Essay: Why is Gentrification a Problem?, Essays (university) of Economics

Download Essay: Why is Gentrification a Problem? and more Essays (university) Economics in PDF only on Docsity!

Why is Gentrification a Problem?

S.Sheppard

Professor of Economics

Williams College

1. Introduction

Social and political concerns with gentrification have waxed and waned since the term was first coined in 1964 to describe the movement of middle class families into the former working-class neighborhoods of London. Since the term “gentrification” was first used, the phenomenon has been a source of debate for both scholars and policy makers in the US, Europe and elsewhere. Some authors have viewed it as a beneficial (or at worst neutral) undoing of the “white flight” abandonment of central city neighborhoods that took place in many cities during the period from the mid-1940s through the late 1960s. Perhaps this gentrification would return some wealth, tax base and a modicum of affluence to urban neighborhoods that had been hard hit by loss of businesses, jobs and tax-payers.

Alternatively, gentrification has been viewed (at a minimum) as an unfortunate desecration of interesting and “authentic” urban neighborhoods, a dilution of vibrant ethnic neighborhoods into something that is bland and uninteresting. At worst, the critics of gentrification have viewed the phenomenon as a major source of disadvantage for low income urban residents who, having established a community with all of its complex social networks must now see it torn apart as they are displaced – either by choice or compulsion – to move to other housing that is less desirable or alternatively remain behind to pay higher rents in a neighborhood they no longer feel is their own.

One other perspective deserves separate mention. This is that gentrification may or may not be unfortunate for the original or displaced residents, but that it is a “natural” or even “organic” part of urban development. Thus Brueckner and Rosenthal (2009) see gentrification as a natural consequence of the process of ageing with a durable housing stock, and present a model that has gentrification as a predicted outcome that can be expected to eventually take place in all cities. A related perspective might accept that gentrification has adverse consequences, but that policies designed to prevent any gentrification would be worse. Such anti-gentrification policies might encourage an urban environment in which economic classes or ethnic subgroups have particular neighborhoods to which they are entitled; and where one ethnic group is entitled others may be excluded. From this, it is feared, it is a short step to say that these are the neighborhoods to which they should be restricted.

Recent history presents a variety of perspectives about who constitutes the “gentrifiers” and the “displaced”. In 1983, for example^1 , a proposal by then New York City mayor Ed Koch to build 117 apartments for artists in the lower east side of Manhattan was defeated after an acrimonious hearing by the city Board of Estimate. One opponent called the proposal “a scam … that would gentrify a neighborhood with the young, the white and the rich.” A supporter, a gallery owner in SoHo, defended the plan to use federal housing funds to build the units saying that “… artists, by their nature, are an integrated race of people.” Almost three decades later, artists living in the area have been mostly pushed out of the neighborhood, and complain about being displaced by gentrification.

Much of the research that has been done concerning gentrification has focused on whether gentrification imposes particular harm on poor households, and whether these households are displaced into worse housing situations. Thus Schill and Nathan (1983) conducted surveys of

(^1) See Carroll (1983)

2. External costs of gentrification

In order to better understand the potential social cost of community instability, consider Figure 1 below. This diagram is constructed to illustrate in simplified form the relationships between efforts to improve neighborhoods and communities (community improvement) and associated dollar value to represent the costs and benefits of these actions.

$ Figure 1

In Figure 1, the upward sloping line labeled “Marginal Social Cost of Action” represents the cost to the community of community improvement actions. These actions require resources (even if they are donated or volunteered) and those resources could be used for other purposes. As more resources are applied to community improvement they become more difficult to find, recruit or purchase so the relationship slopes upwards.

There are two downward sloping lines in Figure 1, one labeled “Marginal Social Benefit of Action” and the other “Marginal Benefit of Action with Gentrification”. The Marginal Social Benefit line represents the value to the community of undertaking community improvement actions. It is the sum of the benefit experienced by all community members, over a lifetime of living in the community, of the community improvement actions. It is downward sloping under the assumption that the community undertakes the highest priorities in the community first, generating the highest value benefits, then the next highest, and so on. As long as the benefit of a community improvement action exceeds the cost to the community of the resources used in the action, it is desirable to undertake the action. The ideal situation for the community is to engage

Marginal Social Benefit of Action

Marginal Social Cost of Action

Marginal Benefit of Action with Gentrification

Area =Social Loss from Reduced Action

q 0 q 1 Community improvement actions

in q 1 community improvement actions, undertaking all those community improvement projects that satisfy this “cost-benefit” test.

