 and F in Equation 2, require additional explanation. Lags of futures contract returns are included in Equation 1 to capture short term shifts in expected returns. Inclusion of indicator variables for the months of the year incorporates effects on returns from seasonal or contract life-cycle effects. Finally, since returns at time t are dependent on risk assess ments, after the first iteration twelve lags of the vola tility estimate from Equation 2 are included as a measure of risk. The F variables in Equation 2 in clude the indicator variables for months of the year and twelve lags of volatility from Equation 1. The motivations for these inclusions differ from those in Equation 1. Including the months of the year is moti vated by Samuelson’s (1965) theory which implies that the volatility of futures prices changes over the life of the contract. Lags of volatility are included to accommodate the persistence of volatility shocks. French, Schwert, and Stambaugh (1987), Poterba and Summers (1986), and Jain and Joh (1988) pro vide evidence for this persistence in asset returns. Finally, it is necessary to iterate the procedure. Iteration is necessary because the hypothesized het eroskedasticity in Equation 1 implies the £( are ineffi cient. The problem can be corrected by using predict ed values from Equation 2 as weights in a weighted least squares re-estimation of Equation 1. Each iteration improves the efficiency of the £ estimates. Davidian and Carroll (1987), using Monte Carlo experiments, find that two iterations are sufficient to resolve efficiency problems. I found that the earlier iterations often produce some negative predictions. To ensure positive weights are used, I iterate five times to avoid this problem. ‘Efficient in the sense that the information set is being used to the fullest extent possible. 11 for less than one-half TABLE 1 month, they would be Summary of tests for the Excess Volatility Argument relatively useless policy _____________ Coefficient t statistics tools. For similar rea sons, I looked at margin Deutschemark contract S&P 500 contract Trading days changes during the 12 after a margin Hedge Speculative Speculative Hedge change positions positions positions positions days following observed volatility in order to test 1 -1.78 -1.08 -2.59 -1.53 the Prudential Exchange 2 -0.41 0.68 -1.71 0.59 Hypothesis. It is reason 0.49 3 0.82 0.98 1.85 able to reject the Pruden 4 2.02 0.60 0.51 -1.20 tial Exchange Hypothesis -0.17 5 -0.80 -0.78 0.59 if exchange responses to 6 -0.05 1.05 -0.63 0.11 increased volatility occur 7 -0.40 2.37 0.57 -0.08 more than twelve busi 8 1.14 1.33 1.13 -0.16 ness days after a substan 0.42 9 0.85 0.13 -0.12 tial increase in volatility. -0.44 10 2.44 0.16 -0.91 The data consist of 11 -1.75 -2.74 -0.49 0.46 daily prices for two finan 12 -0.33 0.32 -1.22 -0.36 cial futures contracts traded at the Chicago Coefficient sums 0.0001 0.0003 0.0001 -0.0008 Mercantile Exchange: the 1.62 F statistic 0.16 0.06 1.12 deutschemark and the (hypothesis that S&P 500 futures contract. coefficient sum equals zero) Sample periods are from June 30, 1974, to Decem (p value) (0.69) (0.80) (0.20) (0.29) ber 31, 1989, for the deutschemark contract and from June 30, 1982, to December 31, 1989, for the S&P 500 con seventeen changes in initial margin for the tract. This sampling interval gives 3,811 obser deutschemark and nineteen changes of initial vations for the deutschemark contract and 1,842 margin for the S&P 500. These margin chang observations for the S&P 500 contract. On any es are expressed as continuously compounded sample date, futures contracts for several deliv rates of margin change. This approach produc ery months trade simultaneously. This implies es zeroes where no margin change has occurred that the prices of any of these contracts might and small positive or negative values else be used to compute returns. Following industry where.23 These data are from the CME clearing norm, I use prices for contracts which are near association. est to delivery. The nearest-to-delivery con Table 1 reports coefficients of margin tract is generally the heaviest traded and, hence, changes before observed volatility used to test regarded as most representative of that day’s the Excess Volatility Argument. Recall that the trading.21 As contracts approach expiration, this Excess Volatility Argument predicts that there procedure requires that expiring contracts be should be a negative association between vola replaced by the subsequent contract. Thus, on tility and previous changes in margin. Individ the last day of the month prior to a delivery ual coefficient t statistics for speculative and month, I roll out of the nearby contract and into hedge positions in both contracts do not support the next delivery month. This procedure avoids the Excess Volatility Argument. For the deut making inferences which are unique to the schemark, one speculative margin change coef delivery month. ficient (lag 4) differs reliably from zero at the Continuously compounded rates of returns conventional 5 percent level, but has the wrong from these price series are matched to the effec sign. Two individual coefficients for the S&P tive dates of changes in initial margin require are significant for both speculative positions ments for speculative and hedge positions.22 (lags 1 and 7) and hedge positions (lags 10 and Over the respective sample periods, there were 11), but these are of opposite sign. Coefficient 12 ECONOMIC PERSPECTIVES sums are examined because the effect of a margin change may be spread across several days, producing a cumulative effect not evident on any one day. Three of the four coefficient sums are positive, indicating that volatility rises following a margin increase. To determine the significance of these coefficient sums, they are tested against 0 with an F test. Asymptotic critical values for this test are: 3.84, at the 5 percent confidence level and 6.63, at the 1 percent confidence level. In each case, the coefficient sums do not differ reliably from 0. Thus, the results do not indicate a negative association between margin changes and vola tility realized after these changes, as implied by the Excess Volatility Argument. An alternative to associating margin changes with the size of price changes is to examine the frequency distribution of price changes. Figure 2 charts the frequency of S&P 500 futures price changes for each level of margin over the sample period. Price changes are categorized as more than 1 percent, more than 2 percent, etc. Thus, horizontal bars in the chart depict the percentage of price changes larger than a given size which were observed for the indicated level of margin. If high vola FEDERAL RESERVE BANK OF CHICAGO tility is more likely when margin levels are low, then a greater percentage of large price changes should be observed in the low margin regions. Examining each price change row, it appears that large price changes are equally likely to occur at each level of margin observed. Thus, the evidence of this test does not show that low margin levels lead to high volatility.24 Table 2 summarizes coefficients on margin changes after observed volatility used to test