unemployment situation was examined separately, four--Detroit, San Francisco, St. Louis, and Newark--had nonwhite unemployment rates of 10 percent or higher. In Cleveland, the rate was 9.5 percent, compared with the national nonwhite rate of 7.6 percent. In four SMSA's--Los Angeles, Chicago, Philadelphia, and Baltimore, nonwhite rates were close to the national rate, and in three areas rates were substantially below the national average: Houston, 5.7 percent; New York, 4.9 percent; and Washington, D.C., 3.1 percent.

Central Cities

For the nine central cities studied, the average unemployment rate was 4.7 percent, higher than either the national rate or the average rate for the same nine SMSA's (4.1 percent). In Washington, D.C., and Houston, city and SMSA rates were extremely close; elsewhere the central city rate was substantially higher than the SMSA rate. Individual city rates varied significantly, however, ranging from a low of 2 percent in Washington, D.C., to a high of nearly 7 percent in San Francisco-Oakland.

Los Angeles-Long Beach and Baltimore, with central city unemployment rates of 6.5 and 5.7 percent respectively, ranked close to the San Francisco high. At the other end of the scale, Houston (3.4 percent) and Washington were the only central cities having rates below the national average. The other four central cities--New York, Chicago, Philadelphia, and Detroit--had unemployment rates between 4 and 5 percent.

Nonwhites

The average unemployment rate for nonwhites at about 7 percent was the same in the nine central cities as in the nine SMSA's; the closeness of these two rates is largely

nonwhites in central city areas. Approximately 85 percent of the nonwhite labor force in the nine SMSA's lived in the central cities, in contrast with 45 percent of the white labor force. The unemployment rate for central city whites was 3.9 percent, higher than the 3.6 percent rate for white workers in the nine SMSA's.

More than 1 out of every 4 nonwhite workers in the country (28 percent) lived in these nine central cities; the comparable proportion for white workers was 1 in 10. The heavy urban concentration of nonwhite persons, with their relatively high unemployment rates, is one reason why the overall central city unemployment rate is higher than in the SMSA. A second contributing factor is that the unemployment rate is higher for whites who live in the central city than in the entire SMSA.

Central Cities vs. Rest of SMSA's

One-half of the total labor force of the nine SMSA's resided outside the central cities, or in the urban fringe. These fringe areas are not limited to residential suburbs. In most cases they contain sizable cities and towns which share many of the urban problems of central cities; for example, the New York ring includes Yonkers and Philadelphia includes Camden, N.J.

The unemployment rate for the fringe areas was 3.5 percent, compared with 4.7 percent for the nine central cities, but this difference was apparent for white workers only. At 3.3 percent, the white rate outside the central cities was substantially below the 3.9 percent in the cities. On the other hand, the nonwhite rate was about 7 percent in both the central cities and the rest of the

SMSA's, an indication that employment opportunities are inferior for nonwhite workers regardless of place of residence.

The area labor force data presented in the accompanying article and tables were developed by the Bureau of Labor Statistics from information collected as part of the Current Population Survey, a national sample survey of 52,500 households conducted monthly by the Bureau of the Census for the Bureau of Labor Statistics. The CPS is fully explained in Concepts and Methods Used in Manpower Statistics from the Current Population Survey, BLS Report 313, which is available from the Bureau of Labor Statistics upon request.

The provision of this labor force, employment, and unemployment information by color for our largest cities is intended to meet some of the urgent needs for demographic information on the part of program planners on a national, State, or local level. Prior to this time, such demographic data for local areas were available only from the Decennial Census or from special surveys.

The figures for metropolitan areas and central cities presented here have a much larger statistical variance than national data, even when averaged over several months. The accompanying tables give the estimated absolute levels and rates, as well as the ranges in which the estimates would occur 9 out of 10 times if a complete census of the area population were taken.

Selection of Areas

The Standard Metropolitan Statistical Areas (SMSA's) and central cities used in the Current Population Survey and in the Bureau's analysis are defined on the basis of their 1960 Census populations. The 15 largest include all SMSA's where the 1967 civilian noninstitution population 16 years of age and over is one million persons or more. These are New York, N.Y.; Los AngelesLong Beach, Calif.; Chicago, Ill.; Philadel

phia, Pa.-N.J.; Detroit, Mich.; San Francisco-Oakland, Calif.; Boston, Mass.; Washington, D.C.-Md. - Va.; Pittsburgh, Pa.; St. Louis, Mo.-Ill.; Newark, N.J.; Cleveland, Ohio; Balitmore, Md.; Minneapolis-St. Paul, Minn.; and Houston, Tex.

