document resume ed 279 947 monheit, alan c.; … · document resume. ed 279 947 cg 019 756. author...

DOCUMENT RESUME

ED 279 947 CG 019 756

AUTHOR Monheit, Alan C.; And OthersTITLE The Employed Uninsured and the Role of Public Policy.

National Health Care Expenditures Study.INSTITUTION National Center for Health Services Research and

Health Care Technology (DHHS/PHS), Rockville, MD.PUB DATE 85NOTE 19p.PUB TYPE Reports - General (140) -- Journal Articles (080)JOURNAL CIT Inquiry; v22 p348-364 Win 1985

EDRS PRICE MF01/PC01 Plus Postage.DESCRIPTORS *Employees; Fringe Benefits; *Health Insurance;

Individual Characteristics; Medical Services;Personnel Policy; *Public Policy; Use Studies

IDENTIFIERS *Employed Uninsured

ABSTRACTThis paper notes that, although most private health

insurance is obtained through the workplace, important gaps remain inthe present system of employment-related coverage. National surveydata are presented which revealed that more than 9 million personswith employment experience were uninsured, who, with their uninsureddependents, accounted for three-quarters of all persons who lackedcoverage. This report examines the circumstances and characteristicsof the employed uninsured, including their opportunity to securehealth insurance fringe benefits, their medical care use andexpenditures, and the benefits available in private insurance that isnot work related. Findings are presented showing that, althoughworkers who are young and poor are the most likely to be uninsured,almost one-half of the employed uninsured are over 30 years old andone-half reside in middle- or high-income households. It is alsonoted that few of the employed uninsured are offered health insuranceat the workplace, and they do not receive higher wages in place ofhealth insurance fringe benefits. Alternative public policy responsesto the problem of lack of health insurance coverage are alsoconsidered, including: (1) mandated employment-related coverage; (2)mandated state insurance pools; and (3) general tax credits forhealth insurance. (Author/NB)

*********************************************************************** Reproductions supplied by EDRS are the best that can be made ** from the original document. *

****************************************************************v.******

National Health CareExpenditures Study

The EmployedUninsuredand the Roleof Public Policy

U.S. DEPARTMENT OF EDUCATIONOffice f Educational Research arid Improvement

ED ATIONAL RESOURCES INFORMATIONCENTER (ERICI

EtJhi5 document has been reproduced asreceived from the person or organizationoriginating it.

0 Minor changes have been made to improvereproduction quality.

Points of view or opinions stated in this docu .merit do not necessarily represent officialOERI position or policy.

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICESPublic Health ServiceNational Center for Health Services Researchand Health Care Technology Assessment

Alan C. MonheitMichael M. HaganMarc L. BerkPamela J. Farley

The EmployedUninsured and theRole of PublicPolicy

Although most private health insurance is obtained through the workplace,important gaps remain in the present system of employment-related coverage.National survey data reveal that more than 9 million persons with employmentexperience are uninsured, who, with their uninsured dependents, account forthree-quarters of ail persons who lack coverage. This paper examines thecircumstances and characteristics of the employed uninsured, including theiropportunity to secure health insurance fringe benefits, their medical care use andexpenditures, and the benefits available in private insurance that is not workrelated. Alternative public policy responses to the problem of lack of healthinsurance coverage are also considered.

Because the workplace is the source of 85% ofall private health insurance coverage, it is usu-ally assumed that employed persons are in-sured through their employers.' This assump-tion has subtly misdirected policy discussionsabout the uninsured in two ways. First, it hasbeet, incorrectly inferred that most unem-ployed persons have lost health insurance cov-erage. This contention was not supported byrecent research that indicated that most of theunemployed either did not have health insur-ance to lose in the first place or had the op-portunity to retain coverage from a previousjob or through the job of a spouse.2 Second,the presumption that employment leads tohealth insurance coverage has shifted attentionaway from the largest component of thosewithout health insurance coverage, the 75% ofpersons uninsured throughout the year who areworkers or theh. dependents. An appropriatepolicy response to the general problem of gaps

in health insurance coverage must thereforeconsider the circumstances that cause em-ployed persons to remain uninsured for sig-nificant periods of time.

In this paper we seek to redirect the policydebate on the uninsured by examining thesecircumstances and calling attention to the fol-lowing points:

0

0

0

The employed uninsured and their depen-dents represent the largest component ofthose without health insurance ccverage.Workers who are young and poor are themost likely to be uninsured, but almost halfof the employed uninsured are more than30 years of age and half reside in middle-or high-income households.Few of the employed uninsured are offeredhealth insurance at the workplace, and theydo not receive higher wag(' s in place of healthinsurance fringe benefits.

Alan C. Monhcit, Ph.D., Michael M. Hagan, B.A., Marc L. Berk, Ph.D., and Pamela J. Farley, Ph.D.,are researchers in the Division of Intramural Research, National Center for Health Services Research andHealth Care Technology Assessment, U.S. Department of Health and Human Services, Rockville, MD20857.

348 Inquiry 22: 348-364 (Winter 1985).0046-9580/85/2204-0348$1.25

3

The Employed Uninsured

Table 1. Employment status of persons uninsured all year, 1977 and 1980

1977 1980

Uninsured persons (1,000s) % of uninsured (1,000s) % of uninsuredTotal uninsured persons 17,175 100.0 16,942 100.0Employed all or part of year

(16-64 years of age) 9,198 53.6 55.6Full-time 6,429 37.4 6,563 38.7Part-time 2,769 16.1 2,861 16.9

Persons not working(16-64 years of age) 3,162 18.4 2,582 15.2

Persons <16 years of age 4,561 26.6 4,436 26.2Other uninsured persons' 252 1.4 500 3.0

Source: NMCES Health Insurance Employer and Household Surveys (1977); NMCUES Household Survey (1980).Includes persons 65 years of age or older and persons with unknown employment status.

0 Benefits and out-of-pocket premiums com-parable to the value of employment-relatedinsurance are not available to the employeduninsured outside the workplace.

0 Lack of health insurance has consequencesfor the health care of the employed unin-sured, who use fewer health services thaninsured workers even after controlling forhealth status.

These findings are based on analyses of the1977 National Medical Care Expenditure Sur-vey (NMCES) and the 1980 National MedicalCare Utilization and Expenditure Survey(NMCUES).3 The NMCES serves as the pri-mary analytic data base because it providesinformation on a person's employment status,insurance status, and use of health servicesthroughout the year, and is the only data setthat describes the premiums and benefit pro-visions of insurance held by respondents. Thelatter information is contained in the HealthInsurance Employer Survey (HIES) compo-nent of NMCES, which also describes theavailability of coverage at the workplace.' Tne1980 NMCUES is used to compare and vJi-date NMCES tabulations of the number of em-ployed uninsured persons, their uninsured de-pendents, and their representation among theuninsured population. Although NMCUESalso contains some of the information avail-able from NMCES, the sample is less than halfthe size of NMCES and does not contain de-tailed information on the availability and pro-visions of employment-related coverage.'

The employed uninsured have been definedas persons 16-64 years of age who are em-

ployed for all or part of the year but are un-insured throughout the year.6 Because suchpersons are working, not eligible for Medicare,and chronically uninsured, they lepresent a keygroup for public policy consideration.

The Employed Uninsm ed and theGap in Health Insurance Coverage

The employed uninsured represent more thanhalf of all persons uninsured throughout theyear. Together with their uninsured depen-dents, they account for more than three-quar-ters of the uninsured population. These find-ings are presented in Table 1, which describesthe employment status of uninsured persons,and in Table 2, which presents data on theuninsured dependents of the employed unin-sured.

