Portland Neighborhood Gentrification Patterns

QUANTITATIVE AND SPATIAL ANALYSIS TECHNIQUES FOR ANALYZING GENTRIFICATION PATTERNS Case Study: Portland, Oregon Author: Gina Clemmer, Independent Research Project, PSU Editor: Prof Gerard Mildner, Urban Studies & Planning, PSU September, 2000 For questions, comments or copies please contact Gina Clemmer at gclemmer@aol.com INTRODUCTION Current debates regarding gentrification and inner city revitalization require better analysis in order to make more informed policy and funding decisions, as well as to engage in a more meaningful community dialogue Community discussions about gentrification are typically based on a subjective reality shaped by a combination of personality and unique experiences,1 and as a result, community dialogue is often fueled by emotion and misinformation Planners and policy makers interested in gentrification will find limited examples of quantitative approaches to analyzing gentrification phenomena.2 Previously, access to relevant and current numeric data at the neighborhood level was limited, and tools with which to synthesize and manipulate large data sets were inadequate However, with improved national data collection efforts and tremendous advances in information technology, neighborhood changes can be studied both more broadly and in more sophisticated ways.3 Three national data sets provide tract level information about key gentrification indicators, as well as several data sets specific to individual localities National data sets include the American Communities Survey (ACS), Home Mortgage Disclosure Act (HMDA) data and business directories In addition, major advances in geographic information systems (GIS) provide the necessary tools to conduct spatial analysis on neighborhood social and economic changes, as well as display neighborhood changes using a visual medium.4,5 Increasingly detailed national and local data sets combined with major technological advances, provide city policy planners and researchers with a powerful set of tools with which to analyze gentrification Spatial analysis of gentrification can aid policy makers by providing information about the intensity and magnitude of change in a given neighborhood, as well as allow for relatively easy comparative analysis across the region.6 This knowledge can be an effective tool in prioritizing how community funds are allocated.7 Moreover, a spatial analysis of neighborhood change can provide a good starting point from which to begin an objective community dialogue around the externalities of gentrification, as well as provide policy makers with a tool to evaluate the impact of housing and economic development programs This paper attempts to provide practical guidance on how information for key neighborhood change indicators may be attained, manipulated, analyzed and displayed Being instructive and empirical in nature, this paper does not promote a particular definition of gentrification nor does it take a position on the morality of gentrification This analysis relies upon available national data sets for information, utilizes GIS as the primary analytic tool and focuses on a case study of Portland, Oregon, a city experiencing many of the symptoms of gentrification in the 1990’s A CASE STUDY: PORTLAND, OREGON Portland, Oregon is often hailed as a model of excellence for efficient and intelligent planning Many believe that a key component of this success is derived from the adherence to an urban growth boundary (UGB) and other protective zoning The UGB, one of the first anti-sprawl stratagems in the nation, encompasses twenty-four cities and is just over 200,000 acres.9 Within Portland, there is certainly a flavorful debate around the merits and drawbacks of the UGB As housing prices soar, making homeownership unattainable for many, a heated debate has ensued over the unintended impacts of the UGB It has been observed that sharp increases in housing prices are a double-edged sword On one hand, rising housing prices is reason to celebrate in many blighted areas, but on the other hand, higher prices mean paying more money for fewer amenities.10 A critical analysis of gentrification in Portland, must recognize that the UGB constrains consumer housing choices, and in doing so forces housing prices up This phenomenon entices people to move to places that they might not otherwise move to if other options were available In Portland thus far there have been several emotional community meetings, a qualitative study commissioned in which residents of inner northeast Portland were interviewed on their feelings about their gentrifying neighborhoods, and intellectuals such as Professor john powell consulted and asked to give lectures on the ills of gentrification And, although Multnomah County was one of four counties in the country to have updated census type information, there has been little emphasis on implementing policy or programs based on an objective data informed analysis Planners and policymakers could benefit from knowing which areas are changing most rapidly (especially in relationship to each other) and in what ways particular neighborhoods are changing Five areas in Portland are commonly thought of as already gentrified (see Figure 1) Area 1, in northwest Portland, began a process of gentrification in the early 1970s and is considered