Medical Progress Report

No. 10 September 2009

Healthier Choice:

An Examination of Market-Based Reforms for New York’s Uninsured

Stephen T. Parente, Associate Professor of Finance, Carlson School of Management, University of Minnesota

Tarren Bragdon, Adjunct Fellow, Manhattan Institute for Policy Research

Executive Summary

EMAIL THIS | PRINTER FRIENDLY

MPR 10 (PDF)

SLIDESHOW

OP-EDS Why Health Care Is So Expensive in New York Steve Parente and Tarren Bragdon, Wall Street Journal, 10-17-09 The New York Solution doesn't even work in New YorkSteve Parente and Tarren Bragdon, Albany Business Journal, October 9, 2009 New York shows how not to do health care reform, Paul Howard, Washington Examiner, 09-23-09 PODCAST Stephen Parente and Tarren Bragdon talk to Paul Howard about "Healthier Choice: An Examination of Market-Based Reforms for New York's Uninsured", listen here. IN THE PRESS Failed experiments on healthcare, Clive Crook, Financial Times Blog, 12-23-09 Health Bill Clears a Hurdle, Schenectady Daily Gazette, 10-14-09 The Lesson of State Health-Care Reforms Peter Suderman, Wall Street Journal, 10-07-2009 Report Recommends Free-Market Reforms To Expand Health Coverage in New York Gerald Silverman, Bureau of National Affairs, 09-23-09 If Big Government Health Care Doesn't Succeed, Try Try Again, Philip Klein, The American Spectator, 09-22-09 RADIO Tarren Bragdon WIBA's "Upfront With Vicki McKenna", 9-22-09 Tarren Bragdon on WLOB's "The Ray and Ted Show", 9-29-09 Tarren Bragdon on WVOM's "The George Hale and Ric Tyler Show", 9-29-09 Tarren Bragdon on WTOP's "The Ray and Ted Show", 9-25-09 FROM THE BLOGOSPHERE With Obamacare on the Ropes, Dems Go After the Insurers, David Gratzer, NewMajority.com, 9-25-09 New York Is No Model for the Nation, NRO Critical Condition Blog, 9-23-09

Table of Contents:

Executive Summary

About the Authors

Introduction

New York’s Current Individual-Insurance Market

The Impact of Regulations on Insurance Premiums and Purchase Patterns: A Literature Review

Summary of Findings from the Survey and Focus Groups Composed of New York’s Nonpoor Uninsured

The Effects of Reform: Projected Take-Up Rates for Plans within a Reformed New York Individual-Insurance Market

Conclusions and Policy Recommendations

Appendix I: Average Premium Differences under Different Micro-Simulation Policy Choices

Appendix II: A Note on the Micro-Simulation Model Used in This Study: Comparing ARCOLA with Simulations from Columbia University and the Urban Institute

Appendix III: High-Risk-Pool Facts and Function

Endnotes

References

Millions of Americans are living without health insurance. Congress is currently considering a variety of insurance market reforms intended to reduce their number. In New York, there are well over 2 million uninsured adults, representing 14 percent of the non-elderly population, a figure just below the national average. The goal of this paper is to estimate the reduction in the number of uninsured New Yorkers that would result from expanding access to unsubsidized, private health insurance.

Bills before both houses of Congress contain provisions similar to New York State laws that mandate guaranteed issue (which prohibits denial of coverage on the basis of health status) and community rating (which requires insurance companies to charge policyholders the same premium, regardless of their age, gender, or health status). Four other states have similar regulations. Yet New York’s individual-insurance market is unique in requiring insurers to offer coverage to all individuals at all times at exactly the same price.

Although New York’s guaranteed-issue and community-rating laws were adopted with the best of intentions, they have not been effective in substantially reducing the size of the state’s uninsured population. In fact, as a result of a significant increase in the cost of private-insurance coverage for individuals, the market for individual health insurance in New York has nearly disappeared, declining by 96 percent since 1994.

Uninsured New Yorkers of all income levels would benefit from access to a reasonably priced private-insurance market. The existence of such a market would ensure that scarce public dollars are reserved for government programs like Medicaid that protect New York’s poorest and sickest citizens.

With data collected from a survey and three focus groups composed of uninsured New Yorkers and conducted by Zogby International, the authors of this study constructed a micro-simulation model to assess the potential impact of four individual-insurance market reforms on the level of premiums that individuals would pay for private-insurance coverage and the potential willingness of the uninsured to purchase coverage voluntarily. This model was first used by the U.S. Department of Health and Human Services to simulate the effect of the Medicare Modernization Act of 2003 on take-up rates of lower-premium, catastrophic-protection health plans in the individual health-insurance market that were compatible with Health Savings Accounts. Such accounts are not available in New York State.

The market reforms that this paper proposes are:

1. Repeal of community-rating and guaranteed-issue laws
2. Approval of Health Savings Accounts for New York’s individual-insurance market
3. Permission to individuals to shop for approved and affordable health-insurance policies across state lines
4. Approval of “mandate-lite” plans, which permit insurers to offer plans with narrower coverage for sale
   in New York

While each of these reforms would have some effect on reducing the number of uninsured, repeal of New York’s community-rating and guaranteed-issue laws would have the greatest impact, potentially reducing the price of individual insurance coverage by 42 percent and encouraging up to 37 percent of the uninsured to buy coverage.

However, as the report also notes, a small portion of the uninsured—those with certain preexisting conditions—could be deemed uninsurable or find individual insurance coverage too expensive. Therefore, the authors recommend a modest assessment on policyholders in the individual and small-group insurance markets, with the proceeds used to fund a guaranteed-access, high-risk pool for this population. The pool would offer portable private health insurance at a subsidized price. Such a program would ensure that all New Yorkers had access to health insurance.

About the Authors

Stephen T. Parente is director of the Medical Industry Leadership Institute and associate professor in the finance department at the Carlson School of Management, University of Minnesota. He specializes in health economics, information technology, medical technology evaluation, and health insurance. Parente has been the principal investigator for a series of evaluations of consumer-directed health plans since 2002. He was a health policy advisor for the McCain 2008 presidential campaign and served as legislative fellow in the office of Senator John D. Rockefeller IV (DWV) during the 1990s health reform initiatives. He received a Ph.D. in health finance and organization from Johns Hopkins University.

Tarren Bragdon is an adjunct fellow at the Manhattan Institute. A former two-term member of the Maine House of Representatives, he served on the House Joint Standing Committee on Health and Human Services and is the youngest person ever elected to the Maine House. He was also health policy researcher for the president of the Maine Senate. In 2007, he testified regarding health-insurance reform before the U.S. Senate’s Small Business and Entrepreneurship Committee. His articles on health issues have been published in the Wall Street Journal, the New York Post, the Buffalo News, and the Albany Times Union, among other publications. Mr. Bragdon received a bachelor’s degree in computer science from the University of Maine and a master’s degree in business from Husson University. Since January 2008, Bragdon has served as chief executive officer of the Maine Heritage Policy Center, a nonprofit, nonpartisan research and educational organization based in Portland, Maine.

Acknowledgments

Support for this work was provided by the New York State Health Foundation (NYSHealth). The mission of NYSHealth is to expand health insurance coverage, increase access to high-quality health care services, and improve public and community health. The views presented here are those of the authors and not necessarily those of the New York State Health Foundation or its directors, officers, or staff.

Introduction

In 2007, the Manhattan Institute’s Empire Center for New York State Policy published “Rx NY: A Prescription for More Accessible Health Care in NY” (“Rx NY”). In that report, we argued that a reasonably priced and widely available private-insurance market should exist for uninsured but solvent individuals, both for their own sake and to ensure that scarce public dollars are reserved for government programs like Medicaid that protect New York’s poorest and sickest citizens.[1]

This paper evaluates the practical impact of four individual-insurance market reforms recommended in “Rx NY”:

1. Repeal of community-rating and guaranteed-issue laws.
2. Approval of Health Savings Accounts for New York’s individual-insurance market.
3. Permission to individuals to shop across state lines for affordable health-insurance policies already licensed in nearby states.
4. Approval of “mandate-lite” plans, which permit insurers to offer narrower coverage by imposing fewer requirements on plans for sale in New York.

We show that just a few policy changes—in particular, the repeal of New York’s community-rating and guaranteed-issue laws—would make private insurance more affordable and could thereby reduce New York’s uninsured population by up to 37 percent. “Guaranteed issue” requires health insurers to enroll all individual-market applicants, regardless of their health status. “Community rating” requires premiums to be uniform, regardless of an applicant’s age and gender. We also note that widespread uptake of insurance polices in a reformed market depends on the cooperation of employers and policymakers, who must educate the uninsured about their new options and then facilitate the adoption of the ones chosen.

At the same time, we recognize that under this model, a small portion of the uninsured—those with certain preexisting conditions—could be priced out of insurance coverage or deemed uninsurable. Therefore, we recommend the imposition of a modest assessment on policyholders participating in the individual-insurance market, with the proceeds to be used to fund a guaranteed-access high-risk pool for excluded individuals (see Appendix III). The pool would offer to such individuals portable private health insurance at an affordable price.

New York’s Uninsured: A Challenging Population to Reach

The uninsured are a diverse population, diverse in outlook, and difficult to reach with any single form of insurance, public or private. However, extending the reach of public plans or subsidies places a substantial strain on public budgets, particularly in periods like today, when the state is running deficits.

The goal of this paper is to estimate the reduction in the number of uninsured New Yorkers that would result from expanding access to unsubsidized private health insurance. Therefore, this paper does not examine the merits of expanding government medical welfare programs such as Family Health Plus.

Although more than 2 million New York residents are uninsured today, they constitute a population that is largely young, healthy, without dependents, and above the poverty level. Over 60 percent of the uninsured earn over $25,000 annually, and one-third earn over $50,000. Many are only temporarily uninsured. Having some disposable income but a relatively small amount of it, the uninsured, as currently constituted, are likely to be highly responsive to a reduction in premium levels resulting from vigorous competition for their business.[2]

The individual-insurance market is the primary unsubsidized private-insurance option for the uninsured. Also known as the direct-pay market, it is properly understood as a residual market for Americans who do not obtain coverage from an employer or qualify for a public program. Since it is a last resort for such people, it is critical that policymakers in New York and other states ensure that this market is as flexible, affordable, and accessible as it can be made to be. Nevertheless, because of regulatory burdens and the costs they impose, the New York market for such policies is atrophying.

This is best demonstrated by the market’s dramatic contraction since the early 1990s, attributable largely to a steep increase in premiums, in contrast to the performance of the national market, which has grown during this same period. As stated in “Rx NY,” which was published in December 2007 (rates have only gone up since then):

In most regions of the Empire State, the monthly individual health-insurance premium (not purchased through an employer) starts at $500 for an individual policy and $1,400 for a family policy. The average premium in the private market is roughly twice the national average. The only cheaper option available to New Yorkers in the private market has been the Healthy NY program, in which the state directly subsidizes premium rates starting at $300. But eligibility for this plan is limited to workers who earn less than $25,300 and have been uninsured for at least a year or have recently lost employer-sponsored coverage. Relatively few workers qualify, and the program has reached relatively few—only 147,000 have enrolled, or less than 0.8 percent of the state’s population.[3]

Many uninsured New Yorkers already qualify for existing public insurance programs such as Medicaid or the State Children’s Health Insurance Program (SCHIP). Almost all uninsured children are eligible for Child Health Plus. About 800,000 uninsured adults are eligible for Family Health Plus or Medicaid.[4] However, for the nonpoor adults who constitute the remaining 1.3 million of the uninsured, the lack of health-insurance plans that reflect their particular needs and preferences—and the size of their pocketbooks—may dissuade them from purchasing any health-insurance policy at all. The stigma associated with dependence on public programs may also drive away even those for whom such programs are intended. Consequently, New York policymakers should consider the potential of additional market reforms to expand coverage more broadly than an expansion of public programs would alone.[5]

Methods in Brief

To measure the potential market impact of our policy recommendations, we retained the polling firm Zogby International to perform a survey of more than a thousand uninsured New Yorkers (defined as those who were uninsured at the time of the survey or who had been uninsured in the two years preceding it) to examine the reasons they were or had been uninsured, to explore their preferences among insurance products, and to solicit their views on what kind of role the state and federal government should play in providing information about or access to health-insurance options.

