Document View

Market surveys, and the accompanying valuation technique known as contingent valuation analysis, have gained increasing acceptance in estimating values of properties or property features, especially when parcels are affected by conditions perceived as undesirable. While earlier articles have noted problems with surveys and the importance of the marginal buyer in determining prices, this discussion examines the marginal buyer in the context of a clientele that responds negatively (or positively) to a particular condition. It then explains how the market's supply side, which is typically omitted from contingent valuation, can be combined with demand side-based survey results so that both sides of the market are considered. [PUBLICATION ABSTRACT]

Introduction

Because relevant data is rarely as plentiful as market observers would like it to be, real estate market analysts are ever on the lookout for new tools that can help them substantiate the various components of their value estimates when the applicable property rights are difficult to value with traditional methods. Consider, for example, adjustment factors for use in the sales comparison approach. Finding matched pairs has almost always been easier said than done, and even the more powerful regression analysis can fall short if the model is not skillfully specified, or if recent transactions involving the relevant type of property have been few.1 As a result, some analysts have attempted to determine adjustment factors in difficult valuation situations by simply asking a large number of potential market participants about their willingness to pay for a property with a particular feature or condition.2 One type of feature that might spark analysts' interests, and lead to the use of contingent valuation analysis (market surveys that pose "what if?" questions regarding real estate values),3 is the presence on the site of such potential disamenities as environmental contamination or high voltage electric transmission lines.

In the discussion that follows, a technique for using survey results to estimate the value of a property or property component is presented. Initially, some problems with surveys are discussed to show why it is critically important to place survey research in the context of market analysis. Subsequent sections describe how the actions of the marginal buyer, as part of a clientele, affect market outcomes. Finally, some brief numerical examples and a graphical representation illustrate how an analyst might combine survey results with supply information, which users of contingent analysis too often ignore,4 in examining the price effects that reflect both sides of the market together. Therefore, while this paper is correctly viewed as a criticism of survey-based valuation work, it can also be seen as a manual for doing, and interpreting, contingent valuation analysis more effectively.

Problems With Surveys

While any process of asking people for their views is fraught with danger, a few common problems encountered in surveys are especially worth noting. One is nonresponse bias: individuals who do not know or care much about the topic of a survey simply may not respond. Those who do respond in such a case are primarily people who feel strongly about the issue, and thus may have an "ax to grind." Analysts, especially in large real estate appraisal firms, have in some instances tried to deal with nonresponse problems by hiring professional surveying companies that directly contact large numbers of desired respondents, especially owners of identified properties. But because this process may elicit answers from those who have not spent time thinking about the issues, the unfortunate choice can boil down to having responses from knowledgeable parties but too few in number to be statistically valid, or having enough responses for statistical validity but with questionable value indicators.

There is also a potential problem relating to the reliability of the answers. A survey respondent might intentionally misrepresent the estimate of the impact that contamination or some other condition has on property values, overestimating the impact if the individual believes that the survey results might help to reduce property taxes, or be used as evidence in a lawsuit that determines compensation to affected land owners, while underestimating the impact if the individual believes that the survey might influence buyers of the property.5 Therefore, even a surveyor who recognizes, and tries to account for, the potential for biased responses might have to wonder whether a respondent's bias would be in an upward or downward direction. The questioner might try to get beyond these obvious problems by surveying only people in the market area whose property would seem not to be affected by the disamenity. Of course, these parties, facing no specific impact from the condition, might be especially prone to being among those who have not thought carefully about the general impact on affected parcels' values. Yet, at the same time, excluding such people from the survey also creates problems, in that those directly affected by a disamenity may systematically have views that are atypical of the market.

A further complication is the fact that some people could feel positively about a condition others seek to avoid. When form follows function in design, the result can be seen as attractive; the repeated theme of transmission line support towers, stretching to the horizon on a rural landscape, can serve as a type of public art. Some cellular telephone towers can have visually appealing features, as well. Aside from mere aesthetics, the acceptance of utility lines could reflect a public spiritedness. Consider that a right of way providing easements for one utility can offer a ready path for new utilities that enhance our collective quality of life; a community can benefit greatly from an established right of way to which access can be resold. Thus, it may not even be clear whether a property feature that seems, at first glance, to be detrimental should be expected to create a "stigma" that has a deleterious impact on transaction prices.