the oldest gentrified area.11 Area 2, Lloyd Center, has undergone rapid changes catalyzed by the development of the Rose Garden stadium and the Lloyd Center mall Area 3, loosely referred to as “inner northeast” located in the Sabin neighborhood, is in an earlier stage of gentrification that the other locations Area 4, known as “Belmont” and located in the Sunnyside neighborhood, is in the middle stages of gentrification and area 5, known as “Hawthorne” is also considered to be in late stage gentrification FIGURE 1: GENTRIFIED AREAS OF MULTNOMAH COUNTY (SELECTED TRACTS), 2000 P al dr the Ca Sunderland Piedmont Woodlawn Arbor Lodge Concordia d In tr us ial est rth w No e kro s Irvington Grant Pk  Lloyd Sullivan's Gulch Madison South Parkrose Heights Montavilla Downtown Goose Hollow Center Buckman  Wilkes City Line Pearl Argay Roseway Rose City Pk Eliot  Alameda Cully Hazelwood Glenfair North Gresham t es k rloo Ove w rth Vernon Humboldt King Boise Sabin Par No Fairview Troutdale Wood Village Rockwood North Central Mt Tabor Mill Park  Hosford-Abernethy Richmond S Tabor Centennial Northeast Centennial East Northwest Powellhurst-Gilbert Powell Valley Asert Reed I Corbett- Forest Park Bmnt-Wlsh ark Kenton Maplewood Multnomah Sellwood-Moreland Hillsdale Woodstock Kelly Creek Southwest Mt Scott Lents Eastmoreland Pleasant Valley Gresham Butte Mt Hood Note: The boundaries reflected here are city designated neighborhood boundaries for Portland and Gresham and city boundaries for Troutdale, Wood Village and Fairview However, the data conforms to tract boundaries indicated by shading where appropriate Tract boundaries not align with neighborhood boundaries Information is presented this way for ease of reference INDICATORS AND DATA SOURCES Multnomah County, Oregon where Portland is located, is one of four ACS test sites Because of this, 1996 tract level information is available In the literature, a wide range of variables characterize gentrification Some variables are related to gentrifiers or changing demographics of an area, and others are related to physical attributes of the area Some noted socio-economic characteristics of gentrifiers includes household income, household type, race, professional status and education level Some noted physical variables describe changes in the neighborhood such as upgrading of the housing stock, commercial enterprise changes, and in the case of Portland, transportation improvements like the addition of bike lanes and enhanced bus and rail service In addition to certain types of variables, other characteristics of gentrification are the rapidity and extremity of neighborhood change Therefore the calculation of existing facts is not the only issue, but the rapidity with which these changes take place, and the comparative extremes that exist within individual neighborhoods is also important information A American Community Survey The ACS is an annual survey for which the Census Bureau plans to replace the long form in the 2010 Census.12 Planners and policy makers can gain much insight into changes in their city by beginning to, or preparing to begin to, utilize the survey information.13 The ACS asks basically the same questions as the decennial census, but as an annual continuous sample survey (meant to provide accurate estimates of demographic, housing, social, and economic characteristics every year) for all localities of 65,000 people or more The sample size for the ACS will be 15% of the population for each county, every year Averages will be calculated, and updates of tract level data will be provided annually on CD-ROM and the internet Information can be downloaded into a database file and imported into any GIS software for analysis and display The collection and dissemination of such massive data will be extremely useful in studying neighborhood level changes The Census Bureau estimates that tract level information will be available for every population group over 65,000 by summer of 2008.14,15 Currently, 1996 tract level test data are available for Multnomah County, Oregon, Rockland County, New York, Brevard County, Florida, and Fulton County, Pennsylvania ACS Key Gentrification Indicators ACS indicators included in this analysis are: 1990-96 changes in median income, 1990-96 changes in percentage of professionals, 1990-96 changes in percentage of adult population with a four year degree, 1990-96 changes in median property value, 1990-96 changes in gross rent and 1991-96 changes in residential mobility A description of the techniques used to derive the information and key findings is presented for each variable Figure demonstrates how individual indicator information can be displayed Figure identifies tracts with higher than average change levels for all six indicators Variable 1: Changes in Income Sharp increases in income can be a key indicator of gentrification Analysts may be tempted to simply map the changes in median household income from one year to the next, however this does not account for inflation or region-wide income changes If done this way, tracts that are getting relatively poorer will actually have positive increases in income.16 Analysts can either adjust incomes according to the Consumer Price Index (CPI) or standardize incomes by calculating a ratio of tract income to county income.17 The latter technique is employed