To further explore the preferences of the individuals in our survey, Zogby then conducted three focus groups with survey respondents who discussed additional issues, such as their openness to interstate insurance sales, their willingness to spend their limited financial resources on health insurance, and whether it was an individual’s responsibility to obtain coverage or the state or federal government’s responsibility to provide it.

Building on the data obtained from the survey and focus groups, we then conducted a micro-simulation that tested the effects of the four market reforms on insurance take-up among the uninsured. The simulation used the Adjusted Risk Choice & Outcomes Legislative Assessment model (ARCOLA),[6] which separately analyzes recent insurance-product innovations such as high-deductible health plans and limited-panel preferred-provider networks. By doing so, the ARCOLA model distinguishes itself from recent models developed at Columbia University and the Urban Institute, which compare private insurance plans in aggregate with a set of public insurance options. Both the ARCOLA and Urban Institute micro-simulation models were initially funded by the U.S. Department of Health and Human Services. For the purposes of this analysis, the ARCOLA model also was able to predict the effects of premium changes on insurance take-up by the uninsured in a reformed New York State individual-insurance market.

While most of our insurance reforms reduced the number of uninsured by significant margins, the repeal of community rating and guaranteed issue had the greatest impact, producing up to a 37 percent decline in the number of uninsured.

New York’s Current Individual-Insurance Market

Market surveys consistently report that New York has one of the most expensive,[7] highly regulated, and restrictive individual-insurance (also known as direct-pay) markets in the country. Indeed, New York is the only state individual-insurance market that requires all insurers to guarantee issuance of all individual-insurance products to all individuals at all times[8] bearing exactly the same premium[9] (known as community rating).

Mandates

This individual health-insurance market—which is the only place where New Yorkers can buy private, unsubsidized health insurance if they are not sole proprietors, or have an employer that does not offer coverage, or are unemployed—allows just two kinds of plans: a health maintenance organization (HMO) plan and a point of service (POS) plan—both of which operate under state mandates specifying the minimum extent of coverage, the maximum deductible, and the maximum co-payment. These mandates have not been significantly modified since their enactment in 1996. The guaranteed-issue law encourages an individual without employer-based coverage to wait until he or she is sick before buying individual coverage, as insurers are forbidden to deny coverage to any individuals applying, including those already ill. However, even in New York, under certain circumstances, individuals with preexisting conditions may be excluded from coverage for up to one year.

(Until March 2003, New Jersey was the only other state with such strict requirements, but its legislature instituted reforms effective on that date and then adopted still-wider reforms that became effective in January 2009 that made more affordable “basic” policies available to residents, as explained further below.)

The experience of Healthy NY, which began in 2001, demonstrates the significance of lower-cost alternatives, given the price sensitivity of the uninsured. Healthy NY is a state-subsidized plan for sole proprietors and low-income single adults and families. For single adults, the monthly income limit in 2008 was $2,167. For families of four, the monthly limit was $4,417. In 2008, the program covered almost 155,000 New Yorkers. It offered two lower-cost plans in addition to the traditional HMO plan: one with no prescription drug coverage; and one that is a high-deductible Health Savings Account (HSA)-eligible plan. The comprehensive Healthy NY HMO plan for single adults averaged $252 in monthly premiums in 2008. The plan with no pharmacy coverage, introduced in July 2003, costs 15 percent less and has attracted about 20 percent of all Healthy NY enrollees. The HSA-eligible Healthy NY plan, introduced in 2007, has a premium that is 22 percent lower than the traditional HMO plan and in just seventeen months increased its enrollment from 600 to more than 6,000 enrollees, representing about 4 percent of the Healthy NY population. The program costs New York taxpayers $122 million a year and has been estimated to have reduced New York’s uninsured population by one percentage point.[10]

In addition to these laws, the New York legislature has enacted fifty-one mandates dictating coverage of certain medical conditions and inclusion of particular categories of providers in insurance plans. The state average is forty-two mandates. Accordingly, the cost of insurance in New York also exceeds the average. Certain mandates, such as coverage of alcoholism treatment and provision of emergency medical services, appear in almost every state.[11] Others, such as ambulatory cancer treatment and hormone-replacement therapy, are found only in New York and two or three other states.

Costs

Premiums for New York’s individual—that is, private and unsubsidized—insurance plans currently range from $753 to $2,655 a month for a single person and $2,205 to $6,770 for a family of four living in New York City.[12] Annualized, New York City’s individual-insurance premiums cost at least $9,036 for an individual or $26,460 for a family. Clearly, such costs are out of reach for all but the wealthy or the very sick, who would presumably choose to obtain coverage, if they could find the money for it, only if they expected to incur costs at least equal to their premium payments.

Since 2002, sole proprietors have been able to buy into New York’s small-group market by paying no more than 20 percent more than the average premium charged to businesses with two to fifty employees.[13] While sole proprietors are not included in New York’s individual-insurance market, as they are in most states, they are included in tables comparing average individual-market premiums in the states. Despite the inclusion of this relatively young and healthy group, New York has a very expensive individual market with a high average age, as shown below.

Enrollment

New York’s individual-insurance market is small by any standard. In 1994, the year after New York’s community-rating and guaranteed-issue laws were passed, its individual market had almost 752,000 policyholders, about 4.7 percent of its non-elderly population.[15] Today, New York’s individual market is just 34,246,16 about 0.2 percent of New York’s total non-elderly population, and a drop of 96 percent in the portion of the state’s population covered by individual insurance.

In this respect, New York State and the nation have diverged. In 1994, about 10.45 million Americans nationwide, or about 4.5 percent of the non-elderly population, were covered by individual insurance.[17] In 2007, that number had climbed to 14.35 million policyholders, as had the portion of the non-elderly population it represented, about 5.5 percent.[18]

In California’s private individual-insurance market, the participation rate in 2007 was even higher: its 2.6 million covered individuals represented 8 percent of the non-elderly population, up from 6.6 percent in 2000.[19]

While the individual market is growing nationally as a share of the population, New York’s well-meaning but costly laws, expansions of mandated coverage, and increases in premiums to cover their cost have effectively undermined the private-sector safety net by forcing ever larger numbers of residents out of the market.

If New York’s individual market were as big as it was in 1994 (4.7 percent of the non-elderly population), it would have 778,000 policyholders today, not the 34,246 it currently has. If its rate of participation were as high as the average U.S. state’s (5.5 percent of the non-elderly population), New York’s market would be 27 times its current size and have 910,000 policyholders today—about 36 percent of New York’s 2.56 million uninsured non-elderly adults.

Lessons from the Garden State

It is instructive to compare New York’s individual-insurance market with that of another large northeastern state. In August 1993, New Jersey began enforcing guaranteed issue and pure community rating in its individual market, just as New York does currently. Unlike New York, however, New Jersey permitted some variation among its standard individual-insurance plans, including a range of deductibles. Before enacting guaranteed issue and community rating, New Jersey had 157,000 policyholders in its individual market. Despite New Jersey’s greater flexibility, this number had dropped to fewer than 86,700 by the end of 2001.[20]

Concerned about falling enrollment, the New Jersey legislature in 2001 passed a law allowing “Basic and Essential” plans to be sold in the individual market. These plans, which went into effect in March 2003, may charge premiums that vary by a ratio of up to 3:1 to reflect a policyholder’s age, gender, and place of residence. Basic and Essential plans offer a limited benefit, which “covers only 90 days per year for hospitalization, $600 per year for wellness services, $700 per year for office visits for illness or injury, $500 per year for out-of-hospital testing, and limited benefits for mental health services, alcohol and substance abuse treatment and physical therapy.”[21] Carriers can sell a rider providing additional benefits.

At the end of 2002, before these Basic and Essential plans began being sold, New Jersey’s individual market had 79,870 policyholders, almost all of them covered by pre-reform standard plans.[22] By the second quarter of 2009, individual-market enrollment had increased to 105,158 (a gain of 32 percent). This increase was solely a result of the popularity of these new Basic and Essential plans. In fact, the number of people in the standard plan dropped from 78,698 at the end of 2002 to just 52,271 by the second quarter of 2009. The number of policyholders with Basic and Essential Plans went from zero, pre-reform, to 52,645 by the second quarter of 2009.[23] Of note, more than 26,000 standard policyholders, a third of the pre-reform market, switched to Basic and Essential plans during this same period.

Surprisingly, New Jerseyans’ enrollment in the individual market kept building during the current recession; 17,417 people signed up for individual coverage (20 percent growth) from the end of 2007 to the second quarter of 2009. During this same eighteen month period, the size of New Jersey’s small-group market dropped by 66,000. In effect, for every four people previously employed by small business who lost coverage, one voluntarily enrolled in New Jersey’s Basic and Essential plan instead of seeking government assistance or going without coverage.

Unfortunately, New York’s uninsured do not have this option. New Jersey, with less than half (45 percent) of the population of New York,[24] now has three times as many policyholders in its individual market. New Jersey’s experience suggests that regulation—and its impact on insurance premiums—matters greatly and that many uninsured individuals will voluntarily buy unsubsidized individual coverage when it becomes affordable.

The Impact of Regulations on Insurance Premiums and Purchase Patterns: A Literature Review

To examine further the impact of regulations on insurance premiums (and the price sensitivity of the uninsured), we conducted a review of articles and scientific papers on the subject. Our review briefly addresses three broad but key questions, and then provides a more detailed summary of issues most relevant to the situation in New York State.

Key Questions for New York State and Responses

The first and most critical question is: What are the effects of regulations—and subsequent premium increases—on the rate of insurance take-up among potential participants in the small-group and individual-insurance markets? Do some actual participants lose or drop their insurance as a result of premium increases?

The Pricing Impact of Guaranteed Issue, Community Rating, Any-Willing-Provider Laws, and Individual Mandates

Insurance rules can increase the cost of insurance. Those having the greatest impact fall into three broad types: community rating, guaranteed issue, and “any willing provider.” In addition, individual mandates require insurers to cover particular medical services or conditions. Of the three major types, guaranteed issue is the most costly for insurers and thus the regulation most responsible for raising premiums and then driving down insurance take-up rates. This finding is based on a comparison of the work of Congdon et al.[25] and Hadley and Reschovsky,[26] which allowed a distinction to be drawn between a guaranteed-issue effect and a community-rating effect. Citing the experience of New Jersey, the authors found that guaranteed issue can increase premiums by as much as 100 percent.

The mandate having the next-largest effect on the size of premiums is community rating (Congdon et al.,[27] Hadley and Reschovsky).[28] It is responsible for a 20–27 percent increase in premiums and is also accompanied by a decline in consumer demand, as such increases usually are (Congdon et al.).[29] The higher premiums reflect actuarially based projections of higher costs. Premium-support programs for high-cost, chronically ill populations should, to some extent, be able to reverse a price-sensitive decline in demand.

