.

Once or twice during the final weeks of the online conference technical problems delayed postings. Nevertheless there was lively discussion and debate. Several issues received significant (often contentious) attention: Is speculation good or bad? Is the Georgist land tax good or bad? Is land speculation useful to society and the economy? Which is better: Torrens or title insurance? Does formalization help or hurt the poor? Should we worry about real estate bubbles? Some participants drew lines in the sand on these issues; others asked for more empirical evidence. There were also interesting contributions about Coase, "horizontal" condominiums, and the experience of transition economies. With one week left to go, moderator Omar Razzaz suggested that the group take up certain additional topics:

The relationship between poverty, asset ownership (formal or informal), and real estate markets. Is there any evidence or experience comparing survival strategies and vulnerability (comparing, for example, the landed and landless poor, the urban and rural poor, the formal and informal sectors).

Real estate's links with the macroeconomy. What evidence or anecdotal experience do we have about the relationship between real property ownership and increasing productivity of workers, households, or firms? Any evidence on investment, labor mobility, and savings rates?

Financial market development. How do financial markets affect efficiency and equity in the real estate market? How do secure, transferable, mortgageable real assets affect development of financial and capital markets? These issues had been discussed somewhat in the U.S. context but not in connection with developing countries.

Emergence of real estate markets. Discussed had focused on advanced economies, with little discussion of how markets evolve enough to allow the exchange, mortgage, or securitization of assets. What are the patterns in African, post-socialist, and other developing economies? What important lessons have we learned about sequencing, gestation periods, and the role of public policy?

The summary that follows is organized by these topics:

• Coase's theorem Costs and benefits of real estate policies
  
• Property cycles
  
• Efficiency in real estate markets
  
• Real estate in transition economies
  
• Real estate's links with the economy
  
• Real estate and environmental risk
  
• Administering land markets
  
• Concentration of land ownership
  
• Reducing population growth
  
• Land value tax vs. speculation

Responding to Malloy's explanation of the different legal positions possible in a condominium project, Mason Gaffney observed that in the rationale for privatization-by-giveaway so commonly advanced by prospective donees in nations emerging from communism, it is a common debating ploy to say they must have equity in land to borrow to build or otherwise improve the land. But Gaffney notices people putting all kinds of valuable buildings on land leased from private parties, and financing the buildings, and buying and selling and refinancing the separate interests thus created. He also notices the financing of improvements on land subject to heavy property taxation, as in Massachusetts, New Hampshire, New Jersey, and Johannesburg. Never underestimate lawyers' creativity in working out ways to protect and value separately the various interests involved. Fred Foldvary thought the term "horizontal condominium" was used for condominiums in which unit owners owned all the vertical space (the land and building) on a particular site as opposed to buildings in which different owners owned different floors. The term "horizontal," he said, was also used for law relating to civic associations generally. The Horizontal Property Act of Virginia, for example, covers homeowners associations and seems to imply an association of co-owners of sites.

COASE'S THEOREM

Austin Jaffe saw Coase's Theorem as an example of a simple proposal that is difficult to understand completely. His favorite illustration was about a parable (contracting between farmers and cattle ranchers -- or was it railroads and farmers?) that never actually happened. Peter Colwell responded with an anecdote about the Presbyterian Church's 21,000-acre Ghost Ranch Conference Center, which surrounds a 12-acre parcel of land in New Mexico, and the adobe house on it, once owned by the late artist Georgia O'Keeffe. Last summer the church exercised a right-of-first-refusal option to acquire the property which, if it had gone forward, would have required matching an outside offer of $3 million to purchase the property. The church wanted restrictions on use of the O'Keeffe property to protect the integrity of its own property and programs at Ghost Ranch. Finally the church agreed to relinquish its option for $350,000 from the Burnett Foundation, a Texas-based family foundation that wanted to preserve O'Keeffe's legacy in the Southwest. The foundation, in an agreement with the property's owner, promised to limit use of the house to O'Keeffe scholars; the church agreed not to build on a nearby 10-acre tract and to provide an easement for a private road to the house. Each was afraid of negative externalities from the other and each provided protection to the other through negotiation. It helped enormously that only two or three parties were involved.

Thanking Yu-Hung Hong for his discussion of Coase's position (see Summary 3), Mason Gaffney said that if Coase believed Stigler had misused his name, he should have broadcast his repudiation publicly, visibly, and often, not just in talks to small groups. In any case, we now have what the world calls the Coase theorem, which is often used as an apology for polluting, converting histories of polluting into property rights based on prior appropriation of the air or water. Coase might reasonably have anticipated this, and Hong's posting (see Summary 3) did not address what Gaffney took to be the major hole in the theorem -- the fact that willingness to accept is greater than willingness to pay (WTA >> WTP). And the example of the polluting industrial plant asks us to assume a false parallel: that the polluter has as much right to pollute the air as the victim of pollution has to air in its natural state. This overlooks the nature of the problem: B, the victim, does not object if A pollutes his own air (the air inside his own walls). B objects only when A invades the air circulating on B's land and might also object to A's polluting common areas, such as parks and streets. B's demands are much more modest and limited than A's demands. A is the one expanding his property right into an easement over the property of others. Yu Hung Hong proposed concluding the discussion of Coase offline as it was off topic.