in this analysis The median income of the tract is divided by the median income of the county and multiplied by 100 This results in a ratio (or percent of regional income) that represents the relative financial strength of a tract The higher the ratio, the wealthier (in terms of income) the tract is, compared to the county as a whole Although this is an interesting enough fact, what is really useful in terms of gentrification analysis is to look at the rate of change 18, that is, the increase or decrease of the percent of regional income over time Combined with GIS this technique provides an easy visual comparative analysis The darker areas in Figure for the central eastside, the northside and the westside of Portland indicate that these experienced more rapid income gains compared to the rest of the county FIGURE 2: PERCENTAGE CHANGE IN THE RATIO OF TRACT MEDIAN INCOME TO COUNTY MEDIAN INCOME, MULTNOMAH COUNTY (SELECTED TRACTS), OREGON, 1990-1996 dr the Ca Change in Ratio of Tract Median Income to County Median Income 1990-96 P al Sunderland Piedmont Woodlawn Arbor Lodge Concordia In al tri k s du we st No rth os e Irvington Grant Pk  Lloyd Sullivan's Gulch Downtown Goose Hollow  Wilkes Madison South Parkrose Heights Montavilla Center Buckman Argay Roseway Rose City Pk Eliot  Alameda Cully 20% 0% -5% -10% -27% North Gresham t es rloo Ove w rth Vernon Humboldt King Boise Sabin Pa rkr No Forest Park Bmnt-Wlsh ark Kenton Avg: 2.6% Fairview Troutdale Wood Village Hazelwood Glenfair Rockwood North Central Mt Tabor Mill Park  Hosford-Abernethy Richmond S Tabor to 45% to 20% to 0% to -5% to -10% Centennial Northeast Centennial East Northwest Powellhurst-Gilbert Powell Valley Reed I Corbett- Asert Maplewood Multnomah Sellwood-Moreland Hillsdale Woodstock Kelly Creek Southwest Mt Scott Lents Eastmoreland Gresham Butte Mt Hood Pleasant Valley Source: Census of Population and Housing, 1990, STF3A American Community Survey, 1996 Notes: (a) The boundaries reflected here are city designated neighborhood boundaries for Portland and Gresham and city boundaries for Troutdale, Wood Village and Fairview However, the data conforms to tract boundaries indicated by shading where appropriate Tract boundaries not align with neighborhood boundaries Information is presented this way for ease of reference (b)The numbered black dots are areas already gentrified as explained in Figure Variable 2: Changes in Residential Mobility In the literature, one of the key elements of gentrification is the mobility of people The ACS provides two measures of mobility: mobility rate and occupancy duration The first way identifies the percentage of people (over five years old) who have moved to an area within the past five years The survey is taken mid year therefore a five year interval is from mid-1991 to mid-1996 The “mobility rate” measures and it gives an idea of neighborhood change in terms of population turnover Occupancy duration is measured by when the moved into the current unit The information is provided for renters and owners separately, allowing for a more detailed analysis As will be discussed later, HMDA data can provide a more detailed and current picture of where homeowners are moving, and is therefore preferable Mobility analysis can provide a broad understanding about where people are moving to, and to some extent where they came from (ie within the metropolitan statistical area (MSA), the state, out of state or out of the country), however, there are some well documented problems associated with residential mobility analysis The three primary confounding variables are the percentage of renters, the dominant age cohort of the tract and household education levels If a tract has a high renter population (i.e university neighborhoods) these tracts will have higher percentage of movers, as renters are more likely than homeowners to move Additionally, younger people tend to move more frequently than older people.19 Finally, very low or very high educational attainment levels also correlate with mobility rates Even with these constraints, a basic review of population flows provides insight into the movement of population, which may affect public and private planning and policy.20 Mobility patterns in Portland show that on average, about half the people living within any given neighborhood have located there within the five years prior to 1996 (“movers”) Out of all movers, 64% moved from one tract to another within the MSA (“local movers”) About 30% of movers were not from this area, and a little over 6% were from outside the United States The county’s highest mobility rates reach 90% to 75% and occurred in the area north of downtown (the Pearl District) as well as in the downtown core.21 The second highest mobility rates were in the central eastside, the area south of downtown (the Corbett-Twilliger neighborhood), parts of outer southeast Portland Third highest were in inner southeast Portland, north Gresham, and the cities of Fairview and Troutdale In addition, mobility data from the ACS measures people that relocated from within the MSA, from within the state, outside of the state and outside of the county In Multnomah County, local movers tended to move to inner northeast Portland, east Portland and eastern part of the county while out of state movers tended to settle in northwest, southwest, and