The latest work to examine the effect of community rating is by Lo Sasso and Lurie (2009). Looking at data provided by all respondents to the Survey of Income and Program Participation (SIPP), the authors found that community rating had small effects on rates of individual coverage. However, the data used were less recent than those used by Congdon et al. (2005), Henderson et al. (2009), and New (2006), having been drawn primarily from the 1990s. In view of the recent growth of preferred-provider organizations (PPOs) offering lower premiums, the age of Lo Sasso and Lurie’s data is a concern, even though their study was published more recently than those of the other authors we have reviewed above.

Several researchers have examined any-willing-provider laws. These laws require managed-care organizations, which favor bilateral negotiated arrangements with selected providers, to compensate all providers willing to accept their payment schedule, and found anywhere from a 1.5 percent to a 10 percent increase in premiums following such laws’ adoption. Congdon et al.[30] and Henderson et al. (2009)[31] collected evidence that was sufficiently credible for use in a micro-simulation analysis, but they are the only two investigators to have done so, owing to the difficulty of measuring the effect. The office of the Assistant Secretary of Planning and Evaluation (ASPE), U.S. Department of Health and Human Services, reviewed the quality of the studies. We purposely did not model any-willing-provider laws, which New York does not have, in order to produce a more conservative estimate.

Individual mandates can still have a large cumulative impact. Specifically, researchers have shown that the average mandate, not counting the three most burdensome ones, produces a premium increase of 0.5 percent (Congdon et al., 2005; New, 2006). While this seems negligible when compared with the impact of community-rating provisions (which may add 20–27 percent to premium price), many states have well over twenty mandates and some considerably more. New York has more than fifty. Forty of these mandates (adding 20 percent to the premium price) can cumulatively equal the financial impact of community rating (Congdon et al., 2005; New, 2006).

Beyond Regulation: Are There Tools to Encourage Insurance Take-Up?

Recognizing the low take-up rates in most individual-insurance markets, policymakers, instead of instituting direct market reforms, have extended subsidies to policyholders in order to bring down their out-of-pocket costs. How have subsidy programs such as tax credits and deductions affected insurance take-up rates in the various individual markets where they’ve been tried?

The health-economics literature shows that subsidies, if they go directly to the policyholder, typically lower the cost of insurance and thus increase take-up. The magnitude of the response will depend on the size of the subsidy provided and whether it is offered in combination with a requirement that individuals purchase insurance, as it has been in Massachusetts. One simulation of California data, by Marquis et al. (2004), showed that a 50 percent subsidy had little effect on insurance take-up, the increase being between 4 and 6 percentage points. An empirical estimation of a subsidy program in Washington State produced less anemic results (Long and Marquis, 2005), especially in enrollment rates for children. A more recent subsidy simulation study (Marquis et al., 2007), using data from the Community Tracking Survey, found that a 20 percent subsidy produced a smaller response than the 50 percent subsidy in the earlier Marquis study.

One explanation offered for why premium subsidies are not more effective is that those who would respond to the price changes have already signed up. Those who have not signed up may need subsidies nearly equal to the entire premium to respond, as the Marquis et al. (2004) study suggested. Price-sensitive people sign up early; the rest don’t like the coverage offered or don’t think they need insurance badly enough to acquire it, even when it becomes more affordable. Thus, marginal return on take-up diminishes, even as subsidies increase. This effect was observed by Ken Thorpe in the New York market seventeen years ago.[32]

Finally, are there any market-based programs like Working Today, an association health plan for freelancers, or eHealthInsurance, an electronic market, that show consumers responding to the greater variety of choices they represent and the wider range of premiums available?

There are fairly limited data on the results of these market-based programs. The National Bureau of Economic Research working paper by Joni Hersch (2003) is the best (though practically the only) example of work in this area. Her paper found a large number of freelance workers who were covered neither by any of the companies for which they did work nor the employer of a family member. Therefore, a program like Working Today, which, by pooling freelance workers, brings rates closer to affordable levels, has promise. Empirical studies of the effect of eHealthInsurance.com on coverage rates do not exist, although that company favors the collection of evidence on this question.

Implications for New York

In short, our literature survey shows very significant effects of guaranteed-issue and community-rating regulations on premium prices in individual-insurance markets, where the uninsured (who are eligible for neither public health insurance nor employer-sponsored coverage) are most likely to find themselves. Mandates probably add additional costs to health insurance, but scholarly sources are divided on the magnitude of the impact. Tax credits, another policy tool that has been widely used to increase insurance take-up, have had mixed effects, according to the literature, but their net effect probably has been relatively minor. Finally, relatively little empirical evidence demonstrates the efficacy of market-based mechanisms like Working Today that bring uninsured individuals into insurance pools. Again, this may be because the market for individual insurance is one of last resort, spends less on marketing, and has higher turnover rates than the large-employer and the small-group markets; for example, individuals in this market may have less information on plan pricing or design than participants in those other markets, a possibility that finds some support in both our survey and focus-group results.

Accordingly, our review of the literature suggests that the cost of coverage is one of the most important factors affecting the decision of whether to obtain coverage. This review also helped determine the questions we asked on our survey and posed to our focus groups. These questions covered such topics as our subjects’ income and health status and the extent of their knowledge of the health-insurance market as it exists today in New York.

Summary of Findings from the Survey and Focus Groups Composed of New York’s Nonpoor Uninsured

Working with Zogby International, we surveyed 1,010 New Yorkers likely to find themselves in the individual-insurance market about the demographic niches they occupied, the health benefits that mattered most to them, and other features they would hope to find in a health plan. The survey firm also conducted focus groups to illuminate issues raised during the survey or to confirm the reliability of certain findings of the survey.

Survey Results

Of those surveyed, 69 percent were uninsured, and 31 percent had been uninsured within the preceding two years but were now insured. When queried about their preferences, the currently uninsured stated the three features of a health plan that they most valued, in order of importance: a) an ability to roll over the unused balance in a health account to the following year; b) access to online tools and resources; and c) an absence of co-payments. To the formerly uninsured, the three most desirable features of a plan were, in order of importance: a) an absence of co-payments; b) a small paycheck deduction; and c) online tools and resources. These results suggest that this population is most concerned about out-of-pocket costs. Premium costs are thus sure to matter to it as well.

We also examined to what extent this population might be willing to shop across state lines for coverage, which they cannot now do, given the existence of a federal law, the McCarran Ferguson Act, which prohibits it.[33] Of the 1,010 individuals surveyed, 52 percent said that they would not consider moving to another state to get better health benefits, but almost a quarter (23 percent) said that they were unsure whether they would, and another quarter (25 percent) said that they would consider it. However, most survey respondents were largely uninformed about restrictions on the interstate purchasing of health insurance as well as rules requiring community rating imposed by the Employment Retirement Income Security Act (ERISA) on large self-insured firms. Most members of the focus groups said that they did not know that they could not purchase out-of-state policies. We also found that the likelihood of purchasing was positively correlated to income. The income group most willing to seek better health coverage encompassed those with salaries between $50,000 and $75,000, low enough to reduce their willingness if premiums rose. Ironically, individuals with lower incomes might be less price-sensitive, since they would have started out less disposed to incur the cost of coverage.

Other salient results of the survey include:

• Of those who were without health insurance in the past two years, 29 percent said that it was because they lost their job, and 15 percent said that it was because they had been covered by someone else’s policy but were no longer eligible. About 12 percent said that they lost coverage because they switched jobs and their new employer didn’t offer health insurance. About 11 percent canceled it to save money.
• Of those who had never had health insurance, a plurality (46 percent) said that it was because they could not afford it, and 26 percent said that it was because their employer never offered it.
• Almost 58 percent said that having an employer that provided health insurance would make them more likely to sign up for it, while 20 percent said that they would be more likely to obtain coverage if the rates for private insurance were cheaper.
• Of those who did not have health insurance at the time of the survey but did have it earlier, 40 percent said that they had been without coverage for less than a year, while 31 percent said that the length of the lapse had been between one and three years.
• The vast majority (82 percent) said that they did not search for a health plan on eHealthInsurance.com in 2008.
• More than half (61 percent) said that they would rate their overall health as “positive,” and 18 percent rated it “excellent.”
  Nearly two-thirds (65 percent) said that neither they nor any of their dependents had a chronic condition, but 32 percent said that they or their dependents did suffer from a chronic condition such as asthma, hypertension, diabetes, or arthritis.
• The health-plan feature that 90 percent of respondents wanted most was coverage of preventive-care services, while 76 percent objected most strongly to co-payments.
• About 69 percent were not sure whether they could buy a health-insurance policy from an adjacent state, but 21 percent said that they believed that they could not.

In sum, the survey reinforced the importance of employment as a gateway to health insurance, and it confirmed the decisiveness of cost in determining whether coverage was acquired. The population surveyed is healthy as a whole, but significant numbers of the uninsured and their dependents suffer from a chronic condition. The population lacked knowledge of insurance laws and regulations affecting the prices of plans in the group and individual markets, including those governing interstate sales.

Given the lack of understanding about key aspects of health-insurance plans, it seems apparent that any effort to reach the uninsured must include outreach and education as well as attractively priced insurance products.

The survey was designed not only to provide information on individuals’ insurance preferences but to serve as the basis of a simulation model that tested the impact of different health-insurance policy reforms on take-up rates. To complete this task, attributes such as age, gender, and the existence of a chronic condition in a person’s household were also weighed.

Focus-Group Results

The survey firm also conducted three focus groups. These were quite helpful in providing information unobtainable through the survey or confirming the reliability of certain findings of the survey.

The focus-group report contains several key findings. First, steep premiums were the most important factor in driving the decision to go without coverage. Zogby summarized the focus-group findings:

Affordability was the clarion call that rang through all three focus groups. It is the reason why many are without health insurance, be it because they lost their jobs and cannot afford to pay COBRA or other private rates, or simply because they choose not to buy into plans that are available to them because they cost too much.

Given the state of today’s economy, most of the participants moved health insurance below paying bills and putting food on the table on their list of priorities. If they could afford coverage, most would like a few different plans to choose from, but they didn’t have much interest in plans that could be customized to meet their particular needs. People just want affordable and basic coverage, plain and simple.

This finding confirms much of the national economics literature on the importance of premium level to health-plan choice and suggests that the solution in New York State to low rates of coverage will have to be some form of premium support (for the uninsured chronically ill) as well as a redesign of the insurance market to lower consumers’ cost of coverage.

The second key finding was consumers’ concern about the scarcity of information on plan coverage and pricing available to them. Zogby notes: “While they may be aware of some government programs, very few have a handle on health care and coverage in New York State, and it was unanimous across all three groups that not enough information is being made available to residents.” Their complaint is consistent with the survey finding that existing government and market-based efforts have done an inadequate job of making consumers aware of the coverage choices available. (Indeed, many in the groups blamed the state and federal governments for their lack of knowledge in this area.) A third finding was that many were unaware that they could not buy insurance across state lines, and, when told that they could not, they became upset, even though they had not previously tried to obtain it. In the conversations, some interest in purchasing insurance from neighboring states, where premiums may be lower, was expressed.

Finally, all the groups thought that the U.S. health-care system was flawed and called on policymakers to fix it, although there was disagreement about the extent of the roles that employers and government should play in the process.

The Effects of Reform: Projected Take-Up Rates for Plans within a Reformed New York Individual-Insurance Market

The survey and focus-group results provided sufficient information to permit the application of the ARCOLA national health-reform micro-simulation model to New York State.