As Table 1 indicates, the largest componentof the 17.2 million persons without health carecoverage throughout 1977 consists of workers16-64 years of age (53.6%, or 9.2 million in-dividuals). Full-time workers (defined asworking 35 hours or more per week) representa third of all the uninsured and 70% of theemployed uninsured, and part-time workersaccount for 16% of all uninsured persons and30% of the employed uninsured. More recentdata from the 1980 NMCUES reveal a strikingstability in the proportion of the uninsuredrepresented by employed workers three yearslater.

Persons 16-64 years of age who are not em-ployed, in contrast, comprise the smallest pro-portion of the uninsured population, account-ing for 18.4% of the uninsured in 1977 and

349

Inquiry/Volume XXII, Winter 1985

Table 2. Uninsured workers and uninsured dependents

Uninsured persons

1977 1980

(1,000s)% of

uninsured (1,000s)% of

un insured

1) Total uninsured persons 17,175 100.0 16,942 100.02) Uninsured workers having uninsured dependents 2,832 16.5 2,355 13.93) Uninsured dependents of uninsured workers 5,504 32.0 5,220 30.8

a) Employed spouses 1,179 6.9 1,154 6.9b) Employed children 16-18 years of age 392 3.3 299 1.8c) Spouses not working 748 4.4 643 3.8d) Children under 19 not working 3,185 18.5 3,124 18.4

4) Uninsured workers, no uninsured dependents 4,794 27.9 5,617 33.25) All other uninsured persons 4,045 23.6 3,751 22.1

Totals6) All employed uninsured persons

(2 + 3a + 3b + 4) 9,198 53.5 9,425 55.67) Employed uninsured persons having Jependents

and their dependcnts (2 + 3) 8,336 48.5 7,575 44.78) All employed uninsured persons and dependents

(2 + 3 + 4) 13,130 76.4 13,192 77.99) Uninsured dependents per emp!oyed uninsured

person with dependents (3 + 2) 1.94 2.34

Source: NMCES Health Insurance Employer and Household Surveys (1977); NMCUES Houselrqd Survey (1980).

15.2% in 1980. This group consists of the de-pendents of the employed uninsured, other un-insured household members, and uninsuredpersons residing in their own households. Theirrelatively small representation among the un-insured suggests that many persons who do notwork secure health insurance as dependents ofinsured workers or through public programssuch as Medicaid. Table 1 also reveals thatchildren less than 16 years of age representedmore than a quarter of all uninsured personsin both 1977 and 1980.

Although the employed uninsured representa minority of all employed persons under 65(8.7% in 1977 and 8.2% in 1980, or just over9 million persons), their numbers far exceedestimates of the number of unemployed work-.ers who lost health insurance coverage in early1982 and 1983.7 Government intervention onbehalf of unemployed workers became thesubject of an intense public policy debate dur-ing this time, with proponents of such inter-vention arguing that it was unfair to allow un-employed workers to lose coverage because ofeconomic fluctuations largely beyond theircontrol, especially in an uncertain period offinancial hardship.

Similar appeals can be made for public pol-icy intervention on behalf of the employed un-insured. Most work for firms that fail to offerhealth insurance fringe benefits as part of their

350

5

compensation, which excludes them, like theunemployed, from the benefits afforded mostworkers. They, too, face higher prices for pri-vate health insurance than do workers withemployment-related group coverage, since theyforgo the tax advantages and economies as-sociated with group coverage. Like unem-ployed workers without health insurance cov-erage, they also face difficulties securing healthcare. The argument for public policy consid-eration becomes even more compelling whendependents of the employed uninsured are in-cluded in measurements of the size of the un-insured population.

Table 2 assesses how the dependents of theemployed uninsured contribute to the size ofthe uninsured population. As the table dem-onstrates, these secondary effects are not in-consequential. In 1977, there were 2.8 millionemployed uninsured persons with uninsureddependents; this number declined slightly to2.4 million in 1980. These workers had morethan 5 million uninsured dependents in bothyears, accounting for more than 30% of alluninsured persons. Together, they and theiruninsured dependents comprise almost half ofall persons lacking health insurance coverage(48.5% in 1977 and 44.7% in 1980). When allemployed uninsured persons (those with andwithout uninsured dependents) and uninsureddependents are combined, they account for

some 13 million persons, or more than three-quarters of the uninsured population in 1977and 1980. This total consisM of more than 9million employed persons and about 4 mglionnonworking dependents each year. If familycoverage were available to and taken by suchworIcers, three-quarters of the gap in healthiosurance coverage could be eliminated. Also,Table 2 suggests that when an employed per-son with dependents lacks health insurance, atleast two other persons are uninsured through-out the year.

What Kinds of Workers Are Most Likely toLack Health Insurance Coverage?

The workers who are most likely to lack healthinsurance tend to be young, in relatively pooreconomic circumstances, and less educated.Workers in in iustries characterized by sea-sonal or transitory employment and in occu-pations requiring less technical skill are alsomore likely to be uninsured. The employeduninsured, however, are a diverse group con-sisting of many prime-age workers and work-ers in favorable economic circumstances.

The figures in Table 3 describe the percent-age of employed workers of selected demo-graphic and economic characteristics fallinginto one of four mutually exclusive insuranceclasses: private insurance only, all year; privateinsurance only, part of the year; public insur-ance, all or part of the year; and uninsuredthroughout the year. Overall, the 9.2 millionemployed uninsured in 1977 represented 8.7%of all persons with employment experience thatyear.

Young workers 19 to 24 years of age are themost likely to be uninsured, and the propor-tion of workers without insurance declines asage inci eases.8 Young workers tend to be inrelatively good health, expect to incur lowmedical care expenditures, and are less likelyto have dependents who require medical care.Many no longer qualify as dependents underthe health insurance of their parents and mayhave to pay for their own insurance out ofrelatively low earnings. They also may not beoffered employment-related coverage becauseof their limited work experience or transitoryemployment status.

The employed uninsured population also in-cludes a significant proportion of &der prime-


age aorkers whose lack of coverage is likely toaffect the insurance status of their dependents.For example, almost half (47%) of the em-ployed uninsured are workers over 30 years ofage, half are married, and 60% are full-timewage earners.9 The limited employment ex-perience and employment opportunities thatmay explain why young workers are uninsuredcannot explain why many older workers areuninsured for significant peliods of time.

Female workers are no more likely to beuninsured than males. However, nonwhiteworkers are more likely to lack health insur-ance than are whites (12.9% compared with8.1%). Health insurance is also associated witheducational attainment, with 12.6% of workershaving less than 12 years of schooling unin-sured compared with only 3.1% of workerswith 16 or more years. Unmarried workers areabout twice as likely to lack health insuranceas are married workers, reflecting the fact thatmarried workers can obtain insurance as thedependent of a working spouse. Table 3 alsoreveals that workers in poor health are lesslikely to obtain health insurance. Of workersreporting poor heahh, 17% are uninsured com-pared with only 7.0% of workers reporting ex-cellent health and 9.3% in good health. Finally,workers in the South and West are twice aslikely to lack health insurance as are workersin the Northeast or North Central states (11.7%and 12.6% of the former compared with 5.5%and 5.4% of the latter). These differences mayreflect the greater degree of industrializationand unionization in e latter regions.