in inner southeast Portland Remarkably, the settlement patterns for local movers and non-local movers were almost in the exact opposite neighborhoods, with the exception of some neighborhoods in inner northeast High residential mobility alone does not indicate that a neighborhood is gentrifying As noted above, a high renter population and a dominant younger age cohort will also show high levels of mobility Additionally, new construction of units or the location of a major employer will affect mobility In isolation, this variable alone cannot indicate a gentrifying neighborhood, but it does provide analyst and communities with a general comparative analysis of which areas are experiencing the transition of newcomers It also can flag neighborhoods to watch or ones that may require more in depth analysis Variable 3: Changes in Occupational Status In The New Middle Class and the Remaking of the Central City, David Ley chose “professional” occupational status as a key indicator for gentrifying areas primarily because it represents both income and prestige associated with the top of the employment hierarchy.22 In this analysis, professional includes jobs in the managerial, technical and sales fields 23 Between 1990 and 1996, Multnomah County experienced an average of 9% increase of professionals Areas that attracted professionals most rapidly (50%-75% increase) were areas in inner northeast Portland, particularly in the neighborhoods of Humboldt, Boise and Eliot The Pearl District also experienced high increases and somewhat surprisingly portions of outer southeast and north Portland in Cathedral Park One notable difference is that the King area is not one of the attractive areas to professionals The north King area increased by 8.9%, however the south King area declined by 5.50% Variable 4: Changes in Educational Attainment Several authors have highlighted changes in educational attainment levels as a key indicator for gentrification activity In order to assess changes in educational attainment, the percentage of the adult population (over 25 years old) that holds a four year college degree must be derived The median for the region was a 19.09% (the mean was 27.6%) increase in the adult population with a four year degree In 1999, Portland was ranked among the top five of the nation’s fifty largest metro areas for educational attainment.24 When this technique is employed some interesting results surfaced Outer southeast, far north in Cathedral Park and Hillsdale in southwest Portland experienced some of the largest gains in educated people Cathederal Park and outer southeast also attracted a high level of professionals, but not necessarily high incomes and only modest increases in newcomers (although the newcomers that did move there were overwhelmingly local) Variable 5: Changes in Property Values Property value changes use the tract to region comparison The median home value for the tract is divided by the county median home value and a ratio is obtain for both 1990 and 1996 Positive increases occurred in the central eastside, inner northeast, inner southeast and bits of the Corbett-Twilliger neighborhood on the inner westside East Portland, the eastern part of the county and southwest Portland had declining property values as compared to the county as a whole This is not to say that these areas have low property values in 1996, in fact southwest Portland had some of the highest property values according to the 1996 data Variable 6: Changes in Gross Rent Changes in rents can either be measured as changes in rents with utilities included (gross rent) or changes in just the rent (contract rent) ACS does not give contract rent by tract for the 1996 data, only gross rent Changes in rent are derived the same way as changes in property values When mapped, sharp increases in average gross rents as a percentage of average county rent appear in the Pearl District, in Irvington which is in inner northeast, Hosford-Abernathy in inner southeast, in some parts of southwest and in east Gresham An interesting comparison to where property values are rising can be made In areas where rents are rising but property values aren’t, it could be assumed that the rental property is being upgraded, or is already in better condition, than the houses.25 This phenomenon is occurring in Cathedral Park, Portsmouth, parts of southwest, East Moreland and east Gresham B Home Mortgage Disclosure Act Data One of the most controversial aspects of gentrification is a perception that as neighborhoods begin to gentrify, outsiders (particularly wealthier, whites) will begin to speculate on the housing market, eventually moving in and taking over a neighborhood Community conversation is often dominated by the fear that current residents will be involuntarily displaced by rising property taxes and rents, or that residents will be encouraged to move through police harassment and predatory lending practices in order to open the housing market to new investors HMDA data is one way to check community assumptions and begin to quantify housing market transitions HMDA variables provide not only more current information than the ACS, but different sorts of information It provides the number of loan applicants and borrowers, their racial composition, and the average loan and income values of borrowers on an annual basis