The primary variable altered in the ARCOLA model is premium price. The focus-group results confirm that price is an appropriate lever for modeling. We use the ARCOLA model to examine the likely impact on insurance take-up rates of four policy options derived from “Rx NY.” The four scenarios to be modeled are:

1. Removing restrictions on underwriting. We model the impact on plan premiums of no longer requiring community rating or guaranteed issue.
2. Allowing Health Savings Accounts into the market. Currently, these high-deductible savings plans may not be sold in the New York State individual market.
3. Allowing the purchase of policies issued by insurers based in and regulated by neighboring states. The ARCOLA model gives the price of premiums and the impact of regulation in all fifty states. The attractiveness of cross-state purchases by the residents of three states—New York, Pennsylvania, and Connecticut—of policies issued by insurance companies based in them will be measured.
4. Allow the sale of “mandate-lite” plans. These plans do not require inclusion of as many types of services, e.g., chiropractic services or alcohol-abuse treatment, as standard policies usually do.

The primary output of these simulations will be a reduction in the number of uninsured as a result of separately measured incremental reductions in premium cost. These results will follow a description of the bivariate findings derived from the survey data.

The methods used by the ARCOLA model are detailed below. The simulation analysis was completed in three steps. First, we drew on the available literature to characterize the regulatory framework of individual states’ insurance markets and to identify its effect on the level of health-insurance premiums. Second, we used empirical data to develop premium estimates for the simulation that reflect case mix as well as differences in the health-care markets of the various states. Third, we used the survey data discussed earlier to complete a set of simulations that identified the relative effectiveness of four different scenarios in achieving New York State market reforms. We summarize these steps below.

Step 1: Characterize Each State’s Individual-Insurance Markets

The first step in this simulation was to describe the regulatory environment of each state and its effect on health-insurance premiums.[34]

Next, we identified the marginal cost of particular regulations, including guaranteed issue, community rating, and any-willing-provider laws, as well as other mandates.

• Mandates are state regulations or laws that require insurers to cover particular services and reimburse certain categories of provider. We decided to count the number of mandates in a state rather than calculate the cost of each mandate. The number of mandates by state was provided by Blue Cross/Blue Shield National Association. Our decision follows the practice of a majority of empirical studies.
• Guaranteed issue requires insurers to sell insurance to all candidates for coverage regardless of their state of health or the presence of a preexisting condition. However, insurers are not prohibited from inserting riders governing preexisting conditions or raising premiums when they are present. Guaranteed-issue laws can be broad (i.e., applying to all products and all consumers at all times) or narrow (i.e., applying only to sharply defined populations or during limited open-enrollment periods). Our coding rules were biased toward states that had fairly broad guaranteed-issue provisions.
• Community rating requires insurers to limit the degree of variation in the premiums that different individuals must pay. We coded a state as having community rating if it was “pure” (no premium differences are allowed) or “adjusted.” We did not consider rating bands (whereby states allow variation in groups’ premiums to be based on factors such as health status or occupation—typically by +/-25 percent) to be a form of community rating.

We reviewed the literature to identify the impact of these state laws and regulations on health-insurance premiums. We drew only on studies of the individual-insurance market. We ruled out studies that focused on the relationship between regulations and premiums in the small-group market (e.g., Simon, 2005).

We utilized estimates from the following four studies: Congdon et al., 2005; Henderson et al., 2009; New, 2006; and Hadley and Reschovsky, 2003. It should be noted that only the Henderson et al. and Hadley and Reschovsky papers have been published in peer-reviewed journals. The other two are working papers. We considered using estimates appearing in only the peer-reviewed papers but found the methods of the other papers sufficiently rigorous to justify including in this analysis. Table 2 on page 13 summarizes the key findings.

To ensure the conservatism of our inferences, we chose to use values at the twenty-fifth percentile of impact from regulation (i.e., we assumed a lower-end impact). Regulations and mandates are responsible for important differences among states’ individual-insurance markets, but other factors may also be important. We note three in particular. First, with regard to regulations governing look-back periods (to determine whether a claim arises from a preexisting condition) and preexisting conditions generally, significant variation exists among states. The impact of regulation on people with chronic or acute illnesses will be similarly variable, with respect to coverage value, prices, and take-up rates. Although we have information from the various states on the permissible extent of look-back periods and the particular preexisting conditions that are reviewable, we know of no studies that model the effect of regulations in this area on premiums. A second difference among states is the effect of premium taxes on insurance take-up, although we have not attempted to determine what the effect might be. Finally, provider costs, plan types, and the market power of provider networks may vary by state, as would their impact on premiums. Any-willing-provider laws might, however, limit some of this variation.

Step 2: Calculate Adjusted Premiums

The second step in the analysis requires calculation of premiums adjusted for the effect of state regulations. The basic idea behind an interstate insurance sales market is that a person living in heavily regulated State A will be able to buy insurance licensed in less regulated State B. Suppose I live in State A, where the premium is $100 per month. This price level reflects the influence of the style of medical practice in my state, as well as the prices charged by local health-care providers (which would not be different if I bought insurance in State B) and the effect of regulation on cost (which would). If I bought insurance in State B, the premium would be $100 minus the effect of fewer regulations in State B.

To implement this step, we relied on the premiums reported by Congdon, Kowalski, and Showalter (2005). These premiums were first adjusted by age and sex to reflect standard actuarial differences in health-care costs, and were then adjusted to reflect the effect of regulation. The adjusted premiums were inputs in the insurance take-up simulation model.

Step 3: Simulation

In the third step, we simulated the effect of a local interstate market (New York, Connecticut, and Pennsylvania) on the take-up of individual health insurance. Adopting a simulation model developed from previous analyses (Feldman, Parente, Abraham, et al., 2005; Parente, Feldman, and Abraham, 2007), we used the New York State survey data collected for this project to develop a set of New York State estimates. The simulation model is capable of generating estimates of health-insurance take-up within both the individual and employer-sponsored (group) markets. For this analysis, we focused on the individual market only.

One distinguishing attribute of the simulation model is the presence of consumer-driven health plans (CDHPs). There are two types of CDHPs: a low-option Health Reimbursement Arrangement (HRA) and a high-option HRA. The low-option HRA is very similar in deductible, coinsurance, and premium structure to a Health Savings Account (HSA) plan. This similarity enabled us to model both HRA and HSA choices in the simulation as well as high-, moderate-, and low-option Preferred Provider Organizations (PPOs) and a Health Maintenance Organization (HMO).

In the simulation, consumers in the individual market have five choices: high-, moderate-, and low-option PPOs; a high-deductible plan with an HSA; or no coverage. The insurance plans, as modeled, are defined as:

• Direct Pay - PPO Low: Restrictive network, high co-pay, 15 percent coinsurance
• Direct Pay - PPO Medium: Lower co-pay and coinsurance than the low PPO
• Direct Pay - PPO High: High option (i.e., open network, lowest co-pay, no coinsurance)
• HSA: Self-paid HSA, no employer contribution

Chronic illness is modeled at the contract level in the simulations. That is, the person choosing insurance, or someone covered by his or her insurance contract, has a chronic illness. This assumption was made because the data used to estimate the health-plan-choice model could be attributed only to contract holders, not the person receiving care. As a result, the chronic illness metric reflects a household’s illness burden, rather than a single individual’s, unless the contract purchased is a single-coverage contract.

We used premium estimates for New York State for each of the plan choices. These were based on our earlier work and are derived from a combination of eHealthInsurance.com estimates and Kaiser/Commonwealth estimates. Their price levels are given in terms of 2009 dollars.

Simulation Results

The simulation results are presented in Tables 2 through 4. The results show the projected impact of our regulatory reform proposals on the individual market and thus New York State’s population as it exists today. Each individual in the market can choose to purchase one of four different types of health plans or go uninsured. The Medicare and Medicaid markets are excluded from this population, as well as those people who are offered insurance by large employers or the Healthy NY program.

Table 2 presents the results of three regulatory changes represented as columns. The impact of the regulations across the three columns is additive. The first column represents the impact on the New York individual-insurance market of rescission of guaranteed-issue laws. In this instance, there would be a reduction of 18 percent in the size of the uninsured population. Given the significant decrease in premiums that result (see Appendix I), this reduction could occur very rapidly, as consumers adjust to new market prices—perhaps in as little as two years. The next column shows the impact of removing both community rating and guaranteed issue. Dominating the growth of the individual market would be high-option PPOs with low cost sharing and higher premiums.[36] The low-option PPO with high cost sharing and lower premiums has the next-greatest effect. The repeal of community rating would result in a 19 percent reduction in the number of uninsured. Combined, the two policies would result in up to a 37 percent reduction in the number of uninsured. Finally, allowing HSAs into the market and encouraging their growth has a very limited effect, reducing the number of uninsured by approximately eight thousand. The HSAs, once introduced, also compete mostly with the high PPO design. This is not surprising, given that people with higher incomes tend to embrace both high-option PPOs and HSAs.

While Table 2 presents the impact of several critical elements of the “Rx NY” proposal, Table 3 shows the effect of allowing New York consumers to cross state lines—specifically, Pennsylvania’s and Connecticut’s. New Jersey was not included because of premium levels for individual policies and a regulatory structure similar to New York’s.

The introduction of interstate health-insurance market competition significantly reduces the number of uninsured in New York State. In the simulation, we modeled two scenarios: one in which 25 percent of the state market considers participating in out-of-state insurance shopping; and one in which 100 percent does so. The survey results indicated that approximately 25 percent of those surveyed would consider crossing state lines to buy insurance. If 25 percent participated, there would be a 17 percent reduction in the number of uninsured New Yorkers. If the entire market participated, there would be a 26 percent reduction. Effective as such a policy would prove to be, reforming New York’s own market would be more so.

In Table 4, we show the impact of introducing a “mandate-lite” plan similar to those offered in Massachusetts. We compare the effect of two different options on the status quo. First, we examine the impact of subtracting twenty mandates from the approximately fifty mandates in force in the New York State market. Second, we look at the impact of subtracting forty mandates. The mandate-lite health plan in our analysis would be available only to adults aged 18 to 45 and is similar to plans offered in other states targeting younger, healthier consumers, who are more price-sensitive.

Forty fewer mandates would reduce premiums 18 percent, and twenty fewer mandates would reduce premiums 9 percent. A study by Parente et al.[37] of individual state mandates, funded by the U.S. Department of Health and Human Services, came up with similar figures.

At most, the mandate-lite option would produce a 9 percent reduction in the number of uninsured, a smaller impact than the other options would have but by no means a negligible one. Getting rid of twenty mandates reduces the number of uninsured by 3 percent; getting rid of forty mandates reduces the number of uninsured by 9 percent.[38]

Limitations of Our Micro-Simulation Analysis

We acknowledge certain limitations of our micro-simulation analysis. Like any predictive statistical analysis, it uses a limited set of assumptions to forecast an unknown future. Given the changing character of the national health-care market due to reform efforts and trends in spending and technology, our results must be viewed as provisional. We note the following important limitations.

First, the ARCOLA model does not directly observe take-up by the uninsured of particular policies. Thus, we make estimates calibrated to existing market conditions. We do so by inputting published estimates of New York State’s current number of uninsured. All subsequent modeling shows a result that is different from the initial baseline figure. This approach is quite similar to that used in other models. Once policy changes are modeled, estimates of the number of uninsured become internally consistent with the status quo. This approach was validated in a study conducted by Feldman et al. (2005).