The nine central cities for which data are presented are: New York, Los Angeles-Long Beach combined, Chicago, Philadelphia, Detroit, San Francisco-Oakland combined, Washington, D.C., Baltimore, and Houston. The criteria for selecting these cities were a 1967 civilian noninstitutional population 16 years of age and over of 500,000 persons or more, and a nonwhite population of 100,000

or more.

The geographic boundaries of Standard Metropolitan Statistical Areas are those established by the Bureau of the Budget with the advice of the Federal Committee on Standard Metropolitan Statistical Areas.1/ The general concept of a metropolitan area is one of an integrated economic and social unit with a recognized large population nucleus. Each SMSA must contain at least one city with 50,000 or more inhabitants, or two adjacent cities totaling 50,000 popultion. The SMSA includes the county of such a city and adjacent counties that are found to be metropolitan in character and economically and socially integrated with the county of the central city. (In New England, the units comprising the area are towns rather than counties.)

The largest city in a SMSA is always considered the central city. One or two additional cities also may be regarded as central cities if they have at least 250,000 inhabitants or meet other criteria with respect to size in

1/ Standard Metropolitan Statistical Areas, Bureau of the Budget: 1967.

city limits are the geographic boundaries for the central cities.

The figures for the SMSA's in the accompanying material have been adjusted to independent population estimates made by the Bureau of the Census for the civilian noninstitutional population 16 years of age and over in these areas as of May 1, 1967, a central point for the January-September averages. This was also done for the four central cities which are coterminous with county boundaries, or which exist independent of a county--New York, Philadelphia, Washington, D.C., and Baltimore. The population adjustment increases the accuracy of total civilian labor force estimates for these areas.

Comparability with BES Area Work Force and Unemployment Statistics

The Bureau of Employment Security and its affiliated State unemployment security agencies also publish work force and unemployment levels and unemployment rates for SMSA's. The labor force and unemployment levels and rates in the BLS report may vary in some cases from those published as part of the BES program. Variances occur because of differences in definition and coverage, sources of information, methods of collection, and estimating procedures. Sampling variability and response errors are additional reasons for discrepancies. There are four basic differences in methods, coverage, and definition between the BLS and BES estimates.

(1) The BES figures are estimated by State Employment Security agencies using insured unemployment as a base for unemployment and then applying a standardized method to estimate the unemployed not included in these counts, such as persons who have exhausted their benefit rights, new workers who have not earned rights to unemployment insurance, and persons losing jobs not covered by unemployment insurance systems (agriculture, State and local gov

ment, unpaid family work, nonprofit organizations, and firms below a minimum size).2/ The BLS estimates come from a sample survey of the households in the population. Although the concepts of unemployment used are the same (with minor exceptions), the two methods of estimation do result in some differences.

(2) The BES estimates of employment are based upon employer payroll records, involving some double-counting of multiple jobholders. Payroll figures count each job, whereas the household survey counts each person only once. There are also problems of place of work versus place of residence. Although both the BES and BLS unemployment estimates are based on the place of residence of the unemployed, the BES employment estimates are based on where people work, while the BLS estimates count employed persons where they live. There is generally net in-commuting into SMSA's of persons who live outside the area but work in the SMSA. The net result is that the BES "work force" tends to be higher than the BLS resident labor force for the same metropolitan areas. The larger denominator in the BES series often provides a lower unemployment rate in these large areas than one based on the BLS resident labor force, even where absolute figures on unemployment may be the

same.

(3) The SMSA boundaries used in the BLS study are those that were in effect in 1960, while BES has updated SMSA definitions for changes made through 1966. Six of the 15 largest SMSA's were redefined from 1960 through 1966, but the boundaries used by the BLS and the BES for only two of these areas differ significantly. In the BES estimates, four counties have been added to the Houston SMSA and one large county removed from the

2/ Handbook on Estimating Unemployment, BES No. R-185; March 1960. Available from the Bureau of Employment Security upon request.