Data Dictionary NOTES 1. The underclass population as defined by Kasarda (1993) is the total number of persons having all five of the following characteristics: both householder and spouse (if present) did not complete high school; householder is either single, divorced, widowed, or separated, and householder's family includes persons under 18; both householder and spouse (if present) worked less than 26 weeks or usually worked less than 20 hours per week during the previous year; at least one member of the household received public assistance income in the previous year; and householder's income was below the poverty line in the previous year. As opposed to Ricketts and Sawhill's (1988) definitions, these figures are for persons in households with the specified characteristics, rather than all persons living in "underclass" census tracts. 2. Includes covered deposits in commercial and savings banks and branches. 3. PUMA boundaries do not match central city boundaries precisely in all cases. See Kasarda (1993), p. 3-41. 4. Actual fiscal health provides a measure of the central city's access to tax revenue compared to the services to which it is committed. The measure is defined as an index equal to zero for the average central city in 1972; deviations from this value are expressed as a percentage of the city's revenue-raising capacity. Values for Washington, D.C. reflect federal assistance through institutions and grants; 1972 index for Cincinnati may understate that city's fiscal health, due to certain biases in the way education responsibilities were defined. 5. Identical to actual fiscal health, except for the standardization assuming a single, market-basket of city fiscal institutions and service-providing responsibilities. Defined as an index equal to zero for the average central city in 1972; deviations from this value are expressed as a percentage of the city's revenue-raising capacity. Values for Washington, D.C. reflect federal assistance through institutions and grants. 6. The dissimilarity index provides a measure of the residential segregation between two different racial or ethnic groups. Given two populations A and B, the index is defined as: D = ( 0.5 Si | Ai - Bi | ) where Ai and Bi represent the respective proportion of members of A and B groups living in each census tract i. The index ranges from zero to 1 (although in the preliminary database 1980 and 1990 figures range from 0 to 100), with 1 (or 100) signifying perfect and complete segregation. The value of the index may be interpreted as the proportion of the minority population that would have to move in order to achieve perfect integration. 7. Wilger's index of dissimilarity ranges from 0 to 100. 8. Excludes federal government physicians. 9. Data derived from the REIS, which is reported only at the county level. For each central city, the county encompassing all or a majority of the city was selected for reporting purposes. Variables derived from Kasarda's database require adjustment for Los Angeles, Santa Ana, Long Beach, and Santa Ana. ST_PLAC8 is columns 351-357, 7.0, and is 69991 for Orange County, 69992 for Los Angeles County. See Kasarda, p. 2.3-7. 13. Carbon monoxide measured as the highest second (non-overlapping) eight-hour concentration. 14. Lead measured as the highest quarterly concentration. 15. Nitrogen Dioxide and Sulfur Dioxide are measured as the highest mean concentration. 16. Ozone is measured as the highest second daily maximum one-hour concentration. 17. Particulate matter (that exceeding 10 micrometers in diameter) is defined as the highest weighted annual mean concentration. 18. Daily particulate matter and sulfur dioxide measurements are defined as the highest second maximum 24-hour concentration. 19. Areal definitions do not correspond with MSA or central city boundaries in most cases. Single metropolitan-wide agency data are reported for several pairs of cities: Minneapolis/St. Paul; Los Angeles/Long Beach; and Boston/Manchester. For details on the areal definitions, see AMSA (1990). 20. Full-time workers only. 21. Operation and maintenance revenues; excludes capital and debt-related items. 22. Percentage of total flow treated at primary (best), secondary, and tertiary levels, respectively. These three items sum to 100 percent in all cases except for plants (or parts of facilities) in which level of service is not specified. 23. Percentage of total flow comprised of residential and industrial wastewater, respectively. Differences between the sum of these two items and 100 percent reflect the proportion of flow comprised of infiltration/inflow. 24. Total annual service cost per one-family residential user. 25. Excessive housing expenditure refers to total housing costs (contract rent for renters, specifed owner costs for owners) that equal or exceed 35 percent of gross household income. 26. Data refer to urbanized areas, and may not correspond to Metropolitan Statistical Area (MSA) definitions. Data for urbanized area used for both twin cities for Kansas City, KS/Kansas City, MO, San Francisco/Oakland, Minneapolis/St. Paul. Data for Norfolk substituted for Virginia Beach. 27. Vehicle kilometers of travel per lane kilometer of the specified roadway type; provides an estimate of traffic flow relative to capacity. 