The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Business Review Federal Reserve Bank of Philadelphia July-August 1991 ISSN 0007-7011 Business Review The BUSINESS REVIEW is published by the Department of Research six times a year. It is edited by Patricia Egner. Artwork is designed and produced by Dianne Hallowell under the direction of Ronald B. Williams. The views expressed here are not necessarily those of this Reserve Bank or of the Federal Reserve System. SUBSCRIPTIONS. Single-copy subscriptions for individuals are available without charge. Insti tutional subscribers may order up to 5 copies. BACK ISSUES. Back issues are available free of charge, but quantities are limited: educators may order up to 50 copies by submitting requests on institutional letterhead; other orders are limited to 1 copy per request. Microform copies are available for purchase from University Microfilms, 300 N. Zeeb Road, Ann Arbor, MI 48106. REPRO D U CTIO N . Perm ission must be obtained to reprint portions of articles or whole articles. Permission to photocopy is unrestricted. Please send subscription orders, back orders, changes of address, and requests to reprint to Publications, Federal Reserve Bank of Philadelphia, Department of Research, Ten Independence Mall, Philadelphia, PA 19106-1574, or telephone (215) 574-6428. Please direct editorial communications to the same address, or telephone (215) 574-3805. Digitized for 2FRASER JULY/AUGUST 1991 IS ACCESS TO CENTER CITY STILL VALUABLE? Richard Voith If "location, location, location" is really all that important to house values, then what makes one residential location more at tractive than another? One answer is accessibility to jobs, shopping, and recreation—amenities usually found in the downtown of a city. But with subur ban employment and shopping centers proliferating, is access to center city still as valuable as it used to be? LESSONS ON LENDING AND BORROWING IN HARD TIMES Leonard I. Nakamura Problem loans and highly leveraged trans actions have brought home a truth about lending that is easily forgotten in good times: loans sometimes fail, with sad con sequences for both borrower and lender. In truth, seizing the collateral of insolvent debtors often harms the lender as much as the borrower. How can lenders ensure repayment by borrowers and yet avoid being too conservative in hard times? FEDERAL RESERVE BANK OF PHILADELPHIA Is Access to Center City Still Valuable? Richard Voith* 1 he old real estate adage—that the three most important factors in house value are "lo cation, location, and location"—may be an ex aggeration. Nevertheless, prices for similar houses vary greatly within metropolitan areas and even more so across metropolitan areas. What makes one location more attractive than another? Some studies have emphasized amenities and the efficiency of local government as im portant determinants of where people choose ^Richard Voith is a Senior Economist in the Urban and Regional Section of the Philadelphia Fed's Research Depart ment. to live and how much their houses are worth. Another major factor, however, is accessibility to employment, shopping, and recreation. And because people prefer to live in neighborhoods convenient to employment and everyday ac tivities, houses in these areas command higher prices. Although we often hear about "accessible" neighborhoods, accessibility is not an easy thing to measure. Before the rapid growth of the suburbs, a city's central business district (CBD) was the focal point of a region's economic activity. Accordingly, economists' early mod els of residential location tended to define ac cessibility in terms of distance from the CBD. 3 BUSINESS REVIEW But suburbanization changed all that, requir ing us to reconsider what makes one location more accessible than another. Since people can work, shop, and find enter tainment in any number of employment cen ters, a neighborhood's accessibility depends not just on how close it is to those various centers, but on the quality of transportation to and from them. Even if close to an employment center, a residential area may not be perceived as accessible if transportation to that center is poor. Another complicating factor is that dif ferent neighborhoods may be convenient to employment and recreation centers that are not equally attractive. A house within easy com mute to a center with a large number of highwage jobs is likely to be more valuable than a house nearby a center with relatively low wages. And since a locale's accessibility and nearby attractions may be greatly affected by eco nomic development and transportation poli cies, it is important to know how much acces sibility affects people's location choices. EARLY MODELS OF RESIDENTIAL LOCATION Urban economists first addressed how ac cessibility influences residential locations and land values by making some simplifying as sumptions. The most important was that busi nesses concentrated in the CBD because being close to one another increased productivity.1 These productivity increases associated with a CBD location were termed "agglomeration economies."2 The only concentration of em 'For a detailed discussion of the monocentric model, see Edwin S. Mills, "An Aggregative Model of Resource Alloca tion in a Metropolitan Area," American Economic Review 47 (1967) pp. 197-210, or Richard F. Muth, Cities and Housing: The Spatial Pattern o f Urban Residential Land Use (University of Chicago Press, 1969). 2See Gerald Carlino, "Productivity in Cities: Does City Size Matter?" this Business Review (November/December 4 JULY/AUGUST 1991 ployment was in the CBD, giving rise to the term "monocentric region." Other common assumptions were that transportation costs per mile to the CBD were equal from anywhere within the metropolitan area, and that only transportation costs for work trips were impor tant. These assumptions implied that the travel costs associated with any location were deter mined solely by its distance to the CBD. Given these assumptions, economists ana lyzed how people trade off commuting costs with what they are willing to pay for housing. They drew three conclusions: 1) the value of land should fall as distance from the CBD increases; 2) population density should fall as distance from the CBD increases; and 3) people choose residential locations that minimize total commute time in the region. Not surprisingly, the monocentric model predicts higher prices for land close to the CBD, which in turn leads to higher house prices for otherwise identical houses. Consumers can avoid some of the costs of commuting by living closer to the CBD, but in doing so, they bid up the prices of houses such that the higher house price just offsets the commuting savings. A direct consequence is higher land costs for locales closer to the CBD and lower land costs for more distant locations. Accordingly, people living closer to the CBD own smaller houses than residents of more distant, less expensive areas. This leads to the second major conclu sion of the monocentric model— that popula tion density declines with distance from the CBD. Finally, the predicted pattern of declin ing house prices and less density with distance from the CBD results in the optimal amount of 1987). This assumption did not preclude employment scat tered in the suburban areas. Of course, the downside of agglomeration is congestion, and if a locale becomes too congested, productivity may decline. Firms locating out side the CBD forfeit the agglomeration economies, but could realize a benefit by lowering their workers' commuting costs, allowing the suburban firms to offer a lower wage. FEDERAL RESERVE BANK OF PHILADELPHIA Is Access to Center City Still Valuable? commuting. That is, given house prices and commuting costs, no two households could exchange locations and both be better off. How do these predictions correspond to the real world? Although there is evidence that residential density declines with distance from the CBD, there is little consistent evidence that house values fall as well.3 Also, residential distances from the CBD and the associated level of com m uting predicted by the monocentric model are much lower than that actually observed. In other words, people tend to live farther away from the CBD than would be expected given the trade-off between house prices and commuting costs to the CBD. Some authors argue there is a great deal of "wasteful" commuting, suggesting that the