With regard to employment characteristics,Table 3 discloses that self-employed personsare far more likely to lack coverage than arewage earners (15% compared with 8%). Thereare also disparities in health insurance cov-erage by industry and occupation. Workers inindustries characterized by seasonal employ-ment, self-employment, or a less technicallyskilled work force, such as agriculture, con-struction, sales, repair and personal services,and entertainment and recreation, are twice aslikely to be uninsured as are workers in suchsectors as manufacturing, transportation, com-munications, and utilities, which offer year-round employment to a specialized and moreunionized labor force. Similarly, where thetechnical and educational requirements of an

351


Table 3. Employment and health insurance coverage: Percentage of employed persons byhealth insurance status according to selected characteristics

Characteristics

% of employed persons

Totalemployed(1,000s)

Private insur-ance all year

Privateinsurance

part of year

Public insur-ance all orpart of year

Uninsuredall year

Employed all or part of year* 106,184 83.7 3.3 4.3 8.7

Age (years)16-18 8,395 81.0 1.9 8.1 9.019-24 19,534 74.4 6.9 5.7 13.025-30 l77 l l 82.0 4.4 5.2 8.531-40 22,018 85.5 2.6 4.3 7.641-54 25,176 89.0 1.6 2.3 7.155-64 13,350 88.8 1.9 2.4 7.0

SexMale 57,922 84.3 3.2 3.0 9.4Female 48,262 83.0 3.4 5.8 7.7

ColorWhite 93,813 85.2 3.3 3.3Nonwhite 12,371 72.4 3.1 11.6 12.9

Education (years)<12 29,830 75.9 3.3 8.2 12.612 38,160 86.4 3.3 2.8 7.413-15 16,110 86.3 3.5 2.2 8.016+ 15,112 92.6 3.2 1.1 3.1

Marital statusNever married 25,964 77.3 4.7 6.0 12.0Married 66,348 88.6 2.6 2.2 6.5Widowed 23,211 79.7 4.0 4.0 12.3Separated 2,774 64.2 4.7 17.1 14.1Divorced 5,782 69.1 4.8 12.1 14.0

Perceived health statusExcellent 50,787 86.6 3.6 2.8 7.0Good 41,723 83.2 2.8 4.7 9.3Fair 9,491 77.3 3.9 8.3 10.5Poor 1,637 66.1 1.8 14.7 17.4

RegionNortheast 21,985 89.8 1.5 3.4 5.5North Central 31,666 87.2 3.6 3.9 5.4South 32,808 80.5 3.5 4.3 11.7West 19,724 76.9 4.6 5.9 12.6

...mployment statusWage earner 91,3l9 84.3 3.4 4.3 8.0

Full-time 72,125 85.9 3.6 3.5 7.0Part-time 19,194 78.1 2.8 7.3 11.8

Self-employed 7,570 80.5 2.4 2.0 15.0Full-time 5,827 81.3 2.4 1.4 15.0Part-time 1,743 77.3 2.4 5.0 15A

Farm worker/unknown 2,103 80.1 2.3 3.1 14.4

IndustryAgriculture, forestry, fishing 2,662 71.2 3.6 3.5 21.7Construct:on 4,980 72.7 4.2 4.9 18.2Manufacturing 16,100 88.6 3.0 3.7 4.8Transportation, communica-

tion, utilities6,166 92.4 1.9 1.3 4.4

Sales 18,898 80.7 3.4 5.1 10.8Finance and insurance 5,185 87.8 3.9 3.3 5.0Repair services 4,922 78.7 3.8 4.3 13.2Personal services 3,262 67.8 4.2 12.0 16.0Entertainment and recreation 1,374 71.2 5.1 5.6 18.1

352

Table 3. Continued

The Employed Unihsured

Characteristics

Totalemployed(1,000s)

% of employed persons

Private insur-ance all year

Privateinsurance

part of year

Public insur-ance all orpart of year

Uninsuindall year

Professional services 12,338 87.6 3.0 3.2 6.3Public administration 2,628 87.7 1.5 1.1 3.7

OccupationProfessional and technical 15,673 91.0 2.7 1.8 4.6Managerial and administrative 8,031 89.9 4.6 1.0 4.5Sales 6,176 86.4 3.7 2.2 7.7Clerical 14,869 88.1 2.6 3.5 S.8Craftsmen and foremen 10,366 83.9 3.3 3.3 9.5Operatives 10,842 84.2 4.0 5.8 6.1Transportation operatives 3,816 84.0 2.1 3.0 10.9Service workers 15,859 74.7 4.0 8.8 12.5Lxborers (nonfarm) 4,812 73.1 3.4 5.6 17.9Farmers (owners, managers) 1,379 85.2 1.5 1.2 12.1Farm laborers and foremen 958 69.4 2.7 7.1 20.8

Family incomePoor/near poor 9,244 50.8 4.4 21.6 23.3Low income 12,317 67.8 6.0 8.7 17.5Middle income 41,282 8518 3.8 2.4 8.0High income 43,400 93.3 1.9 1.1 3.7

Hourly wage rate10,816 67.9 3.7 11.0 17.4

$2.51-$5.00 42,890 79.3 4.0 5.2 11.4$5.01-$7.50 19,346 91.8 2.9 1.1 4.237.51-$10.00 9,058 93.8 1.4 1.7 3.1

$10.01+ 5,640 92.6 2.2 1.9 3.3Income/earnings Mean valueb

Househoid income $23,994 $18,271 $11,398 $14,883(all sources) ($376) ($1,267) ($844) ($753)

Hourly wage rate $5.56 $4.40 $3.63 $3.87(70 (15) (200 (15)

Marginal tzx rate of tax filing 29.8% 25.6% 13.0% 20.6%unit (includes federal, state,and Social Security taxes)

(.24%) (.74%) (.69%) (.63%)

Source: NMCES Health Insurance Employer and Household Surveys (1977).Includes all other ethnic/racial igoups not shown separately, persons with unknown education, marital status, health

status, hours of work, earnings, wages, and industry or occupation.b Standard errors are in parentheses.

occupation are less or its seasonal or transitorynature is greater, the likelihood that a workerwill lack coverage is greater. Thus, profession-al, technical, and managerial workers are lesslikely to lack insurance, whereas service work-ers, laborers, craftsmen, and agricultural em-ployees are among those most likely to be un-insured.

Table 3 also reveals that low income andwages are associated with lack of coverage. Al-most a quarter of workers residing in poor ornear-poor households and 17.5% in low-in-come households are uninsured all year, com-pared with only 8% of workers in middle-in-come households and 3.7% in high-income

families. Of workers who received hourly wagesof $2.50 or less in 1977, 17% were uninsured,as were 11% who earned between $2.50 and$5.00. Less than 5% of workers who earnedmore than $5.00 per hour were uninsured. Onaverage, uninsured workers had household in-comes of $14,883, nearly 40% below those ofworkers who were privately insured all year,and hourly wages of $3.87, 30% below thoseof insured workers.

The government implicitly subsidizes thehealth insurance benefits of most workersthrougll the tax system. If the employed un-insured were to receive health insurancethrough the workplace, they would benefit less

353


Table 4. Availability of health insurance at the workplace arcording to health insurance status

Health insurance statusNumber(11,000s)

Availability of health insurance (%)

Employer notoffering plan

Employer with plan

Employeeineligible

Eligible andnot taken

Eligibleand taken

Total employ ad all or part of year 106,184 27.1 5.9 3.9 63.1Private insurance all year 88,926 22.5 3.7 2.8 71.0Private insurance part of year 3,515 18.4 5.0 3.9 72.8Public insurance all or part of year 4,545 44.0 16.4 11.7 27.9Uninsured all year 9,198 66.'"; 22.0 11.1 0.0

Source: NMCES Health Insurance Employer and Household Surveys (1977).

from this policy than the average worker be-cause of their lower wages. The tax advantagesassociated with employment-related insurancearise because employer contributions to healthinsurance fringe benefits are not treated as tax-able income. Because workers who receive suchbenefits have higher incomes and face highermarginal tax rates than uninsured workers, theyreceive a greater tax break than uninsuredworkers would receive from the same em-ployment benefit. The average tax reductionfor workers with private insurance all year wasabout 300 for every dollar of health insurancefringe benefits, whereas the employed unin-sured would have received a reduction of about200 had they been covered (Table 3, last line).Consequently, low-income workers both lackthe resources to purchase insurance and faceweaker tax incentives to obtain it through theworkplace.