COSTS AND BENEFITS OF REAL ESTATE POLICIES

Continuing a thread about the relative costs and benefits of real estate policies, Austin Jaffe said that the idea of zoning as conferring communal benefits dated back before William Fischel's fine book. The literature generally credits Nelson's ZONING AND PROPERTY RIGHTS (1976) with the idea. There might be others. It was Jaffe's impression that Common Property Resources means different things to different people. To those sympathetic to state regulation and generally hostile to private property regimes, the success of common property schemes is proof positive that at least in some places privatization is not required. To those for whom private ownership is an overriding goal, communal property leads to the tragedy of the commons. In the middle, one can find experts on common property resources arguing that property rights need not be on this spectrum.

PROPERTY CYCLES

About myopia, the real estate industry's tendency to overshoot, and the rationality of developers' behavior, Bertrand Renaud referred participants to the interesting work of Steve Grenadier (1992, 1995). Grenadier used an option pricing framework to evaluate the decision whether to develop or not, faced with a risky, volatile environment and incomplete information. He shows why developers may have a bias toward developing rather than holding up the investment. He also explains why volatility and oversupply differ significantly across segments of the real estate industry. His work has moved the agenda on endogenous cyclical factors forward. (At the request of colleagues, Renaud posted a new paper -- "Property Cycles and Banking Crises: What Have We Learned?" -- an overview of what we have learned about links between macroeconomic policies, banking behavior, and the intrinsic features of real estate cycles during this first decade of global finance. The paper was presented at the seventh IPD Investment Strategies Conference in the U.K. on November 26th, the leading European forum for public and private sector professionals active in the real estate industry.)

Max Kummerow hadn't seen all of Grenadier's papers, but he disagreed with one that said overbuilding is a "rational supply cascade" which tries to maximize the value of development options and capture monopoly rents by being the first project to finish. In Perth, once a number of projects were under construction, the initiative shifted to tenants. Grenadier ignored strategic behavior on the demand side. Enough tenants can afford to wait for lower rents in a falling market so that monopoly profits to earlier finishers probably do not occur once a lot of projects are under construction -- so there is little benefit to being first. To really be first you have to be irrational enough to start really early, when rents are too low to justify construction, and gamble on rent increases. Sydney is demonstrating tenants' strategic behavior right now; order of completion will not matter much. Moreover, it is never rational to lose money. Most projects lose money during an oversupply period -- that is the definition of oversupply -- so it is hard to claim that the overbuilding is rational. It looks more like principal/agent conflict and prisoner's dilemma market failure to Kummerow. Articles about rational oversupply seem like attempts to preserve faith in the idea that markets are never wrong. An interesting reference on that issue: Paich and Sterman's article (1993, in System Dynamics Review) demonstrated in an experimental situation that with delayed feedback, managers have trouble making correct decisions.

Suboptimal and irrational are different constructs, said Richard Green. In the prisoners dilemma, for instance, both parties behave rationally under the circumstances, which forbid them to cooperate, so there is a suboptimal outcome as a result of completely rational behavior. The Grenadier piece on cascading (where outcomes are both rational and suboptimal) is a beautifully crafted piece of theoretical work, but it rests on an implausible assumption: that developers will make decisions sequentially based upon how much private information they have, with the best informed going first.

Omar Razzazz said the issue wasn't rationality: developers and investors behave rationally given the information they possess (bounded rationality). This didn't solve but was the cause of the prisoners' dilemma and suboptimal outcome. The question remains, is there room for institutions to provide better information? If there is demand for such information (including more transparent valuation standards, info on prices, investments, permits, and the like), why has it not been supplied privately in many of the countries the Bank works on? Is the problem one of collection action? of concentrated costs and disbursed benefits? Is it that demand must be generated by big market players (banks and institutional investors, etc.) who can bear the cost, not small players (such as households and developers)? Are there public good elements that should be provided by the state, such as laws, regulations, and disclosure rules?

Richard Green asked if it were not a tautology to say that "the issue wasn't rationality, that developers and investors behave rationally given the information they possess (bounded rationality)." Peter Ellis responded that you can have perfect information and behave irrationally, or have imperfect information and behave rationally. The one is neither necessary nor sufficient for the other to exist. He believed the two concepts were mutually exclusive.

Mason Gaffney said he couldn't think of any objective, verifiable, duplicable test that would refute or confirm the allegation. Whatever the person does, you can call it "rational" by assumption, if you want to. One assumption is that it was based on information the person had, including information about the person's own subjective attitudes or beliefs or forecasts -- something you cannot check on, so that is not very useful. It seemed to Gaffney that some people have set up an imaginary paradigm in which it is useful for model building (and sounds good) to call people "rational," but he didn't see much operational meaning in it. It often went with another assumption, about revealed preference: whatever a person does that appears a little eccentric or self-defeating, it just reveals a preference which (by assumption) the person is pursuing "rationally" (whatever that means). Thus, we may hear a man say, "This land is not for sale at any price." ANY price? That includes an infinite price, such as all the rest of the land in the world, and then some. Is that rational? What remarkable information does this person have? Gaffney asked if he was just missing the point.

Peter Ellis said he could see that assuming rationality might be a tautology, but what would you replace it with? Would you say people are irrational (which assumes that you, on the other hand, are rational)? How else would you model behavior? Don't we all behave rationally given the information we have, even if, in hindsight, we might say it was irrational? Theory is not the real world. It only provides a simplified notion to help us understand the forces driving a process. If we are accurately describing real-world behavior, why call it theory? Equations and diagrams are simply tools to help us understand the real world. Both extremes are problematic: theoreticians who care not about reality give us math no one can understand, and practitioners who blindly cast aside theory give us examples no one can generalize from.