Clientele Effects

Yet despite the severity of the various difficulties with respondents and the interpretation of their answers, a less obvious but more profound problem must be accommodated when an analyst uses survey responses to measure a property feature's expected impact on market value. This problem is that a survey asking "How much more or less would you pay ...?" deals with only the demand side of the market. Price effects can be measured only if demand and supply are considered in unison, so that the position of the buyer at the margin can be analyzed. The buyer at the margin is the individual who is at the break-point between buying and not buying. It is this marginal buyer whose actions determine the transaction prices from which inferences on value are drawn.6

An added challenge that emerges for analysts dealing with this problem is that clientele effects can obscure the position of the buyer at the margin. The term 'clientele effect' is commonly used in finance, but not as much in the broader economics field. A clientele effect exists when the market for a given product or service contains distinct segments that do, or do not, appreciate an important ancillary feature. The clientele that appreciates the important ancillary feature, which might or might not be intentionally provided by the seller, has a propensity to willingly pay a higher price in a transaction that conveys the desired feature. A clientele that dislikes the feature in question has a propensity to avoid transactions that would convey the feature, and an insistence on paying less if the feature is present. Market participants neutral to the feature, and thus not associated with either of these clienteles, would be expected to transact without being affected by the feature, such that the prices they would willingly pay would not reflect the feature's presence or absence. Thus, the provider of an asset or service must figure out whether an existing, or potential, ancillary feature could give rise to various clienteles and, if so, which clientele's views would be most relevant to the setting of transaction prices.

Perhaps the most commonly cited example of clientele effects involves stock market investing. Many households and institutions purchase shares of common stock. Their general desire is to be partial owners of large companies, but an important ancillary feature of stock ownership is the possible receipt of company profits in the form of cash dividends. Some analysts of corporate financial policy feel that regular dividends are more likely to be paid if the stockholder base is largely a lightly-taxed clientele (retirees, Roth IRA investors, charitable foundations, other corporations), whereas profits would be more consistently retained for internal use if the stockholder base tended to face heavier income taxes (a clientele of high-income individuals).7

In the real estate arena, various types of individuals and institutions have long invested in large income producing properties. But within the vast group having some degree of interest in owning commercial real estate, various clienteles have had greater impact on the market than others at different times. For example, prior to the mid-1980s many wealthy individuals who sought to shelter their high incomes from taxes invested through syndications, such as limited partnerships. The LPs became a clientele whose needs were carefully considered by sellers and developers of commercial property, and whose tax-driven willingness to pay had a meaningful impact on market prices. But then passage of the Tax Reform Act of 1986 greatly reduced the tax-sheltering benefits of partnerships,8 and LPs were no longer the most profitable clientele for developers and sellers to serve. Equity REITs had come to the fore as a market-driving clientele by the early to mid 1990s, and continue along with less restricted real estate operating companies (REOCs) as media through which smaller investors can diversify.9 Another relevant clientele group over recent years has been the insurance companies, pension funds, and commingled funds, whose willingness to pay for income producing real estate has been motivated not by income tax issues, but rather by broader portfolio management considerations.

The Role of the Buyer at the Margin

Thus, the idea that specific groups' desires and actions can affect prices is not new to observers of the real estate market. Indeed, while the term 'clientele effect' may not be widely used by economists, urban economists have traditionally held that an area's residents will arrange their locations so as to be close to things they like, and somewhat removed from things they do not like. For example, a family with many children tends to benefit from locating within the boundaries of a good school system. This type of family makes sacrifices in other budget categories so it can outbid childless people, who do not value the school system to the same degree.10 Then, after acceding to the need to pay a premium to buy within the desirable school jurisdiction, the household with children might reasonably be willing to pay yet an additional premium to live in closer proximity to the schools. In fact, studies have confirmed that a price gradient is associated with distance to schools (i.e., the transaction price for an acre of land rises with increased proximity to schools, holding other factors constant)." Therefore, it seems that within the broad group of home purchasers, a clientele of child-raising families would willingly pay higher prices for homes with the important ancillary feature of being located near good schools.