underlying notion that people make residential-location decisions based on a trade-off between com muting and housing costs is fundamentally flawed.4 After all, people may want to be near amenities not available in the CBD and thus may be willing to pay more to locate farther from the CBD. Others suggest that the concept is correct, but that the assumptions about the metropolitan areas are wrong.5 3The density predictions are consistent with the monocentric model, since much of the housing stock in older cities was constructed when the model's assumptions were more consistent with the actual metropolitan struc ture. House prices can adjust much faster than the stock of housing, so the failure of the monocentric model should be observed first in its predictions regarding prices. Examples of recent studies finding either a positive relationship or no relationship between distance from the CBD and house value include D.M. Blackley and J.R. Follain, "Tests of Locational Equilibrium in the Standard Urban Model," Land Economics 63 (1987) pp. 46-61; M.L. Cropper and P.L. Gordon, "Wasteful Commuting: A Re-examination," Jour nal o f Urban Economics 29 (1991) pp. 2-13; and E.J. Heikkila et al., "What Happened to the CBD-Distance Gradient? Val ues in a Poly-centric City," Environment and Planning 21 (1989) pp. 221-32. 4Bruce W. Hamilton, "Wasteful Commuting," Journal of Political Economy 90 (1982) pp. 1035-53. Richard Voith IMPROVING THE SIMPLIFIED MODEL Though the monocentric model is a useful starting point for examining residential-loca tion choices, its basic assumptions are less real istic today than when the model was first pro posed. Most metropolitan areas have not just a CBD but many suburban employment cen ters, and these centers can differ from the CBD in several ways. High-quality automobile trans portation to suburban centers is almost univer sally available, while public transportation to these centers is poor at best. Generally depen dent on the automobile for access, these centers are less dense in their development, which lowers their potential for agglomeration econo mies. On the other hand, most CBDs are accessible by public transportation and by au tomobile, though usually at a higher cost than are the suburban sites. Public transportation allows higher-density development in the CBD than in most suburban centers, increasing the potential for agglomeration economies. Though most suburban neighborhoods have high-qual ity auto access to suburban employment cen ters, not all have high-quality public transpor tation to the CBD. More complicated models that consider sub urban employment centers and differences in public transportation services do not lead to simple conclusions about the relationship be tween distance from the CBD and house value and residential density. These models predict that people "sort" themselves into residential communities that are convenient to specific employment locations. Communities conve nient to an employment center—the CBD, for example— should have a disproportionately high percentage of their residents working in that center. Over time, this sorting process should result in people choosing employment 5See, for example, Michelle J. White, "Urban Commut ing Journeys Are Not 'W asteful,'" Journal of Political Economy 96 (1988a) pp. 1097-110. 5 BUSINESS REVIEW and residential locations that minimize the region's total commuting burden. Recent stud ies provide strong evidence for this sorting behavior. In fact, if differences in the quality of the transportation system and the multicentered nature of regions are taken into account, there is little evidence of "wasteful" commuting.6 But what are the implications for house values? Certainly house prices are no longer strictly linked to their distance from the CBD. Since there are many smaller, similar suburban employment centers, all with relatively good highways and parking, a suburban residential location is likely to be convenient to at least one employment center. A house far from the CBD may not be highly valuable to a CBD worker, but it might be highly attractive to a suburban worker. Though one might expect some differ ences in house prices based on distances from suburban centers, these differences are likely to be small and difficult to observe, requiring detailed geographic and transportation data that are seldom available. Still likely, though, is that higher house val ues would be observed for locations having commuting advantages to high-wage employ ment centers not duplicated elsewhere.7 One such advantage is the availability of high-qual ity public transportation to the CBD. If the CBD has higher-wage employment, differences in the availability and quality of public transpor tation across suburban neighborhoods could cause differences in suburban house values. In communities with good public transportation, higher house values should go hand in hand with a greater fraction of the labor force em- ^ e e White (1988a). 7Note that higher wages can be sustained only if the employment center is more productive. Frequently, this higher productivity depends on the employment center's accessibility to a large, high-quality labor force and agglom eration economies associated with concentrations of busi nesses. Digitized for6 FRASER JULY/AUGUST 1991 ployed in the CBD and with lower auto owner ship, as people substitute public transportation for cars. The extent of residential sorting is important to consider when evaluating policy changes that affect accessibility. Policymakers should take into account not only how a policy change would affect the existing population, but also w hat changes in p opu lation the p olicy w ould induce. A policy that dramatically affects the accessibility of a residential area or the produc tivity of a commercial area could have much larger impacts than expected. For example, suppose a public transit au thority alters its prices or service quality. This will immediately change the demand for its services and ultimately affect accessibility as well. Over time, people will decide to relocate, which magnifies the initial impact of the policy. These sorting impacts may be larger than the direct im pacts, even tu ally affectin g a community's size and house values.8 THE PHILADELPHIA EXAMPLE The Philadelphia metropolitan area is an excellent case study for examining the issues raised by urban models. The Philadelphia region, having multiple employment centers, is fairly decentralized, yet it has a large CBD that has grown along with the suburban subcenters in the 1980s.9 The location and commuting 8For example, Richard Voith, in "The Long-Run Elastic ity of Demand for Commuter Rail Transportation," Journal o f Urban Economics (1991), has estimated that the long-term effects on transportation demand of changing price and quality can be more than twice as large as the short-term effects. A highly readable discussion of these issues is provided by Voith in "Commuter Rail Ridership: The Long and the Short Haul," this Business Review (November/December 1987). 9The Philadelphia CBD is defined as the area bounded by the Schuylkill and Delaware rivers and Vine and South streets. FEDERAL RESERVE BANK OF PHILADELPHIA Is Access to Center City Still Valuable? patterns were examined for evidence of sorting and its effects on residential location, car own ership, and house values.10 Though general evidence is provided on the importance of sorting throughout the metropolitan area, the focus is on where CBD workers live, the role of the suburban commuter rail system in their choice of neighborhood, and the system's ef fects on car ownership and house values.11 Geography and Transportation. Accord ing to the 1980 Census, about 55 percent of the 2.2 million people in the Philadelphia metro politan area labor force lived in the suburban Pennsylvania counties of Bucks, Chester, Dela ware, and Montgomery and in Camden County, New Jersey.12 The extent of employment decen tralization in the region is evident. Only 4.6 percent of the suburban labor force works in the CBD, and fully 17 percent of the suburban census tracts have no residents working in the CBD. Still, the CBD has maintained its impor tance in the regional economy. Some suburban census tracts have as many as 22 percent of their labor force working in the CBD. The CBD's share of the region's total employment has been almost constant at 10 percent in the years from 1976 to 1986. However, while suburban em ployment grew tremendously over the period, the rest of the city did not prosper.13 10See Richard Voith, "Transportation, Sorting, and House Values in the Philadelphia Metropolitan Area," Journal of the American Real Estate and Urban Economics Association (forthcoming), for a detailed description of the analysis. 