Mason Gaffney thought it was a fair question, with no short answer. People are what they are, and our discipline should begin observing them more. Clearly some people are very calculating and their behavior tends to force others to follow suit (for example, when the bank charges you interest at a certain rate). As Keynes pointed out, in securities markets the rationality of the more experienced traders often consists in anticipating the herd behavior of others. In land markets, what do professional valuers or appraisers do when we ask for their opinion? They look for "comparables." Buyers and sellers also look at comparables. So land markets (where value is not constrained by costs of production) experience wide price swings. During a boom, the tax assessor is the only valuing agent whose overestimate of price will dampen the boom. One benefit of a tax on land values (see separate thread) would be the tendency to balance the excesses of other members of the herd, even (or especially) when the assessor shares the herd's overestimate of value. "Rationality" is particularly hard to define when the subject is land valuation. To value land correctly would require perfect forecasts in perpetuity, which is more like omniscience than rationality. Gaffney said he had picked on the word "rationality" mainly to discourage the hubris of economists' private patois.

The core issue in understanding why problems develop with property cycles, said Bertrand Renaud, is to find out what structural improvements are needed to reduce volatility. Grenadier's work is not the millenium, but his work is a step above casual empiricism of the past based on anecdote, and helps account for a good deal of observed behavior.

Generally, he couldn't agree more, said Richard Green, but while assumptions underlying models needn't precisely replicate reality, they should roughly replicate reality, and the results the models produce should be robust to changes in the assumptions. In the case of the Grenadier cascades (and he does greatly admire the paper), Green wondered if the assumptions were too far removed from reality, and whether the model was robust. If they were realistic, he had no objections.

Green had raised an interesting point, said Renaud, but one that has more to do with research methodology than facts. Surely they don't want to start another discussion track in this virtual symposium. Would Green not agree that human beings do not actually behave according to the tenets of economic analysis (total rationality, independence of utility, greed, selfishness, etc)? Most of us think we live in societies, not markets, yet we have found the methodology of economics to be productive in analyzing market outcomes and in searching for sustainable ways to allocate resources - - in a way that is superior to warfare. This brings us back to methodological points Milton Friedman made about predictive power decades ago. Is it unfair to see Grenadier work's according to its interpretative and predictive powers? (It would be interesting to have similar virtual symposium events for a more limited set of policy and research issues. What, for example, do we really know about real estate cycles and sound ways to lower costly volatility?)

In his experience as a lender, said Edward Dodson, some developers, in behaving rationally, will build despite no demonstrated market for what they build, so long as the funds for such buildings come from other sources (such as commercial banks and insurance companies). Peter Ellis responded that the problems seemed to be stemming from credit markets. As long as banks are willing to make the loans, builders will keep on building. Perhaps in trying to identify the market failure we should be concentrating on credit and capital markets.

The week before, Peter Colwell had asked if there might be a market solution to the problem: Suppose by hook or by crook that loan-to-value ratios were lower; would that correct the apparently myopic decision-making associated with "cycles"? Now he said that the point he was trying to make earlier was that the position of the exercise price of the call option depends on the loan-to-value ratio. If the LTV ratio is quite low, the investor owns an asset; the call option feature is not so important. Maybe he should have come right out and said that earlier.

About the decline in property value and exercise of the call option, Man Cho said that the issue of "collateral damage" (a chain of adverse economic shocks, a drop in property price, and damage in collateral value) and its impact on homeowner's exercise of the call option is being examined more in the recent literature (see, for example, a working paper by Andrew Caplin in NYU and others). Cho believed a precipitous drop in property value in a region would have at least three effects on homeowners' exercise of options. It would:

(1) Constrain mobility. Owners could be either stuck in their current residence or forced to choose less housing.

(2) Lose an opportunity for welfare gain. When interest rates decline, homeowners can refinance their mortgages at a lower rate. For those with "damaged" collateral, that opportunity is less likely (because of the increased loan-to-value ratio).

(3) Increase the chance of exercising a put option. A decline in the collateral value would increase the probability of default, either through the borrower's "rational" decision or because of other trigger events (such as job loss) caused by economic shocks.

Such effects from collateral damage can generally be expected in certain geographical areas: Texas in the early '80s, California in the early '90s, and some East Asian countries in recent years. But the mortgage financing system in a given country also plays a role. If there is a prepayment penalty, for example, the second effect becomes irrelevant. The broader issue is what role market and government play in this situation. Carl Case and Robert Shiller have an interesting proposal: a hedging mechanism that can be traded in an exchange. (See Shiller's book MACRO MARKETS for details.) How applicable such a scheme might be will clearly vary from country to country. He hoped other participants could tell him more about this issue.

Peter Colwell said he thought job loss was neither necessary nor sufficient for increasing the chance of exercising a put option. Imagine, if you will, a household with positive equity but no income. Will they default or will they go through an orderly sale (or refinance so they could afford the payments)? They will not default, because that would be a waste of their equity. In fact, there are investors who advertise that they will help keep homeowners with positive equity from defaulting (they strike a deal in which the investors can share the positive equity). On the other hand, imagine a household with plenty of income but with very negative equity on their property (so negative as to swamp the private costs of default). Will this household default? Of course. Trigger events must relate to the prices of property, the subject property, and other properties that provide substitute living opportunities.