However, an accepted tenet of economics is that the price that should be observed in the market is the price that the buyer at the margin is willing to pay. This marginal buyer stands at the intersection of the market's demand and supply functions. Some among the buyers not at the margin might be willing to pay more than the market price, but the price that every buyer pays is determined in such a way that the market clears (i.e., the quantity demanded equals the quantity supplied).12 So it must follow that a particular clientele can be expected to affect the market price, in a transaction characterized by some important ancillary feature, only if that clientele includes the buyer at the break-point between buying or not buying the particular good or service. If a clientele that includes this marginal buyer dislikes (likes) the feature in question, then there should be a price discount (premium) to reflect the aversion (attraction). Alternatively, if a clientele that includes the marginal buyer is indifferent to that feature, then no associated price impact should be observed.

Application

The Basic Idea

Suppose that the market for land affected by high voltage transmission lines can be segmented into potential buyers who dislike the transmission lines for some reason, and those who are indifferent. While different properties in the market would exhibit the types of dissimilarities that real estate typically does, assume that the transmission lines' impact would apply to affected properties in a homogeneous manner. Suppose further that the clientele averse to transmission lines outnumbers, perhaps vastly outnumbers, the group of potential buyers that is indifferent to the lines' proximity. It might seem that the lower price the anti-lines clientele would willingly pay should affect the price paid in market transactions. However, if the proportionate number of properties actually affected by high voltage transmission lines is quite small-smaller than the proportionate number of potential buyers who are indifferent to this seeming disamenity-then the lines' presence should have no impact on the prices ultimately paid for affected parcels.13 After all, under these circumstances the marginal buyer would be indifferent. In other words, if a successful buyer has to outbid at least one person who is indifferent to the transmission lines, then the price paid for a property should not be discounted for the lines' presence.

Think of a market, which is assumed here to be a competitive market in equilibrium (the underlying assumption in traditional real estate valuation techniques), where 10% of the parcels are impacted by high voltage transmission lines, while 20% of those who constitute total demand for the property are indifferent to having the lines nearby (the other 80% have a definite aversion). Thus the marginal buyer of properties impacted by the transmission lines will be someone at the 10th percentile mark. Because everyone up to the 20th percentile is outside the clientele that believes high voltage transmission lines pose a problem, the mere finding that people exist-even vocal people in large numbers-who dislike the lines is irrelevant to the question of whether the lines should be associated with an effect on transaction prices.

It may be that the strong feelings of a few cause there to be, on average, an aversion to land close to the high voltage lines. Indeed, it seems reasonable to expect that at least some people would dislike living or working on land containing any type of contamination or public utility. Yet even a strong general distaste that reduces the average willingness to pay is irrelevant to the expected selling price; the critical issue is whether the buyer at the margin is among those who dislike the potential disamenity. It is quite possible that a condition is disliked, or even feared, by a substantial number of people, and that this clientele's loud complaints attract the attention of journalists and politicians, but that no associated price discount would be observed in market transactions. Such an outcome could be expected if the clientele (the sub-group from among potential buyers) that reacted vociferously did not contain the buyer at the margin, who would determine the price paid for land with the feature in question.

For example, assume that there are 50 potentially interested buyers in a market where 20 available parcels are affected by a potential disamenity, and that survey results can be used to distinguish each possible buyer's aversion. If 16 of the potential buyers are "not very averse" to the disamenity (and would willingly pay $65,000 each) while the other 34 are "highly averse" (willing to pay only $45,000 each), then the average willingness-to-pay would be [(16 × $65,000) + (34 × $45,000)]/50 = $51,400. However, this average figure would not affect transaction prices; the market-clearing price would be the marginal buyer's $45,000 willingness to pay (the four least enthusiastic actual buyers would come from the highly averse clientele). But if 28 potential buyers were "not very averse" (willing to pay $65,000 each) while only 22 were "highly averse" (willing to pay only $45,000 per lot), the average willingness-to-pay would be a higher [(28 × $65,000) + (22 × $45,00O)]/50 = $56,200, yet the market-clearing marginal price would be higher still, at $65,000 (none of the buyers would be expected to come from the highly averse $45,000 clientele).