1’The analysis is based on 1980 Census data. These data are still useful because the issues examined reflect long-run location choices. The factors affecting these choices, espe cially the transportation system and the CBD's relative importance to the region, have changed little in the last 10 years. 12The Philadelphia metropolitan area also includes Burlington and Gloucester counties in New Jersey. We did not examine these counties because they do not have com muter rail service. Richard Voith The transportation system in the Philadel phia area has not changed dramatically in the last 20 years, though the highway system has improved progressively. But these improve ments have barely kept pace with the increase in auto travel due to the region's decentraliza tion. The commuter rail system, now operated by the Southeastern Pennsylvania Transportation Authority (SEPTA), has been a fixture on the Pennsylvania side of the Philadelphia region for most of the century. The Port Authority of Pennsylvania and New Jersey (PATCO) has provided commuter rail service in Camden County, New Jersey, since 1968. The primary function of both systems is to bring suburban commuters to downtown Philadelphia. De spite recent interest in "reverse commuting," these systems are generally not competitive with the automobile for commuting to subur ban employment. With 137 stations combined, they provide service to a large number of sub urban communities. Over 42 percent of the suburban census tracts have access to com muter rail transportation (Figure 1), but the quality of commuter rail service differs consid erably across communities. How Long Do Philadelphians Commute to Work? A powerful piece of evidence for sort ing in the Philadelphia region is that average reported commuting times differ very little across residential locations.14 People have the opportunity to work at an employment center 13See Anita Summers and Peter D. Linneman, "Patterns and Processes of Urban Employment Decentralization in the U.S., 1976-1986," Wharton Real Estate Center Working Paper 75, University of Pennsylvania (1990). CBD employ ment grew about 10.5 percent, but overall employment in the city of Philadelphia fell over 6 percent. Suburban employment rose 33 percent. 14The commuting data are based on the 1980 U.S. Cen sus. 7 JULY/AUGUST 1991 BUSINESS REVIEW FIGURE 1 Commuter Rail Transportation in the Delaware Valley Quakertown BUCKS COUNTY Doylestown Pottstown Newtown >ter Lansdali CHESTER COUNTY West Trenton Trenton Norristown BURLINGTON COUNTY DELAWARE Marcus Hook Wilmington Lindenwold GLOUCESTER COUNTY CAMDEN COUNTY Note: Dashed lines indicate services discontinued since 1980. that is relatively close, regardless of how far their house is from the CBD, and they choose to do so. People in tracts far from the CBD tend to commute the same amount of time as those close to the CBD. Average commute time in the region is about 23 minutes; this figure is re markably consistent across counties, ranging from a low of 22 minutes in Chester and Mont gomery counties to a high of 25 minutes in Delaware County (Figure 2). This contrasts with the dramatic differences in highway com mute times to the CBD across counties, which vary from a low of 36 minutes in Camden County to a high of 77 minutes in Chester County. Even though the average highway commute time from Chester County to the CBD Digitized for 8 FRASER is more than twice that of Camden County, residents of Chester and Camden counties spend nearly the same average time commut ing. It appears that people choose to live in locations relatively close to their work places, and that virtually all suburban residential loca tions are convenient to at least some form of employment. Where CBD Workers Live. Since people choose to live close to their jobs or to seek jobs close to their homes, those having jobs in the CBD should be concentrated in areas from which CBD commuting is relatively less costly. For any location, the greater its accessibility to the CBD, the higher the fraction of its residents that should work in the CBD. And more resiFEDERAL RESERVE BANK OF PHILADELPHIA Is Access to Center City Still Valuable? dents will work in the CBD if the community's accessibility to other work sites is poor. Highway commute time to the CBD is one important factor affect ing a neighborhood's convenience to the CBD and hence its attraction for CBD workers. A look at where suburban ites work shows how strongly highway com muting time influences th eir n eighborhood choice (Figure 3). The fraction of people work ing in the CBD declines dramatically with high way commute time. For example, the percentage of M erion resid en ts working in the CBD, with Merion being just a 25-minute drive from the CBD, is 2.9 times as large as the percentage of workers coming from Paoli, which is 61 min utes away by car. Some differences in the percentage of work ers em ployed in the CBD may result from differences in accessibil ity to other work sites rather than in travel time to the CBD. Consider two communities, both with equal commute times to the CBD; if one has higher average com mute times for all com mutes, including those Richard Voith FIGURE 2 Average Commute Times by County and Highway Commute Time to the CBD Minutes Bucks Chester Delaware Montgomery Camden Source: Average commute time is published by the U.S. Commerce Department's Bureau of the Census. Highway commute time is compiled by the Delaware Valley Planning Commission. FIGURE 3 Percent of the Suburban Labor Force Working in the CBD and Highway Commute Time to the CBD Percent Labor Force Working in CBD Minutes to CBD Source: Based on data from the U.S. Census (1980) and the Delaware Valley Planning Commission. 9 BUSINESS REVIEW outside the CBD, then that locale must have relatively worse accessibility to the non-CBD employment centers. The data suggest that for two communities with equal access to the CBD, increasing travel time to non-CBD employ ment centers by five minutes increases the percentage working in the CBD by 46 percent. For Philadelphia-area commuters, the rail system is an important alternative to the auto mobile. A major difference between the rail system and the highway system, however, is that only some communities have access to the rail system, and it essentially serves only one employment destination—the CBD. Not sur prisingly, for suburban communities with train service,15the fraction of the labor force working in the CBD (5.3 percent) is 29 percent higher than for census tracts without service (4.1 per cent). Part of this difference results from the fact that tracts with service tend to be closer to the city. But even with other factors held constant, the fraction of CBD workers living in census tracts with service is 15 percent higher. The availability of the commuter train also results in fewer purchases of automobiles, even for households of equal income. Households in census tracts with train service own 4.5 percent fewer cars, on average. While this figure ap pears small, it actually is quite significant con sidering that only 5.3 percent of the labor force in these tracts commutes to the CBD. Assum ing that train service is irrelevant for 90 percent of the people in a given census tract (and hence should not affect car ownership), the 4.5 per cent reduction overall implies a household carownership rate for the remaining 10 percent of only 0.97 cars per household, about 60 percent of the average car-ownership rate. Housing Prices and the Commuter Rail System. Does sorting into residential locations 15The designations "with service" and "without ser vice" refer not to the communities themselves, but to resi dences having rail stations in or nearby their census tracts. 