Colwell was right, said Man Cho. The incidence of default is critically related to the amount of equity a borrower has in the collateral. In fact, a key explanatory variable used in default literature is "the option in the money," which is a function of a gap between the current property value and the unpaid mortgage balance. He had used job loss as just one example of a trigger event, all other things being equal. He also expects substitute living opportunities to affect whether owners exercise the put option. Despite possible data/measurement issues, this might be an interesting area for further analysis. He welcomed further comment on the topic. Colwell said he'd written an essay on commercial default for the summer/fall '95 issue of the Illinois Real Estate Letter. He guessed one could find parallels between income effects and alternative housing opportunities. There is a huge literature on residential mortgages, with which Jim Kau is especially familiar.

Austin Kelly said that the paper by Tracey and colleagues, "Collateral Damage," interacts estimated equity with estimated failure to meet underwriting constraints and finds that the collateral constraints inhibit refinancing and raise foreclosure rates. But GAO reports on the VA and FHA programs show that lack of equity inhibits refinancing -- and those programs don't require appraisals or credit/income checks for refinancing. Possibly the authors have simply identified the effect of the default option on inhibiting prepayments (the Kau, Kennan, Kim analysis). But you end up in the same place: falling property prices leading to a wealth transfer (or rather, preventing a wealth transfer) from lenders to borrowers, and encouraging foreclosures, which both lenders and borrowers view as horrendously disruptive and inefficient (although some alternatives might be even worse).

Edward Dodson, who like Man Cho works for Fannie Mae, one of the world's largest investors in residential mortgage loans and issuers of mortgage-backed securities, argued that the U.S. mortgage finance system is the world's most efficient provider of financing because of the size of the secondary market for mortgage loans. Ongoing liquidity is not a problem for U.S. home buyers or home owners, which suggests that isolating causes of disruptions in the housing sector is a somewhat easier task than in other societies.

Referring to how a drop in property values affected mobility, the first of the effects on homeowner's options that Man Cho had listed, Dodson said that one important variable is the amount of actual out-of-pocket equity the home owner has in the property. When properties are purchased at the top of the market, with a minimum down payment and at the margin of housing choices (condominium units in suburban Houston in 1976, for example), the incentives to keep making payments is greatly reduced when a household has to relocate for income reasons. In a few cases, owners of homes that had declined in market value simply turned their keys in to their lender and were able to purchase another home of similar quality at less cost. Anecdotal information suggests that once the title was transferred to the lender they had no difficulty getting financing for the new home. Someone had made an interesting argument to him not long ago: that the New York City MSA has weathered boom-to-bust cycles so well because of the low rate of homeownership. Renters are inherently more mobile, with fewer assets to dispose of if their job changes.

As for the second effect, making it difficult to refinance at a lower interest rate, Dodson said what is interesting is that homeowners whose properties have fallen in value will, if their initial investment is more than nominal, continue making payments so long as they have income and the neighborhood remains stable (the vacancy rate doesn't escalate). Their (often rational) expectation is that given enough time, values will return to what they were and continue to climb. As for the third effect, increasing the chance of exercising the put option, this makes the case for what in the mortgage industry is called "risk-based pricing," which takes into consideration expected losses based on regional differences in economic conditions. As Man Cho would probably agree, the industry is not fully able to price for risk because of the fair-lending implications and concerns about "red-lining."

What interests Dodson about the regional predictability of problems in certain regions is that there is no such thing as a national economy any longer -- and perhaps there never was. Real estate markets (and the underlying land markets) are local -- often very local. The United States has been reasonably fortunate that regional recessions have not expanded into general recessions. Labor and capital, being mobile, quickly relocate to lower-cost areas without much interruption or loss in the quality of goods or services produced. This mobility is facilitated in the U.S. by national standards that have arisen in the secondary mortgage market. Remarkably, plenty of portfolio lenders have found niche markets for themselves by providing highly personalized services.

Dodson wondered if the hedging mechanism Man Cho had referred to amounts to a right of first refusal or the kind of forward commitment our industry now uses under a long-term standby commitment to purchase mortgage assets in the future at a given price?

EFFICIENCY IN REAL ESTATE MARKETS

Taking up an earlier thread (see Summary 3) about secondary mortgage markets in the United States, Robin Paul Malloy said his comments were about efficiency, not discrimination. In thinking about laws, legal systems, and institutions, one should not give primacy to efficiency. What the secondary mortgage market illustrates is that a focus on efficiency can blind us to the consequences of changing the nature of the exchange process, can make discrimination invisible. Efficiency is only one factor -- and not the primary factor -- to consider in the relationship between law and market theory. Law has a different purpose than economics and market theory is informed by more than economics. Economic analysis is most useful when major disagreements within society are about facts; it provides little guidance when the disputes are about underlying values, and unfortunately many of our most pressing social problems are grounded in disputes about values. Law involves mediating between conflicting values in a way that economics does not, so efficiency considerations are of little value. Malloy's primary concern in evaluating legal infrastructure in U.S. and emerging markets is the exchange process. How do different approaches affect the networks and patterns of exchange and how do they influence human practices and values. The process of market choices is interesting to him as an exercise in interpreting not just costs and benefits but rational wealth maximization. We cannot understand markets or property rights with understanding the human practice of exchange. Efficiency analysis looks only at factual disputes, such as how much discrimination is measured by a certain indicator; exchange looks at market processes to identify the way in which market arrangements relate to various social and cultural connections. Property rights and relationships are deeply embedded in cultural, political, and other practices, reflected in the intricate nature of formal and informal law. Economics along cannot explain the use of various forms of zoning, financing, or tenure systems, which have only a secondary connection to efficiency. Understanding real estate transactions in emerging markets requires more than statistics on pricing, sales, and the like. We must map out the web that connects human practices to the legal infrastructure that enables development to promote the culture of exchange.