If the community is well-organized (or perhaps we should say “perfectly” organized) then it will have devised some institutions and methods to support and encourage its citizens to undertake these community improvement actions. It will identify all those persons in the community who stand to benefit from the community improvements and convince them to contribute their own resources, time and efforts towards these actions in an amount that equals their individual marginal benefit of the actions experienced over a lifetime in the community. Even if the community is thus successful in overcoming the “free rider” problem (in which some members of the community do not contribute because they hope to benefit from the efforts and expenditures of others) a problem may arise if many residents are at risk of displacement. Suppose that each private resident believes that there is a 50% chance that he or she will be compelled to leave the community because gentrification forces rents to unaffordable (or unattractive) levels or for other reasons. In such a situation the expected value of the benefits of community improvement actions will be significantly reduced to persons who are at risk of displacement. As a result they will value their own benefits to be received from community improvement at a reduced level, indicated by the Marginal Benefit of Action with Gentrification line.

When a community is subject to gentrification, its residents may value community improvement at less than the true social value. As a result, even if they are persuaded to contribute the full value to them of community improvement, they will only view actions up to amount q 0 as satisfying the cost-benefit test. This is less than the socially efficient amount of community improvement which is represented by q 1. The Marginal Social Benefit of Action represents the “true” social benefit because, even though some existing residents may be forced or induced to leave the neighborhood because of gentrification, they will be replaced by new residents who arrive and will enjoy the benefits of the community improvements undertaken before they arrived. The Marginal Social Benefit of Action takes this benefit received by the “gentrifiers” into account. By undertaking only q 0 community improvement actions rather than the efficient amount q 1 , the neighborhood is losing out on the benefits that could be obtained by adding the q 1

• q 0 actions where Marginal Social Benefit exceeds the Marginal Social Cost. The amount of the loss is the shaded triangular area in Figure 1 labeled Social Loss from Reduced Action. This social loss is why gentrification is (or might be) a problem even without consideration of the distributional impacts of gentrification or the costs of moving. This social cost arises even if the poor are no more likely to be displaced from a gentrifying neighborhood than middle-class residents.

Traditional analysis of gentrification has tended to neglect this potential cost for one or more of three reasons:

 The possibility that higher risk of displacement would lead to undervaluation of community improvement did not occur to the analyst

 The possibility was recognized, but the analyst assumed that all or most of the community members were home owners, and that the value of community improvement benefits would be reflected in – “capitalized into” – the value of the homes. Community members might not continue to live in the neighborhood but if the actions were

despite gentrification that subsequently occurs. The gentrification may occur “organically” through gradual turnover of dwelling units, or through local additions to the housing stock.

Whether this incentive is effective in practice is an empirical question. If increasing the risk of displacement, or increasing the rate of turnover in the local housing market, has little impact on efforts towards community improvements, then gentrification may not generate the type of social cost illustrated in Figure 1 above (or the costs may be very small relative to other inefficiencies in the community).

If, on the other hand, increases in the probability of displacement are associated with significant reductions in efforts devoted to community improvement, then it will be a signal that the problem illustrated in Figure 1 may be working to impose costs on the communities through under-provision of community improvement actions. This would be a cost borne by the neighborhoods and communities affected. It is a real loss in the sense that these neighborhoods are less attractive than they would otherwise be, and the costs required to make them attractive would be less than the value of the community improvement.

Note that if this loss arises it is NOT borne exclusively by households with incomes below the poverty line. It affects all residents of the neighborhood and indeed may affect many neighborhoods that are not currently undergoing gentrification. It would be a cost borne by the entire community or neighborhood, and not only those who are actually displaced or most likely to be displaced.

Where security of tenure is limited and residents understand or believe that they might be put into a situation where they are forced to leave or find it unattractive to remain in the neighborhood because of rent increases, they will have a reduced valuation for the benefits that might be obtained from community improvement. If this reduced valuation is not compensated for by contributions from the owners of the properties where they live, it reduces the amount of community improvement and makes the neighborhoods less attractive. If this reduced valuation is not compensated for by contributions from the owners of the properties where they live, it reduces the amount of community improvement and makes the neighborhoods less attractive than they otherwise would be – in fact less attractive than they SHOULD be. This reduced attractiveness is a cost borne at the neighborhood or community level by all who live there.

