Wednesday, September 12, 2012

Three Things You Need to Know about Today’s Census Uninsured Report

The Census Bureau released their annual report on poverty, income, and the uninsured this morning; the report is now online and can be found here.  Important facts you need to know from the health insurance section of the report:

  1. Uninsured Numbers Fall Slightly:  The total number of uninsured fell to 48.6 million, a decrease of just over 1.3 million compared to 2010’s level of 50.0 million uninsured.  The uninsured rate fell 0.6 percentage points to 15.7 percent.  It’s also worth noting that the coverage gains, while modest, were relatively consistent across all age, income, and demographic categories – so while the Administration will attempt to attribute most or all of the coverage gains to the under-26 insurance mandate, other factors were also clearly involved.
  2. Public Coverage Rises Significantly:  While the number of individuals with private insurance rose by about 800,000 individuals, with most of that increase coming from employer-based coverage, the number of individuals with government-provided health insurance rose appreciably.  Public insurance coverage increased by 3.9 million, led by an increase in Medicaid enrollment of 2.3 million.  The Medicare population also rose by about 2 million, likely reflecting both the retirement of the first Baby Boomers and an increase in disability claims due to the recession. (Note that some people can be enrolled in multiple public programs, explaining why the increase in public insurance coverage is less than the combined increase in Medicaid and Medicare coverage.)
  3. SCHIP Crowd-Out Grows:  Today’s report saw a continuation of existing trend, in which millions of children have lost private insurance coverage and gained coverage through government programs.  Since 2007, the number of children enrolled in private insurance has fallen by nearly 4 million, while the number of children enrolled in Medicaid (which also includes SCHIP) has risen by nearly 5.4 million, or more than 25%.  As a reminder, a significant expansion of SCHIP was enacted in February 2009.  While it’s unlikely that the entire migration from employer-sponsored coverage to Medicaid was caused by individuals voluntarily dropping out of employer coverage to enroll in government-sponsored coverage, it likely contributed to this ongoing migration.

Also of interest is a breakout of the various cohorts of the uninsured:

  • The number of non-citizens (both legal and illegal) without health insurance rose totaled 9.7 million.  Non-citizens comprise just under one-fifth of the total number of uninsured.
  • According to supplemental data, the number of uninsured individuals with household incomes under $25,000 totaled 19.1 million, or nearly 40% of the 48.6 million uninsured.  Many of these individuals may be eligible for public assistance through Medicaid and SCHIP.  Also, as noted in greater detail below, a significant number of these uninsured may in fact already be enrolled in public insurance programs, but their insurance status is not accurately reported by the Census data.)
  • The number of uninsured individuals with household incomes over $75,000 totaled 7.5 million, or just over 15% of the 48.6 million uninsured.  Many of these individuals may be able to obtain coverage on their own, but may choose not to do so if they do not consider the insurance policies offered to be of value to them.

Both a specific and a general caveat on the Census data.  First, the income-based data cited in the two bullets directly above may have reliability issues, particularly when compared to prior year reports.  For instance, last year’s Census report found 16.1 million uninsured in families with household incomes under $25,000.  For this particular group of uninsured to rise by 3 million (from 16.1 million in last year’s report to 19.1 million today), at a time when the uninsured population as a whole was falling by 1.3 million, seems contradictory.  This potential anomaly may explain why the income-based data were relegated to a supplemental spreadsheet on the Census website this year, rather than being placed in the report itself, as was the custom in prior years.

Second, and more generally, while the Census Bureau figure of uninsured Americans is among the most widely reported, it is far from the only measure used – or the most accurate.  Many indicators confirm that the Census survey represents a “point-in-time” snapshot of the uninsured population at any given moment, and does not reflect the number of individuals without insurance for long periods of time – those in most need of assistance.  For instance, while the Census report found 50.7 million uninsured in 2009, a separate study by the Centers for Disease Control found that 32.8 million Americans were uninsured for one year or longer in 2009, and a survey of health spending conducted by the Department of Health and Human Services found 41.3 million non-elderly Americans lacked coverage for all of 2009.  In 2009, the Census survey saw a larger jump in the number of uninsured than the other two reports, which could be a result of methodological flaws, and/or the fact that many of the uninsured lacked health coverage for periods of less than a year. (For a further discussion of these issues, see also a 2006 Kaiser Family Foundation brief comparing the uninsured surveys, as well as a 2003 CBO analysis of the long-term uninsured.)

It is also worth noting that the Census survey relies on individuals to self-report their insurance status, and some individuals may not remember periods of health insurance coverage.  Adding a “residual” question to the Census survey in 2000 – to confirm that those without employer, individual, or government coverage were in fact uninsured – reduced the number of uninsured Americans by 8 percent.  One survey conducted for the Department of Health and Human Services in 2005 adjusted for the number of individuals which the Centers for Medicare and Medicaid Services (CMS) reported were enrolled in Medicaid, but who did not report insurance coverage for the Census survey.  Such adjustments for the Medicaid undercount reduced the number of uninsured by about 9 million – or one-fifth of the total uninsured – and the number of uninsured children by half.  For these reasons, the Census Bureau report itself admits that “health insurance coverage is underreported [in the Census data] for a variety of reasons,” meaning that by Census’ own admission, the number of uninsured is lower than its report indicates.