10 JULY/AUGUST 1991 convenient to work result in higher house prices in neighborhoods with greater accessibility? In particular, are people willing to pay a premium to live in residential neighborhoods that have commuter rail service to the CBD? Median house values in each of the 678 census tracts in the Philadelphia metropolitan area were compared using statistical techniques to adjust for differences in housing quality.16 There was some evidence that houses tend to be more expensive the farther they are from the CBD, contrary to the pred iction of the monocentric urban model.17 But consistent with the idea that most suburban communities are convenient to at least one suburban em ployment center was the finding that average commute times are unrelated to house value. This is not surprising, since sorting has resulted in similar commuting times throughout the region. Even though house prices tend to increase with distance from the CBD and most residen tial locations are equally convenient to some suburban employment, the accessibility to the CBD provided by the commuter rail systems generates significant house value premiums for residents in neighborhoods with service. In fact, if we hold constant other factors, such as highway accessibility and house quality, houses in areas with train service enjoy premiums of 6.4 percent over those without service. This 16A linear regression model was used in which median house value was the dependent variable. House value was a function of its accessibility to the CBD by auto and by train, as well as to other employment centers. Since the theory is developed in terms of a standard unit of housing, character istics of the housing and neighborhood are included in the regression to control for differences in attractiveness that are unrelated to accessibility across tracts. 17House prices may be higher for more distant houses partly because of larger lots. Unfortunately, data on lot sizes are not available. FEDERAL RESERVE BANK OF PHILADELPHIA Is Access to Center City Still Valuable? Richard Voith implies a premium of FIGURE 4 $5,594 for train ser vice.18 House-Value Premiums The house-value for Commuter Rail Service premium associated with train service can $ House Value be used to calculate a net increase in real estate values associ ated with the com muter rail system. A to tal of 258,437 o w n e r-o c c u p ie d houses are in census tracts with train ser vice. This implies that the increase in suburban value asso Entire Region SEPTA (PA) PATCO (NJ) ciated with the train service is about $1.45 billion.19 Source: Based on data from the U.S. Census (1980) and the Delaware Valley Planning Service Quality Commission. and House Values. If people are willing The higher-quality PATCO service has a to pay a premium to live in an area with train service, they would likely be willing to pay much larger positive effect on house values even more for higher-quality train service. Once than the SEPTA service (Figure 4). The pre again, Philadelphia provides a natural test for mium of $6,706 in New Jersey is 10.1 percent of this hypothesis. The two commuter rail sys the average house price in Camden County. tem s servin g the CBD— SEPTA and The $3,437 premium for the Pennsylvania coun PATCO—are very different. PATCO service ties, where the average price of a home is higher is, on average, five times as frequent as than in New Jersey, is 3.8 percent of the average SEPTA's. Furthermore, PATCO enjoys a greater house price. Because these two systems serve time advantage, relative to the automobile, the same destination, the difference in premi than SEPTA. Thus, PATCO generally provides ums very likely reflects consumers' willingness to pay for higher-quality transportation.20 higher-quality service. 1^ h e figures, in 1990 dollars, are based on prices in 1980, inflated by the U.S. Consumer Price Index. 19This assumes that increases in value near stations are not offset by decreases in areas far from stations. Also, about one-third of all riders on the system reside within the city limits; any premiums associated with housing within the city are not included in the figures. CONCLUSION Even in a region with multiple employment 20The difference in premiums could also reflect price differences between the two services. Additionally, the PATCO impacts may be magnified by the lack of alternative employment centers in Camden County. 11 BUSINESS REVIEW centers, access to the CBD remains valuable to suburban residents. A high-quality, CBD-oriented public transportation system can affect suburban residents' choice of neighborhood, the number of cars they buy, and the value of their houses. The house-value premiums asso ciated with the transportation service can be sustained, however, only if service quality is m aintained or enhanced , and if the CBD retains a productive advantage over other employ ment centers. The productivity of the CBD is not indepen dent of the transportation system, as an attrac tion of the CBD is its accessibility to a wide labor pool. However, if other factors—such as local taxes, poor services, or crime— reduce the CBD's attractiveness, the real estate premiums associated with the commuter rail system are 12 JULY/AUGUST 1991 likely to diminish. Additionally, increases in train-service quality are likely to increase housevalue premiums, while eroding service quality will likely have the opposite effect over time. In the Philadelphia area, these effects can be large, as indicated by the estimated $1.45 bil lion premium on suburban real estate values associated with commuter train service. At a d iscount rate of 10 percent, su burban resid ents with train service would enjoy positive finan cial benefits even if they paid up to $145 million a year to support the two rail systems that serve Philadelphia's CBD. This estimate suggests that, despite the region's increasing decentrali zation, over 40 percent of the metropolitan area's suburban residents have a direct interest in the quality of public transportation and the economic health of the CBD. FEDERAL RESERVE BANK OF PHILADELPHIA Lessons on Lending and Borrowing in Hard Times p X roblem loans and highly leveraged trans actions have brought home a truth about lend ing that is easily forgotten in good times: loans sometimes fail, with sad consequences for both borrower and lender. Many existing loans have soured, causing lenders to tighten credit terms on new lending. Meanwhile, borrowers have complained— and policymakers have openly worried—that lenders are refusing sound loans. ^Leonard I. Nakamura is a Senior Economist in the Banking and Financial Markets Section of the Research Department, Federal Reserve Bank of Philadelphia. Leonard I. Nakamura* New theories about lending and about loan contracts emphasize the difficulties lenders face in ensuring repayment of their loans. Accord ing to these theories, the collateral for a loan is not just a back-up source of repayment if the borrower defaults; collateral is also crucial for inducing payments from borrowers who can make them. Cash-strapped borrowers, when their busi nesses sour, will often try to put off lenders and keep paying their employees, suppliers, and landlords. In response, lenders will threaten to seize collateral and declare loans in default to ensure they get their fair share of a distressed borrower's cash flows. 13 BUSINESS REVIEW This threat is a blunt instrument that often harms the lender as much as the borrower. After all, the value of the borrower's collateral, particularly during a recession, may be insuffi cient to repay the loan. But there are other considerations, as well. A foreclosure causes valuable resources to be lost that would not be lost otherwise. Management may lose partial control over the firm becau se of bank ru p tcy rules, or spend too much time in court, strug gling against creditors and other claimants, and too little time running the business. Cus tomer relationships inevitably worsen as cus tomers begin looking for alternative suppliers. And ultimately, if an otherwise viable bor rower is liquidated, valuable relationships be tween management, employees, and custom ers are lost. If a borrower's business is fundamentally sound, its longer-term profitability ought to be the best source to repay the loan. But if the lender forces the borrower out of business, this source of funds is lost. THE DILEMMA OF FORECLOSURE In the tale of the goose that laid the golden eggs, the owner foolishly tried to get the goose's prized eggs more quickly by killing it. Lenders are not so unwise; still, they might have to threaten foreclosure as a way to force borrow ers to repay. When lenders must carry out their threats, they kill the golden goose—and this is the dilemma of foreclosure. Unfortunately, standard economic theory had assumed away this