How can we know if this is likely to be a problem? As noted above, observing that increasing the risk of displacement is associated with decreasing amounts of community improvement actions would be consistent with the hypothesis that social costs of the type illustrated in Figure 1 would be present in communities subject to gentrification. What data are available to us to measure the level of community improvement actions and the risk of housing market turnover? What data are available to us to correct for other factors that might also influence the observed level of community improvement actions? We turn attention to these questions in the next two sections.

3. Measuring community improvement actions

In order to test the hypothesis that increasing turnover or risk of displacement in the housing market is associated with different levels of community improvement actions, we must identify a source of data that is widely available for US communities and provides a plausible measure of such actions. Since “community improvement actions” can include everything ranging from informally organized neighborhood cleanup crews up to large community development organizations and public agencies with budgets in the millions of dollars, finding systematic and reasonably accurate measurements of these activities is likely to be a problem.

Many of these actions take place without the benefit of formal organizations or budgets. Some are undertaken by commercial enterprises working alone (the local merchant who underwrites the cost of new benches or new playground equipment for the park) or in concert (the local chamber of commerce that organizes efforts to improve conditions in an urban plaza). Many are undertaken through the efforts of the public sector through provision of public services in the form of parks and recreation, or efforts organized via local public schools. Each of these poses practical problems as an indicator of the level of such actions. Informal groups are not monitored and their efforts are infrequently reported in the press. Solo or collective efforts of commercial enterprises may be significant but again there is no formal and separate reporting of such efforts. Public agencies or schools are generally required to make public reports of their expenditures, but typically they do not break out the functions of such expenditures in a way that would permit measurement of the expenditures or per-capita expenditures devoted to community improvement.

Many community improvement actions are undertaken by not-for-profit organizations. These will include a wide range of groups including churches and other faith-related organizations, not- for-profit educational organizations (primary, secondary and post-secondary institutions), arts organizations, environmental organizations, clubs, and organizations created specifically for the purpose of neighborhood improvement and community development. These non-profit organizations are of potential interest because with the exception of churches, those organizations with annual budgets exceeding $25,000 are required to submit reports that include total expenditures and total revenues. Data from these annual reports are public records and are available to researchers in computer-usable form beginning in 1988. The data require time for processing so that the most recent data are generally about 2-4 years prior to the current year (there is some variance because different organizations have different fiscal years for reporting and organizations are able to petition for additional time to complete their reporting obligations).

Some of these organizations pose measurement problems that are similar to those encountered with public agencies. The reported budgets are not presented in detail and such details as are available do not always permit determination of the share of expenditures that have been devoted to community improvement. An alternative approach would be to identify those not-for-profit organizations whose mission and core activities are focused on undertaking actions that will improve a specific neighborhood or community. This is the approach that is used for this study.

The Internal Revenue Service must certify any organization that applies for not-for-profit status as being appropriately dedicated to pursuing core activities that are consistent with the law that allows them to be exempt from taxation. In making this determination the IRS assigns each organization a code that places them within the National Taxonomy of Exempt Enterprises

Table 1: Types of organizations used in analysis

NTEE Classification Description of Activity S20 - Community & Neighborhood Development

Organizations that focus broadly on strengthening, unifying and building the economic, cultural, educational and social services of an urban community or neighborhood. Use this code for community and neighborhood improvement organizations other than those specified below.

S21 - Community Coalitions Organizations that are designed to increase citizen participation in local policy issues and thereby improve the overall quality of life in a particular state or community.

S22 - Neighborhood & Block Associations

Organizations whose members are residents of a particular community or neighborhood who have joined together to remedy deficiencies in existing neighborhood conditions or to enhance conditions that are currently satisfactory.

These three categories contain a very wide variety of organizations. In the year 2000 there were over 4100 organizations around the US that were in one of these three categories, and were located within the formal definitions of a Metropolitan Statistical Area (MSA) and had a budget of at least $25000. To get some idea of the type of organizations represented in these categories, consider some specific examples.