It also provides the reasons why people were denied loans, and whether the loan was a conventional, FHA, VA or FmHA Although these variables are not mirrored in ACS data, they can be used as comparative variables to gain a rough understanding of what change might be taking place, or to confirm other research HMDA data is available from the Federal Financial Institution Examination Council (FFIEC) on CD-ROM beginning with 1992 information.26 HMDA Key Gentrification Indicators Variable 7: Changes in Middle/Upper Income White Home Purchases Race and income are often cited as key indicators for gentrification shifts HMDA asks for the loan applicant’s race, income and the tract location of the perspective new home In this analysis, one subset of the data was analyzed: 1996 and 1999 middle and upper income27 white homebuyers Only households that actually received a loan were included Between 1996 and 1999 neighborhoods that experienced the highest increases in percentage of middle/upper income white homebuyers were in the Portland neighborhoods: Northwest, Goose Hollow, Pearl, Sabin, Sunnyside, Buckman and Reed Many other neighborhoods experienced higher than average changes, particularly in the inner eastside, scattered neighborhoods in north and southeast Portland and large sections of Gresham Alternatively, there was disinvestment primarily in outer southeast Portland neighborhoods and pockets of east county All Indicators: A Spatial Query Analyzing individual indicators can provide useful information for specific public policy questions, however combining key gentrification indicators and change rates across a region, provides a much more useful interpretation of emerging gentrification patterns In order to determine where all of these variables come together in higher than average proportions, a spatial query of all seven variables was performed If the selected indicators are associated with gentrification as the literature suggest, the extent to which they overlap and the rapidity with which the overlap happens, must suggest that gentrification is occurring in various degrees in selected areas The mean (or median) for the rate of change was determined for each variable and then tracts where selected that had higher than average change rates for all of the variables and different combinations were tested The results are reflected in Figure FIGURE 3: GENTRIFYING AREAS, MULTNOMAH COUNTY (SELECTED TRACTS), 1996 Tier Gentrification: Shaded areas are tracts that have experienced higher than average increases in the following areas: household median income, perce th Ca ed l Sunderland Piedmont Woodlawn Arbor Lodge Concordia rk Pa Kenton Par kro se Irvington Grant Pk North Gresham st Rose City Pk Eliot Bmnt-Wlsh ial s tr du In k rloo Ove st we rth No Tier Gentrification: Shaded areas are tracts that have experienced higher than average Vernon Humboldt increases in the following areas: household median income, percentage of professionals, King Cully Forest Park percentage of educated adult population, property values, rent plus Argay utilities and home Boise Wilkes Sabin Roseway Alameda whites purchases by middle and upper  income Fairview Madison South Parkrose Heights No rth we Troutdale  Shaded Tier Gentrification: areas are tracts that have experienced higher than average increases Wood Village in the following areas: house  Sullivan's Gulch Lloyd Pearl Buckman Reed Maplewood Multnomah Sellwood-Moreland Hillsdale  Montavilla Hazelwood Glenfair Rockwood North Central Mt Tabor Mill Park  Hosford-Abernethy Richmond I Corbett- Downtown Goose Hollow Center S Tabor Centennial Northwest Powellhurst-Gilbert Woodstock Southwest Mt Scott Lents Eastmoreland Northeast Centennial East Pleasant Valley Powell Valley Tier Gentrification: Shaded areas a Asert average increases in the following ar Kelly Creek professionals, percentage of educated Gresham Butte Mt Hood utilities Tier Gentrification: Shaded areas a average increases in the following ar professionals and percentage of educ Source: Census of Population and Housing, 1990, STF3A American Community Survey, 1996 Home Mortgage Disclosure Act data, FFEIC, 1996 and 1999 Notes: (a)The boundaries reflected here are city designated neighborhood boundaries for Portland and Gresham and city boundaries for Troutdale, Wood Village and Fairview However, the data conforms to tract boundaries indicated by shading where appropriate Tract boundaries not align with neighborhood boundaries Information is presented this way for ease of reference (b)The numbered black dots are areas already gentrified as explained in Figure The “tier” structure in Figure is presented as a way to quantify degrees of gentrification, based on various combinations of variables Neighborhoods identified as “tier 1” exhibited higher than average change rates in all seven categories In the “tier 2” category, the mobility rate was taken out because having higher than average mobility increases as a criterion for selection was very constraining With this variable eliminated, three additional neighborhoods emerged as exhibiting six of the seven selected indicators These neighborhoods are experiencing a lesser degree of newcomers than tier neighborhoods, yet are experiencing positive increases in all other ways “Tier 3” neighborhoods exhibited higher than average change rates