A More Complex Example

The previous example is perhaps overly simple in its implicit supposition that any clientele would be uniform in its like or dislike of the key ancillary feature. The real world tends to be more complex in that views within a clientele group could not likely be characterized simply as "highly averse" or "not very averse" or "indifferent." Actual views would be more likely to run gradually from extreme to mild dislike (or perhaps appreciation). In such a situation, the buyer at the margin might be part of a clientele that dislikes a key ancillary feature, but with some gradation of viewpoint within the group. If the marginal buyer were part of the averse clientele, for example, but had only a modest aversion, then the price paid in a market transaction would be expected to reflect a modest discount from what would be paid for an otherwise similar parcel lacking the disamenity.

Consider a more complex example in which the task is to analyze the likely price impact of a high pressure gas pipeline easement on certain residential lots. First, to better understand the demand side of the market, the analyst could examine a survey of the prices potential buyers would willingly pay for a lot burdened with such an easement. For demand issues to be dealt with thoroughly, the survey would actually have to include both potential buyers and current owners. While the views of current owners are often viewed as affecting the supply side of a market, in a case in which supply is essentially fixed in the short run (as with the number of parcels in a specified market area that exhibit a particular feature), supply would more correctly be viewed in terms of an unchanging "stock." Current owners, who would decline to sell if prices were insufficiently high, indirectly become part of the demand side, essentially reserving for themselves (i.e., choosing not to sell) if others do not offer high enough prices to buy from them. This "reservation demand" on the part of current owners, when added to the demand from potential buyers, constitutes total demand for the type of property in question.14 When the discussion here speaks of surveying "buyers" and the willingness to pay of the marginal "buyer" for a good in fixed supply, it is implicitly including current owners as part of the potential "buying" group.15

The next step would be to rank order the potential buyers (and current owners), based on the price discounts, relative to willingness to pay for a similar lot without the pipeline, they would expect to receive (extend). Next would be identifying the total number of these potential buyers whose willingness to pay fell at or above each specified level. Return to the market with 20 affected parcels and 50 potential buyers. Assume that a ranking of the buyers by preference reveals that the person least troubled by the easement disamenity would willingly pay $79,000 for one of these parcels. The next most interested party would willingly pay $78,000, with willingness to pay dropping by $ 1,000 for each successive buyer, down to $30,000 for the most averse. The average willingness to pay therefore is ($79,000 + $30,000)72 = $54,500, a figure that reflects the existence of people who dislike pipelines, but that does not tell anything about expected prices. Indeed, if all buyers are to pay the same price and the market is to clear, the selling price for each parcel has to be the marginal, or 20lh, buyer's $60,000 willingness to pay (anyone willing to pay $60,000 or more would submit a winning $60,000 bid).16

A Graphical Representation

This process can be described more formally, and more generally, as follows. After reviewing a survey that reveals potential buyers' (and, of course, current owners') willingness to "pay" for property exhibiting a particular feature, the analyst derives the density function suggested by the survey. This density function should identify the percentage of potential buyers whose required price discount (relative to willingness to pay for a similar property without the feature) falls at each of several specified levels. (If the results reflect the randomness and objectivity that characterize good survey design, then the computed density function should indicate the true impact of various clienteles on the market.) The next step is deriving the cumulative distribution function, which shows the larger percentage whose required price discount falls at or below (such that willingness to pay falls above) a specified level. The next step is to rotate a graph of the distribution function clockwise 90 degrees, so that potential buyers who would be willing to pay the highest prices (seeking only a minimal discount, or perhaps willingly paying a premium) are shown at the top; this representation can be interpreted as a traditional downward sloping demand curve. The only remaining step is to determine the disamenity's proportional impact on the market value of an affected parcel. The expected impact can be read as the vertical distance between the point of indifference (a discount of 0% or less required) and the discount associated with the percentage of potential buyers that equals the percentage of lots exhibiting the disamenity.17

Assume that the three figures shown below reflect a survey of potential buyers of lots affected by the high pressure pipeline easement discussed above. Exhibit 1, a graph of a density function, shows that almost 30% of buyers would expect to pay 7%-8% less than for a similar lot without the easement burden, whereas only about 3% would expect discounts in the range above 12%, while only 5% would willingly pay within l%-2% of the price that would characterize an otherwise similar lot with no easement problem.