the Prudential Exchange Hypothesis. Recall that the Prudential Exchange Hypothesis predicts a positive association between volatility and subsequent margin changes. Signs of individu al coefficients are mixed and their magnitudes are generally insignificant. No important dif ferences appear to exist between speculative positions and hedge positions, indicating that exchange responses to volatility do not differ between these two classifications. Coefficient sums for the deutschemark contract are posi tive. This is indicative of a positive association between past volatility and margin changes as predicted by the Prudential Exchange Hypothe sis. However, F test results indicate these coef ficient sums do not reliably differ from 0. Co efficient sums for the S&P 500 contract are 13 The negative signs for the S&P are opposite Summary of tests for the Prudential Exchange Hypothesis those expected. This Coefficient t statistics motivates further exami nation of volatility and Deutschemark contract S&P 500 contract Trading days S&P margin levels. prior to a Hedge Speculative Hedge Speculative margin change positions positions positions positions Volatilities obtained from the above iterative 1 0.62 1.22 0.35 -2.35 procedure are restated to -1.37 -2.44 2 0.76 0.20 obtain the dollar volatil-1.41 -0.66 2.05 0.38 3 ity per day of the S&P 4 -0.37 1.10 2.38 0.63 contract. These volatili-0.47 0.77 0.03 -2.10 5 ties and the level of 0.44 -0.26 -1.40 -0.33 6 speculative margin are 0.37 -0.24 0.01 0.58 7 graphed in Figure 3. 0.44 1.34 2.43 0.83 8 The graph shows that 1.57 -0.06 0.39 -0.96 9 the level of required -1.32 -2.67 -0.07 -0.33 10 margin has remained 1.24 0.22 -1.21 11 1.10 high while volatility for -0.44 -0.47 -0.67 -0.16 12 most of the period after 1987 fell to 1986 levels. 0.0004 -0.0001 -0.0009 0.0003 Coefficient sums Dividing margin re 1.24 0.89 0.99 1.65 F statistic quirements by dollar (hypothesis that volatility gives the level coefficient sum of coverage obtained by equals zero) the exchange. Compar (0.34) (0.27) (0.32) (0.20) (p value) ing the pre-1987 period with the post-1987 peri od, margin levels since negative, but F test results indicate they do not October 1987 provide the exchange with 51 significantly differ from 0. F test results fail to percent greater coverage than previously. This support the Prudential Exchange Hypothesis. greater coverage lessens the need of the ex- 14 TABLE 2 ECONOMIC PERSPECTIVES change to raise margin in response to volatility increases. In other words, the exchange does not need to raise margins in response to higher volatility because margin requirements are already high enough to cover its increased risk. This may explain the lack of evidence for the Prudential Hypothesis. Another way to test the Prudential Ex change Hypothesis is as follows. Prudential exchanges can be expected to set margin levels for contracts according to the risk of losses from insolvency. Since high price volatility places the exchange at greater risk, levels of margin required for contracts should rise with the anticipated price volatility of these con tracts. One way to observe anticipated volatili ty is to use the volatility implied by observed prices on futures options. Thus, I hypothesize that margin levels will be positively associated with implied volatilities. To demonstrate this approach, implied volatilities were computed for closing prices on futures options traded on September 9, 1991. The contracts used were: soybean, com, and Treasury bonds from the Chicago Board of Trade; and S&P 500, live cattle, Swiss franc, deutschemark, and Japanese yen from the Chi cago Mercantile Exchange. Volatilities are stated on a per day, dollar basis.25 This gives, in dollars, the largest up-or-down change which can be expected in a single day with probability .33. Thus, setting margin levels at three times this volatility provides these exchanges with 99 percent confidence that margin balances will be sufficient to cover losses realized in one day by either long or short positions. Figure 4 graphs margin required for these contracts on our vola tility estimates. The predicted positive associa tion is demonstrated by the graph. The simple correlation between margin levels and volatility is .92 which does provide some evidence for the claim that margin levels are positively asso ciated with the level of exchange risk. The evidence from a single sample date presented in this article is not sufficient for a test of the Prudential Exchange Hypothesis, however, the positive result suggests that further testing may provide stronger evidence. Summarizing the evidence, my tests for the link between futures margin and volatility do not support the Prudential Exchange Hypothe sis. However, this result may be due to the relatively higher margin requirements after FEDERAL RESERVE BANK OF CHICAGO 1987. My tests do produce further evidence against the Excess Volatility Argument. C onclusions The Excess Volatility Argument implies that higher margin can be used to control spec ulation resulting from excessive volatility. This article presents several arguments suggesting that this argument is flawed, as well as new evidence indicating that the volatility of futures prices is not reduced by raising futures margin. The evidence that changes in futures mar gin do not lead to changes in volatility is quite compelling, consequently, the Excess Volatility Argument should not be a consideration in the government regulation of margins. It is clear that private interests in setting margins do exist. I have described the prudential interests of the futures exchanges. These interests provide some support for the view that exchanges are motivated to set margins at prudential levels. Effective public oversight of margin set ting for futures contracts requires policymakers to identify the interests which are best served by changing margins. Otherwise, financial markets risk being encumbered by unnecessary regulation. Margin regulation is unlikely to reduce the volatility of futures prices. Howev er, other roles for margin, including the public’s interest in the safety of futures clearing houses and the payments system, warrant additional research. 15 FOOTNOTES ‘The Brady Report is the name generally given to a report prepared by the January 1988 Presidential Task Force on Market Mechanisms headed by Nicholas Brady, Secretary of the Treasurer. 2Since 1934, the Federal Reserve Board of Governors has set margins for stock by specifying the initial margin required for stock purchases. Margin regulation is motivat ed by the Excess Volatility Argument which is explained later. At this point, it is important for the reader to realize that, in addition to this regulatory activity, private interests are also at work in determining margin. 3T o avoid a technical problem, I oversimplify by assuming the cost of carry for the cash asset is zero. Costs of carry are the financing costs net of returns from holding the cash asset. They determine the difference between futures prices and current prices for the cash asset. For the purposes of this example, they can be ignored. 