Although poor and near-poor workers aremost likely to be uninsured, it is important tonote that the employed uninsured populationincludes some workers in relatively favorableeconomic circumstances. For example, morethan half of all employed uninsured personsare in middle- and high-income households(twice the poverty level or above) and about14% have earnings in excess of $10,000 (in1977 dollars). Consequently, a policy directedat the employed insured that failed to recog-nize differences in economic circumstancesmight subsidize workers capable of financingtheir own health care coverage.

Uninsured Workers and the Availability ofHealth Insurance at the Workplace

An important prerequisite for assessing thefairness of the current system of employment-related coverage is to determine ythether in-

354

sured and uninsured workers have the sam,:topportunity to purchase comparable health in-surance at the workplace. If employment-re-lated insurance is available but uninsuredworkers decline to purchase it, the argumentfor public policy intervention may be less com-pelling. The NMCES data in Table 4, whichexamines the relationship between insurancestatus and the availability of employment-re-lated coverage, clearly demonstrate that thisrarely happens. The NMCES data, of course,cannot disclose whether the employed unin-sured could have obtained coverage throughan alternative choice of employment or underwhat conditions.

The most striking finding in Table 4 is thecontrast in the availability of empioyment-re-lated health insurance to insured and unin-sured workers.") Of workers with private in-surance, more than 70% obtained coverage attheir workplace, whereas only a fifth were notoffered insurance by their employers." How-ever, the overwhelming majority of uninsuredworkers some 89% were unable to obtaininsurance through their employers. Two-thirdsof uninsured workers were employed at firmsthat did not offer coverage and 22% were in-eligible for the health insurance fringe benefitsthat were offered. Only 11% of uninsuredworkers declined to select health insurancebenefits made available by their employers.Note that 60% of workers covered by publicinsurance for all or part of 1977 did not ha veemployment-related coverage as an alterna-tive; of these, 44% were employees with publicinsurance who were working for firms with noinsurance plan and 16% were ineligible for anycovr:rage offered. Only 12% of workers withpublic insurance failed to elect employment-related coverage that was available, suggesting

9

that public insurance was a relatively smallfactor in decisions to forgo private coverage.

To determine what types of workers or firmsare systematically associated with the lack ofheahh insurance benefits, we used two logitequations to examine the likelihood, first, thata worker is employed by a firm that offers healthinsurance and, second, that a worker will beeligible for such benefits.12 Our empirical re-sults (available on request) indicate that low-wage workers, those employed in small firms(fewer than 50 employees) or who are self-employed, or those who work part-time or inindustries characterized by seasonal or tran-sitory employment (such as agriculture, con-struntion, sales, personal services, and enter-tainment and recreation) are less likely to workfor firms that provide health insurance bene-fits. These results are consistent with findingspresented elsewhere.'3 Workers with less labormarket experience, who receive low hourlywages, are emn1 oyed part-time, and are em-ployed in agrh:ultural and construction sectorsare less likely to be eligible for their firm'shealth insurance benefits. These Endings sug-gest that employers are less willing or able toincur the costs of extending such benefits tomarginal workers with limited bargainingpower (given the ready availability of substi-tute labor) and with a relatively high rate ofturnover.

Do Wage Increases Compensate Workers NotOffered Fmployment-Related Coverage?

Another issue related to the availability ofhealth insurance fringe benefits at the work-place is whether workers who fail to receivesuch benefits are compensated by higher wagesinstead. This possibility is suggested by thetheory of competitive labor markets. In suchmarkets employees receive the full monetaryvalue of their productivity and employers areindifferent between providing the same totalcompensation in wages alone or through acombination of wages and fringe benefits.Workers can then substitute wages for fringebenefits by selecting employers who offer thecompensation package they prefer. Amongworkers with the same productivity in the samelabor market, those preferring wages to fringebenefits will receive higher money wages to


equalize the value of the compensation pack-age that combines wages and fringe benefits.The available empirical research on this issueprovides mixed empirical support for this the-ory.'4 Furthermore, even if the theory is cor-rect, similar workers need not receive com-pensating wage differentials to make up fordifferences in fringe benefits if they are in dif-ferent markets and face different demand fortheir services.

Because the argument for policy interven-tion may be less compelling if workers not of-fered insurance are compensated by higherwages, we empirically examined the relation-ship between hourly wages and the availabilityof employment-related coverage, holding pro-ductivity constant. We used a reduced-formregression equation with the natural logarithmof hourly wages as the dependent variable.'5The explanatory variables included factors re-flecting productivity differences among work-ers (education, labor market experience) sexand color (to reflect disparities related to pos-sible labor market discrimination), census di-vision and urban/rural locale (to reflect labormarket and cost-of-living differences), and in-dustrial sector (to capture variation in union-ization, reasonality of employment, and pro-ductivity related to differences in capital stock).We included a dichotomous variable to indi-cate that an employee was not offered insur-ance (individuals offered insurance were thereference group). A positive relationship be-tween this variable and hourly wages wouldsupport the equalizing-wage-differences hy-pothesis (i.e., workers not offered health in-surance, with productivity and other differ-ences held constant, receive higher wagesinstead). We restricted the sample to full-timeworkers and estimated separate equations byoccupation.

The regression results (available on request)uniformly failed to support the hypothesis thatworkers excluded from health insurance ben-efits receive compensating wage increases. Withthe exception of professional and technicalworkers, workers not offered insurance havelower wages than workers receiving such ben-efits. Although professional and technicalworkers not offered insurance receive higherwages, this relationship is small and not sta-tistically significant. Instead of supporting the

0

355

Inquiry/Volume XXII, Winfer 1985

Table S. Premiums mad benefits of privatelyinsured workers by type of insurance

Employ-ment-

relatedinsurance

Not emp3oy-ment-related

insurance

No. of insured workers(1,000s) 83,390

Premiums'Mean dollars $409

8,990

$247% paid by employer 73.3 -% paid by employee 23.8 96.4% paid by other 2.9 3.6

% with benefitType of coverage

Any HMO 4.5 1.9Basic only 9.1 51.7Major medical only 15.2 13.0Basic and major medical 70.6 29.8Other/unknown 0.5 3.6

3readth of coverageDental care 27.9 1.5Vision care 8.7 2.7Outpatient prescriptions 87.8 36.0Physician office visits 88.5 43.5Routine physical exam 6.1 3.0Outpatient psychiatric care 77.4 24.5

Hospital room and boardNo deductible,

semiprivate' 72.2 34.9Deductible, semiprivate 5.7 5.9Less than semiprivate 21.0 55.2No coverage 1.1 4.0

Major medical out-of-pocket limit'$750 or less 33.7 11.4$751 or more 20.2 10.7Unlimited 33.7 20.8No major medical 12.5 57.1

Source: NMCES Health Insurance Employer and House-hold Surveys (1977).

Premium per insured family member.b Full semiprivate or daily benefit of $90 or more.Refers to the limit applicable to most services under a

policy.

equalizing-wage-differences hypothesis, thefindings suggest that workers who receive healthinsurance benefits may be getting a total com-pensation package that is superior to that ofworkers not offered such fringe benefits.

Is Private Health Insurance Comparable toEmployment-Related Coverage AvailableOutside the Workplace?

