U.S. Bureau of Labor Statistics

Compensation and Working Conditions Online

The New Health Participation and Access Data from the National Compensation Survey

by Jason L. Ford
Bureau of Labor Statistics

Originally Posted: October 26, 2009

New data on participation in and access to health care benefits allow for the study of the correlation between health care cost and health plan participation. This article compares this correlation between two occupational groups: management, professional, and related workers and service workers. Although participation is a significant factor in determining the estimates of average employer costs for health care benefits, other factors, such as annual hours worked, mix of jobs and industries, and the percentage paid by the employer can also have substantial effects.

Introduction

The National Compensation Survey (NCS) collects data from employers on the prevalence, or incidence,¹ of employer-provided benefits and employer costs of wages and benefits.² The NCS provides a rich source of data on health care benefits that were recently enhanced to provide even more information than previously available.

The NCS has published estimates on both health insurance coverage rates and data on the cost of health care for many years.³ Until recently, estimates on health care costs and health insurance coverage were not derived from the same sample of employers and health plans. The NCS had provided data on employer costs of health insurance premiums separately from its data on employee access to and participation in various health care plans.⁴ Furthermore, among the health care access and participation data, estimates were available for medical, dental, vision, and outpatient prescription drug plans, but none were available for the entire category of “health care.”

The NCS used the March 2008 data to make the link between employer health care costs and worker access to and participation in health insurance plans. This article describes how NCS data on employer costs for health insurance coverage and new NCS statistics on worker access to and participation in health insurance plans can be used together to gain greater insight into the status of health care coverage in the United States than was previously possible.

For example, if one group of employees has health care costs that are twice as high as another group but also has twice as high a percentage of workers receiving health care, the difference in cost might be due to the difference in the percentage of covered workers. If the costs and employee participation rates are not proportional, however, factors other than participation could be driving the costs.

One such factor is the coverage available through the health care plan; separate categories include comprehensive medical care, dental care, vision care, and drug plans. NCS has recently added two new series called health participation and health access data. These data reveal the percent of employees that are covered by at least one health care plan.

The NCS began producing these health care data using the data from March 2008. Users can study these series in terms of correlation between costs and participation. While the standard errors are available for the health care data in tables 1 and 2 (see below), standard errors for the remaining estimates presented in this article are not available; therefore, comparisons implied cannot be validated by a statistical test.

Types of health care coverage data in NCS

To understand the new health care data, users should be familiar with the types of health care coverage data available in the NCS:

• Access, which indicates the percentage of employees who are in jobs where the employer offers to pay part or all of the cost of health care coverage. Some of the employees with access to health care may not choose to participate in a health care plan.



• Participation, which indicates how many employees are actually enrolled in a health care plan.



• Take-up rates, which indicate the percentage of employees with access that actually participate in the health care plan.

These concepts are related. The number with access multiplied by the take-up rate will equal the number who participate. When comparing data on access to health care plans with data on access to medical, dental, vision, or prescription drug benefits, participation and take-up statistics are important. Because employees do not always enroll in available plans, access alone is not sufficient to determine the extent of health care coverage.

A comparison of health coverage to medical coverage for civilian workers

In many cases, health care coverage data are similar to data for “medical” coverage, which the NCS currently produces. Medical coverage refers to any health care plan that provides coverage for the core areas of health care such as doctor's visits and hospitalization. (Plans that only cover dental or vision or prescription drug plans are not considered medical care plans by the NCS definition.)

Data from March 2008 indicate that access to health and medical care benefits are virtually identical for all published breakouts of workers. When employers offer a health care package, they virtually always offer a medical care plan. Theoretically, employers could offer dental, vision, or drug coverage without offering a medical plan. In practice, employers rarely do so.

Participation is a different matter, however. Here, health care coverage tends to be higher than medical coverage. The reason is that some employees who participate in dental or vision plans do not participate in medical plans. These employees are counted as participating in health care but not in medical care. Overall, 60 percent of workers participate in at least one health care plan, compared with 56 percent who participate in a medical plan.

In small establishments, the differences in participation between health and medical care are not statistically significant. Among employers with 500 or more employees, however, 79 percent of employees participate in health care, compared with 72 percent who participate in medical care plans.