A second potential area of concern is that our model finds greater responsiveness to premium-price changes than other micro-simulation models, including those used by the Congressional Budget Office (CBO) and several academic researchers. For instance, the CBO uses Medical Expenditure Panel Survey (MEPS) data, which, unlike the ARCOLA model’s, do not include more recent health-insurance designs such as narrow-network PPOs and high-deductible health plans. In one case, Glied et al. used a micro-simulation model that estimated national take-up of HSAs not exceeding 1 million people by the end of 2005, whereas the ARCOLA model correctly predicted a take-up of 3 million people by then (Feldman et al., 2005). It is difficult to judge the relative accuracy of our prediction of PPO take-up volumes, since the other simulations, with the exception of Glied et al.’s analysis of high-deductible plans, compare only the broad categories of public and private insurance and their success at promoting take-up and do not investigate particular types of private insurance offerings.

In particular, ARCOLA’s findings in this paper contrast with those of two recent models that have been used to examine the impact of similar policy changes proposed for the New York market—including reforms borrowed directly from the 2007 “Rx NY” report. These models were used by Columbia University researchers Glied, Tilipman, and Carrasquillo in Analysis of Five Health Insurance Options for New York State[39] as well as in the recent report prepared by the Urban Institute on behalf of the New York State Department of Health and Insurance: Achieving Quality, Affordable Health Insurance for All New Yorkers: An Analysis of Reform Options.[40]

The Columbia model was used to estimate the size of the reduction in the number of individual New Yorkers without health insurance. The figure it arrived at was 100,000 to 130,000 individuals. The Urban Institute estimated that the uninsured population would be reduced by 15.4 percent, with approximately 400,000 new entrants into the individual-insurance market, under market-oriented policy changes. The Urban Institute estimate is much higher than Columbia’s, but both estimates are lower than those generated by the ARCOLA model, which foresees a reduction of up to 37 percent in the proportion of uninsured, with 816,000 individuals purchasing insurance in the individual-insurance market.

Unlike the models used by the Urban Institute and Columbia University, ARCOLA allows individuals to choose among several different private insurance options with widely varying premiums, including newer consumer-directed health plans such as Health Savings Accounts. The Columbia and Urban Institute models also at least partly utilize cell-based models that predict the behavior of groups rather than, as in ARCOLA’s case, individuals. If, as we predict, the uninsured are highly sensitive to changes in the price of premiums, the differences between ARCOLA and the other models explain why ARCOLA predicts a greater take-up of private insurance plans among the uninsured. (For additional details on the differences between these models and ARCOLA, see Appendix II.)

Our third limitation is that we cannot control for the impact of individual mandates. Clearly, every mandate has a different impact. However, only two studies in the economics literature estimate the effects of mandates, and neither provides any insight into how their presence or absence affects premiums. An actuary could help with the solution to this question, but only to the point of more fully informing the assumptions underlying the estimates.

Finally, we have not modeled the interaction effects of our reforms, such as employer crowd-out or their impact on enrollment in public health-insurance programs. To model these interactions, we require a fuller depiction of the New York State insurance market, both small-group and large-employer, than we can obtain from our Zogby survey and focus groups.

Conclusions and Policy Recommendations

New York State has a long history of providing generous health-care benefits to poor and indigent residents. In recent years, Medicaid has been expanded to cover as many as one in five New Yorkers. However, the state must continue to cope with its substantial number of uninsured people (approximately 14 percent of its population, slightly below the national average) and rapidly rising budgets for its programs.

Health-care reform, which must entail both expanding access to health care and restraining runaway costs, has been championed in Albany as well as the nation’s capital. But recently, state policymakers have acknowledged that although state spending is well above average, New Yorkers’ health outcomes are not.

At the same time, legislators, with the best of motives, have expanded regulation of private unsubsidized insurance markets—most notably, through the adoption of community rating and guaranteed issue, which have significantly raised premiums and discouraged many young and healthy residents from obtaining coverage. Unfortunately, individuals forced to participate in those markets—because, unlike the “young invincibles,” they are too old and sick to risk forgoing coverage—have had to deal with skyrocketing costs.

Our study corroborates the economic literature on price sensitivity, and our survey and focus-group findings highlight the challenges confronting policymakers who seek to expand access to unsubsidized private insurance. These are our three key recommendations:

• Allow underwriting of individual-insurance premiums. Each policy reform that we identified (repealing community rating and guaranteed issue, access to HSAs, cross-border sales, and mandate-lite plans) would reduce the number of people without insurance. But allowing underwriting to operate effectively in the market by taking account of individuals’ particular risk profiles would have the greatest impact, reducing the number of uninsured by up to 37 percent.
• Broaden education and outreach efforts. Knowledge of the substance of insurance policy coverage and the spectrum of options available is very low among the uninsured. Improving the level of understanding would probably promote insurance take-up.
• Create a separate insurance pool for the chronically ill or the otherwise uninsurable. They or their dependents will probably require some amount of subsidy to obtain private insurance they can afford. We recommend that the legislature create a high-risk pool (discussed in greater detail in Appendix III) for the chronically ill in a reformed New York State individual-insurance market.

We believe that the findings of the survey and focus groups indicate a need for innovative education, outreach, and enrollment programs that are directed at the uninsured. Our suggestions are:

1. A state health exchange, perhaps in combination with a market actor like eHealthInsurance.com, where individuals could evaluate prices and health-insurance options could be created.
2. Community and civic organizations (unions, churches, health clubs, chambers of commerce) could be offered a small “finder’s fee” for enrolling members in creditable health-insurance programs.
3. Child Health Plus and Family Health Plus could offer individuals premium quotes for unsubsidized coverage.
4. Brokers should receive a residual annual commission for the first three to five years of coverage maintained by previously uninsured individuals. (The standard commission at present is based on direct-pay customers’ first year of coverage.) Multiyear commissions, paid by the insurer, are already built in to group health coverage. Their availability to brokers selling individual insurance would motivate them to promote such policies.

Connecting young, healthy New York residents to coverage could offer them benefits beyond protection from crushing hospital bills resulting from catastrophic illness. Because insurers will want to hold on to these customers, these young, healthy people should be able to keep their rates down by continuing their coverage rather than periodically acquiring and dropping it as the need arises, or waiting until they are older to obtain it for the first time. An increase in both the number of holders of portable health insurance and the duration of their coverage should also encourage insurers to focus more on wellness and prevention efforts.

For the past several decades, New York policymakers have focused on making publicly subsidized programs available to new segments of the population rather than ensuring the affordability of private plans and thus their availability. This approach may be tempting in flush economic times, but it tests government budgets when, as now, tax receipts are declining and the numbers of the needy are rising. Public programs are an essential safety net, particularly in difficult economic times, but when the economy is functioning as it should, high-quality, affordable health insurance provided by the private market should exist for all but the poor.

Public officials should not assume that public subsidies are the only way to reduce the number of uninsured citizens. Rather, as this report suggests, an effectively regulated private market can serve the needs of a significant percentage of New Yorkers currently without health insurance.

APPENDIX I

APPENDIX II

A Note on the Micro-Simulation Model used in This Study: Comparing ArCOlA with Simulations from Columbia university and the urban Institute

This appendix explains the micro-simulation model used in this study to estimate the effects of policy changes on New York’s individual-insurance market, as well as to compare it with other similar models used in related work.

This study used the Adjusted Risk Choice & Outcomes Legislative Assessment (ARCOLA) model to estimate the impact of health-policy proposals. The model predicts individual adult responses to proposed policy changes and can be used to generalize to larger population groups (by state or nationally) the proposals’ impact on the number of lives covered as well as the cost of that coverage.

ARCOLA is a predictive micro-simulation based on multivariate regression analysis that predicts individuals’ plan choices under different parameters such as income, health status, the extent of benefit cost sharing (e.g., the size of co-payments and deductibles and the presence of coinsurance), benefit design, and premium. Note, especially, that it is the behavior of individuals under these varying conditions that lies at the heart of this model.

ARCOLA Model Background

The model was developed for the Office of the Assistant Secretary of Planning and Evaluation (OASPE) of the U.S. Department of Health and Human Services (DHHS) and was used to simulate the effect of the Medicare Modernization Act of 2003 (MMA) on take-up of high-deductible health plans in the individual health-insurance market (Feldman, Parente, Abraham, et al., 2005; Parente et al., Final Technical Report for DHHS Contract HHSP233200400573P, 2005).

Later, the model incorporated the effect of prior health status on health-plan choice, a refinement that strengthened its predictive power. The latest version of the model also incorporates actual claims data to develop premium estimates (i.e., claims expenses multiplied by a loading factor) and then predicts choices again, using the newly calculated premiums. The choice model then iterates until premiums and choices converge at an equilibrium state. A subsequent change to the model permitted state-specific predictions of the impact of particular policy changes as well as the total national impact.

Model Components and Data Sources

The three major components to the ARCOLA model are: 1) model estimation; 2) choice set assignment and prediction; and 3) policy simulation. As illustrated in Figure A-1, often more than one database was required to complete each task. Integral to this analysis was the use of consumer-directed health-plan data from four large employers working with the study investigators.

The model estimation involved several steps. In the first step, pooled data from four employers offering CDHPs were used to estimate a conditional logistic plan-choice model. In the second step, estimated choice-model coefficients were used to predict health-plan choices for individuals in the Medical Expenditure Panel Survey Household Component (MEPS-HC). In order to complete this step, it was necessary to assign to each respondent in the MEPS-HC the number and types of health-insurance choices that are available. For this purpose, ARCOLA used the smaller but more detailed MEPS Household Component–Insurance Component linked file, which contained the needed information. The third step was to generate HSA premiums and benefit designs. The final step was to apply plan choice model coefficients to the MEPS data with premium information from independent sources such as eHealthInsurance.com. In this last step, care was taken to make sure modeling of the tax treatment of health insurance benefits reflected current regulations in order to get final estimates of take-up and subsidy costs. 

The econometric specification of the choice model driving the ARCOLA simulations took the form of a conditional logistic regression model. Here we consider utility to be a function of personal attributes such as health status, health-plan attributes (such as the out-of-pocket premium), and the interaction of premium and health status, formally stated as:

Uij = f(Zj,Yi,Xij), where i is the decision-making employee choosing among:

• j = health-plan choices
• Yi = employee personal attributes
• Zj = health-plan attributes
• Xij = interactions between alternative-specific constants and personal attributes

Because any plan attribute relying on employer data that was used in the plan-choice model also had to be available in the MEPS data, its key variables were:

• SCALEDPREM After-tax premium paid by the employee
• CLB The amount of money in the employee’s HRA, if any
• CUB The difference between the employee’s plan deductible and the HSA’s
• COIN Coinsurance rate
• CHRONIC Employee or dependent has a chronic illness=1, else 0
• AGE Employee’s age (years)
• FEM Employee’s gender (1=female, 0=male)
• FAM Employee has a two-person or family contract=1, else =0
• INC Employee’s annual wage income

Also included in the regression were alternative-specific constants (intercepts) for each of the possible health-plan choices. These intercepts are used to capture plan-specific features not represented by other identifiers of plan design. They are also included as interaction terms, along with age, gender, family status, and income.

The simulation model adjusts premiums according to the tax treatment of health insurance offered by employers in the group market. Specifically, premiums are adjusted to take into consideration the federal marginal tax rate as well as the Social Security tax burden. The capability to adjust for state-tax effects may exist but was not exercised in this model so that the pure effects of differences in insurance regulations by state could be identified.

ARCOLA, used previously by the Congressional Budget Office and by author Parente and his University of Minnesota colleague Roger Feldman in published peer-reviewed scholarly work,[41] was used in this report as well. It found that up to 816,000 individuals might buy into a reformed New York individual-insurance market.