4Further, resources would be expended to make this deter mination. Thus, the ability of counterparties to avoid this cost will weigh in their assessment of the worth of futures contracting. correction. Note that this is an assumption that exit can be perfectly timed. Relaxing the perfect-timing assumption introduces the risk of being late and incurring losses during the correction. Risk averse investors will take on this risk only if it is compensated. Since the risk is costlessly avoid ed by not participating in the bubble, it is not compensated. Thus, if investors are risk averse, bubbles are not possible. 13The presence of performance guarantees offered by the exchange makes these costs more explicit. The member ship is contractually obligated to make good on defaults of its nonperforming members. 14The effectiveness of regulating margin becomes depen dent on the relative costs of leverage obtained through margin loans and leverage obtained from other sources; that is, homemade leverage. If homemade leverage is relatively costly, then raising margin requirements increas es the cost of obtaining leverage and may decrease specula tive activity. 15A clear case of extending the Excess Volatility Argument to futures markets can be found in the Brady Report. 16Salinger (1989,Table 1) also makes this point. 5This description is somewhat oversimplified. Edwards (1982) goes into more detail. Essentially, the clearing association guarantees payments between the clearing members of the exchange. Were the hypothetical contract made through a single clearing member, Ms. Long would face the risk that the clearing member would be unable to make good on the payment should Mr. Short be insolvent. The clearing association is not obligated to fulfill commit ments between a clearing member and any other party. 6Fenn and Kupiec (1991) point out that increasing the frequency of marking contracts to market serves as a substitute for raising the level of margin. 7This loss may be further reduced by proceeds from the sale of assets going to the exchange. 8The Chicago Mercantile Exchange presently determines margin requirements of positions using its Standard Portfo lio Analysis of Risk (referred to as “SPAN”). The system evaluates the risk of the individual after netting out posi tions in several markets and determines the level of margin required for the net position. 9See Kindleberger (1989) and Chance (1990). I0Federal Reserve Regulations T, U, X, and G state current margin requirements. " Kindleberger’s (1989) history provides an excellent description of the events preceding and following the 1929 crash from an Excess Volatility perspective. Similar arguments have also been made regarding the role of margins on stock index futures in the 1987 crash. For an example, see the Brady Report. ^Alternatively, it might be argued that these investors all believe they can exit the market prior to the necessary 16 17Chance (1990) and France (1990) review the literature of the relationship between volatility and stock and futures margin. 18Hardouvelis (1988) is a notable exception. Hsieh and Miller (1990) point out that the Hardouvelis procedure is susceptible to problems with persistent variance. Kupiec (1988, Table 5) replicates the Hardouvelis procedure. He finds that much of the effect traces to the last half of the 1930s. 19That is, it is assumed that changes in volatility due to shocks are permanent, not temporary. For example, if the volatility of futures prices increases due to an oil crisis, the assumption is that volatility will remain at the new level until another shock occurs. This assumption is important for determining the cause of observed changes in volatility. For example, if volatility responses to shocks were tempo rary rather than permanent, then an observed change in volatility might be the result of volatility returning to its previous level after a temporary response, rather than a response to a new shock. This assumption is supported by the evidence from Schwert (1989). Additionally, the results from the specifications used in this paper support volatility persistence. 20For transitory effects from margin changes to be useful, regulators must be willing to change margin requirements frequently. 21France and Monroe (1991) investigate the effects of futures margin on the less heavily traded contracts expiring on later delivery months. This approach investigates the importance of liquidity on the margin-volatility association. ^Continuously compounded rates of change are computed as the difference in the log of prices. I am indebted to ECONOMIC PERSPECTIVES Bjorn Flesaker who suggested this approach to obtain symmetry between rates of increase and decrease. 24Using margin as a percentage of contract value in place of margin levels does not change this conclusion. 23Signed dummy variables were also tried in place of percentage changes of margin. The results were similar to those reported here, however, the level of significance was lower. This suggests that the amount of margin change provides information in addition to the information that margin changed and the direction of that change. ^Volatilities were implied using the Black-Scholes option model for options on futures nearest to expiration and at the money. This procedure obtains an annualized volatility for rates of change. Annualized volatilities were restated to dollars per day by dividing them by the square root of 365 and multiplying by the dollar value of the contract. REFERENCES Chance, Don M., “The effects of margin on volatility of stocks and derivative markets: a review of the evidence,” manuscript, Virginia Polytechnic and State University, 1990. Davidian, Marie, and Raymond J. Carroll, “Variance function estimation,” Journal of the American Statistical Association 82, 1987, pp. 1079-1091. Edwards, Franklin, “The clearing association in futures markets: guarantor and regulator,” Conference paper presented at the Industrial Organization of Futures Markets: Structure and Conduct, Columbia University, New York City, November 4-5, 1982. Fenn, George, and Paul Kupiec, “Prudential margin policy in a futures-style settlement system,” Finance and Economics Discussion Series Paper No. 164, Federal Reserve Board, Washington, D.C., 1991. Furbush, Dean, “The regulation of margin levels in stock index futures markets,” working paper, Office of Economic Analysis, United States Securities and Exchange Commission, 1988. Hardouvelis, Gikas, “Margin requirements and stock market volatility,” Federal Reserve Bank of New York, Quarterly Review, 1988, pp. 8089. Hseih, David, and Merton Miller, “Margin regulation and stock market volatility,” Journal of Finance 45, 1990, pp. 3-29. Jain, P., and G. Joh, “The dependence be tween hourly prices and trading volume,” Jour nal of Financial and Quantitative Analysis 23, 1988, pp. 269- 283. Kindleberger, Charles P., Mania, panics, and crashes: a history of financial crises, Basic Books, Inc., 1989. Fishe, Raymond P. H., and Lawrence G. Goldberg, “The effects of margins on trading in futures markets,” Journal of Futures Markets 6, 1986, pp. 261-271. France, Virginia Grace, “The regulation of margin requirements: a survey,” University of Illinois at Urbana-Champaign Working Paper No. 90-1670, 