One reason for the difference is that large establishments are more likely to have stand-alone dental plans and vision plans. A stand-alone plan is one that offers only dental or vision care, as opposed to a plan in which dental or vision is included in a comprehensive medical care plan. If a stand-alone dental or vision plan exists, it is likely that some employees will participate in this plan but not the medical plan and therefore cause health participation rates to be higher than medical participation rates. A common reason that employees make this choice is that they can be covered under their spouses' medical plan, but their spouses do not have a dental or vision plan. Table 1 shows the relationship between the size of the establishment and the incidence of stand-alone dental or vision plans:

Since stand-alone dental and vision plans are more common in large establishments, it is not surprising that health participation is higher than medical plan participation in these establishments.

A similar pattern emerges when looking at workers by the level of earnings. Table 2 shows that for those in the lowest 10^th percentile of earnings, the difference between health and medical participation is not statistically significant; however, for those in the highest 10^th percentile of earnings, health participation is 83 percent and medical participation is 76 percent.⁵ These highly-paid workers are more likely to have a stand-alone dental or vision plan. Table 2 shows the relationship between stand-alone dental and vision plans and workers' earnings⁶:

Comparison of health cost and participation in occupational groups: an example of analysis using health care data

Health care data allow for an analysis of participation rates and costs. This section looks at an example of such analysis by occupational group. Occupational groups provide an interesting area for contrast because of the large differences in health costs among various groups of workers.

As table 3 shows, of the five major occupational groups as defined by the Standard Occupational Classification (SOC) system, management, professional, and related workers have by far the highest health care costs on a per hour basis.

The modeled hourly cost per participant is an approximation, calculated by multiplying the hourly cost by the inverse of participation. The result is an estimate of the average hourly cost for only those who actually participate in a health plan.

Many factors affect the average cost for hourly health insurance costs. Other than participation and the underlying cost of health care, factors include the number of hours worked, the percentage of the cost the employer and employee pay, and whether the establishment is in the private sector or the State and local government sector. The level of health coverage and the amount of health utilization are also factors, but these are beyond the scope of this article.

To discover possible reasons why one group of workers has a higher health cost than another group of workers, one must consider all the factors. Consider the differences in participation and health costs between two groups: management, professional, and related workers and service workers.⁷ Management, professional, and related workers have health care participation rates that are 100 percent higher than that of service workers (76 percent compared with 38 percent), but their per-hour health cost is 157 percent higher ($3.32 compared with $1.29.) What accounts for the extra 57 percent in higher costs?

Hours worked

One explanation is that in State and local government, full-time workers in management, professional, and related occupations work fewer hours on average than do full-time service workers, which results in higher cost per hour worked for management, professional, and related workers. Table 4 shows the average annual hours worked, by sector, for each of these occupational groups:

Another explanation is that teachers and other educational personnel, who are included in the management, professional, and related workers category, tend to work 9 months per year, which brings down the average annual hours in this occupational category in local government.

Participation rates

Annual health costs allow for comparisons that are not affected by annual hours worked. Annual health costs for management, professional, and related workers are $5,412.84, as compared with $2,233.42 for service workers. Annual costs are thus 142 percent higher for management, professional and related workers than for service workers. As mentioned, the participation rate is twice as high for management, professional, and related workers as it is for service workers. Thus, the difference in participation still does not account for all of the difference in health care cost.

Occupational composition

A second issue that helps explain the cost difference is that management, professional, and related workers are more likely to be employed in the State and local government sector, where the employer's health costs are higher. As table 5 shows, State and local government has an average employer health cost of $6,340 per employee, as compared with $3,526.93 for the private sector.

Some reasons that State and local government health care is more expensive include the following:

• State and local government has higher participation in health care as compared to private industry (78 percent versus 57 percent.)



• State and local government employers pay a higher percentage of premiums (89 percent vs. 78 percent.)

If these two factors were the only factors, health care costs would be 56 percent higher in State and local government than in private industry. This number can be computed by multiplying the ratio of participation and the ratio of premiums between government and private industry. These are not the only factors, however; the actual difference between the annual health costs of the government and private sectors is about 80 percent.

The difference in health costs between the management, professional, and related workers and the service workers are larger in the private sector than in the government sector. Table 6 breaks out annual health costs for management, professional, and related workers and services workers by the sector of the economy:

Much of this difference can be explained by participation. Table 7 shows the participation rate for health insurance for these same categories:

Table 8 provides a modeled annual cost per participant by multiplying the cost by the inverse of the health care participation: This cost is an approximation.⁸

State and local government costs for health care are similar for management, professional, and related and service workers. Private industry costs for health care, however, are about 31 percent higher for management, professional, and related workers compared with service workers. Part of this difference can be explained by the fact that private industry employers pay a slightly higher percentage of the health costs for management, professional, and related workers (77 percent) than for service workers (73 percent.)