Micro-Simulation from Glied et al. at Columbia University

Columbia University’s Sherry Glied provides distributions of costs across a wide range of population groups using cell-based data. In order to increase the precision of her team’s estimates, Glied pooled Current Population Survey data for New York State for 2004–06. With this increased sample size, Glied was able to estimate subpopulations by age and primary insurance coverage more precisely and to further sort the population by additional factors, such as work status, employer size, industry, and so on.

Estimates of national per-capita health expenditures stratified by age, gender, and type of service were obtained by using the 2004 MEPS-HC. To obtain New York State health-expenditure estimates that conform to the estimates from the Centers for Medicare and Medicaid Service’s (CMS) National Health Accounts, Glied adjusted the MEPS data using aggregate New York State–level expenditure data from CMS.

Relying on these baseline data, Glied then calculated the effects of each proposal for coverage and expenditures by the state on each population group (age, insurance coverage, poverty level, and so on). In addition, she applied assumptions and parameters uniformly to each applicable policy proposal, permitting attribution of most variations in the effects to the nature of the proposal. However, the variation—or outcome—being measured was coverage outcome, i.e., whether an individual obtained coverage. Glied did not examine the type of coverage offered to individuals, which may affect the decision of whether to purchase insurance.

The Urban Institute’s TRIM3 Model

The Transfer Income Model, version 3 (TRIM3), is a comprehensive micro-simulation model developed and maintained at the Urban Institute under primary funding from the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation (HHS/ASPE). TRIM3 simulates the major governmental tax, transfer, and health programs that affect the U.S. population, and it can produce results at the individual, family, state, and national levels. It is also a cell-based model but, like ARCOLA, is capable of making individual predictions. Since the first TRIM model became operational in 1973, TRIM models have been used to generate potential outcomes of public-policy changes in the areas of welfare reform, tax reform, national health-care reform, and so forth.

TRIM3’s annual “baseline” simulations (simulations of actual program rules) are used to correct for the underreporting of transfer program benefits in the survey files used as input to TRIM3 and to create other variables—such as program eligibility indicators—unavailable in the input data. Registered public users can access the micro-level variables produced by the model’s baseline simulations. They can also access the model’s historical library of program rules.

A baseline simulation applies the actual program rules that were in effect in a particular year to the input data for that year. A baseline simulation is performed for almost every year that the simulated programs examined. In simulating transfer programs, TRIM3 identifies units eligible for assistance under each program and selects participants who will enable it to match its administrative targets as to size and composition of the caseload. In simulating tax programs, TRIM3 calculates each unit’s taxes under the rules in effect in the year being simulated and assumes that the unit pays the full amount of taxes that are due.

Baseline simulations serve three main functions. First, they augment the input data (usually each year’s March CPS data) by creating additional micro-level variables that are not present in the input data. For example, the model adds variables that indicate whether a given individual or household is eligible for each of the main governmental transfer programs and whether a given tax unit is eligible for various types of tax credits. Second, the baseline simulations correct for the underreporting of transfer-program income that is prevalent in survey files. By simulating for each transfer program a caseload that matches the actual caseload in size and other key characteristics, the model creates a data set that can be used in place of the input data, when underreporting would pose a problem for a particular analysis. Third, baseline simulations serve as the comparison point for “alternative” simulations—simulations of proposed or hypothetical program rules.

ARCOLA in Comparison

The primary differences between the Columbia and ARCOLA micro-simulations are:

• ARCOLA uses 2006 insurance plan-choice information from data in which actual individual health-plan benefit choices have been made. The Glied model uses cells and data from 2004 for plan choices.
• ARCOLA interacts premium information with other household attributes to fashion the basic econometric model used for prediction.
• ARCOLA breaks down private insurance coverage to more granular levels by offering different benefit designs, including: HMOs; PPO high, medium, and low options; Health Savings Accounts; and Health Reimbursement Arrangements. The Glied model identifies private insurance only as a single coverage category. This difference may be significant because of the substantial premium and cost-sharing differences between benefit designs as expensive as high-option PPOs and as economical as HSAs and HMOs.
• The Glied model uses specific CPS estimates from New York State; the ARCOLA model does not. This is an asset of the Glied model. The ARCOLA model compensates by calibrating its baseline estimates to published New York data on the uninsured and direct-pay populations, as well as by using a 2009 New York–specific Zogby survey targeted at the individual-insurance market in the state and providing New York–specific statistical weights for common socio-demographic factors such as age and gender

The primary differences between ARCOLA and TRIM are:

• ARCOLA uses data on actual individual health-plan benefit choices; TRIM uses cell data. This means that ARCOLA can take a survey respondent and specifically assign him or her a probability based on a set of consumer attributes as well as a set of consumer-attribute interactions. A cell-based approach does not have to assign specific person-level probabilities and account for their interactions. This distinction is important because the individual effects of premium price and chronic illness might be quite different from the effects of chronic illness interacted with premium price. Someone with a chronic illness may be much more sensitive to premium price because he or she needs some sort of coverage. A regression model like ARCOLA will pick up these effects automatically.
• ARCOLA interacts premium information with other household attributes as part of the basic econometric model that it uses for prediction.

The Relative Utility of the ARCOLA, Columbia, and Urban Institute Models

Each of the models depends upon assumptions about the plan choices offered in the individual-insurance market. The primary difference between ARCOLA and the Columbia / Urban Institute models is the ARCOLA model’s ability to observe and predict the response of individuals faced with four types of private insurance contracts widely varying in premium price.

The Columbia / Urban Institute models are well suited to estimate the impact of reforms on public insurance programs. But we believe that the ARCOLA model is best suited for analyzing our area of interest: a private, unsubsidized individual-insurance market for New York. We believe that that model will equip policymakers to evaluate how reform of New York’s existing insurance regulations, which substantially increase costs in the individual-insurance market and limit the variety of insurance plans available, might positively affect New Yorkers’ decisions about whether to purchase insurance or to remain uninsured.

APPENDIX III

High-Risk-Pool Facts and Function

This paper shows that up to 782,600 more New Yorkers would buy individual health insurance in a market featuring greater choice and more flexible regulation—specifically, one no longer burdened by community rating and guaranteed issue and allowing the introduction of Health Savings Account–eligible plans.

However, a flexible and robust individual-insurance market is not necessarily open to all, inasmuch as not all individuals are “insurable,” particularly those with serious chronic illness that predates their application for insurance.

How Many People Could Be Refused Coverage in New York’s New, Flexible Market?

According to a recent study of 1.9 million individual-insurance applicants undergoing medical underwriting, 89 percent were offered coverage (with 79 percent of applicants offered coverage at or below standard rates and just 10 percent offered coverage at higher than standard rates). Of the 89 percent offered coverage, about one in twelve had a “condition waiver” or exclusion for a specified condition. One in twenty-five of those offered coverage faced a condition waiver and a higher premium.[42] In sum, 11 percent of all applicants were denied coverage, and up to another 10 percent faced higher premiums and/or condition waivers. Thus, some individuals seeking coverage in the individual market for high-cost medical conditions may require subsidies if they are to afford private health insurance.

To provide for the so-called uninsurables, most states with competitive individual-insurance markets have created subsidized high-risk insurance pools. If New York is to have a competitive individual-insurance market, it, too, should sponsor a high-risk pool for the excluded.

What Is a High-Risk Pool?

A high-risk pool is typically a state-chartered nonprofit that runs a health-insurance program designed to serve the medically “uninsurable” population by providing it access to affordable private insurance. It does so by subsidizing premiums. These subsidies are often financed by assessments of insurers.

According to the National Association of State Comprehensive Health Insurance Plans (NASCHIP), high-risk pools have two primary purposes: “they provide a means for guaranteed access to insurance, which enables individuals to protect themselves from catastrophic medical bills; and they are increasingly recognized for the role they play in helping to keep the individual insurance markets viable for companies to continue to compete in.”[43]

Who Is Served by a High-Risk Pool?

NASCHIP has collected actual cases of people who fell through the cracks:

Joanne’s Story

Joanne is a Redmond, Oregon, resident and a leukemia patient. She faced an uncertain future when she learned last April that her health coverage was ending the day before she was scheduled to enter the hospital for treatment.

Representatives from the Oregon Medical Insurance Pool (OMIP) enrolled Joanne in the state’s high-risk insurance pool. OMIP, which became operational in 1990, was designed for residents turned down by private insurers. She pays a monthly insurance premium of $577 for a plan with a $500 annual deductible and a maximum out-of-pocket expenditure of $1,000, after the deductible has been paid.

Currently, Joanne is waiting for a donor so that she can undergo a stem-cell transplant, which costs $250,000. If she had not been enrolled in Oregon’s high-risk pool, she could not have afforded it.

Betty’s Story

After being employed for fifteen years at various companies that provided health-insurance benefits, Betty became self-employed and lost her coverage. She applied for an individual health-insurance policy and was turned down because she had diabetes.

Fortunately, Betty lives in Minnesota, where there is a high-risk insurance pool: the Minnesota Comprehensive Health Association (MCHA). She secured coverage through MCHA and has been insured since 1988. She has MCHA’s federally qualified High Deductible Health Plan, which is a qualified plan for a Health Savings Account. Betty’s premium payment is $447.31 a month, and she must discharge a $3,000 annual deductible before she receives 100 percent coverage.

“Without MCHA, I am not sure what I would have done,” said Betty. “By having MCHA, I have been able to have peace of mind that I will not be financially ruined and have therefore been able to work and pay taxes. In those twenty-plus years, I have only missed a few days of work. The high-risk insurance pool insurance has permitted me to keep my diabetes under control.”[44]

How Large Are High-Risk Pools, and How Large Would New York’s Be?

According to 2007 data available from the Kaiser Family Foundation, thirty-three states with a high-risk pool open to new enrollees reported that an average of 1.9 percent of all those in the individual-insurance market had enrolled in the high-risk pool. (Specifically, 10.6 million individuals participate in the individual market in these states, 199,320 of whom inhabit the high-risk pools.) The percentages by state range from 0.5 percent to 8.4 percent.[45]

If it were to incorporate the reforms that we suggest, New York’s individual-insurance market could have up to 816,000 enrollees. Therefore, it is reasonable to assume that a New York high-risk pool would have between 15,520 (at 1.9 percent, the average size) and 68,616 enrollees (at 8.4 percent, the largest pool). Healthy NY, by comparison, covers 157,000 individuals.

Texas, with its 23.7 million residents, has 1.2 million individuals who obtain coverage in its individual-insurance market. Since 1998, Texas has operated a high-risk pool that had 27,733 individuals in it as of December 2007.[46]

How Expensive Are High-Risk Pools, and How Expensive Would New York’s Be?

Again, according to 2007 data available from the Kaiser Family Foundation, thirty-three states with a high-risk pool open to new enrollees reported costs for premium subsidies totaling $742.2 million to cover 199,320 lives. This works out to a subsidy of $3,742 per enrollee, reflecting a cost to the 10.6 million individuals in the individual-insurance markets of these states of $70 per individual, per annum, or $4.34 per individual if all 171 million people in these states were assessed. Minnesota, with the highest percentage of the individual market and the general population enrolled in these pools, reported a subsidy of $4,282 per enrollee in the high-risk pool; an assessment of $360 per enrollee in the individual market; and $24 per state resident.[47]

On the basis of the aforementioned information, if New York had a high-risk pool that was of average size and cost, it would need to raise over $58 million to underwrite its premium subsidy. This sum could be reached by imposing an assessment of about $6 per member, per month (PMPM) on those in the reformed individual-insurance market. If the assessment were extended to the 1.6 million in the small-group market, the per-month assessment would be just $2.