Workers who are not offered employment-re-lated coverage may, as an alternative, purchaseother private health insurance for themselves

356

and their dependents directly from an insur-ance company. Such plans, however, are hn-perfect substitutes for those availble at theworkplace because they do not offer the ad-ministrative economies of group insurance andtypically offer far less generous benefits. In fact,benefits comparable to those found in em-ployment-related coverage are rare, as shownin Table 5, which describes the differences inpremiums and benefits.

Premiums for employment-related insur-ance were, on average, $409 per year per in-sured family member in 1977, compared with$247 for coverage that was not employmentrelated. This difference reflects the broadercoverage obtained through the workplace. De-spite higher total premiums, workers with em-ployment-related coverage pay less than aquarter of total premium costs, or some $97out of pocket, with their employers paying al-most three-quarters. In contrast, workers whoobtain health insurance outside the firm payalmost two and one-half times as much out ofpocket, or some $238. Employees covered byemployment-related plans also receive the im-plicit tax subsidy on premiums paid by em-ployers, equal to $96 per insured family mem-ber. The inability of workers to obtain similarcoverage through the workplace means thatthey forgo a tax subsidy of some $77 per in-sured family member.'6 As noted earlier, thisfigure is lower because of the lower incomesand tax rates of workers not offered an em-ployer plan.

The most pronounced difference in benefitsbetween employment-reated and other pri-vate coverage is the extent of supplementarymajor medical coverage. About 71% of work-ers with employment-related insurance haveboth basic and major medical coverage, com-pared with only 30% of workers without em-ployment-related plans. The majority of thelatter (52%) have only basic benefits. More thanhalf of all workers with employment-relatedcoverage have limits on their out-of-pocket ex-penditures under a major medical plan. Theseexpenditures are limited for only 22% of work-ers with major medical coverage obtainedthrough private insurance. The lack of majormedical coverage also means less coverage forservices typically insured in major medicalplans. For example, employment-related plans

1 .1

are far more likely to insure physician officevisits and outpatient psychil-.4.ric care. Em-ployment-related insurance also covers rou-tine exams and dental care more frequently,and hospital benefits are far more generousthan those available in private plans. Morethan 70% of workers covered by employment-related plans face no deductib/e for hospitalcare and are ftilly covei.ed for a semiprivateroom or have daily benefits of $90 or more.In contrast, only 35% of workers with otherprivate coverage have such benefits. Half ofsuch workers are covered for less than thesemiprivate room rate.

In sum, the fact that most uninsured workersare not offered health insurance through theworkplace means that they are at a disadvan-tage in the purchase of private health insur-ance. Private health insurance plans that arenot work related typically provide more re-strictive benefits at significantly higher out-of-pocket costs. Employment-related plans offerthe economies of group purchase, rating of pre-miums on a group rather than an individualbasis, and tax subsidies that effectively lowerthe price of insurance.

Do the Employed Uninsured Use Fewer HealthServices?

The data in Table 6 demonstrate the reper-cussions on workers who lack health insur-ance. As a group they use fewer health servicesthan insured workers. Almost 77% of the in-sured reported at least one physician visit dur-ing the year compared with 62% of the unin-sured.' 7 Among those with physician visits, theinsured averaged about five visits comparedwith about four visits for the uninsured. Therewere also large differences between insured anduninsured workers in the use of hospital ser-vices and the use of prescription drugs. Of theinsured, 61% used prescription drugs com-pared with 49% of the uninsured. More than11% of the insured were hospitalized as op-posed to less than 5% of the uninsured.

There are also large differences between in-sured and uninsured workers with similarhealth status. The differences are large amongthose workers experiencing poor health. Ofthose with fair or poor perceived health status,the insured have about two more physician


visits per year than the uninsured, comparedwith a difference of oaly .7 visits between theinsured and uninsureo who consider them-selves in good health. The same pattern emergeswhen disability days are used to control forhealth status. Similar utilization patterns arealso found for prescription drugs. Amongworkers in fair or poor health, only 65% of theemployed uninsured used prescription drugscompared with 81% of the inzured. Insuredworkers in fair or poor health who used pre-scnption drugs also had a higher mean numberel prescriptions. Insured workers in good orexcellent health were also more dkely to useprescription drugs than were uninsured work-ers of the same perceived health status.

Perhaps the most striking findings are thosepertaining to hospitalization. Hospitalizationis generally considered less discretionary thanother health services, so we did not expect lackof insurance to have the same effect on hospitaluse as for other care. Nonetheless, employeduninsured individuals are only half as likely asinsured workers to be hospitalized, a patternthat holds when either perceived health statusor disability days are used to control for healthstatus. The differences are most pronouncedamong those in poor or fair health, with only6.7% of the uninsured hospitalized during thesurvey year compared with 20.4% of the in-sured, a threefold difference.

Given theie differences, the insured havesubstantially higher medical expenses overall($510) than do the uninsured ($367). Yet theuninsured pay about $40 more out of pocketthan the insured. This varies according tohealth status, with no difference in out-of-pocket expenditures between the insured anduninsured with less than eight bed days but avery large difference for eight or more days ofhospitalization. The sick uninsured paid $638out of pocket in 1977, whereas those with in-surance paid only $376.

These data strongly suggest that the em-ployed uninsured use fewer health services thaninsured workers even when they encounter se-rious health problems. Such utilization datacannot be used to determine whether the higherlevels of use among insured workers are moreappropriate, but disparities of this magnitudeseem to warrant concern if equitable access tohealth care is an important social goal."

1 2

357

fable 6. Use and expenditures for health services by employed persons according to insurance and health status.Health Insurance and health status

Il

(1,000s)

% with phy-

sician visit

Mean no, of

visits for

those with

at least one

% with

prescriptiondrugs

Mean no. ofprescriptions

for those

with atleast one hospitalized

Mean expend-

itures for thosewith services

($)

Out-of-pocket

expenditures

for those with

services

($)

Employed uninsured 9,198 62,0 4.1 48.9 5.3 4.7 367,16 216.71Employed insured' 88,926 76.9 5.0 61.3 6.2 11.2 509.88 179.54

Perceived health statusb

Fair or poor (total) 9,698 84.8 7.5 79.0 11.4 18,6 908.77 264.33Employed uninsured 1,282 69.5 5.7 64.9 7.7 6.7 610.74 372.88Employed insured 8,417 87.2 7.7 81.1 11.8 20.4 924.89 257.59

Good or excellent (total) 86,139 74.5 4.6 57.9 5.3 9.7 454.39 168.87Employed uninsured 7,436 60,6 3.9 45.5 4.9 4.3 319.24 195.43Employed insured 78,703 75.9 4.7 59.1 5.3 10.2 458.90 170.49

Disability days

days (total) 11,900 96,2 8.9 88.4 9.4 32,0 1,899.22 381.37Employed uninsured 1,031 88.5 7.1 77.8 9.8 28.0 1,357.27 638.25Employed insured 10,869 96.9 9.1 89.4 9.3 52.2 1,932.35 375.62

<8 days (total) 86,186 72.7 4.2 56.2 5.4 5.2 281.25 145.59Employed uninsured 8,152 58.6 3.5 45.2 4.3 1.8 203.62 147.49Employed insured 78,034 74.1 4.3 57.3 5,5 5.5 286.81 147.96

iource: NMCES Health Insurance Employer and Household Surveys (1977),Privately insured all year.

Totals may not sum because of missing values.