dilemma, maintaining that the interests of borrowers and lenders could always be aligned through loan agree ments. Consequently, economists believed that inefficiencies associated with loan default were small and that liquidation decisions were al ways sound. After all, wasn't it true that only firms having no value as going concerns were liquidated? More recent theories offer less optimistic conclusions about lending. They show that Digitized for 14FRASER JULY/AUGUST 1991 firms having value as going concerns may well be liquidated and that inefficiencies associated with loan default can have important conse quences for aggregate economic activity. In particular, the implication is that some loans that would ordinarily be made in good times would not be made in uncertain times. These newer theories are more realistic about the potential for conflicts betw een b o rro w er and lender; accordingly, they are useful guides—to all parties—for anticipating, and thereby less ening, the pain associated with hard economic times. Two theories in particular have emphasized the importance, and difficulty, of maintaining the borrower's incentive to repay. The idea is disarmingly simple: if given a choice of how much to repay, a borrower who wishes to maximize profits will always choose to repay the smallest amount. One theory, originated by Robert Townsend,1 underscores the lender's ignorance, relative to the borrower, of the borrower's net worth. The other theory, origi nated by Oliver Hart and John Moore,2*empha sizes the borrower's control over cash flows (the revenues that flow to the borrower from sales of products and services). TOWNSEND: LOAN CONTRACTS REDUCE INFORMATION COSTS Townsend's model stresses the cost to inves tors of obtaining financial information about borrowers. Before granting a loan, outside financial investors must first obtain detailed information about the firms seeking finance. This information extends to the firm's products and services, the customer base, marketing R o bert M. Townsend, "Optimal Contracts and Com petitive Markets With Costly State Verification," Journal of Economic Theory 21 (1979) pp. 265-93. 201iver Hart and John Moore, "Default and Renegotia tion: A Dynamic Model of Debt," MIT Working Paper (August 1989). FEDERAL RESERVE BANK OF PHILADELPHIA Lessons on Lending and Borrowing in Hard Times data, advertising plans, management, alterna tive financial resources, plant and equipment, labor resources and costs—in short, a detailed financial analysis and forecast. And until the financing is actually in hand, the firm has a strong incentive to provide investors with satisfactory information. But once the investment is made, the inves tors may not be well positioned to keep in formed about a firm's net worth. Information gathering is a costly procedure, requiring, at a minimum, an audit of current assets and liabili ties, an explanation of variances between planned income and expenses and the results achieved, and an evaluation of future profit prospects. The loan contract, according to Townsend, minimizes this informational cost by specifying a fixed dollar amount that the borrower agrees to pay; as long as repayment is made, no further financial investigation is required. If the bor rower fails to repay, however, the lender inves tigates the borrowing firm, learns of its net worth, and seizes its assets up to the value of the debt plus the cost of the investigation. A solvent borrower will have a strong incentive to repay, as long as the costly investigation following default makes a solvent borrower worse off. The loan contract thus minimizes the cost of post-investment financial investigation while preserving the incentive to repay.3 The Defaulting Borrower Pays a Penalty. 3Townsend's model can best be understood by compar ing the loan contract with a venture-capital contract, whereby the investor expects to receive a share of the venture's net worth. This financing contract repays the investor an amount depending on the firm's net worth; if the investor is ignorant of the firm's position, the owner, in an effort to minimize the repayment, will likely claim that the firm has low net worth. As a result, this type of contract requires the investor to always know the firm's net worth— which is costly informa tion to obtain— and is likely only when the investor takes a large stake in the firm and the venture shows potential for substantial returns. Venture capitalists invest relatively large stakes in small start-ups and follow them closely. Leonard I. Nakamura In practice, a borrower who fails to make timely repayments faces the threat of loan foreclosure and seizure of collateral. (See the box on p. 16 for a discussion of collateral.) Although the borrowing firm can partially protect itself by seeking bankruptcy protection, its business and plans become subject to legal restrictions and scrutiny by the lender. Such constraints, not to mention the loss of reputation and goodwill that bankruptcy may entail, can hurt the firm. The key consequence of default, as required by Townsend's theory, is that the borrower pays a penalty:4 a loss of asset value. The penalty can be imposed on borrowers through various methods—loan workouts, liquidation, takeover of the firm by an outside administrator acting on behalf of creditors, or seizure and selling of collateral. (The practical steps on the road to liquidation are briefly defined in the box on p. 17.) Let's take, as an example, an investment in a fictitious computer chip manufacturer, Cus tom Chip. Custom Chip's value is only par tially its factory and inventory of materials and chips; much of its value is its new ideas for chips. Only an expert in the computer chip business can know how much Custom Chip's value increased—or decreased—in a given period. One way to find out might be to auction off Custom Chip's patents, its chip-design depart ment, and its manufacturing plant (as would happen in a liquidation). But doing that would destroy the firm. If Custom Chip owes its lender $2 million, then as long as the firm's true value is greater than $2 million, the owners will have a strong incentive to repay the debt rather than risk having the firm thrust into bankruptcy. The 4For a precise specification of how the losses of collateral associated with liquidation relate to the optimal debt con tract, see Jeffrey M. Lacker, "Collateralized Debt as the Optimal Contract," Federal Reserve Bank of Richmond Working Paper 90-3 (March 1990). 15 BUSINESS REVIEW JULY/AUGUST 1991 bankruptcy. The high cost of default is most obvious when the lender seizes collateral. The collateral is then no longer available to the borrower, who was actively using it, and it goes to a lender, for whom it has no direct use. The borrower loses by not having use of the collateral, which is often Are You Sure It's Collateral? necessary to doing busi Collateral may be any asset of the borrower. Physical assets would ness. In addition, the be land, plant, equipment, and inventory. Financial assets would lender incurs costs in include receivables (customers' promises to pay) and financial securi seizing, storing, and sell ties (stocks and bonds). ing the collateral. And as However, collateral is of value only to the extent that the lender can the lender has no special actually claim, seize, and dispose of it in the event of default. For most expertise with the collat borrowers, collateral is property that is a functional part of the eral, its value may dete business, and its value varies with the business's ups and down. Then riorate further while in there are other important assets—customer goodwill and other future the lender's possession. profit opportunities, for example— that are intangible and cannot be used as collateral because the lender cannot seize and sell them. Lender Must Carry Establishing a clear claim to collateral is not always easy. Lenders Out the T h reat. In must follow procedures, set forth in the Uniform Commercial Code, Townsend's model, the to ensure that their claim is valid. In essence, this requires clearly story ends there. Once identifying the collateral, making sure that no one else has a prior claim default occurs, the lender to it, and making public the lender's claim. This process is called must carry through the securing and perfecting collateral. If not crucial to the borrower's threat of foreclosure and business, the collateral may actually be held by the lender. However, seize