Earlier in the conference Steve Malpezzi had suggested a discussion of regulation and other public interventions in land and real estate markets, but that thread got lost in the shuffle. He now resumed the discussion, taking up ideas covered in his survey paper on regulation. Most of the online discussion, he said, had focused on the "big picture"; he wanted to get more specific. Most of his comments (and most of the research readily available) focused on housing and land; the effects of regulation on commercial property are woefully understudied.

Regulation is neither good nor bad, said Malpezzi. What matters are the costs and benefits of specific regulations under specific market conditions. Regulation, which he would discuss most, is only one of many instruments for public intervention in real estate and other urban markets. Governments around the world intervene through:

o the definition and enforcement of property rights

o taxation

o subsidies

o direct public provision.

In a sense these interventions can be treated as substitutes. Certainly each can be valued, and their incidence can often be studied, but in other senses they are not equivalent. Much of the environmental literature, for example, suggests that tax policies are superior to command and control regulation in most circumstances. The World Bank's study of Malaysian housing markets, "Getting the Incentives Right," illustrates how to study interventions in a unified framework -- how, using simple but defensible assumptions, a set of taxes, subsidies, and regulations can be treated as functionally equivalent to study their net effect on outcomes in urban real estate markets.

Why governments intervene: types of market failure. One classic rationale for public intervention is a public good, which economists define as existing where there are no rivalries for consumption of the good; where consumers cannot be excluded from consumption of the good, once provided; and (in most definitions) where no method exists for determining consumers' true willingness to pay (but see discussion of Tiebout in Malpezzi 1998). National defense is a classic public good. It is generally impossible for a government to defend some of its citizens and not others. At the same time, individual taxpayers have an incentive to understate their willingness to pay for defense, since an individual either consumes the entire defense package or leaves the country. True public goods in this strict sense are rare in urban areas.

Another classic reason for public intervention -- the existence of a natural monopoly because of decreasing or increasing returns to scale for the entire relevant range -- is generally cited in discussions of public utility regulation and the public provision of infrastructure. Baumol (1982) showed that markets could work well even under such conditions so long as the markets were contestable (entry and exit were free). Barriers to entry, or general conditions of entry and exit, have long been recognized as critical, although the literature has long recognized that market entry or exit can be impeded by regulation, whether intended or not. Other sources of market failure include the absence of clearly defined and enforceable property rights (the definition of property rights and the adjudication of disputes is essential for market transactions); large transaction costs, especially those resulting from information failure (especially asymmetric information); and externalities (costs -- or benefits -- imposed from outside the transaction), the most oft-cited rationale for regulating land and real estate markets.

MAJOR TYPES OF URBAN REGULATION

Land and real estate development is governed not only by planning processes, but by zoning regulations; restrictions on converting rural land to urban uses; land use regulations such as those governing road widths, setbacks, and floor area ratios; building codes; rent controls; impact fees; and many regulations affecting the provision of infrastructure and the transport network needed for real estate development. (Reviews of the literature on such regulations can be found in Fischel 1990, Pogodzinski and Sass 1990, 1991, Malpezzi and Ball 1991, and Malpezzi 1996.

Zoning, greenbelts, and restrictions on land use conversion. The mechanisms for regulations that mandate or limit how a parcel of land can be used include zoning, greenbelts or "urban service boundaries," and restrictions on converting land from agricultural to urban uses. The mechanisms for such regulations vary. In U.S. zoning, for example, it is common to follow a set of land use codes keyed to a map of parcels; in much of Europe, planners make decisions about land use. The way zoning can correct for externalities is clearly laid out in a number of theoretical papers (including Crone 1983). The first question is, Are there such externalities in land use -- or are they "large" and can regulation mitigate them? Somewhat surprisingly, several studies in the U.S. and Canada conclude that such externalities are often not very large. Mark and Goldberg (1986), Crecine, Davis and Jackson (1967) and Grether and Mieszkowski (1980) undertake to measure spillovers (how much a parcel's value is affected by nearby parcels), typically using some variation of hedonic models (based on a regression of property values against characteristics of the property and also, in this case, of nearby properties). They generally conclude that externalities are surprisingly small and (Mark and Goldberg) that the effects vary over time. Generally these studies have suffered from the endogeneity of the zoning decision. In fact, if zoning works well -- internalizing the externalities -- and is provided "on demand" to current homeowners, one will not observe many instances in which nonconforming uses reduce values, because zoning will prevent their occurrence.

Other studies (e.g., Lafferty and Frech 1978 and Li and Brown 1981) find that spillovers do matter. Lafferty and Frech, for example, develop an index of dispersion of nonresidential uses within cities and use this index in a model of housing prices. Cities with concentrated nonresidential uses do indeed have higher property values than cities where these uses are spread out among residential neighborhoods.

Only a few studies have tried to estimate the net costs and benefits of zoning. Peterson (1974) examined costs and benefits from Boston landowners' viewpoint and found that large-lot zoning conferred local benefits from the positive externalities associated with being in a richer neighborhood and from a lower fiscal burden. But under large-lot zoning, land is used less intensively so in a sense landowners bear a cost from developing less densely than optimal. In that case, the cost of prohibiting more intensive development greatly outweighed the two benefits. (The exercise did not consider the costs and distributional implications of restricting housing supply for households of modest means.)