Neighborhood Development Corporation of Des Moines

Founded in 1999, the Neighborhood Development Corporation of Des Moines, Iowa is a modest organization with activities focused on neighborhoods in Des Moines. The NDC has been working closely with the City of Des Moines with a particular focus to improve and revitalize specific neighborhoods in the city, particularly the East Grand, 6th^ Avenue Corridor and the Forest Avenue Corridor.

Figure 1: Forest Avenue Corridor in Des Moines, Iowa

Along these streets the NDC has purchased several properties, and is investing to repair and improve the properties to create a “friendly environment through mixed-use residential and commercial development.” The neighborhoods where NDC has been active are generally low to moderate income areas that contain structures that are usable but in need of investment. The NDC describes itself as a “… community-focused organization that revitalizes distressed neighborhoods and encourages neighborhood sustainability …. by offering commercial and residential options through building rehabilitation, new construction, and in-fill development.”

Neighborhood Development Center, St. Paul, Minnesota

The Neighborhood Development Center was founded in 1993 and describes itself as an organization that “… works in the low-income communities of Saint Paul, Minneapolis and surrounding suburbs … to help emerging entrepreneurs develop successful businesses that serve their communities, and to help build stronger neighborhood economies.” They collaborate with other partners in the community to offer a 16 week entrepreneur training course in 20 different ethnic communities and neighborhoods. Since their founding they have provided training to thousands of entrepreneurs, provided financing, consulting and

capacity-building for small businesses, and operated several business incubators in targeted inner-city neighborhoods.

Figure 2: Plaza Verde, an NDC incubator

relatively standard community development corporations, through training and business education and consulting services, provider of educational services and community resources, to organizers of neighborhood cleanup and small local enterprises. What they all have in common is a focus on mobilizing resources to improve, in one way or another, economic and social conditions in specific neighborhoods and communities within the cities where they operate. The analysis below will use the sum of expenditures by all such organizations within an urban area or, the total number of such organizations, per 1000 residents as a measure of the level of private community improvement actions within the city.

4. Data for analysis

Across the US there are some very large organizations of the types discussed in the previous section, but the majority of them are of modest size with budgets much less than $1 million per year. The MSA or Census-defined Metropolitan Statistical Area will serve as our unit of observation for analysis. The data are aggregated across all organizations within a given MSA and the total expenditures per 1000 thousand residents and total number of organizations per 1000 residents is used as the central measure of the level of community improvement actions in each city. Just as the distribution of organization expenditure is skewed, the distribution of cities by expenditures per thousand persons is highly skewed, with most cities having modest levels of a few thousand dollars, but a few cities having very high levels. The distribution is illustrated below in Figure 5.

Figure 5: Distribution of MSAs by expenditures of neighborhood organizations per 1000 persons

The distribution of MSAs according to the number of neighborhood improvement organizations is also skewed, although not as much as the distribution by the expenditures per 1000 persons shown in Figure 5. Figure 6 below shows the distribution according to the number of such organizations.

The number of such organizations in an MSA averaged 6 in 1990 but increased to 15 by 2000. This is consistent with a nationwide trend of increasing numbers of not-for-profit organizations as well as increasing numbers of such organizations whose budgets come to exceed the filing threshold due to inflation and the fact that the threshold itself is not indexed for changes in the price level. The inflation adjusted expenditures per thousand residents in the MSA also increased from 1990 to 2000, growing at an inflation-adjusted annual rate of 8.25%. This compares to

then new residents will be able to move in without necessarily causing an existing resident to move and fail to realize the expected benefit of community improvement. Thus the social cost identified above arises only if there is displacement, but in contrast to the arguments put forward in Vigdor (2002) it is not necessary that this displacement be involuntary. What is required is that it must be in some sense anticipated.

A resident who seeks to form an expectation about the probability that he or she is likely to have only a relatively brief period of residency in the neighborhood will have several potential signals to observe. Perhaps the most readily observed is simply the amount of turnover in the local housing market. Urban areas that have ongoing processes of gentrification will have higher proportions of dwellings occupied by persons who have moved to those dwellings recently. Perhaps to distinguish gentrification from a general process of growth of the metropolitan area, the resident might consider the share of dwellings that are occupied by someone who has recently moved from somewhere else in the same MSA. Fortunately, both measures are available to us as part of the Census data. Figure 7 below shows the distribution of MSAs by the rate of housing market turnover. Figure 8 shows the distribution by rates of “local” moves.

