We recently received an inquiry from a Seniors Count! community partner, asking:

Interesting questions, all. But the answer to even such a simple question as the first can be peppered with caveats. The definitive population count for every place in the United States comes from the Decennial Census, but as the name suggests, that only occurs every ten years. Between those times, estimates are produced by a number of different sources. For example, the American Community Survey, also a U.S. Census Bureau product, uses its sample to extrapolate a population estimate. The better estimates, however, come from the Census' Population Estimates Program. Here are their estimates in answer to question a above:

So there we have our population estimates in answer to question a. Question b is a much trickier question, as is any question involving projections into the future. The answers to these questions always rest on a large number of assumptions, and their accuracy ultimately depends on the correctness of those assumptions. For example, in- and out-migration assumptions made to predict future population when economic conditions in an area are good may differ significantly from those made in times of economic struggle. When subdividing the population by criteria such as age, even more tenuous assumptions must be made. So the big caveat behind all projections is that they are only as good as the assumptions upon which they rest.

The Census Bureau makes population projections by age group, but only does so on a state-by-state basis and not at the county level. More info on the Census Bureau's state population projections is available here:

http://www.census.gov/population/www/projections/stproj.html

The State of Michigan does not produce its own projections, but the Southeast Michigan Council of Governments (SEMCOG), another Seniors Count! community partner, does produce projections for the seven counties under its purview. SEMCOG's projections are the best county-level projections available, but they are somewhat out-of-date and times have changed since they were produced. SEMCOG's projections overestimate the region's current population, probably because the economic downturn of the last couple of years was not anticipated and therefore not accounted for in their development. However, they are probably the best available projections of anticipated growth in different segments of the population at county level. SEMCOG projects that between 2010 and 2015 the 65+ population will grow 15.335% in Macomb County, 17.389% in Monroe, 21.058% in Oakland, and 10.876% in Wayne. This represents an overall growth rate of about 15% in the older four-county population. More details on the SEMCOG projections are here:

http://www.semcog.org/Data/Apps/regional.forecast.data.cfm?mcd=8999 (Pull down the menu at the top of the page to see the data for different counties.)

The U.S. Census Bureau projects growth in the 65+ population in the entire State of Michigan to be about 13% over the same period. See details here:

http://www.census.gov/population/www/projections/projectionsagesex.html

However, considering the significant variation by county, the danger of committing an ecological fallacy by applying state projections to counties is too great and we had better base our growth estimates on the SEMCOG projected growth rate (although not raw numbers, since SEMCOG's 2010 projections differ so significantly from the Census' more accurate 2009 estimates). Keep in mind that all such projections vary significantly depending upon the assumptions on which they are based, and it is foolish to put too much stock in the absolute accuracy of any of them. Predicting the future involves nearly as much art as it does science, and sometimes the best available projection has to be good enough. It is advisable to favor the Census Bureau's numbers over others, estimates over projections, and projections based on more geographically appropriate assumptions over those applied from larger areas to smaller areas. Therefore, I would tend to use the Census 2009 population estimate as the base, and apply SEMCOG's five-year growth rate pro-rated for six years to get very rough 2015 population projections for those aged 65 and older. Which gives us:

Which brings us to question c, another tricky question. Don't forget, there are 4 "parts" to Medicare, and the coverage rates between them vary dramatically. According to the 2009 Statistical Supplement of the Social Security Administration Office of Retirement and Disability Policy, 95% of people age 65+ are eligible for Medicare (http://www.ssa.gov/policy/docs/statcomps/supplement/2009/medicare.html). It would be best if we had access to CMS data on actual Medicare enrollment in our four counties, but those data are not public and we currently do not have access to them (something for the wish list, perhaps?). The same publication says that 37.5 million people age 65+ had Medicare Part A coverage in 2008, and since the Census estimates the 65+ population at that time was 38,799,891, that would indicate that about 96.65% of older people were covered by Medicare Hospital Insurance. We can drill down further to the state level using Table 8.B3a, which shows that in 2008, there were approximately 1,298,000 Michigan residents age 65 or over enrolled in Medicare Part A and/or Part B (http://www.ssa.gov/policy/docs/statcomps/supplement/2009/8b.html#table8.b3a). Considering the Census estimate of 1,304,322 for the age 65+ population in Michigan in 2008 (http://www.census.gov/popest/states/asrh/SC-EST2008-01.html), this represents a coverage rate of about 99.5%. How do we reconcile these different enrollment rates? I find it helpful to grasp such a dilemma by both horns; let's calculate a likely range of the increase future Medicare beneficiaries by county, using 95% enrollment rate as our floor and a 99.5% rate as our ceiling. We can expect Medicare coverage rates to stay relatively stable in the absence of significant statutory changes, which would be impossible to predict anyway, so we assume the growth rate of Medicare recipients to be about the same as the growth rate of the older population. Therefore:

So, in answer to question c, what does all this tell us? It tells us that if all of our assumptions are correct, and if all of the assumptions underlying the estimates we employ are correct, there were probably around half a million to 525,000 older people on Medicare Parts A and B in the four-county area in 2009. If SEMCOG's population projections and the way we've employed them here are correct, there will be about 90,000 to 95,000 more older Medicare recipients in 2015, including around 37,000 to 39,000 more in Oakland County alone. According to our rough back-of-the-envelope calculation, there will most likely be upwards of 600,000 four-county area residents age 65 and over on Medicare Parts A and B by 2015.

Until we get better enrollment data, more accurate population estimates and projections, or a better crystal ball, we'll have to rely on phrases such as "probably," "about," "approximately," "roughly," "most likely," and others when discussing these projections.