Policy Implications

Our empirical analyses demonstrate that theemployed uninsured and their dependentsconstitute some three-quarters of the 17 mil-lion persons uninsured throug'; lout either 1977or 1980. Analyses of NMCES data also revealthat workers most likely to lack coverage earnlow wages, are less educated, and have house-hold incomes significantly below those of in-sured workers. Workers in industries andoccupations characterized by seasonal em-ployment and less technical or administrativeskill are also more likely to be uninsured, asare young workers, part-time employees, andthe self-employed. Nevertheless, the employeduninsured also include many typical main-stream workers with dependents. Almost halfare between 30 and 64 years of age, half aremarried, 60% are full-time wage earners, andmore than half reside in middle- or high-in-come households.

Lack of insurance for employed persons doesnot simply reflect decisions made by youngworkers who are relatively healthy and withoutinsurable dependents to forgo purchasing in-surance. It probably reflects a number of otherinfluences as well, including poor economiccircumstances, relatively weak bargainingpower (little or no union representation, readi-ly available labor substitutes for low-skillworkers), and an unwillingness of employersto insure seasonal or marginal workers. In con-trast to insured workers, the majority of un-insured workers do not have the opportunityto purchase employment-related coverage. In-stead, their only choice is generally to payhigher out-of-pocket premiums for less gen-erous benefits than are commonly availablethrough health insurance obtained in the work-place. Many uninsured workers can ill affordsuch coverage.

These observations have important impli-cations for a public policy that seeks to expandhealth insurance coverage in order to improveaccess to health care and implement a morecoherent system for financing the uncompen-sated care of those presently uninsured. Be-cause the employed uninsured and their de-pendents represent the largest component ofthe gap in health insurance coverage, policyinitiatives on their behalf could significantly


reduce the size of the uninsured population.Any such policy intervention, however, mustrecognize the diversity of the employed un-insured. At the same time, it must not imposecosts on employers that would jeopardize theemployment of many marginal workers. Itshould also not penalize part-time or seasonalworkers.

Public policy initiatives must also confronta political environment characterized by fiscalausterity and a keen interest in constrainingoutlays for health care. Concern over the fed-eral budget deficit, the desire to shift more re-sponsibility and discretion for health care fi-nancing to state governments, and the need toguarantee the solvency of the Medicare TrustFund are among the factors likely to precludeproposals for a new federal entitlement pro-gram for health insurance coverage)9 Sub-stantial changes in the eligibility standards of'existing public health insurance programs arealso unlikely. As interest in tax reform and taxsimplification mounts and as attempts aremade to expand the federal revenue base byreducing tax expenditures on health insurance,tax incentives io promote the purchase of pri-vate health insurance become less realistic.

Nevert ieless, although Lie focus of healthcare policy since the 1960s and 1970s has shift-ed from one of broad concern over access tocare to interest in cost containment and pro-competitive reform of the health care sector,a number of issues still demand that attentionbe paid to the uninsured.20 These issues, wherethe public has clearly expressed its concern,include uncompensated hospital care, trans-fers of uninsured patients from private to pub-lic hospitals, and the loss of health insuranceassociated with unemployment. Moreover,even as the Medicaid program has been cutback, the poverty rate remains at a high level(14.4% in 1984, the third highest level since1966) and certain sectors of the U.S. economycontinue to experience depressed levels of in-come and employment.

The spectrum of available policy optionsranges from nonintervention to broad-base:::national health insurance initiatives that wouldinclude all uninsured persons. The latter ap-proach, prohibitively costly to implement, haslong been debated, with no sign of consensus.Some intermediate approaches, such as ex-

359

1nquir.v/Volume XXII, Winter 1985

panding Medicaid eligibility to all uninsuredpoor or providing federal grants to state orlocal governments to insure their indigent pop-ulations, target the poor and the unem-ployed.2' However, they ignore low-incomepersons above the poverty line and the rest ofthe employed uninsured, who, along with theirdependents, account for a large part of the =-insured population.

Policy options that are particularly relevantto the working population, and that need notadd to the federal budget, have already beenproposed. Three of them, along with the pos-sibility of maintaining the status quo, are con-sidered below. They are:

O requiring employers to make insuranceavailable to all workers;

O encouraging the formation of state insur-ance pools to make coverage available touninsured workers; and

O severing the tie between employment andhealth insurance benefits while using in-come tax incentives to promote health in-surance purchases.

These strategies rely, rezpectively, on the factthat the employed =instil ed do have employ-ers to offer them coverage, they represent apopulation healthy enough to be insurable andhave some income to pay at least part of thepremium, and they earn income and file taxreturns, making use of an altered tax system afeasible alternative.

Nonintervention by Policy Makers

One option is to retain the present system ofemployment-related coverage. In a competi-tive labor market with a wide choice betweenwages and fringe benefits for a given occupa-tion, it can be asserted that some workers lackhealth insurance by preference. Our empiricalevidence fails to support this assertion. Almost90% of uninsured workers are not offered healthinsurance, and only about 1% of all workerswho are offered insurance choose to be unin-sured. Thus, it is difficult to infer that mostworkers are exercising their preferences. More-over, there is no indication that uninsuredworkers receive higher wages in place of healthinsurance benefits. The low educational andwage levels of most uninsured workers suggestthat they face limited labor market alternati ves

360

1 5

and have little bargaining power to negotiatecompensation arrangements. Noninterven-tion, however, is likely to remain the short-run policy as concern grows over the federalbudget deficit and health care cost contain-ment.

Mandated Employment-Related Coverage

Mandated health insurance benefits have fre-quently been discussed in the context of na-tional health insurance proposals. As definedby Mitchell and Phelps, such an approachwould "require by law that all employers makeavailable health insurance policies of specifiedbenefits and characteristics to cover all em-ployees and their dependents."22 The man-dated approach would legislate the eliminationof existing gaps in employment-related cov-erage and would meet equity concerns by im-posing minimum benefits.

Because a legislative mandate would imposeadditional labor costs on firms, however, itmay be an ill-suited means of extending cov-erage to uninsured workers, especially for smallemployers, who are less likely than large em-ployers to offer coverage. Phelps has suggestedthat the large fixed costs of acquiring and man-aging a health insurance plan make it too ex-pensive to insure only a few employees.23Phelps also speculates that small firms expe-rience higher labor turnover, resulting in ad-ditional administrative costs for enrollment.Because small firms are troie iii.ely to employuninsured workers, the7 would disproportion-ately bear the costs of mandated coverage. Ifthese and other firms that do not offer healthinsurance are marginally profitable, the addedcosts of mandated coverage could pose a ihreatto their long-run survival.

For firms to survive in a competitive envi-ronment, the cost of mandated coverage re-quires a compensating decline in wages to keeptotal labor costs equivalent to worker produc-tivity. Because these wage adjustments wouldnot be instantaneous, workers employed byfirms that previously did not offer coveragemight suffer some short-run unemployment.Even in the long run, wages cannot adjustdownward for workers at or near the minimumwage. Thus, because the employed uninsuredare concentrated among low-wage earners,mandated coverage might lead to a permanent

decline in employment for a significant portionof this group.24 That is, a mandated approachcould perversely affect the very goup it in-tends to assist.