Greenbelts are an extreme form of zoning: certain areas of the city are off-limits to development. Seoul, Korea, which has a large and stringently enforced greenbelt, also has very high housing prices and a strange development pattern for a market-based city -- more like the outcome from a command and control city such as Moscow (see Bertaud and Renaud). Kim (1991) presents a general equilibrium model that simulates the effect of relaxing Seoul's greenbelt. Under plausible assumptions, a 1-percent increase in Seoul's land supply (equal to a 1.2 percent decrease in the greenbelt) leads to a 1.4 percent decline in the price of land and a 0.2 percent decline in the asset price of housing.

Subdivision regulations. Subdivision regulations affect development at the project level. They cover things like setbacks; standards for roads, sewers, and other onsite infrastructure; and so on. Lowry and Ferguson show that in a lightly regulated county (Orange, Florida), finished building lots cost four times as much (per hectare) as raw land already zoned for residential development. In Sacramento County, California, which has much more stringent subdivision codes, the ratio is closer to 9 to 1.

Regulating density with floor area ratios. Using Ahmedabad as an example, Bertaud and Cuenco (1996) show how inappropriate levels of the floor area ratio (FAR) -- called floor space index (FSI) in India -- adversely effect development. The FAR is the ratio between the area of a land parcel and the floor space allowable. A parcel with a FAR of 1 can build 1 square meter of floor space for every square meter of land. With a FAR of .5, one could build a two-story house with 50 square meters on each floor on a 200-square-meter plot. One key to appropriate density regulation is to permit the FAR to vary with location within the city. Bertaud points out that the appropriate FAR varies in different locations of large cities. Ratios vary by 20 to 1, and sometimes even 50 to 1, between central business districts and suburbs in cities worldwide.

Indian planning regulations have a near-uniform FAR, ranging from 1 to 1.5 in most cities. In Ahmedabad, the FAR is as high as 3 in the old city area, which is high by Indian standards. But FARs in similar-sized cities around the world run as high as 5 to 15 in the central business district, falling to as low as 0.2 in suburban areas. The extremely low and relatively uniform FAR in Ahmedabad is shown to lead to greater total consumption of land and general urban sprawl. A low, uniform FAR increases the cost of land in the aggregate without making it more productive. Bertaud and Cuenco show that FAR should be allowed to vary within the city to mirror the changes in population density to be expected in a city that size. Land use would be better rationalized, infrastructure costs would fall, the city would be more compact, and traffic and parking problems would be reduced. They also show how an impact fee could be developed to cover the cost of additional infrastructure needed for such dense development, internalizing the currently external costs of such development.

Impact fees. Impact fees are effectively taxes on development. Generally, new development imposes marginal costs beyond outside costs. Among types of fees and exactions in kind are land and infrastructure dedicated for schools and the like; cash impact fees for community facilities, trunk infrastructure, and so on. Linkage fees can be used as a tax to bring private costs in line with public costs (to ameliorate problems such as congestion). There have been several studies of impact fees (such as Delaney and Smith 1989 and Gyourko 1991) but considering its potential importance this type of regulation is understudied.

Malpezzi and others use the "Bertaud Model" to numerically compare the costs and benefits of alternative project designs. The model can be used to analyze both specific projects and the costs and benefits of land use regulation generally. It costs out a proposed project design, simultaneously considering a wide range of design parameters, such as road width and design, floor area ratio (FAR), land required for public uses (such as schools and parks), infrastructure standards (offsite and on), minimum plot sizes and setbacks, site preparation costs, and design and administrative costs, among others. (Details about all inputs and calculations can be found in Carroll and Bertaud 1986, Bertaud 1981, and Bertaud and others 1988.) A simple example illustrates the model's value and how it works. Take the development of a plot for a housing unit with 10 meters frontage, for which the price of land is assumed to be $10 per square meter. Assume a requirement that the road in front of the house be 7 meters wide. If the house in question faces an identical unit, we would say that the land required for the road adds roughly .5x7x10x10 or $350 to the unit cost of development. Of course, the road provides an offsetting benefit but to set the "correct" road width, we want to know the cost-benefit of a meter of required width, on the margin. Bertaud and Malpezzi (1998) discuss exact and approximate methods for such analysis. Generally, exact methods require more information about the demand for public goods than one can get. The Bertaud model relies on an approximate method, best explained by example. Suppose that by studying the market in question we determine that existing developments can be found with households similar to the target market for this development, with roads averaging, say, 5 meters in width; that no evidence can be found that units with slightly wider roads command a higher price; and that no significant unpriced external benefit from wider roads can be established. Using one of the computer implementations of the Bertaud model, the user would enter a 5-meter road as a baseline case, and a 7-meter road as an alternative for comparison. Under the assumptions given, changing to a 5-meter road would yield a savings of $100 per unit. If it were determined that the wider road offered some offsetting (private or public) benefit, the amount of offset could be readily entered.