28. The roadway congestion index (RCI) compares existing traffic flow characteristics with specified threshold values identified with congested conditions. The index includes separate components for each roadway type (freeway and principal arterial). Index values exceeding 1.0 signify congested conditions areawide, while values below 1.0 indicate no systemic congestion. The RCI value is based on flow values aggregated across each urban area; thus the index is an overall measure, an urban area with an index less than 1.0 may have congested sections of road. 29. Represents the percentage of daily vehicle-kilometers of travel on each roadway type during the peak period operating on congested conditions. 32. The limiting factor in estimating gross metro product (GMP) is the availability of data on output. In the U.S., the smallest areal unit for officially released data on output is at the state level (gross state product or GSP). We derive GMP by assuming that sector-specific productivity levels are constant across all parts of each respective state -- i.e., that the GSP-to-earnings ratio for a state prevails within each of the state's metropolitan areas. We calculate GMP estimates by multiplying metropolitan earnings by sector-specific productivity levels at the state level and summing across sectors. Both GSP and earnings data are available at approximately the two-digit SIC level. We adjusted these datasets for comparability, yielding a total of 60 industrial sectors for which productivity estimates were calculated. Consequently, the GMP figures not only account for state differences in productivity, but also for differences in the industrial mix between the metro and non-metro portions of each state. 33. Area in acres devoted to agricultural, business/commercial, residential, and transportation uses, based on aggregations of prime-use categories in the U.S.D.A. database. These categories do not sum to the total Metropolitan acreage, the difference reflects land devoted to recreation, reservation and dedication, waste land, and unspecified uses. 34. Excludes undeveloped rural areas. 35. Excludes water-based transportation. 36. Total pounds of pollution released into the environment (air, land, water, underground) by facilities in 1987 Standard Industrial Classification codes 20 to 39 with 10 or more full-time employees. These variables exclude establishments releasing less than 50,000 pounds in each respective reporting year; these smaller releases comprise about 11 percent of the national total reported in 1992 (1.57 million tons). Data cover more than 300 chemicals; thus figures on total pollution conceal wide variations in type of pollution and degree of threat to health or environment. 37. Data refer to urbanized area. 38. Consist of Medicare payments, medical vendor payments, and CHAMPUS payments. 39. Includes general assistance, emergency assistance, refugee assistance, foster home care payments, earned income tax credits, and energy assistance. 40. Monthly bill assuming 500 c.f. use, with a 5/8" meter. 41. Monthly bill assuming 50,000 c.f. use, with a 2" meter. 42. Labor force and unemployment data reported for MSAs in New England; exercise extreme caution when comparing unemployment rates with metropolitan variables reported by NECMA for Hartford, CT, Portland, ME, Boston, MA, Manchester, NH, Providence, RI, and Burlington, VT. 43. Average annual unemployment, weighted by total labor force in each respective year. 44. Deflated to constant 1987 dollars, using the GDP price deflator. 45. This index was calculated by CUPR using data from the REIS, published by the U.S. Department of Commerce's Bureau of Economic Analysis. For each year, from the CD we extracted both earnings and employment data at the division level (one-digit SIC designation) for each pertinent metropolitan statistical area and the nation. Subsequently, for the metropolitan areas we calculated the average earnings per job for each division. Then, we calculated the average earnings for each metro area assuming that its industry mix was the same as the nation's. Finally, we took these average earnings and divided them by the national average earnings per job and multiplied by 100. 46. These totals do not distinguish Hispanic/non-Hispanic origin within each racial category (e.g., 'White' includes both non-Hispanic Whites and Hispanic Whites. 47. Dependency ratio is defined as the percentage of all persons aged under 18 or over 64. 50. Includes families of each respective type with related children under 18. 51. All civilian labor force figures exclude persons in armed forces; participation rate denotes percentage of persons aged 16 and over in the labor force. 52. Exercise extreme caution when comparing educational attainment across census years. Data are not strictly comparable, for two reasons. First, decennial census tabulations of CC9HSCH, CM9HSCH, CC8HSCH, and CM8HSCH report high school educational attainment of persons aged 18 and over. For the 1970 census tabulations the universe is persons aged 25 and over. Second, 1980 and 1990 decennial census figures reflect changes to the census questionnaire in 1990. For 1980, "high school graduates" refers to those persons completing four years of high school (as well as those completing one or more years of college); for 1990, tabulations distinguish between persons earning a degree (or equivalency) and those completing grades 9 to 12 without receiving a degree. CC9HSCH and CM9HSCH report the percentage of all persons aged 18 and over who earned a high school diploma or equivalency (including all persons attaining higher levels of education). For 1980, "college graduates" refers to those persons completing four or more years of college; for 1990, the tabulations distinguish between persons who earned a bachelor's degree and those completing one or more years of college without receiving a degree. CC9COLL and CM9COLL report the percentage of all persons aged 18 and over who earned bachelor's degrees (including those completing graduate or professional degrees). 