Figure 7: Distribution of MSAs by proportion of population moved

We use both the share of population that have moved from any other location within the past 5 years and the share who have moved “locally” – that is moved from another location within the same MSA – within the past 5 years. The hypothesis we wish to examine is whether the level of housing market turnover appears to have a statistically significant negative relationship with the level of community benefit actions.

Figure 8: Distribution of MSAs by proportion of “local” movers

Of course the level of community benefit actions are also affected by other factors. They may be affected by the level of need in the community, and also by the level of income and other resources available to deal with these needs. Finally, as discussed in section 2 above, they may be affected by the nature of housing tenure in the city, with higher rates of renter-occupation putting more households at risk of displacement. Because the displacement can be as much of a problem even if the departures are voluntary, it is not clear a priori how much of an impact we should expect renter-occupation to have.

To take into account these additional factors, we controlled for the poverty rate (to reflect the level of need), the average household income (to reflect the available resources), and the share of the housing stock that is renter-occupied. We obtained these data for 1990 and 2000 using Census figures for each MSA, and matched them with the expenditure and number of organizations data. As mentioned above, all dollar values were adjusted to 2000 price levels.

Table 4: Descriptive statistics for in 2000

Variable Mean

Std. Dev. Min Max Obs Expenditures per thousand 12807 20454 0 132852 275 Organizations per thousand 0.018 0.012 0.000 0.083 275 Pct moved 0.474 0.066 0.286 0.679 275 Pct moved locally 0.256 0.034 0.141 0.349 275 Ave household income 51027 7309 35591 83525 275 Pct renters 0.302 0.058 0.131 0.509 275 Poverty rate 0.129 0.044 0.052 0.359 275 Number of organizations 15 40 0 457 275

Tables 2, 3 and 4 above present descriptive statistics for each of these variables for the entire sample, for 1990 and for 2000 respectively. It is worth looking at these values – at least for the entire sample – to form a sense of what might constitute a “large” change in values of the variable.

One convention is to regard a one standard deviation as a reasonably important change, and a two standard deviation change as large. By these terms a large change in the rate of neighborhood stability would be to move from the sample mean value of 0.48 to a rate below 0.33 or above 0.62. Both of these values are within the range of observed values in the sample. A large change in the rate of neighborhood stability would be a 30% increase or decrease in the rate of recent movers. For local movers, a 27% increase or decrease would be a large change. For the outcome of expenditures or numbers of organizations per 1000 persons, a large increase would be a 378% increase in expenditures or a 184% increase in numbers of organizations. A large decrease would be a decline to zero. Once again both ranges are within the range of observed data for US cities.

We now turn attention to estimating the relationships.

5. Estimated impact

To test the hypothesis that displacement risk could be associated with significant reduction in community improvement actions, we estimate four different models that consider the two possible measures of the level of community benefit actions (expenditures per thousand and number of organizations per thousand) and the two possible measures of displacement risk (percent of the population above age 5 who have moved within the past 5 years, and the percent who have moved within the past 5 years from another location within the MSA). To facilitate comparisons between the estimates, we estimate the relationships in “elasticity” form: ( ) ( ) ( ) ( ) ( )

or

( ) ( ) ( ) ( ) ( )

By estimating a relationship that is linear in the logarithm of the variables, the parameter estimates can be interpreted as elasticities – meaning the percentage change in the expenditures or numbers of organizations generated by a one percent change in the variable of interest. Thus the estimate of will provide an estimate of the percentage impact of a one percent change in the measure of neighborhood stability. Table 5 presents the estimates, with the column header indicating which measure of community benefit actions is being used and the rows associated with the explanatory variables. Standard errors of the parameter estimates are in parentheses below the parameter estimates themselves, and the number of asterisks indicates a level of statistical significance, with *** signifying that the estimate Is significantly different from zero at the 1% level, ** indicating the 5% level, and * indicating the 10% level. In every case the standard errors of the estimates have been clustered by MSA and calculated to be robust to heteroscedasticity and model specification errors.