Various versions of the Bertaud model have been applied to roughly 30 countries, as diverse as India, Peru, Senegal, Russia, and Thailand. Malpezzi described how the model was used to analyze land use regulations in Malaysia in the late 1980s. The Malaysian government had promulgated new land use standards for low-cost housing, but under the new regulations Bertaud found that public uses took up 56 percent of land, only 44 percent was saleable, and the overall floor area ratio was only 0.23 -- although the regulations had been designed to give relief from an even more restrictive regulatory baseline. There were areas for potential savings, however. Requirements included wide roads (internal roads, 8 meters wide; distributor roads, 12 meters); back alleys 6 meters wide; and large setbacks on corner plots. Specific changes suggested by analysis with the Bertaud model included reducing internal roads to 7 meters and distributor roads to 10 meters, eliminating back alley requirements, reducing corner setbacks to 2 meters, and not reducing regular setbacks. Under the revised standards, the plot floor area ratio (FAR) increased significantly: to .78 for standard plots, .6 for corner plots, and .41 for the site overall. With the higher FAR the estimated profit per hectare rose to $193,000, a 17% increase over the baseline middle-income alternative development. Clearly changes in regulations can tilt profitability back toward the low-income market, which in Malaysia is most of the market. Density was increased to 378 people per hectare.

The model (and some of the lessons learned) are portable to other markets and have been applied to several countries moving from a command and control approach to a more market-oriented approach to land use (including South Africa and the post-socialist economies). In countries of the former Soviet Union, detailed master plans allocate land between various land uses and include a street design layout often to the tertiary level. The model has been used to test the consistency of permitted land use with current market price for each type of construction included in the plan. It can then be used to "back out" the value of undeveloped land. In assuming land uses based on parts of the master plan, using current construction and market sales prices, the calculated value of land is often negative. Using the model, the plan can be adjusted to reach positive land values. Russian municipalities, which both control land use and sell land, have an incentive to adjust standards to maximize land value. Parameters that can be readily improved upon include type of land use, road right of ways, infrastructure standards, and standards for community facilities. Politically, it is often easier to amend the current plan to make it consistent with market prices and customer preferences rather than recommend discarding the plan because it is inconsistent with market forces. So the amended master plan is progressively transformed into a zoning plan, allowing much more flexibility in land use and reducing the costly overdesign of infrastructure Malaysia experienced. Important model input includes current market price per m2 for housing, offices, and commercial facilities in different locations, so municipalities willing to use this method will put considerable effort into monitoring land and property markets. This monitoring has the positive effect on the market of disseminating real estate price information. Previously such price data were often only reluctantly shared by various government entities in charge of land sale and lease.

Rent controls. Among real estate-related regulations, rent controls have been studied most by economists, from the well-known partial equilibrium analysis of rent control as a tax on housing capital to more sophisticated models such as those of Olsen (1969) and Arnott (1991). Until recently, most of the literature on rent control has been theoretical rather than empirical (with such exceptions as Olsen 1973). Several years ago the World Bank undertook a research project examining the cost and benefits of rent controls in a number of markets. One part of the study was analysis of the type and nature of controls across some sixty countries (see Malpezzi and Ball 1993, 1994). Cost-benefit studies were also done for markets in Kumasi, Ghana; Bangalore, India; Rio de Janeiro, Brazil; and Cairo, Egypt.

The widely held notion that rent control trades efficiency losses for equity gains is not born out by Bank research (or most empirical) research. Certainly there are efficiency losses associated with controls (see Malpezzi and Ball 1993, for example, for evidence that housing investment declines with more stringent forms of controls). But research findings also show the deleterious distributional outcomes of controls. Although the median or "typical" renter in a rent-controlled market often had experienced some benefit from controls, for example, in each case study market there was an enormously wide distribution of household benefits. Some households received very large benefits; but in many other households the benefit from a lower rent was greatly outweighed by a welfare loss from disequilibrium in the consumption of housing services. Moreover, the costs and benefits were never well targeted by income and were sometimes perversely targeted.

In several countries it was possible to directly compare the incomes of landlords and tenants and the data suggested that rent controls were also not a very progressive redistribution mechanism. In all cases, landlords were richer than tenants, but not that much richer. The rule of thumb that emerged was that typically about a quarter of tenants would be richer than the median landlord, and about a quarter of landlords would be poorer than the median tenant, in the markets for which they undertook comparisons.

There are two general types of empirical study about housing and real estate prices and regulation. One type, a case study of one market or a few markets examines a rich set of local regulations in detail. (Asabere and Colwell 1984, Schuetz and White 1992, and Green 1997 are examples.) But studies of a single market may not be open to generalization. The other (complementary) type of study, less detailed but easier to generalize from, is the cross-market study.

Among cross-market studies that tried to measure "regulation" across U.S. markets, a few have examined the effects of regulation on land and housing prices. Segal and Srinavisan (1985), for example, surveyed planning officials and collected their estimate of the percentage of undeveloped land in each MSA rendered undevelopable by land use regulations. Using a simple OLS model of house prices, they found that the percent of developable land removed by regulation had the hypothetical effect on house prices. In the same journal issue, Black and Hoben (1985) categorized MSAs as restrictive, "normal," or permissive, based on a survey questionnaire of planning officials. They appeared to base this on a series of questions from which they scaled "areas most openly accepting growth" as +5, and those where growth was "most limited" as -5. They found a simple correlation of -.7 between their index and 1980 prices for developable lots. Chambers and Diamond (1988) used data apparently based on the ULI questionnaire in a simple supply and demand model for land. They found mixed results. In their equation explaining 1985 land prices, for example, average time of development project approval had a positive and significant effect on land prices but a negative and insignificant effect in the 1980 regressions. In another paper using the ULI data, Guidry and others (1991) found that the average 1990 lot price in 15 "least restrictive" cities was $23,842 but that in 11 "most restrictive" cities the average was $50,659.