the collateral. Thus, very often the collateral is integral to the borrower's business and Townsend's theory pre cannot conveniently be held by the lender. Numerous anecdotes attest to the problems that can arise with dicts that costly bankrupt collateral. In one instance, the collateral was salad oil, held in vats. cies will arise from the When default occurred, the vats turned out to contain water with a thin existence of debt con film of oil on top. In another instance, collateral was mineral rights and tracts— and that firms a car. But the borrower, it turned out, had never bought the mineral having more value as go rights, and when the lender came to collect the car, he found that it had ing concerns than in liq already been sold. uidation may be liqui A cattle rancher pledged five steers as collateral for a loan, but none dated solely because they of the steers was specifically identified as such. Just before the rancher defaulted, five steers left the herd and, caught in a lightning storm, cannot pay their debts. If sought shelter under a tree. The tree was struck by lightning and the lenders chose instead to five steers died. The rancher was able to argue successfully that the renegotiate the terms of bank's claim was to the five dead steers. the loan, then borrowers Collateral often deteriorates in value when the firm's lines of would lose their incen business deteriorate. When oil prices slumped in 1986, drilling rigs fell tive to repay. Unfortu in value. When retail sales slumped in 1990, the value of unsold nately, by foreclosing on merchandise declined along with them. If a firm's sales falter because borrowers who are po its customers are in financial straits, the firm's receivables will turn out tentially viable, the lend to have little value. ers may lose their best threat of foreclosure enforces the loan repay ment and means that the lender need not pay computer consultants to analyze Custom Chip's value. However, if Custom Chip cannot or will not repay the $2 million, the lender may have to declare a default and thrust Custom Chip into FEDERAL RESERVE BANK OF PHILADELPHIA Lessons on Lending and Borrowing in Hard Times Leonard I. Nakamura source of repayment: the borrower's value as cash flows and can always divert them from investors by, say, using cash to pay workers an ongoing business. A partial parallel for the lender's dilemma and suppliers instead. The only commitment entrepreneurs can can be found in the famous Bible story about King Solomon. The wise king was able to make is collateral—and lenders can seize col discern which of two women claiming to be a lateral if fixed payments are not made. This baby's mother was telling the truth when he threatened to cut the The Road to Liquidation: child in two. Similarly, the lender must threaten to destroy the firm Some Terminology in order to learn the owners' true assessment of its worth. In both Bankruptcy - A debtor is afforded relief from its debt under cases, the threat must be made in the provisions of the Bankruptcy Code either through a liquidation (Chapter 7 of the Code) or rehabilitation (Chapter order to learn information. King 11 for commercial enterprises and Chapter 13 for individu Solomon, at least, had the advan als). In a liquidation proceeding, the assets are collected and tage of knowing that his threat distributed by a trustee. In a bankruptcy, lenders cannot was only a th reat. But in seize assets or attempt to collect payments; secured lenders Townsend's model, the lender are entitled, eventually, to payments equal in value to their may discover that the firm cannot collateral, but unsecured lenders often receive little. Reha repay and that the threat will have bilitation and emergence from bankruptcy proceedings typi to be carried out. And so, a tem cally involve the consent of creditors and equity holders. porary cash shortage can result in Collateral - Any property of the borrower that secures the business failure when the lender debt to a lender. In the event of default, a lender may seize cannot verify that the borrower's the borrower's collateral; in bankruptcy proceedings, a se problems are indeed temporary. cured lender has first claim to proceeds from collateral. HART AND MOORE: COLLATERAL MAKES RENEGOTIATION POSSIBLE A more recent model, by Hart and Moore, explores loan renego tiation as an alternative to liqui dating the firm. But unlike Townsend's model, this one as sumes that investors have no dif ficulty maintaining good infor mation about borrowers—only trouble controlling them contrac tually. Hart and Moore assume that investors and entrepreneurs be gin with the same information and that they always learn new in form ation sim u ltan eou sly. However, entrepreneurs control Default - A borrower's violation of the loan's terms. Failure to make timely payments or to fulfill other terms, such as providing timely and accurate financial data, constitutes a default. The lender's response— foreclosure of the loan— typically includes the right to demand full loan repayment and the right to seize any collateral specified in the loan contract. Loan workout - A business plan by which a borrower tries to resolve a problem loan. The business plan is typically an agreement arrived at by the lender and the borrower in an effort to avoid bankruptcy proceedings. Renegotiation of loan payments is often a part of a loan workout. Liquidation - The collection and disposal of a borrower's assets. Renegotiation - Resetting the terms of a loan contract, typi cally involving a delay of payments and often a reduction in interest or principal. 17 BUSINESS REVIEW collateral, however, is worth more when left in the hands of the entrepreneur. And if collateral falls in value, as often happens in recessions, the lenders' ability to collect payment decreases. This theory rests on the idea that the variety of possible events that can affect a business is simply too large and complex to be captured in a contract. Moreover, as contract provisions become more complicated, both writing and interpreting the contract become increasingly expensive. Lenders thus keep financial con tracts in a form as simple as possible in order to enforce them at low cost. This allows them control over specific types of collateral, but not over details about cash flows. By Hart and Moore's reasoning, the owner of Custom Chip will repay the loan as long as the manufacturing plant and inventory of com puter chips (as distinct from anticipated future profits) remain valuable. However, if the plant and inventory fall in value, the owner can divert cash and ideas to start up a new firm, defaulting on the original loan, even if current cash flows would suffice to repay it. Threat of Loss Enforces Repayment. An other example of the role collateral plays in enforcing payment can be found in the mort gage market. Consider Robin House, who is buying a $200,000 house with a 10 percent down payment of $20,000; her debt is therefore $180,000. Initially, the value of the collateral— the house— exceeds the value of the debt by $20,000. The threatened loss of home easily enforces Robin's mortgage payments. But sup pose the housing market deteriorates and the home falls in value to $150,000. If Robin values her credit reputation (including assets the mort gage lender might be able to seize) at only $20,000, she has an economic incentive to de fault on the mortgage: her debt exceeds the value of the collateral plus bankruptcy cost. She may refuse to make mortgage payments even though she can afford them. Borrowers who lose their incentive to repay when their collateral falls in value frequently Digitized for18 FRASER JULY/AUGUST 1991 do default. It is also true that when borrowers are unable to repay, their collateral is often low in value—and both situations occur for the same reason: a weak economic environment. Consider inventory as collateral. When a firm fails, its inventory will consist of those goods it could not sell at close to the original price. Loans that are overcollateralized when made may be severely undercollateralized when fore closed on. Yet, this does not mean that the collateral serves no purpose; indeed, it helps ensure repayment during periods in which the borrower can repay. While lenders would prefer collateral with an unshakable value, it is extremely hard to find. Indeed, it is not always easy to put the proper value on collateral in the first place. Collateral may