1990. France, Virginia Grace, and Margaret A. Monroe, “The effects of margin requirements on futures trading,” mimeograph, University of Illinois at Champaign-Urbana, 1991. French, K.R., G. W. Schwert, and R. F. Stambaugh, “Expected stock return and vola tility,” Journal of Financial Economics 19, 1987, pp. 3-29. FEDERAL RESERVE BANK OF CHICAGO Kupiec, Paul, “Initial margin requirements and stock return volatility: another look,” Journal of Financial Services Research 3, 1988, pp. 287-301. ___________ , “Futures margins and stock price volatility: is there any link?,” Finance and Economics Discussion Series, Board of Governors of the Federal Reserve System, 1990. Moser, James T., “Evidence on the impact of futures margin specifications on the perfor mance of futures and cash markets,” Federal Reserve Bank of Chicago Working Paper No. WP-90-20, 1990. ___________ , “The implications of futures margin changes for futures contracts: an inves 17 tigation of their impacts on price volatility, mar ket participation and cash-futures covariances,” Review of Futures Markets, forthcoming. regulation of stock index futures,” Journal of Financial Services Research 3, 1989, pp. 121-138. Poterba, J. M., and L. H. Summers, “The persistence of volatility and stock market fluctua tions,” American Economic Review 76, 1986, pp. 1142-1151. Samuelson, Paul A., “Proof that properly an ticipated prices fluctuate randomly,” Industrial Management Review 6, 1965, pp. 41-49. Presidential Task Force on Market Mecha nisms, “Report of the presidential task force on market mechanisms: January 1988,” excerpt reprinted in Robert J. Barro, Robert W. Kamphius, Jr., Roger C. Kormendi, and J. W. Henry Watson, eds., Black Monday and the Future of Financial Markets, Homewood, IL, Irwin, 1989, pp. 127-203. Salinger, Michael A., “Stock market margin requirements and volatility: implications for 18 Schwert, G. William, “Business cycles, finan cial crises and stock volatility,” CarnegieRochester Conference Series on Public Policy 31, 1989, pp. 83-126. Schwert, G. William, and Paul Seguin, “Heteroskedasticity in stock returns,” Journal of Finance 45, 1990, pp. 1129-1150. Telser, Lester G., “Margins and futures con tracts,” Journal of Futures Markets 2, 1981, pp. 225-253. ECONOMIC PERSPECTIVES State and local governm ents' reaction to recession Richard H. M atto o n and W illiam A . Testa Despite the recent economic malaise, state and local gov ernments in the Seventh Dis trict largely avoided drastic tax hikes and spending cuts. Instead, governments have drawn down their reserves, trimmed spending, and deferred pay ment of bills in the hopes that robust economic recovery will make tax hikes unnecessary. The same strategy was attempted during the re gion’s economic troubles of the early 1980s, and it is a common strategy for state and local governments during a contractionary period. Nevertheless, this strategy failed District gov ernments in the early 1980s when major tax rate hikes ultimately became necessary and were subsequently implemented after the U.S. economy hit bottom during the fourth quarter of 1982. This time around, owing to the extended economic weakness, the history of the early 1980s may repeat itself in the form of severe belt tightening and significant tax hikes during fiscal years 1992 and 1993. In this article we examine the behavior of the state and local sector during the business cycle, paying particular attention to those dis cretionary actions such as tax hikes and spend ing cuts that are typically taken by state and local government to maintain fiscal balance in response to business contractions. In particular, we focus on the discretionary fiscal actions of the five Seventh District states (Illinois, Indi ana, Iowa, Michigan, and Wisconsin). No two business cycle episodes are identical, especially for the state and local sector which must cope with sharp changes in the direction of federal FEDERAL RESERVE BANK OF CHICAGO grant-in-aid programs. This time around, state and local fiscal pressures are arising from spend ing pressures as much as from lagging revenues. In response, solutions to budgetary stress are likely to focus on spending cuts as well. The secto r's response d u rin g th e business cycle Business cycle contractions are usually ac companied by escalating state and local govern ment fiscal stress and budget crises. Much of the budget stress is taken on willingly by state and local governments as they try to maintain spending commitments without heaping new taxes onto overburdened workers and faltering businesses. In this way, the tax and spending behavior of the state and local sector helps to cushion business cycle contractions; govern ments build up reserves during business cycle expansions and draw down these reserves or borrow during contractions. Often, taxes are ultimately raised and spend ing cut during the later stages of a business cycle contraction or during the recovery period. State and local governments often cannot or will not build up sufficient reserves to see them all the way through business downturns. Further more, their ability to take on debt to fund opera tions is limited so that stop gap fiscal measures become exhausted as contractions wear on. In the aggregate and on net, state and local behavior has been countercyclical during every Richard H. M attoon is a regional econom ist and W illiam A. Testa is a senior regional econom ist and research officer at the Federal Reserve Bank of Chicago. The authors thank David R. Allardice for his comments. 19 business cycle contraction since TABLE 1 World War II. In examining state State and local responses to business cycles and local expenditures from peak (Annualized percent change) to trough over each contraction, Contractions expenditures rise relative to receipts peak to trough Expenditures Receipts Grants (see Table 1). This comes about as the rate of revenue growth de 1948:4 to 1949:4 15.2 8.7 4.5 clines more than the rate of expen 1953:3 to 1954:2 10.7 4.8 0.0 diture growth. 1957:3 to 1958:2 11.9 8.9 44.4 Revenues have tended to slow 1960:2 to 1961:1 9.9 7.8 12.1 or decline immediately following 1969:4 to 1970:4 14.6 11.1 16.8 the peak in the cycle. State and 1973:4 to 1975:1 14.7 9.6 16.9 1980:1 to 1980:3 8.3 8.2 7.4 local government revenue sources 1981:3 to 1982:4 6.4 5.7 -1 .6 are highly sensitive to economic 1990:3 to 1991:1" 7.4 4.7 21.4 aggregates such as spending, prof its, and income.1 Receipts from Expansions such tax sources as personal and trough to peak corporate income quickly turn 1949:4 to 1953:3 7.9 10.3 5.8 sluggish following the peak of 1954:2 to 1957:3 10.9 10.8 15.4 business conditions. In the case of 1958:2 to 1960:2 6.6 10.0 8.9 corporate income, the profit decline 1961:1 to 1969:4 14.9 16.1 21.3 which typically accompanies a 1970:4 to 1973:4 14.4 11.6 19.6 downturn in the business cycle 1975:1 to 1980:1 11.3 13.5 14.7 1980:3 to 1981:3 6.9 8.9 -3 .0 translates into a precipitous decline 1982:4 to 1990:3 10.5 9.7 6.9 in corporate tax revenues. In the case of personal income, recession"Trough not yet determined for this period. related declines in employment and SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts (NIPA), Table diminished payrolls translate into 9.4, "State and local government receipts and expenditures." slower personal income growth and sluggish state income tax revenues. Because state and local govern ments try to maintain expenditures in the face obvious reason for this is that the underlying of declining receipts during contractions, their tax bases accelerate along with the economic liquid reserves are frequently exhausted or recovery. This is reflected by the historic close to exhaustion toward the trough of a busi growth in the national economy during the first ness downturn. For this reason, state and local full year of recovery (see Table 2). GNP governments quickly rebuild budget balances in growth in the first four quarters following the the quarters following the recession, as is re trough of a recession has been very robust, flected by the inverse relation between receipts particularly following the 1975 and 1982 reces and expenditures (see Table 1 and Figure 1). sions. But a second reason is that if tax rate Annual expenditure growth generally slows during the expansionary period following a TABLE 2 recession. Expenditure cuts and spending con Real GNP following business cycle trough trols put in place to relieve state and local fiscal (Annualized percent change) stress during the recession tend to take hold at the tail end of the contraction or during the 02 Trough Q1 03 Q4 early recovery, thereby reducing the rate of M arch 1975 4.0 6.8 5.6 7.6 expenditure growth. Also, demand for social programs such as Medicaid and General Assis J u ly 1980 5.2 7.6 -1 .2 1.6 tance tend to abate with the recovery. N o vem b er 1982 3.6 9.2 6.0 7.2 But a far greater contribution toward re SOURCE: U.S. Department of Commerce, Bu building reserves is exerted from the revenue reau of Economic Analysis, N a tio n a l In c o m e a n d side as receipts grow much faster during the P ro d u c t A c c o u n ts o f th e U .S ., 1957-88. expansion than during the contraction. One 20 ECONOMIC PERSPECTIVES FIGURE 1 State and local expenditures and receipts quarterly percent change quarterly percent change quarterly percent change quarterly percent change NOTE: Two-quarter moving average, nominal dollars. SOURCE: Bureau ot Economic Analysis, Department of Commerce, NIPA. hikes have been adopted at the tail end of a recession, the rate hikes often take hold after the recession. The combination of robust eco nomic growth in the tax base and higher tax rates often lifts state tax receipts dramatically. Explaining the sector's behavior The federal government has a legislated policy to ease the impact of cyclical swings in the economy,2 while the state and local sector has no such legislative requirement. Neverthe less, state and local governments are responsi ble for public health and related functions which are heavily demanded during business cycle contractions.3 As the economy sours and unemployment rises, demands for Medicaid, General Assis tance, and other state aid programs increase. FEDERAL RESERVE BANK OF CHICAGO State and local transfer payments grew at a nearly 14 percent annual rate during the 197375 recession and again during the 1980 reces sion. Rising Medicaid expenses, which can be attributed to both recessionary demands and rising program costs, have proven particularly unyielding and will continue to exert pressure as they eat up larger and larger shares of state spending. The relentless climb in Medicaid costs is demonstrated by state Medicaid spend ing per $100 of personal income, which more than doubled from 1976 to 1990 (see Figure 2). The jump in Medicaid spending has continued with FY92 state general fund spending estimat ed to increase by nearly 22 percent.4 There are also a host of institutional rea sons which help to explain state and local coun tercyclical behavior. States are often unable to 21 FIGURE 2 State spending per $100 of personal income—Medicaid dollars adopted. For example, a reported estimate of $25 billion5 in combined spending reductions and tax increases by state governments did not take effect until FY92, four quarters after the peak in the business cycle. H o w s ta te and local g o ve rn m en ts spend beyond th e ir m eans •Estimate. SOURCE: U.S. Health Care Financing Administration. retrench at the outset of a business cycle con traction because the length and nature of the budget cycle permits little adjustment to unan ticipated changes in economic conditions. Thus, when recessions are short in duration or unexpected, state and local governments have difficulty adjusting in a timely fashion. One reason for the readjustment problem is that state budgets are based on economic forecasts which are generated many months before the fiscal year begins. But while the economic condi tions underlying the budget may change, com ponents of the budget such as wages and pro gram expenses may be locked in by contracts, agreements, and program plans. One example of the impact of the state budget cycle on fiscal adjustment occurred with the onset of the recent contraction. Having begun in July of 1990, the contraction’s onset coincided with the first month of the new fiscal year for 46 states. Few of the state budgets, which had been developed and submitted in the late winter or early spring of 1990, had antici pated a downturn, so that state spending plans were adopted without any significant revision. Once the downturn began, the budget cycle made it difficult to make more than marginal adjustments to the spending plan which was already underway. Accordingly, unless the economic decline is unusually steep in a given state, major fiscal adjustment is often put off until the next fiscal year, when programs can be evaluated and budget cuts and tax hikes can be 22 Governments have a variety of tools at their disposal which allow them to spend more than they receive in revenues for a limited period of time. These range from explicit mea sures, such as drawing down fund reserves, to less visible actions such as various types of fiscal accounting maneuvers. Such measures allow governments to buy time before having to make fundamental adjustments to their spend ing and revenue systems. Short of tax rate hikes or spending cuts, the most straightforward method governments have to bridge deficits is to draw down available fund balances. This can take two forms. First, state and some local governments can use accu mulated general fund reserves to help close budget gaps. Some states are required to run a surplus in their general fund on an annual basis. Wisconsin, for example, is required to end each fiscal year with at least a 1 percent general fund balance. This surplus can provide a cushion if an unexpected downturn arises. Other states try to maintain informal cash balance targets. Illi nois, for example, tries to maintain a general fund cash balance of $200 million as a reserve to pay for budget gaps. Nevertheless, few states seem able or willing to maintain the suggested reserve level of 5 percent of general fund expenditures recommended by the bond rating agencies and investment banks. During the generally strong fiscal years of the late 1980s, year end general fund balances as a percentage of