Local Area Employee Benefits Estimates for 15 Metropolitan Areas
Originally Posted: September 28, 2009
This article presents experimental estimates for access to retirement benefits, medical care benefits, and life insurance for the 15 largest U.S. metropolitan areas. The results for December 2008 show that most of the estimates for the areas do not differ from the corresponding national estimate for March 2008 by an amount much larger than their standard errors.
Incidence and detailed provisions of selected employee benefit plans are one of the four key products estimated using data from the National Compensation Survey (NCS). The incidence measures are presented as the percentage of employees who have access to, or participate in, certain types of benefits, such as paid vacations and holidays, disability insurance (short-term and long-term), life and health insurance, and retirement plans.
The NCS reports these incidence statistics by a wide range of worker and establishment characteristics, as well as by geographic area, for workers in civilian,1 private, and State and local government establishments. The geographic areas by which NCS statistics have been available historically consist of the broad groups of States that make up the Census regions and Census divisions.
This article introduces a set of experimental benefit incidence measures for civilian workers in several large metropolitan areas; the data have a reference period of December 2008. These estimates include worker access to three of the main types of employee benefits for which the NCS publishes estimates: retirement benefits, medical care benefits, and life insurance benefits. Because of the relatively small sample sizes for the metropolitan areas, this article reports only access statistics. Access, as defined by the NCS, is having the benefit available for use. Access statistics are generally more reliable than participation statistics--that is, they generally have the higher response rates and the smaller standard errors for a given benefit.
Local area benefits for the 15 largest metropolitan areas
The local area estimates for access to benefits were produced using the same formulas as the national benefit estimates,2 and using data with a reference period of December 2008. The areas covered for this analysis are the 15 largest metropolitan areas in the United States as ranked by total population in 2000: Atlanta-Sandy Springs-Gainesville, GA-AL CSA; Boston-Worcester-Manchester, MA-NH CSA; Chicago-Naperville-Michigan City, IL-IN-WI CSA; Dallas-Fort Worth, TX CSA; Detroit-Warren-Flint, MI CSA; Houston-Baytown-Huntsville, TX CSA; Los Angeles-Long Beach-Riverside, CA CSA; Miami-Fort Lauderdale-Pompano Beach, FL MSA; Minneapolis-St. Paul-St. Cloud, MN-WI CSA; New York-Newark-Bridgeport, NY-NJ-CT-PA CSA; Philadelphia-Camden-Vineland, PA-NJ-DE-MD CSA; Phoenix-Mesa-Scottsdale, AZ MSA; San Jose-San Francisco-Oakland, CA CSA; Seattle-Tacoma-Olympia, WA CSA; and Washington-Baltimore-Northern Virginia, DC-MD-VA-WV CSA.3 However, due to a combination of sample sizes and response rates, not all estimates for all areas are reported.
To evaluate the significance of the access estimates for the metropolitan areas, they were compared with corresponding national estimates. If the estimates for the areas did not vary much from the national estimates, or if they are so imprecisely estimated that a data user cannot conclude much from the differences, then the area estimates might not provide much additional useful information.
Statistical tests were performed for the comparisons. Suppose the access rate for the local area population equals the access rate for the national population as a whole. Then, if all possible samples from the local area population were selected, in about 90 percent of them, the access estimate for the area would differ from the national estimate by no more than 1.6 times the standard error of the difference.4 Therefore, a difference between the estimates that is larger than 1.6 times the standard error is considered statistically significant at a 90-percent confidence level. It implies that the access rate for the local area population does in fact differ from the access rate for the national population. A similar approach can also be used to test whether the difference between the access estimates for any two areas is statistically significant.5
Note that the combined standard error must be used with the distance between the two estimates to determine whether the difference is statistically significant. For example, the estimate for one area may be relatively far from the national estimate, yet the difference is not statistically significant because of a large standard error, whereas the estimate for another area is relatively close to the national estimate, but the difference is statistically significant because of a smaller standard error.
Retirement benefits collected in the NCS include defined benefit pension plans and defined contribution retirement plans.6 Access is defined as a worker having at least one of these two plan types available for use.
Table 1 shows the estimates for access to retirement benefits for the 15 areas, along with the corresponding standard errors. For purposes of comparison, the national estimate for access to retirement benefits among civilian workers in March 2008 is 66 percent, with a standard error of 0.6.7 The magnitude of the standard errors is in the range of 2 to 7 percentage points among the 15 areas, which means that most of the access estimates are not statically different from the national average. Nonetheless, both Detroit and Atlanta, with relatively high percentage-point estimates of 76 and 74, respectively, were significantly different from the national estimate of 66 percent at a 90-percent confidence level. Los Angeles, with a value of 60, is well below the national access estimate, and it is significantly different from the national estimate at a confidence level of 90-percent.
Medical care benefits
The NCS considers an employee as having access to medical care benefits if the benefits are available for the employee’s use or will be available once a service requirement has been met. A medical care plan under NCS definition must provide medical care through at least one of two provisions: hospital care covering inpatient hospital charges; or physician or surgical care.