in four out of seven variables Changes in property values, rents and mobility were eliminated and in doing so three more neighborhoods experiencing signs of gentrification emerged In tiers and 5, selection criteria was relaxed even more which identified several scattered areas exhibiting subtle signs that gentrification processes are at work Tier areas include: the Pearl District, Old Town (many new apartments and condominiums have been built on the waterfront which is the western border of Old Town), Eliot, Lloyd Center, Overlook and the eastern part of Buckman These areas had proportionally higher change rates in all seven variables than any other areas in Multnomah County From a policy perspective, a clearly defined prioritization scheme emerges, with tier neighborhoods most rapidly and markedly gentrifying Tiers and might be considered to be on an upward climb toward full fledged gentrification, assuming gentrification patterns continue until a neighborhood is fully gentrified, the end result being the likes of NW 23rd avenue Tiers and are not experiencing higher than average newcomers and in the case of tier 3, property and rents have not risen as much as in tier and neighborhoods Portland neighborhoods included are: the western part of Hosforth-Abernathy, parts of Richmond and Sunnyside, Forest Park (many new condominiums and luxury apartments have been built in this area), Humbolt, parts of the Grant Park area and Beaumont-Wilshire, and the northern part of Rose City Tiers and include Portland neighborhoods: University Park, Cathedral Park, the southern part of St Johns, Portsmouth, parts of Arbor Lodge, Boise, Roseway, the south Rose City, Roseway, Kerns, Mt Tabor, southwestern Richmond, parts of Brentwood-Darlington, Lents, Foster-Powell and Troutdale An interesting note on Lents, Brentwood-Darlington, Foster-Powell and Troutdale is that these areas are not relatively close to the city center, which many authors have suggested is a key criterion of gentrification This may be attributed to the fact that housing prices are less expensive in these locations than in the central city From a policy perspective, these are areas that housing and economic development policy makers should begin to work with in terms of putting mechanisms in place to stem the tide of gentrification displacement C Business Directories Gentrification encompasses both housing and commercial sector changes Commercial changes often included the dissipation of “seedy” establishments and give way to boutiques, galleries, coffee shops, restaurants and pubs.28 Additionally, the creation of social gathering places produces increased social capital instilling new civic pride.29 In turn, civic pride may contribute to the growth of a gentrifying area 10 Business directories are primarily used by businesses to target consumers Several such directories are available including Cole’s Business Directory, Dunn and Bradstreet, ReferenceUSA and PhoneDisc Essentially, these directories are a form of computerized yellow pages on CD-ROM or via the Internet Usually, they are updated quarterly by SIC classification These resources are very useful in analyzing general business changes Business Directory Key Gentrification Indicators Increases and Decreases in different kinds of businesses One of the most visible signs of gentrifying neighborhoods is evidenced in changing consumption patterns Often a sprouting up of coffee shops, bookstores and pubs are signifiers that gentrification is on its way Typically gentrifying areas will experience increases in businesses such as: galleries and arts organizations, coffee shops, restaurants, cafes, pubs, book stores, flower shops, housewares and gift shops And simultaneously, areas will experience a decrease in businesses such as pawnshops, check cashing businesses and adult book and video stores In many cases this is true and business directory information can confirm suspected changing consumption patterns The commercial sector might be quantified to depict changes over time (as with all of the other variables) or simply as a point in time reference The latter technique is employed here In Figure 4, coffee shops, pubs and florist have been mapped by address (geocoded) This additional information gives more insight into the character of the gentrifying neighborhoods Many of the identified areas indeed show a clustering of these businesses What is also interesting to note is that some not, for example, Humboldt and Overlook in north Portland This demonstrates an important point Policymakers and local residents can not legitimately make assumptions on which areas are undergoing rapid change based on what is happening in the commercial sector As the data demonstrates, there are some areas that are undergoing changes associated with gentrification that not have the visual image of a gentrifying area This type of analysis could be modified to assess economic development needs and priorities for funding 11 FIGURE 4: GENTRIFYING AREAS WITH COFFEE SHOPS, PUBS AND FLORIST th Ca l ed Sunderland Piedmont Woodlawn Arbor Lodge Concordia Pearl ro s e Irvington Grant Pk  Lloyd Sullivan's Gulch Downtown Goose Hollow Center Buckman  Argay Roseway Wilkes North Gresham No rth w  Alameda Cully Rose City Pk Eliot est ial str du In Vernon Humboldt King Boise Sabin Par k t es k rloo Ove w rth No