Rose (1989a,b) constructed an index that measured land removed from development by natural constraint and Rose (1989b) used the number of governments a la Hamilton as a proxy for regulatory constraint. City by city, Rose carefully measured area removed from development by natural constraint (mainly water), and used a simple monocentric city model to account for the fact that an acre removed close to the central business district has a greater effect than an acre further out. Using FHA and ULI land price data for 45 cities, he found that the natural and contrived restrictions explained about 40% of variation in land prices, of which about 3/4 reflected natural restrictions and about 1/4 reflected regulation.

States as well as local governments regulate land use. In the '70s the American Institute of Planners collected a great deal of information about state land use and environmental regulations (American Institute of Planners 1976). Shilling, Sirmans and Guidry (1991) found that cities located in states with more restrictive land use regulations had higher land prices. The elasticity of price with respect to state land use controls was estimated to be 0.16. Malpezzi (1996) constructed a cross MSA regulatory measure from data collected by Linneman and Summers (1990) on:

(1) The change in approval time (zoning and subdivision) for single-family projects between 1983 and 1988.

(2) Estimated number of months between application for rezoning and issuance of permit for a residential subdivision less than 50.

(3) Time for single family subdivision greater than 50 units.

(4) How the acreage of land zoned for single-family use compares with demand.

(5) How the acreage of land zoned for multifamily compares with demand.

(6) Percentage of zoning changes approved.

Scale for adequate infrastructure (roads and sewers).

Malpezzi found that house values are strongly affected by regulation, albeit in an apparently nonlinear way. Additional research by Malpezzi, Chun, and Green (1998) shows that the strong relationship between regulation and housing prices is robust with respect to choice of model and measure, especially if an instrumental variable is used to correct for possible simultaneity bias.

Of course, regulations produce benefits as well as costs; why else would we regulate? Regulations are presumably put in place to tackle such externalities as congestion, environmental costs, excessive infrastructure costs, fiscal effects, or neighborhood effects. External benefits might also be found in measures of productivity and employment, health benefits, racial or economic integration, or externalities associated with home ownership. Malpezzi (1996) tested for price effects in both owner and rental markets, as well as for the effects of regulation on tenure choice, neighborhood rating, racial segregation, and congestion. He found there was little benefit to offset costs, once past the inflection point apparent in the attached figure. Home ownership rates declined (which one could view neutrally, or as an additional cost if one believes that such asset ownership is important). The only measured positive effect was a slight reduction in commutes.

This cross-market work has been pushed furthest using U.S. data, but has its roots in international comparisons. To facilitate international comparisons (see figure), the dependent variable is the ratio of typical house prices in the largest metropolitan area of each country to the median income of the same metro area. The measure of regulation is actually drawn from work on price distortions from government interventions more generally by Agarwala (1986). The positive relationship between "bad" regulatory environments (overvalued foreign exchange markets, rationing finance by directed credit rather than by price, distorted labor markets, etc.) clearly affect the housing market. Malpezzi (1990) guessed that this could be because some such policies affect real estate markets directly, but it could also be because there is a correlation among policy environments. In other words, countries that have distorted regulatory economic policy environments tend also to have distorted real estate and other urban development regulations (see Malpezzi and Ball 1993 for confirmation and discussion, and recent work by Angel and Mayo (1996) with data from the Housing and Urban Development Indicators Project, confirms these qualitative results.)

Recent events in Asian real estate markets. The topic of the hour is the role real estate played, or didn't play, in the recent East Asian crisis. Most people studying this issue are too close to events, with too little real research available, to make confident pronouncements, but Malpezzi offered a few comments and/or conjectures. News reports suggest that the trigger event may have been the failure of Thai banks to properly underwrite real estate lending (among other things). A perusal of property company stock prices suggests that they led other stock market declines in several countries. What does this have to do with regulation? The U.S. S&L crisis, Credit Fonciere, and other past financial episodes have demonstrated the macroeconomic costs of inadequate prudential regulation, adverse banking incentives, and the failure of banks and policymakers to understand or act properly on the risk/return characteristics of real estate. Drawing on previous experiences, we can obviously design regulatory systems that reduce the frequency and severity of such episodes. Malpezzi especially argued for a well-designed system of prudential regulation and the introduction of true underwriting of real estate (and other) lending, solutions far superior to calls for blunt limits on lending for real estate or other activities. Heavy-handed controls are ultimately just a way of allocating capital bureaucratically instead of according to return. But we have to solve the systemic problems that have led banks to make mistakes as big as any bureaucrat might make.

Research under way, or soon to be, will focus, among other things, on the links between real estate prices and the crisis. The dynamics of real estate prices are still poorly understood, despite their importance (real estate represents more than half the world's tangible capital stock). Even basic relationships between, for example, real interest rates and asset prices are surprisingly hard to pin down, even in countries with well-developed data systems. The research under way in Asia is especially hampered by data problems. The Bank should take a leading role in collecting better, more timely data. Malpezzi's conjecture, unencumbered by actual results, is that simple Granger/time series models of real estate prices and exchange rates will yield models that permit us to predict (following Samuelson) "10 of the next 3 crises." The same will be true of any monocausal model, such as one focusing on short-term external debt. After all, Korea and Malaysia -- to name just two Asian countries -- have weathered extreme boom-bust cycles in real estate (closely related to their difficult development regulation environments) several times during the last two decades, without triggering anything like the current crisis.

NEXT