53. Excludes workers working at home. 54. Single-family housing excludes attached units; medium-density housing includes attached single-family units, and housing in structures with 2, 3, or 4 units; high-density housing includes units in structures with 5 or more units. 55. Figures for Indianapolis, Jacksonville, and Nashville refer to the respective counties with which these cities have been consolidated. 56. Total population living within a 500-mile radus of each respective central city. 57. Payments and number of participants during December of each respective reporting year. 58. Geographical coverage of permits data varies for several cities. Anchorage permit system covers all of the Aleutians East Borough, Aleutians West Census Area, Bethel Census Area, Dillingham Census Area, Kenai Peninsula Borough, Lake and Peninsula Borough, Matanuska-Susitna Borough, Nome Census Area, Northwest Arctic Borough, Valdez-Cordova Census Area, and Wade Hampton Census Area. Miami city permits are measured using Dade County data. Indianapolis permit system covers the entire county of Marion except Beech Grove, Lawrence, and Southport cities; and Speedway town. Kansas City, KS permit system also covers the portion of the unincorporated area of Wyandotte County, KS not covered by Bonner Springs city. New Orleans is coextensive with Orleans parish. Detroit permit system also covers property owned by the city outside the corporate limits of Detroit. Billings permit system also covers a 4 - 1/2 mile radius beyond the city limits. Omaha permit system also covers a 3-mile radius beyond the city limits. Portland, OR, permit system also covers permits issued for the portion of the unincorporated area of Multnomah County not covered by the Gresham and Troutdale offices (1994). Memphis permit system also covers Germantown and Lakeland cities, Arlington town; and unincorporated areas of Shelby County. Cheyenne city permit system covers the flood-prone area within a 5-mile radius around the city. 60. Users should exercise caution when comparing employment figures drawn from BEA and U.S. Census sources. All BEA employment figures describe workers by place of work, while census figures describe workers by place of residence. 61. Defined as the number of occupied housing units in which there are more than 1.01 persons per room. 62. Exercise extreme caution when examining metropolitan-level migration measures for 1980; all such measures are based on SMSA boundaries in force for the 1980 census enumeration (hence the "MSA1" designation). Metropolitan-level migration measures for 1980 are based on SMSAs in New England, rather than NECMAS; therefore, metropolitan migration totals for 1980 are incomparable to most other metropolitan area measures. 63. Workers 16 years and over does not include those who worked at home. 64. The degree day normals are used to determine relative estimates of heating and cooling requirements for the buildings. Each degree that the average temperature for a day is below 65 degrees F produces one heating degree day. For example, if the maximum temperature is 70 degrees F and the minumum temperature is 52 degree F, the average temperature for the day is 61 degrees, resulting in four heating degreee days. Cooloing degree days are calculated in a similar fashion for each degree that the average temperature is above 65 degree F. 65. Figures for Buffalo refer to the Buffalo-Niagara Falls, NY Consolidated Metropolitan Statistical Area (CMSA). 66. Users should exercise extreme caution when comparing government financial information across central cities. These figures refer only to city government operations, and exclude other units of local government such as school districts or water management districts. Variations in state and local government organization and jurisdiction, therefore, may be responsible for variations in city financial information reported here. For detailed information, see U.S.B.O.C. (1993b). 68. Governmental administration expenditures include expenditures on financial administration, general control and general public buildings; expenditures on environment and housing include parks and recreation, housing and community development, and natural resources; expenditures on health services include hospital and health expenditures; expenditures on transportation include highways, air and water transportation, and parking, but exclude capital outlays; expenditures on sewerage and sanitation exclude capital outlays; expenditures on education include expenditures on libraries; expenditures on public safety include police, fire, and corrections. 70. Data excludes teachers reported as having working in school district offices rather than individual schools. 