Forest Park Bmnt-Wlsh rk Pa Kenton Madison South Parkrose Heights Montavilla Fairview Troutdale Wood Village Hazelwood Glenfair Rockwood North Central Mt Tabor Mill Park  Hosford-Abernethy Richmond S Tabor Centennial Northeast Centennial East Northwest Powellhurst-Gilbert Powell Valley Reed I Corbett- Asert Maplewood Multnomah Sellwood-Moreland Hillsdale Woodstock Kelly Creek Southwest Mt Scott Lents Eastmoreland Gresham Butte Mt Hood Pleasant Valley Source: Census of Population and Housing, 1990, STF3A American Community Survey, 1996 Home Mortgage Disclosure Act data, FFEIC, 1996 and 1999 PhoneDisc, 1997 Notes: (a) The boundaries reflected here are neighborhood boundaries for Portland and Gresham and city boundaries for Troutdale, Wood Village and Fairview However, the data conforms to tract boundaries indicated by shading where appropriate Tract boundaries not align with neighborhood boundaries Information is presented this way for ease of reference (b)The numbered black dots are areas already gentrified as explained in Figure (c) This analysis is based on seven neighborhood change indicators See Figure for a detail explanation (d) Businesses were mapped for Portland only OTHER USEFUL DATA SOURCES AND METHODOLOGIES Thus far only information that can be derived from inexpensive, national data sets has been used, however many localities collect all sorts of data that can be used to supplement ACS, HMDA and business directory information In fact, the Urban Institute in conjunction with many cities across the nation, has been working on establishing “neighborhood indicators” using a wide variety of data sets in localities coast to coast.30 Some examples of other frequently used data sets are free lunch programs,31 residential mobility based on school records.32, tax assessor records, building permits, community asset directories, department of motor vehicles, geocoded crime records and high speed internet access connections Also, a note on the methodology is necessary There are other indicators that are compelling and different combinations of variables will yield different results Researchers and policymakers 12 may want to weight different variables according to community values and questions Additionally, localities may want to seek qualitative analysis to complement these statistical outcomes in order to validate and texturize research results However, this type of spatial analysis is a good starting point from which to begin community discussions and strategic plans CONCLUSION There is a wealth of information available to analyze changing neighborhoods and gentrification in particular Many communities will have access to ACS data over the next couple of years and all communities will have access to the 2000 Census data in the near future (the same techniques can be employed for Census data) HMDA data and business directories are available every year Moreover, combining these data sets with local quantitative or qualitative data can provide policymakers and citizens with a basic understanding of emerging gentrification phenomena In the case of Portland, the change rates of seven selected variables provided insight into key indicators typically associated with gentrification Moreover, a spatial query allowed for the identification of which tracts experienced higher than average change rates in all or some of the variables This provided a handful of tracts that exhibited varying degrees of the gentrification process By applying the techniques presented here, researches can begin to concretely identify changing neighborhoods This knowledge can be an effective tool in prioritizing how community funds are allocated It could also be used to establish community goals, identify community problems and solutions and create community specific programs 13 14 End Notes Greenburg, Michael R (1999) Improving Neighborhood Quality: A Hierarchy of Needs Housing Policy Debate 10 (3) p.605 Wyly, Elvin K, Hammel, Daniel J (1998a) Modeling the context and contingency of gentrification Journal of Urban Affairs, 20(3), p.304 Kingsley, G Thomas (1998) Neighborhood indicators: Taking advantage of the new potential Washington D.C.: The Urban Institute October, pp.2-4 Kingsley, G Thomas, Coulton, Claudia J., Barndt, Michael, Sawicki, David S and Tatian, Peter (1997) Mapping your community: Using Geographic Information to Strengthen Community Initiatives Washington D.C.: U.S Department of Housing and Urban Development October pp.17 Can, Ayse (1998) GIS and Spatial Analysis of Housing and Mortgage Markets Journal of Housing Research, 9(1), p 62 Ley, David (1996a) The New Middle Class and the Remaking of the Central City Oxford, NY: Oxford University Press, p 82 For example, if an area is showing signs of gentrifying is likely to attract private investment If private investment can replace public funds, policy makers may want to reprogram funds to areas that have a greater need, or they may want to make modifications to the types of programs offered in that area Phillips, Justin and Goodstein, Eban (1999) Growth Management and Housing Prices: The Case of Portland, Oregon Contemporary Economic Policy, p.334 Benfield, F Kaid, Raimi, Mathew D., Chen, Donald D.T (1999) Once There Were Greenfields: How Urban Sprawl is Undermining America’s Environment, Economy and Social Fabric Washington D.C.: Natural