If New York had a high-risk pool that was the country’s largest and highest-cost, it would require over $453 million in premium subsidy, which could be collected by assessing those in the new individual market about $46 per month. If the assessment were extended to the 1.6 million in the small-group market, the per-month assessment would be just $15.63.

Healthy NY costs taxpayers $122 million a year, covers 155,000 people, and is estimated to reduce the proportion of uninsured by one percentage point. It is estimated that a New York high-risk pool would cost between $58 million and $453 million if it were to cover between 15,500 and 105,900 people and would be part of a policy reform agenda that would reduce the number of uninsured by up to 37 percent.

It is possible for any state to set up a high-risk pool. It should have a board of directors representing health care’s stakeholders: citizens, legislators, the insurance industry, employers, and the medical community. Following Minnesota’s example, the board would contract with insurers to cover plan participants. Insurers perhaps should be required to cover innovative programs designed to reduce individuals’ costs, such as disease-management programs, individual case management, and health and wellness programs. The high-risk program would offer a choice of competing plans charging premiums no greater than 150 percent of those charged by the comparable standard-risk plans sold in the state.

The high-risk pool would receive premiums, the proceeds of fines for insurers’ noncompliance, and other receipts associated with operation of the plans. Eligibility would result from outright denial of coverage or restriction of the applicant to coverage that falls well short of meeting the cost of a serious medical condition. The state should also establish timely procedures for confirming denial of coverage and making a referral to the high-risk pool. The process must allow for the adjudication of claims of inappropriate denial of coverage, which, if found to have occurred, should result in a fine.

What Criticisms or Problems Are Associated with High-Risk Pools?

A frequent criticism of high-risk pools is that the premiums charged in such pools, although capped, may still be a significant financial burden or simply unaffordable. (That is why we recommend capping premiums at 125–150 percent of standard market rates, with state funding to subsidize these lower premiums.) Also, when health-care costs reach a certain level, some states may begin restricting eligibility to their high-risk pools, limiting the time that individuals may remain the pool, capping claims payments, or increasing policyholders’ share of costs. Florida’s high-risk pool, for example, is supported by dollars from the state’s general funds and has been closed since 1991. Indeed, critics have argued that this form of funding is insufficiently stable and broad-based to meet the financial challenges of the chronically ill.

These are legitimate concerns. Policymakers should carefully consider disbursing additional subsidies to those individuals of limited means who do not qualify for public programs such as Medicare and Medicaid. But the long-term health of the individual market depends on well-functioning high-risk pools’ receiving adequate funding. As we have proposed here, funding for high-risk pools should come from a flexible PMPM assessment that can be adjusted to meet demand and not from general revenues, which are subject to wide fluctuations. Any other funding stream considered should be just as stable and long-term. Federal policymakers should also consider increasing funding of high-risk pools that meet minimum criteria of affordability, access, and program scope.

Endnotes

1. Tarren Bragdon, “Rx NY: A Prescription for More Accessible Health Care in NY,” Empire Center for New York State Policy (December 2007), p. 8. Available at: http://www.empirecenter.org/Documents/PDF/Rx_11_07.pdf (accessed July 2, 2009).
2. Ibid., p. 7.
3. Ibid., p. 8.
4. New York Department of Health, “Uninsured New Yorkers Urged to Sign Up for Health Insurance during ‘Cover the Uninsured Week’ ” (March 23, 2009). Available at:
   http://www.health.state.ny.us/press/releases/2009/2009-03-23_cover_the_uninsured_week.htm (accessed July 22, 2009).
5. As noted, there are 2.1 million uninsured in New York State, of which approximately 800,000 are eligible for public programs. However, if the policy changes modeled in our report are enacted, they would reduce the uninsured population, possibly allowing some of the eligible population to receive private, unsubsidized coverage.
6. ARCOLA is a micro-simulation model designed to estimate the impact of health-policy proposals at the federal and state level. The model predicts individual adult responses to proposed policy changes and suggests what the impact on the number of people covered and the premiums they pay would be if those changes were applied nationally. This model was first used by the Office of the Assistant Secretary of Planning and Evaluation (OASPE) of the Department of Health and Human Services (DHHS) to simulate the effect of the Medicare Modernization Act of 2003 (MMA) on take-up of high-deductible health plans in the individual health-insurance market (Feldman, Parente, Abraham, et al., 2005; Parente et al., “Final Technical Report for DHHS Contract HHSP233200400573P,” 2005). The model was later refined to incorporate the effect of prior health status on health-plan choice—a necessary step if one wants to predict enrollment more accurately. The latest model also used insurance expenditures derived from actual claims data to reset premium levels and then predicted their impact on take-up rates. The model then iterated consumers’ choice of plan until consumers got the plan they wanted at a price they could afford. A further change to the model permitted state-by-state predictions of consumer behavior per state and nationwide.
7. “New Study Shows Half of Individual Health Insurance Policy Holders Paid Less Than $130 per Month,” eHealth (November 12, 2008). New York is cited as the highest average individual-insurance premium, at $388 per month. Available at: http://www.marketwire.com/press-release/Ehealth-Inc-NASDAQ-EHTH-919521.html (accessed July 2, 2009).
8. Kaiser Family Foundation, “Individual Market Guaranteed Issue (Not Applicable to HIPAA Eligible Individuals), 2008” (data as of December 2008). Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=353&cat=7 (accessed July 2, 2009).
9. Kaiser Family Foundation, “Individual Market Rate Restrictions (Not Applicable to HIPAA Eligible Individuals), 2008” (data as of December 2008). Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=354&cat=7 (accessed July 2, 2009). New Jersey allows Basic and Essential individual insurance plans with premiums that vary based on age, gender, and geography at a rating band of 3.5:1. Carriers began offering these plans in New Jersey in March 2003.
10. Navigant Consulting, “2008 Annual Report on Healthy NY” (January 2009), pp. I-1 through I-4. Available at: http://www.ins.state.ny.us/website2/hny/reports/hnynav2008rep.pdf (accessed July 2, 2009).
11. Victoria Craig Bunce and J. P. Wieske, “Health Insurance Mandates in the States 2009,” Council for Affordable Health Insurance (May 2009). Available at: http://www.cahi.org/cahi_contents/resources/pdf/HealthInsuranceMandates2009.pdf (accessed July 2, 2009).
12. “Premium Rates for Standard Individual Health Plans: July 2009 for New York County,” New York State Insurance Department. Available at: http://www.ins.state.ny.us/hmorates/html/hmonewyo.htm (accessed July 2, 2009).
13. “N.Y. Gov. Pataki Signs Sole Proprietor Health Bill into Law,” Insurance Journal (September 23, 2002). Available at: http://www.insurancejournal.com/news/east/2002/09/23/23248.htm (accessed July 2, 2009).
14. “Cost and Benefits of Individual and Family Health Insurance Plans,” Forrester Consulting on behalf of eHealthInsurance.com (November 2008), p. 18. Available at: http://www.ehealthinsurance.com/content/expertcenterNew/eHealthCBreport2008FINAL.pdf (accessed July 2, 2009).
15. “Private Health Insurance: Millions Relying on Individual Market Face Cost and Coverage Trade-Offs,” General Accounting Office (November 1996), p. 71. Available at: http://www.gao.gov/archive/1997/he97008.pdf.
16. Geoffrey C. Sandler, “New York Individual Market Rules” (February 19, 2009), slide 4. Available at: http://www.nhpf.org/library/handouts/Sandler.slides_02-19-09.pdf (accessed July 2, 2009).
17. “Private Health Insurance: Millions Relying on Individual Market Face Cost and Coverage Trade-Offs,” p. 72.
18. Kaiser Family Foundation, “Health Insurance Coverage of Nonelderly 0–64, States (2006–2007), U.S. (2007).” Available at: http://www.statehealthfacts.org/comparetable.jsp?typ=2&ind=126&cat=3&sub=39 (accessed July 2, 2009).
19. California HealthCare Foundation, “Snapshot: California’s Uninsured 2008” (December 2008), p. 4. Available at: http://www.chcf.org/documents/insurance/UninsuredSnapshot08.pdf (accessed July 2, 2009).
20. State of New Jersey Department of Banking and Insurance, “Individual Health Coverage Program: Historical Comparison of Enrollment” (June 2, 2009). Available at:
    http://www.state.nj.us/dobi/division_insurance/ihcseh/enroll/1q09historical.pdf (accessed July 2, 2009).
21. Idem, “NJ Individual Health Coverage Program Buyer’s Guide.” Available at: http://www.state.nj.us/dobi/division_insurance/ihcseh/ihcguide/benefits.html#be (accessed July 2, 2009).
22. Idem, “Individual Health Coverage Program: Historical Comparison of Enrollment.” Available at: http://www.state.nj.us/dobi/division_insurance/ihcseh/enroll/2q09historical.pdf (accessed September 2, 2009)
23. There was a small number of people in hold-over plans from before Standard plans were first introduced.
24. Kaiser Family Foundation, “Total Number of Residents, States (2006–2007), U.S. (2007).” Available at: http://www.statehealthfacts.org/comparemaptable.jsp?ind=1&cat=1 (accessed July 22, 2009).
25. W. J. Congdon, A. Kowalski, and M. H. Showalter, “State Health Insurance Regulations and the Price of High-Deductible Policies,” unpublished manuscript (2005). Accessed at http://www.byu.edu.
26. J. Hadley and J. D. Reschovsky, “Health and the Cost of Nongroup Insurance,” Inquiry 40, no. 3 (2003): 235–53.
27. Congdon, Kowalski, and Showalter, “State Health Insurance Regulations and the Price of High-Deductible Policies.”
28. Hadley and Reschovsky, “Health and the Cost of Nongroup Insurance.”
29. Congdon, Kowalski, and Showalter, “State Health Insurance Regulations and the Price of High-Deductible Policies.”
30. Ibid.
31. J. W. Henderson et al., “Estimating the Impact of State Health Insurance Mandates on Premium Costs in the Individual Market Using the Community Tracking Survey,” Journal of Insurance Regulation (June 2009).
32. K. E. Thorpe et al., “Reducing the Number of Uninsured by Subsidizing Employment-Based Health Insurance: Results from a Pilot Study,” JAMA 267, no. 7 (1992): 945–48.
33. In 1944, in United States v. South-Eastern Underwriters Association, the Supreme Court held that while insurance companies are certainly engaged in interstate commerce, which confers federal jurisdiction, federal law need not preempt state regulation. In response, Congress soon passed the McCarran Ferguson Act in 1945, which explicitly granted to the states regulatory authority over the insurance industry, including the providers of health insurance, which the states had already been exercising. If policies were sold across state lines, the federal government would have jurisdiction over them, which the statute does not authorize. ERISA, enacted in 1973, and the Gramm-Leach-Bliley Act, enacted in 1999, curtailed the scope of McCarran Ferguson somewhat.
34. We used several secondary sources for this description, including Blue Cross / Blue Shield for state mandates; the Georgetown University Health Policy Institute for guaranteed issue and community rating; and Thomson-West’s Netscan / Health Policy Tracking Service (“Major Health Care Policies, 50 State Profiles, 2003/2004”) for any-willing-provider laws.
35. Although the New York market does not currently offer PPO plans (POS, or Point-of-Service plans, are offered), there are minimal differences in plan designs between PPOs and POSs, with most of the difference in premiums attributable to cost sharing of out-of-network services and the size of the base network of physicians and hospitals. However, many PPOs are effectively structured as POSs so that they can provide coverage for out-of-network services. Consequently, we have used baseline enrollment in existing POS plans for our PPO estimates. None of these differences, however, would affect our estimates of take-up rates or changes in the price of premiums.
36. It may seem surprising that consumers gravitate to plans with higher premiums. The ARCOLA model and our prior research show that, given a discount (as in this case), individuals prefer richer benefit plans (with lower out-of-pocket costs) even after lower-priced plans become available. This reflects the strong preference of most consumers for low cost-sharing plans, once they become more affordable as the result of the repeal of guaranteed issue and community rating.
37. S. T. Parente et al., “Consumer Response to a National Marketplace for Individual Health Insurance,” contract HP-07-024 from the Office of the Assistant Secretary of DHHS (submitted July 2008). Available at: http://www.hsinetwork.com/National_Marketplace_7-21-2008%20FINAL_Blind.pdf.
38. The response of the uninsured to a drop in the price of premiums resulting from a reduction in the number of mandates is nonlinear, i.e., a reduction in the number of mandates from forty to twenty has a bigger effect on take-up than a reduction in the number of mandates from twenty to zero.
39. Available at: http://www.nyshealthfoundation.org/userfiles/file/Microsoft%20Word%20-%20Analysis%20of%20Five%20Health%20Insurance%20Expansion%20Options%20for%20NYS_FINAL.pdf
40. Available at: http://www.urban.org/health_policy/url.cfm?ID=411925
41. R. Feldman et al., “Health Savings Accounts: Early Estimates of National Take-up from the 2003 Medicare Modernization Act and Future Policy Proposals,” Health Affairs (November 10, 2005).
42. AHIP Research, “Individual Health Insurance 2006–2007: A Comprehensive Survey of Premiums, Availability, and Benefits” (December 2007), pp. 4 and 11. Available at:
    http://www.ahipresearch.org/pdfs/Individual_Market_Survey_December_2007.pdf.
43. Lynn R. Gruber, “How State Health Insurance Pools Are Helping Americans: An Important Safety Net for Persons with Chronic Medical Conditions,” National Association of State Comprehensive Health Insurance Pools (January 6, 2009), p. 2. Available at: http://www.naschip.org/Position%20Paper%20NJRv25.pdf.
44. Ibid., pp. 4–5.
45. Kaiser Family Foundation, “State High Risk Pool Programs and Enrollment” (December 2007). Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=602&cat=7); idem, “State High Risk Pool Cost per Participant” (December 2007). Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=673&cat=7.
46. Ibid.
47. Ibid.

