Projecting the Effect of
the Recession on
Health Insurance Coverage for Nevadans
by Decision Analytics, Inc.
February 9, 2009
The Study of Nevada’s Uninsured report (2002 – 2008) is based on estimates reflecting the insurance situation prior to the difficulties in the financial system and the collapse of the housing market. An immediate question arises: What impact should we expect to see in 2009 and 2010?
In order to answer this question, we developed a limited time-series model relating some key drivers to the percentage of persons uninsured. We examined aggregate trends over time, demographic factors, poverty rates, unemployment rates, part-time working rates, and known “recessionary” periods (as defined by NBER), over the period 1987-1998. We used annual data for this effort. Because of the simple nature of our model, and the wide variety of uncertainty surrounding current economic conditions and Federal and state governments’ response to them, we refer to these results as “projections” rather than “forecasts.”
Some factors, such as demographic components (e.g., percent Hispanic, percent male aged 18-34) are known to correlate with uninsurance rates across different geographic locales. That is, they correlate over space. However, because they change relatively slowly relative to economic factors, they correlate less well with uninsurance over time. Thus, we assumed that demographic factors would be stable.
One important consideration when modeling effects over time is the existence of “lagged” effects. Put simply, an economic recession (affecting many factors in the economy) may take some time to appear in other data series. This is certainly true for unemployment, which is a “lagging indicator”; unemployment rates tend to increase approximately a year after a recession takes hold of the economy.
Uninsurance rates are similar to unemployment in this way: They appear to lag unemployment by approximately 18 months. (Notably, this corresponds to the period of “Cobra” coverage, should a former employee choose to use it.) Thus, unemployment is not an immediate driver of uninsurance; it takes some time for the impact to be felt.
Our limited modeling effort takes such lags into account. For the purpose of projecting 2009 and 2010 uninsurance rates, we developed three scenarios: A “main scenario,” which represents a middle ground projection of part time rates, unemployment rates, and poverty rates; a “high scenario,” which assumes that these rates will achieve near-historical-highs; and a “low scenario,” which assumes that these rates will not rise as high as the middle scenario.
In all cases, we assume that the recession of 2008 will continue into 2009 and at least partially into 2010. With the exception of recession indicators, the data came from the March Current Population Survey (CPS). Our most recent data are listed below:
|
|
Recent CPS data |
|
|
|
2007 |
2008 |
|
Percent male 18-34 |
12.7% |
11.3% |
|
Percent Hispanic |
23.2% |
24.1% |
|
Percent Part-Time Worker |
13.7% |
16.4% |
|
Percent in Poverty |
9.7% |
10.7% |
|
Percent Unemployed |
4.8% |
6.7% |
|
Recession Period |
No |
Yes |
|
Percent Uninsured |
17.4% |
17.5% |
Our scenario assumptions, and the projected uninsurance rates derived from these assumptions, are described in the following table:
|
|
Low Scenario |
Main Scenario |
High Scenario |
|||
|
|
2009 |
2010 |
2009 |
2010 |
2009 |
2010 |
|
Percent male 18-34 |
11% |
11% |
11% |
11% |
11% |
11%
|
|
Percent Hispanic |
24.5% |
24.5% |
24.5% |
24.5% |
24.5% |
24.5% |
|
Percent Part-Time Worker |
17% |
18% |
18% |
20% |
19% |
21% |
|
Percent in Poverty |
11% |
13% |
12% |
14% |
13% |
16% |
|
Percent Unemployed |
9% |
9% |
9% |
10% |
9% |
10% |
|
Recession Period |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
|
Projected Percent Uninsured |
18.3% |
20.4% |
18.4% |
20.6% |
18.5% |
20.7% |
How does this projection translate into numbers of persons uninsured? If we multiply the projected percent uninsured times the State Demographer’s projected Nevada total population[1], we obtain:
|
|
2009 |
2010 |
|
Population Projection |
2,878,288 |
2,963,812 |
|
Low Scenario |
526,727 |
604,618 |
|
Main Scenario |
529,605 |
610,545 |
|
High Scenario |
532,483 |
613,509 |
Thus, we project a range of 527,000 to 532,000 medically uninsured citizens in 2009, and a range of 605,000 to 614,000 in 2010, as compared to an estimate of 478,000 in 2008. This projection is, of course, dependent on the accuracy of the population projection that is here simply taken without changing any projection assumptions.
