Notes on the 2014 ACS 5-Year Estimates Eric Guthrie, Michigan s State Demographer December 3, 2015 The U.S. Census Bureau has released the 2014 American Community Survey (ACS) 5-year estimates. The 5-year ACS estimates represent the most current and comprehensive dataset available describing a wide range of socioeconomic and housing characteristics, which allow community, state, and business leaders to make better, data-driven decisions. Since the 5-year estimates have a larger sample size than the 1- year estimates, there is not a size limit on the geographies which can be accessed, which means that these estimates contain data for all counties, cities, towns, villages and other minor civil divisions in Michigan. The 2014 5-year estimates, which cover the 2010 to 2014 period, are the first set of 5-year estimates that have a non-overlapping predecessor with which comparisons can be made since the program s start in 2005. Like all ACS estimates, the 2014 5-year estimates are a period estimate, which means they describe socioeconomic characteristics over a span of time rather than a specific point. The Census Bureau recommends that comparisons not be made between datasets that have overlapping periods, which mean that this is the first set with a period that does not cross over the first 5-year estimates series produced for the period 2005 to 2009. This release establishes the new cycle envisioned with the creation of the ACS estimates which is two points in a 10-year period for every geography that is contained in the ACS data. This is a significant improvement from when there was only access to data at one point every 10 years, as was the case with data from the decennial censuses. In addition to being able to make comparisons between two periods in the 5-year estimates, this dataset is now the only place where data for geographies with a population under 65,000 can be found outside of the population figures that will be produced through the decennial census program. Previously the U.S. Census Bureau produced the 3-year estimates, which contained estimates for geographies with populations between 20,000 and 65,000, but in response to shrinking federal budgets, that product has been discontinued. The table below looks at the geographies in the 5-year estimates. Geography Number in the 5-year Estimates Unique to the 5-year Estimates Place (City, Town, CDP, Village) 692 676 Township 1,240 1,234 County 83 54 Zip Code 989 989 School District (Elementary and Unified) 550 524 From this table we can see that most of the areas that are contained in the 5-year estimates are not reported elsewhere in the ACS data. The difference between the geographies contained in the 5-year estimates and those unique to the 5-year estimates are the data contained in the 1-year estimates. It is interesting that while the majority of people in Michigan can find themselves in the 1-year ACS data, either through their city or county, most geographic areas that people are concerned with are not
contained anywhere except the 5-year estimates. This highlights the urban nature of the majority of our population and also the value in being able to have a 5-year estimates series with which we can see changes over time. The list above contains the geographies that I am asked about most frequently and is not an exhaustive list of the geographies contained in the ACS data. Poverty in Michigan One topic that is always of interest when discussing new ACS statistics in Michigan is the percent of persons experiencing poverty. There are many sources for poverty statistics, but none allow for the analysis at the small geographic and population levels that is afforded by the ACS. The first map below looks at the percent of persons who were in poverty in the previous twelve months over the period 2010-2014. The second map looks at the percentage point difference between the 2005-2009 and the 2010-2014 estimates. The second map also indicates which percentage point differences were significant at the 90 percent level. One striking point to be gleaned from the second map is that every county that had a statistically significant difference between the two periods had an increase in the percentage of persons in poverty.
Given these county-level data, it is not a surprise that poverty in the state overall has increased between the two periods, moving 2.4 percentage points from 14.5 in the 2005-2009 period to 16.9 in the 2010-2014 period. Median Household Income A topic that also garners much attention when new data are released is household income. The first map below looks at the median household income in the various counties. The top three counties in the 2010-2014 period were Livingston, Oakland, and Washtenaw, while the counties with the lowest median incomes were Lake, Oscoda, and Clare. These rankings are consistent with past trends. A possible area of concern may be how the 2005-2009 period compares to the 2010-2014 period when we are talking about real incomes. When incomes are adjusted to 2014 dollars, it is apparent that incomes in all but two counties have gone down between the two periods. This can be seen in the second map below. Houghton and Crawford counties saw very small increases in median household income between the periods. This matches trends at the state level where a decrease of over $6,000 in 2014 dollars was observed.
Educational Attainment The percentage of the population 25 years and older with a bachelor s degree or higher is also a key metric that people track and are interested in with a new data release. The first map below looks at that population in the various counties. We can see that Washtenaw, Livingston, and Leelanau are at the top of this group. Other areas with above average educational attainment are either more urban, have significant institutions of higher education, or both. The second map below looks at the percentage point difference between the periods for the percent of the 25-year and older population with a bachelor s degree or higher. While 29 counties saw a statistically significant increase in the percentage of the population in question, no county that had a statistically significant decrease in this population. The data show that 69 counties recorded an increase while 14 had a decrease in the population 25 and older with a bachelor s degree or higher. This is also consistent with the trend in the state overall, which registered a 1.9 percentage point increase in the population 25 and older with a bachelor s degree or higher.
My discussions above have been at the county level due to the time and space constraints of this format, but part of the real value of the 5-year estimates product is its ability to delve in to the smallest geographies in the state. All of the analysis that I have provided above can also be done at the all of the geographic levels I mention in the first table and even smaller areas down to the block group level. These data and a great deal more are now available on the Census Bureau s website, www.census.gov. Easy links to a variety of census data are also available on our website, www.milmi.org. As always, I am here to assist you with your demographic data needs. I can be reached at guthriee@michigan.gov or 517-241-8667.