To understand Exhibit 2, which is a graph of a cumulative distribution function, look at the bar for the "7% or 8%" range. Because this bar's height, as measured along the vertical axis, corresponds to 76%, it can be seen that just over three-quarters of all potential buyers would be content to receive discounts of no more than 7% or 8%, meaning that this large proportion of potential buyers would willingly pay 92% or more of the price they would expect to pay for an unaffected lot. Someone who asserted, based on the expressed views of vocal opponents, that the pipeline easements would reduce values by 10% or more would be proven wrong by the survey data, as long as there are at least 24% fewer affected lots than there are survey respondents, if it can be assumed that the survey truly reflects total demand.

Enlarge 200% Enlarge 400%

Enlarge 200% Enlarge 400%

Enlarge 200% Enlarge 400%

To genuinely understand price effects, however, so that market values can be estimated, both the demand and supply relationships must be examined for the property attribute being studied. Exhibit 3 permits an examination of both demand and supply, the latter in terms of a stock of properties with the specified characteristic. Note first that the figure is a 90° clockwise rotation of Exhibit 2. Connecting the "tops" (the high end of each range) of the bars with a line provides an estimate of the demand curve. The demand curve, by itself, tells nothing about expected market price, but if supply information is combined with this demand information, an idea about the price impact of the subject attribute can be developed. Exhibit 3 incorporates a supply-side assumption that the attribute is present in 5% of the properties (see the dotted vertical line) in the market that was surveyed; an analyst with strong knowledge of the market would not find it difficult to estimate the relevant magnitude. The market clearing price is then found where demand intersects this 5% stock, which suggests approximately a 1 % or 2% discount relative to the price expected for an unaffected but otherwise similar property.

Those who would willingly pay an undiscounted price are too few in number to purchase all the affected properties, but that group, plus those who would seek a discount of only 1% to 2%, would number more than enough to own all the affected parcels. Therefore, the marginal buyer of the affected property type will be someone who insists on receiving a discount, but of only 1 % or 2%, so the relevant clientele consists of those for whom the necessary discount is 2% or less. (If the attribute were present in 20% of the properties, the market price would be expected to be about 4% less than for a similar, unaffected parcel; if the supply relationship showed 76% of the market's properties to constitute the affected stock, then the expected discount would be about 8%, as seen earlier.) Some among this clientele (those who would set the needed discount at 0% or less, thereby showing a willingness even perhaps to pay a premium price) will be able to make purchases for less than they would have been willing, if necessary, to pay. The amount by which their willingness to pay exceeds the market price is commonly referred to in the economics literature as consumer surplus.

Conclusion

Contingent valuation analysis, based on surveys of market participants, has gained some acceptance in recent years as a means of justifying value estimates when direct market evidence is lacking.18 While the technique can be useful, a problem that can accompany asking current or potential owners their views on value-even when sufficient, reliable responses are obtained (an obviously difficult proposition in many cases)-is the resulting focus on the demand side of the market only. To develop a conclusion regarding price impacts (e.g., adjustment factors) associated with a given property attribute, it is necessary to analyze both the demand and supply sides of the market. Those who have attempted to use contingent valuation without addressing supply issues in the market have clearly been working with a technique they do not understand.

The analysis discusses how the actions of a clientele that contains the marginal buyer will determine the price paid in a market transaction; the existence of numerous vocal critics does not assure that properties with a particular feature (e.g., high voltage power lines) would sell for less than similar parcels unaffected by the perceived disamenity. Simple examples then show how an analyst can use the results of a survey (demand side), along with knowledge of the stock of various types of properties in the relevant market area (supply side), as a potentially useful tool for estimating value in the presence of a specified adverse-or desirable-property condition.

[Author Affiliation]

* University of Illinois, Champaign, IL 61820 or pcolwell@uiuc.edu.

** Illinois State University, Normal, IL 61790 orjwtrefz@ilstu.edu.