The mandating of employer-related cover-age by states is currently precluded by the fed-eral Employment Retirement Income and Se-curity Act (ERISA) of 1974, which limits stateregulation of employee benefits such as healthinsurance. To date only Hawaii, which ob-tained an exemption from ERISA in 1981, re-quires employers to provide health insuranceof minimum standards to full-time employeesin 1574 legislation. Minnesota, in 1976 legis-lation, required employers offering coverage toprovide minimum benefits or lose state in-come tax deductions for health insurance pre-miums. The Minnesota law, however, waspreempted by ERISA in a 1980 court deci-sion.25

Mandated State Insurance Pools

States have the option of creating risk poolsto make coverage available to the uninsured.26Because many of the uninsured are workerswho lack insurance not because they are poormedical risks but mostly because group insur-ance is unavailable to them and other privateinsurance may be unaffordable, a state poolthat can attract a large proportion of the em-ployed uninsured could offer reasonable cov-erage at a reasonable price. A precedent forthis kind of activity has been established in sixstates (Connecticui, Minnesota, Wisconsin,Indiana, North Dakota, and Florida) that makeinsurance available to high medical risks whoare otherwise uninsurable. Connecticut's leg-islation also extends the risk pool to all stateresidents except those eligible for Medicare. Ingeneral, the states that sponsor such plans re-quire all insurers doing business in the state tocooperatively offer coverage specified by thestate and to share in !any underwriting losses.

The attractiveness and feasibility of such arisk pool would depend on its ability to attractsufficient numbers of the employed uninsuredwith coverage that is less expensive or morecomprehensive than existing alternatives, andto avoid or minimize underwriting losses. Todate, most pools face difficulties in both areas,since all but Connecticut's are limited to highrisks." Connecticut's broad-based pool, con-


sisting of the Health Reinsurance Association,comprising all private commercial carriers, andthe Residual Pool organized by Blue Cross andBlue Shield of Connecticut, attracted only 8.4%of the state's uninsured population in 1981.The Health Reinsurance Association attractedhigh risks and operated at a loss, as the com-mercial insurers enrolled the more favorablerisks in their own plans, and the Residual Poolappeared to enroll persons who sought indi-vidual policies but missed the annual BlueCross and Blue Shield Plan open enrollmentperiod. Connecticut's small enrollment con-sists largely of young people in transition andothers seeking insurance protection betweenjobs. Its limited success may be the result ofpremiums that are too high for many of theuninsured (between 125% and 150% of the av-erage group rate for individuals with similarcharacteristics).

Experience thus suggests that the affoi-da-bility of state-sponsored insurance dependsmost importantly on the size of the pool andon attracting good risks so that experience-rated premiums can be created that are moreattractive than those currently available fromprivate insurers. A successful state pool mustalso balance considerations of premium costsand affordability with comprehensiveness ofcoverage. To keep premiums low may requirehigh deductibles and coinsurance provisions,which may be necessary to encourage enroll-ment, especially by good risks. To keep pre-miums low in relation to benefits may requirelimited subsidies, which could be worthwhileif large numbers of the uninsured obtainedcoverage at mostly their own expense. In thefinal analysis, the formation ofstate pools mightbest be facilitated by extending the favorabletax treatment of employer-sponsored plans tothose enrolled in other types of plans (includ-ing the state's), as discussed next.

General Tax Credits for Health Insurance:The Enthoven Approach

Enthoven suggests replacing the present exclu-sion of employer-paid premiums from eachworker's taxable income with a refundable taxcredit available to anyone who buys insur-ance.28 The tax credit would be 40% of com-bined individual or employer payments up toa limit of $150 per month for family coverage

1 6

361


and $60 per month for individual coverage (in1983 dollars). The tax credit would limit thetax reduction and thereby reduce incentives topurchase very comprehensive health insurancebenefits that contribute to rising health carecosts. The tax credit would also offer the sametax break to individuals in different tax brack-ets. In contrast, the current tax break is re-stricted to employer-paid premiums, is open-ended, and is proportional to the individual'smarginal tax rate. Enthoven argues that hisapproach would encourage even low risks topurchase private insurance, lessening prob-lems of adverse selection outside employergroups. The cost to the federal budget wouldbe reduced by making employer premium con-tributions subject to income and payroll tax-ation.

The flexibility and broad base ofthe tax creditapproach is appealing. Workers who are noto ffered employment-related coverage wouldface the same tax incentives as others who areoffered coverage. Household members couldpool their earnings to obtain more benefits.Furthermore, additional labor costs would notbe imposed on firms that failed to offer healthinsurance fringe benefits (as in a mandatedstrategy). By cutting the tie to employer-paidbenefits, a general tax credit would providesimilar incentives for individuals in differentcircumstances the self-employed, full-timeand part-time workers, the unemployedtosecure coverage.

As Enthoven observes, the tax credit ap-

proach is not perfect. Persons unable to securegroup coverage and those considered high risks(i.e., the elderly, those with preexisting medicalconditions, and those in risky occupations)would still face higher premiums. The creditcould be adjusted to relect these differences,however, and insurance companies could becompensated for the added risks of coveringsuch persons. The tax credit approach couldalso be integrated with state risk pooling effortsfor the uninsured. It is also consistent with costcontainment policies directed at the overlygenerous benefits of some workers while it pro-vides a way to make health insurance afford-able to others.

In sum, a mechanism to enable uninsuredworkers to secure health insurance would sub-stantially reduce the number of persons chron-ically without coverage. Such a policy initia-tive is likely to require an innovative approachthat departs significantly from the current sys-tem of employment-related coverage as asource of insurance. Given current concernsover uncompensated care, transfers of patientsfrom private to public hospitals because oflackof insurance, and the loss of health insuranceassociated with unemployment, there is grow-ing interest in the problems of equitable accessto health care and affordable health insurance.Serious public policy efforts to address theseproblems will be incomplete unless they con-sider the circumstances of the employed un-insured.

Notes

An earlier version of this paper was presented at theannual meeting of the American Public Health Asso-ciation, Anaheim, CA, Nov. 13, 1984. The views ex-pressed herein are those of the authors. No official en-dorsement by either the National Center for HealthServices Research and Health Care Technology As-sessment or the Department of Health and HumanServices is intended or should be inferred.

1 P. J. Farley and G. R. Wilensky, "Options, Incentivesand Employment-Related Health Insurance," Ad-vances in Health Economics and Health Services Re-search 4 (1983): 57-82.

2 A. C. Monheit, M. M. Hagan, M. L. Berk, and G. R.Wilensky, "Health Insurance for the Unemployed: IsFederal Legislation Needed?" Health Affairs 3 (Spring1984): 101-111.

3 The 1977 National Medical Care Expenditure Survey(NMCES) was cosponsored by the National Center for

362

Health Services Research and the National Center forHealth Statistics. The 1980 National Medical CareUtilization and Expenditure Survey (NMCUES) wascosponsored by the National Center for Health Sta-tistics and the Health Care Financing Administration.A description of each data base is available on request.

4 For a description of HIES, see S. B. Cohen and P. J.Farley, Estimation and Sampling Procedures in the1VMCES Insurance Surveys, Instruments and Proce-dures 3 (Rockville, MD: Department of Health andHuman Services, National Center for Health ServicesResearch, May 1984).