The NCS data suggest that the differences between cost and participation are not caused by individual establishments treating their management, professional, and related workers differently than their service workers. Eighty-five percent of workers are in jobs where the establishment does not differentiate between the sampled jobs in terms offering health care plans. For this 85 percent, either all the sampled jobs get offered the same health care plans or none of the sampled jobs get offered a health care plan.

NCS samples a set of jobs in each establishment. This sample is usually four, six, or eight jobs, depending on the size of the establishment. If all sampled jobs get the same plans, it is probable (although not certain) that the establishment provides one set of health plans to all employees.

If individual establishments treat their workers similarly, what accounts for the differences in costs of health care between management, professional, and related workers and service workers? The ratio of management, professional, and related workers to service workers varies from establishment to establishment. In all likelihood, the establishments that employ a higher percentage of management, professional, and related workers relative to service workers spend more on health care costs.

Comparison of health cost and participation for industry breakouts

On the industry side, the NCS publishes data for health costs and health participation for both goods-producing industries and service-providing industries. Average costs per hour worked in civilian goods-producing industries are 24 percent higher than in service-providing industries ($2.68 vs. $2.17), while participation is 26 percent higher in goods-producing industries (73 percent vs. 58 percent.) Thus, participation and cost are roughly equivalent.

Conclusion

Health participation and access data will allow for the study of the correlation between health care cost and health plan participation. This article provides an example by comparing this correlation among management, professional, and related workers with service workers. While participation is a significant factor in determining the estimates of average employer costs of health care benefits, other factors, such as annual hours worked, mix of jobs and industries, and the percentage the employer pays can have significant effects.

For more data on health insurance benefits, see tables A, B, and C; for standard errors, see tables A (SE), B (SE), and C (SE).

Jason L. Ford
IT Specialist, Division of Directly Collected Periodic Surveys, Office of Technology and Survey Processing, Bureau of Labor Statistics.
Telephone: (202) 691-6267; E-mail: Ford.Jason@bls.gov.

Notes

1 For a discussion on the concepts of incidence, access, participation and take-up rates on health insurance coverage in the NCS, see Carl B. Barsky, “Incidence benefits measures in the National Compensation Survey,” Monthly Labor Review, August 2004, on the Internet at http://www.bls.gov/opub/mlr/2004/08/art3full.pdf.

2 For a list of all NCS publications of survey estimates, and publication of survey estimates from BLS programs that directly preceded the NCS, see http://www.bls.gov/ncs/ncspubs.htm.

3 See William J. Wiatrowski, “Documenting benefits coverage for all workers,” Compensation and Working Conditions Online, on the Internet at http://www.bls.gov/opub/cwc/cm20040518ar01p1.htm.

4 See Michael Lettau, “New statistics for health insurance from the National Compensation Survey,” Monthly Labor Review, August 2004, available on the Internet at http://www.bls.gov/opub/mlr/2004/08/art6full.pdf.

5 For an analysis of the relationship of earnings and health coverage, see: Allan Beckmann, “Comparing employer-provided medical care benefits for lower and higher wage full-time workers,” Compensation and Working Conditions Online, on the Internet at http://www.bls.gov/opub/cwc/cm20071214ar01p1.htm.

6 Averages for occupations within an establishment were used to produce estimates for worker groups averaging hourly pay within the six earnings percentiles: Below the 10^th percentile, 10^th to under the 25^th percentile, 25^th to under the 50^th percentile, 50^th to under the 75^th percentile, 75^th to under the 90^th percentile, and the 90^th percentile and greater. Individual workers can fall into an earnings category different from the average for an occupation and establishment into which they are classified. The percentile breakouts are based on the average wage for each occupation surveyed, which may include workers both above and below the threshold. The percentile values are based on wages published in the National Compensation Survey: Occupational Earnings in the United States, 2007, U.S. Department of Labor, June 2008, bulletin 2704. Values corresponding to the percentiles used in the tables are as follows:

7 Management, professional, and related occupations are those in the Standard Occupational Classifications (SOC) 110000 to 299999. Service occupations are those in the range of SOC categories from 310000 to 399999. A full list of the SOC categories can be found at http://www.bls.gov/SOC/.

8 These data are not the direct result of a statistical sample. Instead, these data are derived by multiplying two types of data together. These are the rounded annual cost and the rounded participation.