not be as difficult to evaluate as the value of an ongoing concern, but it still may not be straightforward. (See the box at right for difficulties in determining how much the home underlying a mortgage is worth.) Collateral Is Key to Renegotiation. In Hart and Moore's model, lenders can renegotiate a loan instead of seizing collateral. In a renego tiation, lenders may allow payments to be stretched out or even reduced so as to avoid the losses from seizing collateral. But since only collateral can enforce repayment, the lender will be willing to do this only if the borrower can offer immediate cash and future collateral that are at least as good as what the lender can gain through immediate seizure. If future col lateral is inadequate, the lender will foreclose and a viable firm may be lost. Thus, renegotia tion only partially solves the dilemma. Suppose Custom Chip is unable to repay a loan during a period in which profit margins decline because of a recession in the computer industry. The lender has two options. One, it could seize the plant and its inventory of chips. Or two, it could renegotiate— permit Custom Chip to stay in business, accept an incomplete payment of the loan, accept the owner's beach condo, say, for additional collateral, and agree FEDERAL RESERVE BANK OF PHILADELPHIA Leonard I. Nakamura Lessons on Lending and Borrowing in Hard Times Who Assumes the Risk in Offers to Pay Closing Costs? Real estate ads sometimes include the come-on "Seller will pay closing costs." This practice creates the innocent appearance of a generous home seller helping the prospective buyer who otherwise would have trouble making the down payment on the house. But is this practice innocent from the perspective of the mortgage lender? No, because the seller is really being generous with the lender's money. An offer to pay closing costs actually inflates the house's selling price. To see this, consider a house priced by its owner at $100,000 but whose true market value has fallen to $92,000. In method 1, the standard method, the owner straightforwardly lowers the price by $8,000, to $92,000. In method 2, the owner offers to pay the borrower $8,000 up front by paying the buyer's closing costs. House price Down payment Mortgage loan Closing costs Buyer puts up Seller gets Method 1 $92,000 $9,200 $82,800 $8,000 $17,200 $92,000 Method 2 $100,000 $10,000 $90,000 $8,000 $10,000 $92,000 The only difference in the bottom line is that the lender has loaned $7,200 more to the borrower; in both cases, the seller winds up with exactly the same amount of money. But suppose the house falls in value by 10 percent, to $82,800. In the first case, the lender is fully protected, and the borrower has no incentive to default on the mortgage. But in the second case, the lender is likely to take a substantial loss if the borrower defaults on the mortgage— and now the borrower may have an incentive to default because the collateral that the borrower loses is less than the debt the borrower would otherwise have to pay. to a partial write-down of the remaining debt. However, if Custom Chip cannot come up with some combination of current cash and future collateral that is more valuable than existing collateral, the lender will go with the first op tion and seize the collateral. So, in this case, although Custom Chip might have a good chance of substantial future earnings, it is un able to realize them because it cannot promise the lender an adequate share of future earnings. Renegotiations preserve the firm's value. eral, with which And loans that make re negotiation more possible by preserving repayment incentives are attractive to borrowers as well as lenders. In renegotiation, loans in which borrowers have uncom m itted re sources to offer the lender are preferable to loans in which borrowers have no negotiating room. In the Custom Chip example, the fact that the owner's beach condo can be put up as collateral helps keep the firm alive. If the owner lacked this resource, re negotiation would be less attractive to the lender. In 1989, M ichael fensen5argued that highly leveraged transactions would not result in bank ruptcies because lenders would always be better off renegotiating. In ret rospect, fensen's argu ment appears incorrect. One reason may be that, in many highly leveraged transactions, the borrow ers had very little cash margin, or extra collat to renegotiate. LENDING DURING A RECESSION Lenders' most difficult decisions are made during recessions. For the prospective bor rower, access to additional financing may be 5Michael Jensen, "Is Leverage an Invitation to Bank ruptcy? On the Contrary— It Keeps Shaky Firms Out of Court," Wall Street Journal, February 1,1989. 19 BUSINESS REVIEW crucial to survival. But for the lender, recession financing is treacherous. In recessions, the probability of bad economic outcomes is higher than at other times, and inefficient, costly bank ruptcies and liquidation are more likely. Unless lenders have established procedures for commanding cash flows from troubled bor rowers, they will be unable to lend profitably during recessions, when cash flows become more questionable. Collateral is crucial—both as an ultimate source of repayment and as a threat to command repayment. But in reces sions, collateral— unfortunately—becomes less reliable. Loan Contracts Are Less Efficient in Reces sions. According to both of the models just discussed, firms with going-concern value may be shut down if loan repayments cannot be made. This is more likely to occur during recessions, when demand falls and cash flows dry up. As a consequence, loan contracts are more likely to lead to inefficiency during bad times than in good times, since bankruptcies and liquidations are more likely. Thus, the practice of making fewer loans in a weak economy is consistent with these theories. Several other points about lending during a recession fall out of these models: 1. More collateral will be required to further ensure repayment, although this makes borrowing more difficult. During a recession, the increased risk that collateral will fall in value means that lenders will need larger amounts of it to main tain the borrower's incentive to repay. Inevita bly, more potential borrowers will find that they lack the collateral necessary for the loan they're seeking. 2. More documentation will be presented, and past lender-borrower relationships will be more important. Lenders should attempt to know more about borrowers during recessions be cause default is more likely— and more expensive— when lenders are relatively igno rant. This makes it doubly hard on borrowers whose normal lenders themselves become cash JULY/AUGUST 1991 constrained; for borrowers to exchange a lender who knows them well for one who does not will be expensive, if not impossible. Detailed and accurate record-keeping may make the differ ence in whether new financing is obtained. 3. Noncredit terms on loans will tighten. Tight ening noncredit terms for borrowers may make it harder for them to qualify for loans, but at least lenders will be able to continue making profitable loans in hard times. For example, in a weak real estate market, lenders should re quire higher down payments on mortgages and be particularly wary of techniques home sellers may use to foist greater risk on the lender. In addition, lenders may demand more cov enants to their loans. Loan covenants are legal conditions added to the loan contract that per mit the lenders to declare loans in default. Some covenants constrain managerial discre tion; others specify standards of continued creditworthiness. Covenants increase the lender's ability to seize collateral while it re tains much of its value. RENEGOTIATION IN RECESSION Hart and Moore's model also has implica tions for what borrowers and lenders can ex pect from loan renegotiations during a reces sion. The lender's purpose in renegotiating a loan is to achieve new combinations of cash and collateral that leave the lender better off than under the previous agreement.6 For example, a lender will write down an unsecured loan, 6This article assumes that a firm's lenders are acting in concert. A natural tension between lenders often emerges in loan renegotiations, and the presence of many independent lenders may complicate renegotiation outside the frame work of bankruptcy court. A lender acting independently should be cautious about infringing on the rights of other