Table 1 also shows the estimates for access to medical care benefits for the 15 areas. For purposes of comparison, the national estimate for access to medical care benefits among civilian workers equals 74 percent in March 2008, with a standard error of 0.7.8 As with retirement benefits, the medical care benefits standard errors are generally in the range of 2 to 7 percentage points. This also results in the difference between the access estimate and the national estimate not being statistically significant for most of the areas at a 90-percent confidence level. Three exceptions are Atlanta, Seattle, and Los Angeles. Atlanta and Seattle, both with point estimates of 84 percent, are significantly different from the national average at a 90-percent confidence level. Los Angeles, with an access estimate of 70—which is four percentage points lower than the national average—was also significantly different from the national average at a 90-percent confidence level.
An employee is considered to have access to life insurance in the NCS if his or her employer provides access to a plan that provides cash payments to beneficiaries upon the death of the employee or cash payments to the employee due to dismemberment.9 The NCS does not include plans that provide money for disabilities other than dismemberment, nor does it include a death benefit based on credits in a defined benefit plan.
The farthest right-hand columns of table 1 show the estimates for access to life insurance for the 15 areas. The national estimate for access to life insurance among civilian workers equals 62 percent in March 2008, with a standard error of 0.7.10 Like retirement and medical care benefits, the difference between the access estimate and the national estimate is not significant at a 90-percent confidence level for the majority of the areas. Two areas were identified as having access rates significantly higher than the national average at a 90-percent confidence level: Atlanta and Chicago, with access rates of 75 and 70 percent, respectively. Similarly, Los Angeles was significantly lower than the national average at a 90-percent confidence level for access to life insurance.
Access estimates by census division
Currently, the NCS includes access estimates by census division as part of its annual report of employee benefits in the United States.11 The nine census divisions are groupings of States based on their location in the country. Table 2 reports estimates for access to retirement benefits, medical care benefits, and life insurance for the census divisions in the same format as the estimates in table 1, although the reference period is now March 2009.12 As might be expected, the differences among the estimates for the census divisions tend to be smaller than those among the estimates for the metropolitan areas because the census divisions are larger sample units. However, the advantage of the larger sample units is that the access estimates by census division almost always have lower standard errors than the estimates by metropolitan area. As with the estimates by metropolitan area in table 1, the estimated difference between access to the benefit between the census division and the nation as a whole is not statistically significant at a 90-percent confidence level for most of the census divisions.
Census division will remain the most specific geographic unit for which estimates for access to employee benefits will be reported on a regular basis. However, the NCS program will continue to monitor the potential quality of its estimates for these and other benefits by metropolitan area, and it will consider reporting them in the future.
This article presents estimates for worker access to employer-provided retirement benefits, medical care benefits, and life insurance for the 15 largest U.S. metropolitan areas. Among the various types of incidence and provision statistics that the National Compensation Survey program reports, access statistics are generally the most reliable--that is, they have the highest response rates and the smallest standard errors. The results for December 2008 show that most of the estimates for the areas do not differ from the corresponding national estimate for March 2008 by an amount much larger than their standard errors.
1 Civilian includes workers in the private nonfarm economy except those in private households, and workers in the public sector, except the federal government.
2 See the Chapter 8 of the BLS Handbook of Methods for details about the calculation of access statistics. Also, see National Compensation Survey: Employee Benefits in Private Industry in the United States, March 2003 for details about the introduction of access statistics by the NCS program.
3 Note that some of these areas are Consolidated Statistical Areas (CSAs) and others are Metropolitan Statistical Areas (MSAs). The NCS is in its second year of a 6-year transition from a sample of areas based on the December 1993 Office of Management and Budget (OMB) area definitions to a new sample of areas based on the December 2003 area definitions. The NCS is phasing in new Metropolitan and Micropolitan areas as defined by OMB and county clusters defined specifically for the NCS; at the same time, some areas under the December 1993 OMB definitions are being phased out of the sample. For more information on metropolitan area definitions, visit the U.S. Census Bureau Metropolitan and Micropolitan Statistical Areas page on the Internet at http://www.census.gov/population/www/metroareas/metrodef.html.
4 See BLS Resumes Estimation of Sample Errors for Benefit Measures for information on how to interpret the standard error for an estimate. σ(Combined) is the standard error of the estimated difference. It is equal to the square root of the sum of the squared values of the standard errors of all the estimates involved in the comparisons. If estimates X and Y are used, then σ(Combined) = [(σX²) + (σY²)]0.5. For simplicity, this calculation of the combined standard error ignores any overlap between the samples used to calculate the estimates X and Y, even though there is overlap between the NCS sample for an area estimate and the NCS sample for the national estimate.
5 For a comparison of the estimates between two areas, denoted by A and B, σ(Combined) = [(σA²) + (σB²)]0.5, where (σA) is the standard error of the estimate for area A and (σB) is the standard error of the estimate for area B.
6 See NCS Glossary of Terms.
7 See Table 1 of Employee Benefits in the United States, March 2008.
8 See Table 2 of Employee Benefits in the United States, March 2008.
9 See NCS Glossary of Terms.
10 See Table 5 of Employee Benefits in the United States, March 2008.
11 See Employee Benefits in the United States, March 2009.
12 The March 2009 estimate for access to retirement benefits for the entire United States in Table 1 is higher than the corresponding estimate for March 2008 in Table 1 primarily because of a change in how the NCS defines access. See New Definitions of Employee Access to Paid Sick Leave and Retirement Benefits in the National Compensation Survey for details.