state general funds averaged 1.7 percent in 1987, 2.0 percent in 1988, and 1.0 percent in 1989.6 In addition to attempts to build a surplus directly from the general fund during periods of robust economic growth, 37 states have moved since the late 1970s to adopt so-called “rainy day” funds. These funds are often patterned after Michigan’s “Counter-cyclical Budget and Economic Stabilization Fund.” As originally designed, the fund was to permit deposits and withdrawals based on growth in Michigan’s adjusted personal income. Money would be paid into the fund when the annual rate of per ECONOMIC PERSPECTIVES sonal income growth exceeded 2 percent. For funds to be withdrawn from the account, an economic contraction would have to be severe enough to cause personal income growth to fall below zero or the unadjusted unemployment rate to exceed 8 percent.7 Rainy day funds have proven to be a disap pointment to some observers. Budget stabiliza tion funds are seldom sufficient to provide long term fiscal relief.8 While 37 states technically maintained rainy day funds in FY89, nine of the funds contained no reserves. Furthermore, only 5 of the states had built up reserves as large as 5 percent of state expenditures.9 By the end of FY91, virtually all of the funds were exhausted.10 Sometimes states prefer not to run down their fund balances at the outset and turn to socalled accounting maneuvers to relieve immedi ate fiscal pressure. One such maneuver is a fund transfer, which usually entails transferring the liability for a particular expense from the general fund to a dedicated fund such as transportation or infrastructure. When transfers to dedicated funds are unavailable (often due to legal restric tions or the insolvency of the dedicated fund), states often reclassify certain operating expenses as capital expenditures, thereby using bond money to pay for the expense rather than tax or other revenue. Another stop gap measure is to change the actuarial assumptions underlying state pension fund contributions. This permits the state to reduce its level of pension contribution, thereby freeing revenues for other purposes. Two other popular techniques involve deferring expendi tures and accelerating tax payments. By defer ring spending liabilities, states act to roll over expenses incurred in one fiscal year into the next fiscal year. This allows the state to end a fiscal year with a balanced budget even if it has out standing bills. In the case of accelerating tax payments, the schedule for taxpayer payment or user fees is moved up. Annual payments become quarterly, quarterly become monthly, and so on. This technique improves cash flow and can add a one time extra payment during the fiscal year in which the change is made. With some limitations, states can also re lieve fiscal pressure by issuing short term debt. This can improve a state’s immediate cash flow while a state is waiting for revenues. States can also take actions to increase non-tax revenues such as fees, permits, and user charges which can often be increased less visibly because they im FEDERAL RESERVE BANK OF CHICAGO pact only particular constituencies. Finally, states sometimes sell specific assets in order to raise cash. Often these asset sales consist of selling a state asset to a quasi-government agency which then leases the facility back to the government. Why the sector sometimes falters Despite the extent of both explicit reserves and implicit reserves which are tapped during business cycle contractions, there often comes a time when states exhaust their reserves. At that time, discretionary behavior switches to re building government surpluses through tax hikes and spending cuts. The particular timing of this transition from maintaining spending levels and tax rates to rebuilding surpluses is dependent not only on the extent of the business cycle contraction, but also on special conditions such as trends in federal aid and the disparity in regional conditions. Evidence that the state and local sector can spend beyond its means for only a limited time can be seen from the aggregate behavior of expenditures and receipts following the trough of the contractions. In all three cases beginning with the 1973-75 recession, receipts have shown rapid growth in the first several quarters following the trough while expenditure growth either flattens or turns down (see Figure 1). A number of conditions explain if, how, and when a state or local government moves from expan sionary to contractionary behavior. Federal aid Federal aid has sometimes acted as a counter balance to falling own-source revenues during business cycle contractions. However, the behavior of federal aid over the last four contractions has been far from consistent (see Table l).11 During the current contraction, this behavior has taken an about face with grants up over 21 percent largely due to a surge in federal Medicaid support. Election cycle Another special factor influencing the state and local response to recession is the election cycle. Some analysts claim that the odds that tax increases will be passed in a given year depend in part on where the year falls in the state election cycle.12 Assuming that elected officials behave as incumbency maximizers and that tax increases are unpopular with voters, tax increases will be approved in those years in 23 which approval will have the least repercussion on incumbency. In terms of the election cycle it means that tax increases are most likely in the year following the gubernatorial election. The next most likely choice is the year following the mid-term legislative elections. The 1 9 9 0 -9 1 recession During the recent recession, state and local discretionary actions with regard to revenues point up yet another set of special conditions which influence the timing and extent of state and local behavior with respect to contractions. Unlike the previous two downturns, state gov ernments made significant discretionary moves to raise tax rates, expand tax bases, and raise user fees in the year preceding the 1990-91 recession. According to estimates, $5 billion in discretionary revenues were raised in both fiscal 1989 and fiscal 1990.13 A skewed and out-of-sync deterioration in regional economies accounts for much of this behavior. The U.S. economy began to slow in 1989, especially (and earlier still) in the New England and MidAtlantic regions. Moves to hike tax revenues were undertaken at the begin ning of calendar year 1990. A distinct North east incidence of discretionary revenue moves can be discerned. Massachusetts, New Jersey, New York, and Vermont all expanded individu al income tax rates or bases, or accelerated withholding. Discretionary sales tax actions were undertaken by New Jersey, Massachusetts, Rhode Island, and New York. Rapidly rising costs of health care in the later 1980s also helped to create the need for discretionary revenue hikes. The cost of pro viding public health care through Medicaid rose inexorably along with the costs of providing health care as a fringe benefit to public employ ees. Finally, the prison population doubled during the 1980s so that expanded prison capac ity could no longer be delayed. The build-up