Resources Defense Council, p 153 10 Staley, Samuel R., Edgens, Defferson G and Mildner, Gerard (1999) A Line in the Land: Urbangrowth Boundaries, Smart Growth, and Housing Affordability Policy Study No 263 Reason Public Policy Institute, p.8 11 Oesterle, Sabrina (1994) An Empirical Assessment of the Gentrification Process in Northwest Portland, Oregon Portland, OR: Portland State University, Master Thesis, p 13 U.S Bureau of the Census (2000a) American Communities Survey Website: About the American Community Survey http://www.census.gov/acs/www/index_main.htm 12 Taeuber, Cynthia M., Lane, Julia and Stevens, David (2000) Meeting State and Community Needs: Social, Economic and Housing Information: The Why, What, and How of Converting Program Records and Summarized Survey Data to State and Community Information Systems Paper presented at a conference on “Developing Public Policy Applications with Summarized Survey Data and Community Administrative Records,” p 13 Data has been collected for 36 counties nation wide For these sites, tract level information will be available by summer of 2002 2000 Census information will be available shortly after and a comparative analysis will be conducted The 38 counties are Pima (AZ), Jefferson (AK), San Francisco and Tulare (CA), Brevard (FL), Upson (GA), Lake (IL), Miami (IN), Black Hawk (IA), De Soto Parish (LS), Calvert (MD), Hampden (MA), Madison (MS), Iron, Reynolds and Washington (MI), Flat Head and Lake (MT), Douglas (NB), Otero (NM), Bronx and Rockland (NY), Franklin (OH), Multnomah (OR), Fulton and Shuykill (PA), Sevier (TN), Fort Bend, Harris, Starr and Zapata, (TX), Petersberg (VA), Yakima (WA), Ohio (WV), Oneida and Viles (WS) (U.S Census, 2000b) 14 15 McGuinn, Larry, Senior Administrator Interview, July, 2000 U.S Bureau of the Census 16 Tract 11.01 in the Buckman neighborhood provides a good example of the usefulness of this technique In 1990, the median household income was $15,164 and in 1996 it was $16,432, representing an unadjusted percentage increase in income of 8.36% However, in actuality in 1990 the income of that tract was 56.3% of the county but in 1996 it was 50.2%, representing an actual decline of nearly 12% of percent of county income 17 Myers, Dowell (1992a) Analysis with Local Census Data: Portraits of Change San Diego, CA: Academic Press Inc., p 239 The rate of change is measured by the total at one point in time subtracted by the total of an earlier point in time and then divided by the total of the earlier point in time, multiplied by 100 18 19 Myers, Dowell (1992b) Analysis with Local Census Data: Portraits of Change San Diego, CA: Academic Press Inc., p 191 Bourne, Larry, Lay, David (Eds) (1993) The Changing Social Geography of Canadian Cities McGill-Queens University Press p 188 20 21 In the case of downtown there are many homeless and transitional housing units situated there, as well as Portland State University The transitional nature of these populations accounts for a large degree of the mobility in downtown however there has also been a substantial amount of condominiums, lofts and apartments built in the Pearl District and close to the waterfront Ley, David (1996a) The New Middle Class and the Remaking of the Central City Oxford, NY: Oxford University Press, p 82 22 The percentage change in professionals is derived by taking the number of professionals in the tract and dividing by the number of people in the labor force and multiplying by 100 23 24 Oregon Progress Board (1999) 1999 Benchmark Performance Report 25 Myers, Dowell (1992c) Analysis with Local Census Data: Portraits of Change San Diego, CA: Academic Press Inc., pp 140-41 26 1996 and 1999 data were used for this piece of the analysis due to the timeliness of the data Middle and upper incomes are based on percentages of the HUD Annual Median Family Income (HAMFI) figures Middle income is defined as 80% to 150% of HAMFI (1996: $35,52-$66,600, 1999: $41,921-$78,600) Upper income is defined as anything above 150% of HAMFI (1996:$66,601, 1999: $78,601 and above) 27 Henig, Jeffery (1982) Gentrification in Adams Morgan Washington D.C.: GW Washington Studies, p 48 28 Irwin, Michael and Sharkova, Irina V (1998) Social Capital of Local Communities Portland, OR: Metroscape, Spring Volume, p 13-19 29 Kingsley, G Thomas (1999) Building and Operating Neighborhood Indicator Systems: A Guidebook Washington D.C.: The Urban Institute March, p.1 30 31 Orfield, Myron (1997) Metropolitics: A Regional Agenda for Community Stability Washington D.C.: The Brookings Institute, p 48 Thompson, Maxine (1999) Student Mobility and Its Effects on Student Achievements: A preliminary Study Prepared for the Leaders Rountable Portland, OR: The Leaders Roundtable, pp.1-29 32 ... ways particular neighborhoods are changing Five areas in Portland are commonly thought of as already gentrified (see Figure 1) Area 1, in northwest Portland, began a process of gentrification. .. the Sabin neighborhood, is in an earlier stage of gentrification that the other locations Area 4, known as “Belmont” and located in the Sunnyside neighborhood, is in the middle stages of gentrification. .. northeast Portland, east Portland and eastern part of the county while out of state movers tended to settle in northwest, southwest, and in inner southeast Portland Remarkably, the settlement patterns