Bragdon, Tarren. (2007). “Rx NY: A Prescription for More Accessible Health Care.” Empire Center for New York State Policy. Available at: http://www.empirecenter.org/Documents/PDF/Rx_11_07.pdf.

Bunce, Victoria Craig, and J. P. Wieske. (2009). “Health Insurance Mandates in the States 2009.” Council for Affordable Health Insurance. Available at:
http://www.cahi.org/cahi_contents/resources/pdf/HealthInsuranceMandates2009.pdf.

California Healthcare Foundation. (2008). “Snapshot: California’s Uninsured 2008.” Available at:
http://www.chcf.org/documents/insurance/UninsuredSnapshot08.pdf.

Congdon, W. J., A. Kowalski, and M. H. Showalter. (2005). “State Health Insurance Regulations and the Price of High-Deductible Policies.” Unpublished manuscript. Accessed at http://www.byu.edu.

eHealthInsurance.com. (2008). “New Study Shows Half of Individual Health Insurance Policy Holders Paid Less Than $130 per Month.” Available at:
http://www.marketwire.com/press-release/Ehealth-Inc-NASDAQ-EHTH-919521.html.

Feldman, R., Parente, S.T., Abraham, J., Christianson, J., and Taylor, R. (2005). “Health Savings Accounts: Early Estimates of National Take-Up from the 2003 Medicare Modernization Act and Future Policy Proposals.” Health Affairs 26, no. 6: 1282–1591.

Forrester Consulting. (2008). “Cost and Benefits of Individual and Family Health Insurance Plans.” On behalf of eHealthInsurance.com. Available at:
http://www.ehealthinsurance.com/content/expertcenterNew/eHealthCBreport2008FINAL.pdf.

General Accounting Office. (1996). “Private Health Insurance: Millions Relying on Individual Market Face Cost and Coverage Trade-Offs.” Available at: http://www.gao.gov/archive/1997/he97008.pdf.

Glied, S., N. Tilipman, and O. Carrasquillo. (2009). “Analysis of Five Health Insurance Options for New York State.” Available at: http://www.nyshealthfoundation.org/userfiles/file/Microsoft%20Word%20-%20Analysis%20of%20Five%20Health%20Insurance%20Expansion%20Options%20for%20NYS_FINAL.pdf.

Gruber, Lynn R. (2009). “How State Health Insurance Pools Are Helping Americans: An Important Safety Net for Persons with Chronic Medical Conditions.” National Association of State Comprehensive Health Insurance Pools. Available at: http://www.naschip.org/Position%20Paper%20NJRv25.pdf.

Hadley, J., and J. D. Reschovsky. (2003). “Health and the Cost of Nongroup Insurance.” Inquiry 40, no. 3: 235–53.

Henderson, J. W., T. LaPierre, C. J. Conover, J. A. Seward, and B. A. Taylor. (2009). “Estimating the Impact of State Health Insurance Mandates on Premium Costs in the Individual Market Using the Community Tracking Survey.” Journal of Insurance Regulation (June).

Hersch, J. (2005). “A Workers’ Lobby to Provide Portable Benefits.” In Emerging Labor Market Institutions for the Twenty-First Century, 207–28. Ed. Richard B. Freeman, Joni Hersch, and Lawrence Mishel. Chicago: University of Chicago Press. Available at: http://www.nber.org.floyd.lib.umn.edu/chapters/c9955.

Insurance Journal. (2002). “N.Y. Gov. Pataki Signs Sole Proprietor Health Bill into Law.”September 22. Available at: http://www.insurancejournal.com/news/east/2002/09/23/23248.htm.

Kaiser Family Foundation. (2007a). “Health Insurance Coverage of Nonelderly 0–64, states (2006–2007), U.S.” Available at: http://www.statehealthfacts.org/comparetable.jsp?typ=2&ind=126&cat=3&sub=39.

———. (2007b). “State High Risk Pool Programs and Enrollment, December 2007.” Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=602&cat=7.

———. (2007c). “State High Risk Pool Cost per Participant, December 2007.” Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=673&cat=7.

———. (2007d). “Total Number of Residents, States (2006–2007), U.S. (2007).” Available at: http://www.statehealthfacts.org/comparemaptable.jsp?ind=1&cat=1.

———. (2008a). “Individual Market Guaranteed Issue (Not Applicable to HIPAA Eligible Individuals), 2008.” Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=353&cat=7.

———. (2008b). “Individual Market Rate Restrictions (Not Applicable to HIPAA Eligible Individuals), 2008.” Available at: http://www.statehealthfacts.org/comparetable.jsp?ind=354&cat=7.

Long, S. H., and M. S. Marquis. (2005). “State Efforts to Insure the Uninsured.” Rand Corporation report. Available at: http://www.rand.org/pubs/research_briefs/RB4558-1.

Lo Sasso, A. T., and I. Z. Lurie. (2009). “Community Rating and the Market for Private Non-Group Health Insurance.” Journal of Public Economics 93, nos. 1–2: 264–79.

Marquis, M. S., and S. H. Long. (2001). “Effects of ‘Second Generation’ Small Group Health Insurance Market Reforms, 1993 to 1997.” Inquiry 38, no. 4: 365–80.

Marquis, M. S., et al. (2004).“Subsidies and the Demand for Individual Health Insurance in California.” Health Services Research 39, no. 5: 1547–70.

Marquis, M. S., et al. (2007). “The Role of Product Design in Consumers’ Choices in the Individual Insurance Market.” Health Services Research 42, no. 6: 2194-2223.

Navigant Consulting. (2009). “2008 Annual Report on Healthy NY.” Pp. I-1–I-4. Available at: http://www.ins.state.ny.us/website2/hny/reports/hnynav2008rep.pdf.

New York Department of Health. (2009). “Uninsured New Yorkers Urged to Sign Up for Health Insurance During ‘Cover the Uninsured Week.’ ” Available at:
http://www.health.state.ny.us/press/releases/2009/2009-03-23_cover_the_uninsured_week.htm.

New York State Insurance Department. (2009). “Premium Rates for Standard Individual Health Plans: July 2009 for New York County.”Available at: http://www.ins.state.ny.us/hmorates/html/hmonewyo.htm.

New, M. J. (2006). “The Effect of State Regulations on Health Insurance Premiums: A Revised Analysis.” Center for Data Analysis Report #06-04, Heritage Foundation. Accessed at http://www.heritage.org.

Parente, S. T., et al. (2005). “Final Technical Report for DHHS Contract HHSP233200400573P.” Office of the Assistant Secretary of the Department of Health and Human Services.

Parente, S. T., R. Feldman, and J. Abraham. (2007). “Continuation of Research on Consumer-Directed Health Plans: HSA Simulation Model Refinement.” Final report on Contract #HHSP23320054301ER, Office of the Assistant Secretary of the Department of Health and Human Services.

Parente, S. T., R. Feldman, J. Abraham, and X. Yu. (2008). “Consumer Response to a National Marketplace for Individual Health Insurance.” Contract HP-07-024 from the Office of the Assistant Secretary of the Department of Health and Human Services. Submitted July 2008. Available at:
http://www.hsinetwork.com/National_Marketplace_7-21-2008%20FINAL_Blind.pdf.

Sandler, Geoffrey C. (2009). New York Individual Market Rules.” Available at: http://www.nhpf.org/library/handouts/Sandler.slides_02-19-09.pdf.

Simon, K. (2005). “Adverse Selection in Health Insurance Markets? Evidence from State Small-Group Health Insurance Reforms.” Journal of Public Economics 89, nos. 9–10: 1865–77.

State of New Jersey Department of Banking and Insurance. (2007). “Individual Health Insurance 2006–2007: A Comprehensive Survey of Premiums, Availability, and Benefits.” AHIP Research. December. Available at: http://www.ahipresearch.org/pdfs/Individual_Market_Survey_December_2007.pdf.

———. (2009a). “Individual Health Coverage Program: Historical Comparison of Enrollment.” June 2. Available at: http://www.state.nj.us/dobi/division_insurance/ihcseh/enroll/1q09historical.pdf.

———. (2009b). “The New Jersey Individual Health Coverage Program Buyer’s Guide.” Available at: http://www.state.nj.us/dobi/division_insurance/ihcseh/ihcguide/benefits.html#be.

Thorpe, K. E., A. Hendricks, D. Garnick, K. Donelan, and J. P. Newhouse. (1992). “Reducing the Number of Uninsured by Subsidizing Employment-Based Health Insurance: Results From a Pilot Study.” JAMA 267, no. 7: 945–48.

Urban Institute, on behalf of the New York State Department of Health and Insurance 2009. “Achieving Quality, Affordable Health Insurance for All New Yorkers: An Analysis of Reform Options.” Available at: http://www.urban.org/health_policy/url.cfm?ID=411925.

Thank you for visiting us. To receive a General Information Packet, please email support@manhattan-institute.org
and include your name and address in your e-mail message.

The Manhattan Institute, a 501(c)(3), is a think tank whose mission is to develop and disseminate new ideas
that foster greater economic choice and individual responsibility.

Copyright © 2013 Manhattan Institute for Policy Research, Inc. All rights reserved.

52 Vanderbilt Avenue, New York, N.Y. 10017
phone (212) 599-7000 / fax (212) 599-3494