in fiscal pressures prior to the contraction, along with continuing regional problems in the Northeast and expanded fiscal travails in defense oriented states such as Cali fornia, ensured that discretionary revenue hikes were once again undertaken during fiscal 1991. An estimated $5 billion in additional revenue hikes were carried out in fiscal 1991. A widely accepted prognosis for a tepid economic recov ery all but ensures that states will embrace discretionary measures again in 1992. Short of 24 an unexpected robustness occurring during the economic recovery, the recent period will be remembered as one in which the discretionary revenue actions of state governments were carried out prior to, during, and subsequent to the recession. But even more than revenue actions, the extended length of fiscal stress has induced discretionary spending cuts by state and local governments. Perhaps this should not be sur prising given that rising program costs have been an important source of fiscal stress; gov ernments have attempted to short circuit fiscal pressures from the very programs that have been rising the fastest. For example, the State of Michigan has eliminated General Assistance aid to nearly 80,000 state residents and Massa chusetts has trimmed the number of Medicaid benefits it offers. While it is too early to be definitive, there is reason to believe that a fundamental change in direction has taken place once again for the sector. Much as the federal government has moved away from the idea that it should be all things to all people, state and local govern ments may be looking toward an era of shrink age rather than expansion. Payroll employment has levelled off over 1991, while many more state and local governments are planning future cutbacks (see Figure 3). In response to an eco nomic recovery period characterized by weak overall job growth, the citizenry will continue ECONOMIC PERSPECTIVES to look to state and local government to provide services, but their willingness to pay for those services will be closely guarded. As a result, there will be greater pressures on governments to provide existing services with cheaper deliv ery mechanisms or to come up with more inno vative services themselves. Seventh District reactions to the 1990-91 recession With the onset of the recession, budgetary stress in the District ranged from severe in Michigan to mild in Wisconsin. When the current contraction began in the third quarter of 1990, four of the five District states had already begun their 1991 fiscal year. As budgetary stress began to accelerate, potential fiscal ac tion focused on changes in the FY92 budget. After considerable discussion, none of the five District states passed any major tax increases. Taxes such as income and sales were largely untouched. For example, Illinois’ most signifi cant tax increase was an extension of the state’s personal income tax surcharge (raising the permanent rate from 2.5 to 3 percent) which had been in effect since 1989.14 Iowa’s only tax increase was a hike in its state cigarette tax. Indiana, aside from transferring some program expenses to bond funds, enacted no major tax hikes although it did renew a vehicle tax that had been set to expire. Wisconsin passed a biennial budget for FY92 and FY93 which calls for no significant tax increases. Even Michi gan, whose economy has been hard hit by the slump in the auto industry, has adopted a FY92 budget that does not contain major tax increas es. District fiscal actions so far appear to mir ror District state behavior during the 1980 and 1981-82 recessions, when District states put off major tax hikes until the second half of FY83.15 The behavior of the District states during FY91 and the first half of FY92 appears to represent the early stages of adapting to the contraction. Having initially been less impact ed by the recession than other parts of the na tion, most of the states entered FY91 with rea sonable budget reserves and fiscal flexibility. As conditions worsened, most of the states turned to accounting and other fiscal maneuvers to balance their books. Illinois for example chose to roll Medicaid expenses and some other vendor payments into the 1992 fiscal year. Indiana transferred $40 million in prison expen ditures from the General Fund to bond funds. FEDERAL RESERVE BANK OF CHICAGO Michigan has favored employee furloughs to balance expenditures. All of the states adopted hiring freezes and travel restrictions. Critical budgetary pressures are now build ing for the sector in general and are forcing some states to revise their FY92 budgets. Fis cal conditions were deteriorating prior to the falloff in general business conditions so that the reserve position of state governments is weak or nonexistent. District states budgeted for 1992 under the assumption that economic recovery, however modest, would help to lift revenues and maintain expenditures. However, recent indications from statehouses are that 1992 reve nue projections have been too sanguine, so that elected officials are mapping out a change of course.16 How will District governments respond in the future? District states would like to refrain from draconian spending cuts and major revenue hikes during the coming months. The results of regional and, in particular, District government action during the early 1980s suggests that, even under severe fiscal stress, state and local governments can sometimes forestall such budget-balancing actions through one or more fiscal years. Despite both economic stress and sharply falling aid from the federal sources in the 1980s, for example, Seventh District gov ernments refrained from the most dramatic discretionary moves until well after the down turn’s trough. Nevertheless, the recent environ ment suggests that District governments have more than likely breached the thresholds which require more profound fiscal remedies. Many governments in the District are already cutting payrolls and programs so as to preserve a mini mum of fiscal integrity. Today’s budgetary pressures in the District differ from those of the early 1980s. To a greater extent, pressures are arising from the spending side of the ledger as much as from revenue shortfalls. Accordingly, as budget remedies become necessary during 1992 and beyond, spending cuts rather than tax hikes will be favored. Either cuts in spending or increases in state and local revemues are likely to act as a drag on the rate of recovery in the District during 1992. Some of the spending pressures, such as spiralling health care costs and the need for suitable prison space, are partly beyond state 25 government control. Even so, governments will need to identify and reach a concensus on spending programs that can be reduced. Never theless, as spending cuts become deeper and therefore more difficult to agree on, such a strategy may prove inadequate. At that time, District governments will once again consider major tax hikes and other revenue enhance ments in order to balance their budgets. FOOTNOTES ‘Crider (1978), p. 9. Also, Shannon (1985) p. 341. zHumphrey-Hawkens, “Full Employment and Balanced Growth Act of 1978.” The bill defines the role of federal policy as encouraging full employment. 