5 Although economic conditions and insurance coveragechange over time, recent survey data yield estimatesof the proportion of the population uninsured and theuse of health services by the insured and uninsuredthat are similar lo those of NMCES. However, theylack the same breadth of information available in

1 7

NMCES. Although the Current Population Survey 10(CPS) questions respondents about insurance status inthe prior year, respondents appear to describe their 11

status at the time they are interviewed. Estimates usingthe March 1980 CPS to gauge the number and pro-portion of the population uninsured resemble pointestimates obtained in NMCES (28.6 million persons 12uninsured or 14.4% of the population using the CPS,compared with 26.2 million or 13.8% at the firstNMCES interview). See K. Swartz, "How DifferentAre Four Surveys' Estimates of the Number of Amer-icans Without Health Insurance?" unpublished manu-script (Washington, DC: Urban Institute, Decemberi 984). Published data for the fourth quarter of 1983from the Survey of Income and Program Participation(SIPP) indicate that an average of 15.2% of the pop-ulation is uninsured during each month, comparedwith 11%-12% uninsured for a somewhat longer timeperiod (roughly a quarter) in NMCES. See U.S. Bureau 13of the Census, Current Population Reports, P-70-83-4, Economic Characteristics of Households in theUnited States: Fourth Quarter 1983 (Washington, DC:Government Printing Office, 1985), p. 5. Neither CPSnor SIPP contains any data on the use of health ser-vices. The 1982 National Access Survey, a survey of3,014 adults between May 10 and September 27, 1982,sponsored by the Robert Wood Johnson Foundation,reports 8.7% of the population uninsured, the same as 14the all-year NMCES estimate. Utilization data are alsosimilar, with 32.9% of the uninsured in the accesssurvey having no physician visits and those with visitsaveraging 4.4 visits, compared with 38.0% and 4.1visits for the NMCES emph:yed uninsured. See "Up-date Report on Access to Health Care for the AmericanPeople," Special Report I (Princeton, NJ: Robert WoodJohnson Foundation, 1983); and L. Aday, G. Fleming,and R. Andersen, Access to Medical Care in the U.S.:Who Has It, Who Doesn't? (Chicago: Pluribus Press, 151984).

6 The experience of persons uninsured for part of theyear is more difficult to characterize, since the durationof lapses in coverage cannot be precisely measured ineither NMCES or NMCUES. Such lapses may also beof a transitory nature, reflecting changes in employ-rnent and personal circumstances that, once resolved,lead to a continuation of covetage at a subsequentpoint during the year.

7 The number of unemployed persons losing health in-surance coverage was estimated by Monheit et al. (note2) to be 1.4 million in March 1982. The CongressionalBudget Office (CBO) provided a more pessimistic es-timate of 5 million persons in February 1983. CBOalso estimated that these unemployed workers had anadditiclal 5 million uninsured dependents. See U.S.Congressional Budget Office, "Health Insurance forthe Unemployed: A Comparison of CBO Estimateswith NCHSR Estimates" (staff memorandum, March1984).

8 Unless otherwise indicated, differences discussed aresignificant at the .05 level for a two-tailed test.

9 Young adults age 19 to 24 remain the largest propor-tion of the employed uninsured (28%). The compo-sition of the employed uninsured for any subgroup inTable 3 is derived as follows: Multiply the total em-ployed in the subgroup cell by the percentage unin-sured in that cell, sum to obtain the total employeduninsured, and divide the number of employed un-insured in each cell by the total. 21

16

17

18

19

20


Self-employed workers who were uninsured were clas-sified as not offering insurance to themselves.Insured workers not offered coverage by their ownemployers obtained insurance either as dependents ofa working spouse or by purchasing private insurancethat was not work related.Explanatory variables in the firm equation includedemployee sex and color, full-time or part-time status,the hourly wage rate, whether the worker was a wageearner or self-employed, firm size, census division, andindustry with manufacturing as the reference group.Variables in the eligf)ility equation included an esti-mate of the worker's overall labor market experience(age less years of education less six), the hourly wagerate, color and sex, and whether the worker was em-ployed full- or part-time. Establishment size, location,and industrial sector were also included with manu-facturing as the reference group.S. Malhotra, K. M. McCaffree, J. M. Wills, and J.Baker, Determinants of the Decisions by Establish-ments to Offer a Group Health Plan (report to theLabor Management Services Administration and theAssistant Secretary for Policy Evaluation and Re-search, U.S. Department of Labor, June 1980); W. S.Mellow, "Determinants of Health Insurance and Pen-sion Coverage," Monthly Labor Review 105 (May1982): 30-32.S. A. Woodbury, "Substitution Between Wage andNonwage Benefits," American Economic Review 73(March 1983): 166-182; A. Leibowitz, "Fringe Ben-efit; in Employee Compensation," in The Measure-ment of Labor Cost, ed. J. E. Triplett (Chicago: Uni-versity of Chicago Press for the National Bureau ofEconomic Research, 1983), pp. 371-389; R. S. Smithand R. G. Ehrenberg, "Estimating Wage-Fringe Trade-Offs: Some Data Problems," in The Measurement ofLabor Cost, pp. 347-367.The specification conforms to an earnings functionsuggested by human capital theory. See J. Mincer,Schooling, Experience ana' Earnings (New York: Na-tional Bureau of Economic Research, 1974).This figure is computed by applying the marginal taxrates of workers without employment-relate .1 coverageto the average premium paid by employers for insuredworkers (total premium times percentage paid by em-ployer).Differences reported in the text are significant at the.05 level for a two-tailed test.See, e.g., President's Commission for the Study of Eth-ical Problems in Medicine and Biomedical Research,Securing Access to Health Care, vol. 1 (Washington,DC: Government Printing Office, March 1983).For discussions of the changing federal role in healthcare policy, see J. Feder, J. Holahan, R. Bovbjerg, andJ. Hadley, "Health," in The Reagan Experiment, ed.J. Palmer and I. Sawhill (Washington, DC: Urban In-stitute Press, 1982), pp. 271-305; and L. Ethridge,"Reagan, Congress, and Health Spending," Health Af-fairs 2 (Spring 1983): 14-24.See G. R. Wilensky, "Solving Uncompensated Hos-pital Care: Targeting the Indigent and the Uninsured,"Health Affairs 3 (Winter 1984): 50-62; S. Mulstein,"The Uninsured and the Financing of UncompensatedCare: Scope, Costs, and Policy Options," Inquiry 21(Fall 1984): 214-229; and R. J. Blendon, D. E. Altman,and S. M. Kilstein. "Health Insurance for the Un-employed and Uninsured," National Journal, May 28,1983, pp. 1146-1149.See Wilensky (note 20).

18

363


22 B. M. Mitchell and C. E. Phelps, "National HealthInsurance: Some Costs and Effects of Mandated Em-ployee Coverage," Journal of Political Economy 84(June 1976): 554.

23 C. E. Phelps, "National Health Insurance by Regula-tion: Mandated Employee Benefits," National HealthInsurance: What Now, What Later, What Never? ed.M. V. Pauly (Washington, DC: American EnterpriseInstitute, 1980), pp. 52-73.

24 Ibid.; W. K. Viscusi, "Comment on Phelps," in Na-tional Health Insurance (note 23), pp. 80-82.

364

25 See J. Needleman, M. Anderson, and R. Jaffe, StateOptions for Addressing Catastrophic Health Expenses,vol. 2 (Washington, DC: Lewin & Associates for theNational Center for Health Services Research, April1983), chap. 5.

26 This discussion draws on ibid., chap. 4, and on Wil-ensky (note 20) and Mulstein (note 20).

27 See Needleman et al. (note 25), chap. 4.28 A. Enthoven, "A New Proposal to Reform the Tax

Treatment of Health Insurance," Health Affairs 3(Spring 1984): 21-39.

Yale Executive Program in Health Care ManagementJune 8-13, 1986

Yale's School of Management is offering its ninth executive programon health care management with the support of the Henry J. Kaiser FamilyFoundation. Hospital administrators, physicians, nurses, government of-ficials, and corporate executives will use original case studies to examinecomplex problems encountered in the health care delivery system.

The learning process combines structured and free-form sessions. Full-time participation by faculty members and a high faculty-to-participantratio ensure a high degree of interaction and intensive exploration ofproblems.

The program is held in the facilities of the Yale School of Management.For further information and application materials, call or write:

Dianne CesaroniCoordinator, Executive ProgramsYale School of ManagementBox lANew Haven, CT 06520(203) 436-0808

1 9

document resume ed 279 947 monheit, alan c.; … · document resume. ed 279 947 cg 019 756. author...

Documents