71. Includes all categories of high school completions, such as GEDs. 72. Denotes the total amount of revenue transfers from federal government to each respective school district. 73. Expenditures by local school districts only. Excludes expenditures by state educational agencies made for local school districts. 74. Current expenditures per pupil based on fall enrollment as enumerated by the Bureau of the Census. 75. Educational data on local school districts are representative of central cities except for the following: Charleston, WV (represented by Kanawha County); Columbia, SC (Richland Co.); Honolulu, HI (HI Department of Education); Jacksonville, FL (Duvall Co.); Las Vegas, NV (Clark Co.); Louisville KY (Jefferson Co.); Miami, FL (Dade Co.); New Orleans, LA (Orleans Parish School Board); Phoenix, AZ (Mesa Unified School District); Tampa, FL (Hillsborough Co.); Washington, D.C. (D.C. Public Schools). 76. The urban stress index indicators provide measures of the relative hardship experienced by urban regions. Each index is a composite of variables in five categories: employment (unemployment rate, labor force participation rate); demographic (single-parent family households as share of all households; dependency ratio); education (high-school, college graduation rates); housing (ratio of median rent to median household income, rate of excessive housing expenditures); social (death and crime rates); and income and poverty (city-to-suburb per capita income ratio, poverty rate, and gini coefficient of household income inequality). We used principal components analysis to combine these variables for each of the two time periods (1980 and 1990), and two levels of aggregation (city and metropolitan area) to yield four static indices. We also calculated two trend indices to measure changes in each of the constituent variables. For all metropolitan-level measures, the city-to-suburb per capita income ratio was replaced with per capita income. The relative contributions of variables to the final principal components differ for each index, so users should exercise extreme caution when comparing indices over time or across different levels of aggregation. 77. Household expenditure variables represent the estimated proportion of total household spending for each respective category of goods and services. These indicators were derived from 1993 Consumer Expenditure Survey (CES) data. For some metropolitan regions, separate tabulations were not available in the CES file; for these areas, we matched the local demographic composition to national-level expenditure estimates reported for different demographic categories. The Consumer Expenditure Survey (CES) data on housing expenditures estimate costs facing the typical resident, which in turn incorporate wide variations in factors not directly related to housing costs (e.g., length of tenure, differences in the remaining balance of the typical mortgage, etc.). Since we wanted housing costs to reflect the relative prices facing households moving to each local area, we performed several adjustments to the CES housing data. For renters, we substituted fair market rent data for a three-bedroom home as estimated by the U.S. Department of Housing and Urban Development. For owner-occupied housing costs, we inflated median home values reported in the 1990 Census to 1993 dollars; we then assume a 20 percent downpayment, and take one percent of the resulting mortgage balance as the estimated monthly payment. The composite cost of living index incorporates these revised housing estimates along with all of the other listed consumption commodities. For each category, we calculated the product of the ratio of local-to-national expenditures and the standard deviation of expenditures across all metropolitan areas in the study. We then applied a Cobb-Douglas production function to this set of figures, using the local expenditure proportion for each commodity to represent its share in the household consumption function. The resulting figure was subsequently standardized by the equivalent result for the nation. 78. Census Bureau midyear population estimates. Estimates for 1990-94 reflect county population estimates available as of October 1995. 79. "Other" consists of the number of jobs held by U.S. residents employed by international organizations and foreign embassies and consulates in the United States. 80. 1969-74 based on 1967 SIC. 1975-87 based on 1972 SIC. 1988-94 based on 1987 SIC. 81. Consists of medicare payments, medical vendor payments, and CHAMPUS payments. 82. Includes general assistance, emergency assistance, refugee assistance, foster home care payments, earned income tax credits, and energy assistance. 83. Generated using the 15 income categories from Table 75, 4th Count BOC (1970). 84. Generated using the 17 income categories from Table 73, BOC (1980). 85. Generated using the 25 income categories from the STF3C, Table 107, BOC (1992). 86. All measures for suburban areas are generated by subtracting the central city from the MSA. In the case of the three MSAs with two central cities both central cities are subtracted from the MSA. For example, "suburban Los Angeles" equals the Los Angeles MSA minus the central city of Los Angeles minus the central city for Long Beach. Suburban Minneapolis, MN and suburban Kansas City, MO are figured in the same way. Notes to Variables Version 2.2a January 19, 1998