lenders; indeed, obtaining a preference over other lenders can be reversed if bankruptcy actually occurs. Worse yet, if the borrower is viewed as being in a lender's control, that lender may become liable to other lenders for the borrower's debts. FEDERAL RESERVE BANK OF PHILADELPHIA Lessons on Lending and Borrowing in Hard Times forgiving part of the debt to obtain collateral and immediate cash under a new agreement. Conversely, borrowers should realize that in times of a weak economy, failure to repay a loan is lik e lie r to have serious co n seq u en ces— collateral may be seized, for example. In assessing their possibilities for a successful renegotiation, borrowers should review those assets that may be used for cash and collateral. Lenders are best off pushing for low-risk operation of the firm. A debt-burdened bor rower has a strong incentive to divert funds at the lender's expense.7 To counter this, the lend er w ill— in what is called a "loan workout"—actively negotiate the borrower's business plan to maximize the probability of receiving cash flows. In the loan workout, the lender should push to err on the side of safety and carefully monitor the borrower's expenses and receipts to see whether the borrower is adhering to plan. (A bank that handles a borrower's transactions is often well positioned to conduct a loan workout because it can best observe the borrower's behavior.) Cutting costs to conserve cash should almost always be part of a workout plan. A borrower who must give up something during renegotiation is less likely to default frivolously. Both parties to the renegotiation should rec ognize the fundamental importance of good information. A strong relationship between lender and borrower and full, open communi cation are crucial to sound loan renegotiations. In a renegotiation, lenders often demand more information than in the initial loan process. Borrowers should recognize that, lacking good information, lenders ought not to make conces sions in a renegotiation. 7See Leonard I. Nakamura, "Loan Workouts and Com mercial Bank Information: Why Banks Are Special," Federal Reserve Bank of Philadelphia Working Paper 89-11 (Febru ary 1989). Leonard I. Nakamura A final but crucial point following from the logic of Hart and Moore's model is one more pertinent to planning for the next recession than surviving the current one. When embark ing on a relationship with a lender, borrowers too often care only about the short term, believ ing that all will be fine if only the lender grants the loan request. But borrowers ought to be forward-thinking, too, and ask themselves whether the lender will be helpful in hard times or force them to turn elsewhere when loan funds tighten generally. Just as lenders must look for sound borrowers, so should borrowers seek out sound lenders. CONCLUSION Recent theories on lending and the loan contract build on the idea that borrowers may lack adequate incentives to repay lenders. One conclusion they share is that loan defaults can have important economic consequences and lead to the failure of otherwise viable busi nesses. Another conclusion is that noncredit terms of loans can be expected to tighten in recessions. In a downturn, credit terms to new borrow ers normally tighten. The models attribute this tightening to the inherent conflicts that inten sify between lender and borrower during re cessions. Consequently, lending becomes less efficient and is more likely to lead to foreclo sures and real economic losses. By tightening up lending practices, lenders may be able to increase their confidence in repayment and perhaps avoid being excessively conservative in hard times. And by anticipating potential credit problems, borrowers may be better able to minimize them. Tighter credit terms are unpleasant for the borrower and may reduce the borrower's ac tivity from the original plan. But they may be crucial for borrowing to continue in a tough economic environment. 21 Philadelphia/RESEARCH Working Papers Institutions and libraries may request copies of the following working papers, written by our staff economists and visiting scholars. Please send the number of the paper desired, along with your address, to WORKING PAPERS, Department of Research, Federal Reserve Bank of Philadelphia, 10 Independence Mall, Philadelphia, PA 19106. For overseas airmail requests only, a $2.00 per copy prepayment is required; please make checks or money orders payable (in U.S. funds) to the Federal Reserve Bank of Philadelphia. 1990 No. 90-1 Gerald Carlino and Richard Voith, "Accounting for Differences in Aggregate State Productivity." No. 90-2 Brian Cody and Leonard Mills, "The Role of Commodity Prices in Formulating Monetary Policy." No. 90-3 Loretta J. Mester, "Traditional and Nontraditional Banking: An Information-Theoretic Ap proach." No. 90-4 Sherrill Shaffer, "Investing in Conflict." No. 90-5/R Sherrill Shaffer, "Immunizing Options Against Changes in Volatility." No. 90-6 Gerald Carlino, Brian Cody, and Richard Voith, "Regional Impacts of Exchange Rate Move ments." No. 90-7 Richard Voith, "Consumer Choice With State-Dependent Uncertainty About Product Quality: Late Trains and Commuter Rail Ridership." No. 90-8 Dean Croushore, "Ricardian Equivalence Under Income Uncertainty." No. 90-9 Shagil Ahmed and Dean Croushore, "The Welfare Effects of Distortionary Taxation and Government Spending: Some New Results." No. 90-10 Joseph Gyourko and Richard Voith, "Local Market and National Components in House Price Appreciation." No. 90-11 Theodore M. Crone and Leonard O. Mills, "Forecasting Trends in the Housing Stock Using AgeSpecific Demographic Projections." No. 90-12 William W. Lang and Leonard I. Nakamura, "Optimal Bank Closure for Deposit Insurers." No. 90-13 Loretta J. Mester, "Perpetual Signaling With Imperfectly Correlated Costs." (Supersedes 89-22.) No. 90-14 Sherrill Shaffer, "Aggregate Deposit Insurance Funding and Taxpayer Bailouts." No. 90-15 Dean Croushore, "Taxation as Insurance Against Income Uncertainty." No. 90-16/R Loretta J. Mester and Anthony Saunders, "When Does the Prime Rate Change?" No. 90-17 James McAndrews and Richard Voith, "Regional Authorities, Public Services, and the Location of Economic Activity." No. Digitized for90-18 FRASERSherrill Shaffer, "A Test of Competition in Canadian Banking." No. 90-19 Brian J. Cody, "Seignorage and the European Community: Is European Economic and Monetary Union in Danger?" No. 90-20 John F. Boschen and Leonard O. Mills, "The Role of Monetary and Real Shocks in NearPermanent Movements in GNP." No. 90-21/R Mitchell Berlin and Loretta J. Mester, "Debt Covenants and Renegotiation." No. 90-22 Richard Voith, "Transportation, Sorting, and House Values in the Philadelphia Metropolitan Area." No. 90-23 Gerald A. Carlino and Leonard O. Mills, "Persistence and Convergence in Relative Regional Incomes." No. 90-24 Sherrill Shaffer, "Stable Cartels With a Cournot Fringe." No. 90-25 Ahmed Mohamed, "The Impact of Domestic Market Structure on Exchange Rate Pass-through." No. 90-26 Loretta J. Mester and Anthony Saunders, "Who Changes the Prime Rate?" No. 90-27 Sherrill Shaffer, "The Lerner Index, Welfare, and the Structure-Conduct-Performance Linkage." No. 90-28 Sherrill Shaffer, "Regulation and Endogenous Contestability." No. 90-29 Brian J. Cody, "Monetary and Exchange Rate Policies in Anticipation of a European Central Bank." No. 90-30 Leonard I. Nakamura, "Reforming Deposit Insurance When Banks Conduct Loan Workouts and Runs Are Possible." 1991 No. 91- 1 Dean Croushore, "A Measure of Federal Reserve Credibility." No. 91-2 James J. McAndrews and Leonard I. Nakamura, "Worker Debt With Bankruptcy" No. 91-3 William Lang and Leonard I. Nakamura, "Housing Appraisals and Redlining." No. 91-4 Paul S. Calem and John A. Rizzo, "Financing Constraints and Investment: New Evidence from the U.S. Hospital Industry." No. 91-5 Paul S. Calem, "Reputation Acquisition, Collateral, and Moral Hazard in Debt Markets." No. 91-6 Loretta J. Mester, "Expense Preference and the Fed Revisited." No. 91-7 Choon-Geol Moon and Janet G. Stotsky, "The Effect of Rent Control on Housing Quality Change: A Longitudinal Analysis." No. 91-8 Dean Croushore, "The Short-Run Costs of Disinflation." Digitized for 91-9 FRASER Leonard I. Nakamura, "Delegated Monitoring With Diseconomies of Scale." No. 3 FEDERAL RESERVE BA NK O F PHILADELPHIA BUSINESS REVIEW Ten Independence Mall, Philadelphia, PA 19106-1574 Address Correction Requested