3Few would argue that state and local governments should carry the primary responsibility for economic stability. Insofar as the benefits of local fiscal action spill over local boundaries, such a decentralized system could easily result in an inadequate countercyclical stimulus. Nevertheless, state and local governments in many states and regions are reportedly accelerating capital spending plans as an eco nomic growth measure which is intended to address the current economic sluggishess. See Enos (1992), p. 1. 4National Conference of State Legislatures, State Budget and Tax Actions 1990. sState Policy Research, Inc., 1991. 6National Association of State Budget Officers/National Governors Association, Fiscal Survey of the States, 1988, p. 15. ’Advisory Commission on Intergovernmental Relations, Significant Features of Fiscal Federalism, Vol 1, 1991. 8For further discussion of rainy day fund behavior see Testa and Mattoon (1992). ’National Association of State Budget Officers/National Governors Association, Fiscal Survey of the States, Sep tember 1989. 10State Policy Research Inc., 1990. "See Gold (1991). ,2See Mikesell. "National Association of State Budget Officers/National Governors Association, Fiscal Survey of the States, various years. "Illinois made permanent the 1989 income tax surcharge which had been scheduled to expire. The 20 percent surcharge may be reduced to 10 percent in FY94. Revenue raised through the surcharge will be distributed to educa tion, municipalities, and the state. The surcharge on the state’s corporation business tax was also extended. lsFor more on Seventh District behavior in the 1980 and 1981-82 recessions see Testa and Mattoon (1992). 16For example, the Illinois legislature recently approved the Governor’s plan to close an impending gap in the 1992 budget. Under the plan, the state will cut $273 million in services, and it borrowed $500 million to pay a backlog of unpaid bills. Aside from transfers, no revenue features were included in the plan. REFERENCES Advisory Commission on Intergovernmental Relations, Significant Features of Fiscal Fed eralism, Vol. 1, 1991. Mikesell, John, “Election periods and state policy cycles,” Working Papers in Public Fi nance and Economics, Indiana University. Crider, Robert A., The Impact of Recession on State and Local Finance, Academy of Con National Association of State Budget Officers/National Governors Association, Fiscal Survey of the States, various years. temporary Problems, Urban and Regional De velopment Series No. 6, Columbus, Ohio, 1978 National Conference of State Legislatures, Enos, Gary, “Governors: bring back the WPA,” City & State, January 27, 1992. State Budget and Tax Actions 1990, Denver, Gold, Steven, “Changes in state government finance in the 1980s,” National Tax Journal, Vol. XLIV, No. 1, 1991, pp.1-19. __________________ , State Bud 26 1990. get and Tax Actions 1991, Denver, 1991. ECONOMIC PERSPECTIVES Shannon, John, “The impacts of the cyclical behavior of the economy on the finances of federal, state and local government,” Federal and State Fiscal Relations, Department of the Treasury and the Advisory Commission on Intergovernmental Relations, Washington, D.C., September 1985. State Policy Research, Inc., State Policy Reports, Volume 8, Issue 8, 1990. _________________ , State Budget and Tax News, Volume 10, Issue 18, 1991. Swonk, Diane C, “The Great Lakes economy: a regional winner in the 1990s,” in William A. Testa, ed., The Great Lakes Economy Looking North and South, Federal Reserve Bank of Chicago, April 1991, pp. 148-157. Testa, William A., and Richard H. Mattoon, “State and local government and the business cycle,” in Issues in State and Local Government Finance, Collected Works From the Federal Reserve Bank of Chicago, William A. Testa, ed., Federal Reserve Bank of Chicago, 1992, pp. 13-24. Related publications Issues in State and Local Government Finance: Collected Works from the Federal Reserve Bank of Chicago This compendium of research in the area of state and local finance is both timely and relevant to the current policy debate. Issues examined range from tax policy and the value of tax incentives in business development to factors influencing the behavior of state and local governments. The Great Lakes Economy Looking North and South This volume examines the region’s economic structure and outlines challenges and opportuni ties for the future. To order please w rite to: The Federal Reserve Bank of Chicago Public Information Center P.O. Box 834 Chicago IL 60690-0834 or telephone 312-322-5111 FEDERAL RESERVE BANK OF CHICAGO 27 Credit Markets in Transition The 28th Annual Conference on Bank Structure and Competition, May 6-8,1992 T h e F e d e r a l R e s e r v e B a n k o f C h i c a g o w i l l h o l d its 2 8 t h A n n u a l C o n f e r e n c e on B a n k S tru c tu re a n d C o m p e titio n at th e W e s tin H o t e l in C h i c a g o , Illin o is , M a y 6 -8 , 1 9 9 2 . A tte n d e d each y e a r by several h u n d re d a c a d e m ic s , re g u la to rs , an d fin a n c ia l in s titu tio n e x e c u tiv e s , th e c o n fe r e n c e s e rv e s as a m a jo r f o r u m fo r th e e x c h a n g e o f id e a s re g a rd in g p u b lic p o lic y t o w a r d th e fin a n c ia l s e rv ic e s in d u s try . The 1 9 9 2 c o n f e r e n c e w ill fo c u s o n th e c h a n g in g s tru c tu re o f c re d it m a r k e t s a n d its i m p l i c a t i o n s f o r t h e b a n k i n g i n d u s t r y a n d t h e e c o n o m y . F e d e ra l R e s e rv e C h a ir m a n A la n G r e e n s p a n w ill s e rv e as th e k e y n o te sp eaker. T o p ic s fe a t u r e d a t th is y e a r 's c o n f e r e n c e w ill in c lu d e : p e r fo r m a n c e o f n e w c re d it m a r k e t in s tr u m e n ts c re d it m a r k e ts d u rin g th e " c re d it c ru n c h " th e c re d it m a r k e t ■ th e ■ th e p e rfo rm a n c e o f ■ t h e f u t u r e r o l e o f b a n k s in ■ t h e s ta t u s a n d f u t u r e o f t h e life i n s u r a n c e in d u s t r y ■ m a n a g in g th e c o n s o lid a tio n o f th e b a n k in g in d u s try . T h e firs t d a y o f th e c o n fe r e n c e w ill b e d e v o t e d to te c h n ic a l p a p e r s o f p r im a r y in te re s t to an a c a d e m ic a u d ie n c e w h ile th e fin a l t w o d a y s a re d e s ig n e d to a p p e a l to a m o re g e n e ra l a u d ie n c e . In v ita tio n s to th e 2 8 th B a n k S t r u c t u r e C o n f e r e n c e w i l l b e m a i l e d in m i d - M a r c h . If y o u a re n o t c u r re n tly on o u r m a ilin g lis t o r h a v e c h a n g e d y o u r a d d re s s a n d w o u ld lik e t o re c e iv e an in v ita tio n to th e c o n fe r e n c e , p le a s e sen d y o u r n a m e a n d a d d re s s to : Public Information Center - 3rd floor. Federal Reserve Bank of Chicago, P.O. Box 834, Chicago, Illinois 60690-0834. ECONOMIC PERSPECTIVES BULK RATE P u b l i c I n f o r m a tio n C e n t e r Federal Reserve Bank of Chicago P.O. Box 834 Chicago, Illinois 60690-0834 U.S. POSTAGE PAID CHICAGO, ILLINOIS PERMIT NO. 1942 Do Not Forward Address Correction Requested Return Postage Guaranteed F E D E R A L R E S E R V E O F C H IC A G O B A N K