A Review of the Latest Census Data: Are Housing Values Keeping up with Inflation?

Rising housing prices are drawing the attention of the media, homeowners, and home shoppers across the country. In North Carolina, there is a lot of talk about low housing inventories, accelerating prices, and a lack of affordable housing in urban, suburban, and rural areas. 

It takes time to assemble the data, do analysis, and determine actual numbers to quantify what we see and hear anecdotally. Outside of the major cities, it can take months or years to get data on all of NC’s counties. While we await the publication of 2020 and 2021 home values by the U.S. Census Bureau, we can examine trends leading up to the COVID-19 pandemic to see what has happened across the state over the last decade. This will add some perspective to the current condition. 

To fully understand what has happened to housing values over the last decade, we must consider inflation. To do this, the median value estimates were analyzed using owner-occupied housing data from data.census.gov (specifically the U.S. Census Bureau’s American Community Survey, Table B25077). Note that the Census Bureau releases 1-year estimates for large geographies such as major metro areas, states, and the U.S., but to identify conditions for all 100 NC counties, five-year estimates is the best data available.

Statewide Results

Looking at the Census 5-year estimates for North Carolina, median housing values have indeed increased after adjustment for inflation. The increase from two five-year estimates has been modest between 1.2% and 1.3%. 

This increase may not be as large as you might expect for several reasons. First, as addressed above, the most recent information available to analyze is 2019 data. When the 2020 and 2021 estimates become available, we may see a more significant increase. Second, we are using five-year estimates that include housing values during the housing peak in 2006 – 2008. Home values in NC didn’t really start to decrease until 2010 and took a while to recover from the depths of the Great Recession.

To test the second theory and see if we can get a better idea of when housing values decreased, the Census 1-year estimates starting in 2006 and ending in 2019 were reviewed. This data showed that median housing values, after being adjusted for inflation, increased year-over-year from 2006 till 2008, held steady in 2009, and fell from 2010 until 2012. The median value held steady in 2013 and 2014 and began to increase starting in 2015. It was not until 2018 when values returned to the peak experienced in 2009. The chart below shows a graphical representation of the roller coaster ride North Carolina median housing values experienced over the 14-year period.

Looking at Counties

An analysis of the 5-year estimate ranges for each of the state’s 100 counties yielded an interesting finding. After adjusting for inflation, median home values for sixty-six counties remained steady (no statistical difference) from 2010 to 2019. During the same period, seventeen counties showed a statistically significant decrease while another seventeen counties saw an increase.

As you can see from the list and map above, the counties that showed statistically significant growth in housing values are concentrated in the mountains, the Charlotte area, and the Triangle. The average increase in these counties varied tremendously from as much as 30% in Mitchell County to 4% in Gaston County. Only two counties with higher median housing values – Montgomery and Mitchell -- did not experience population growth between 2010 and 2019. Another interesting fact is that 14 of these counties are located in either a Metropolitan Statistical Area or a Micropolitan Statistical Area.

Looking at the counties with lower values, 12 of the 17 are in the eastern part of the state with three located in the Piedmont Triad. The average decrease in these counties ranged from nearly 25% in Dare County to 5% in Guilford County. County population growth was seen in 10 of the 17 counties. The more populous counties, such as Guilford; New Hanover; and Forsyth, saw prices decrease the least. The counties located in the northeast and west, specifically Dare; Avery; Washington; Camden; Pasquotank; and Macon Counties, saw the most significant decreases.

Conclusion

Evaluating housing values using Census data is a tricky proposition. While it may not necessarily provide info about the current housing market conditions, it does help shed some light on the trends.

1. Upon review of the 5-year data, 66 of NC’s 100 counties experienced no significant change in median home values after being adjusted for inflation over the past 9 years.
2. Changes in housing value in a county with a smaller population is likely to be more extreme compared to a county with a larger population.
3. While it makes sense that housing values would, for the most part, increase in counties that are experiencing the most growth in population, statistical analysis shows a moderate correlation. This suggests there are other factors affecting housing values beside population growth. The chart below plots the 34 counties with a statistically significant change in median housing values against population increases. This shows that most counties with increasing housing values – 15 out of the 17 -- are also experiencing an increase in population.
4. The 1-year Census data shows that median home values – on average statewide – have trended up since 2015 after a 5-year period of price decreases and stagnation. The 5-year data will take longer to show the significant growth we already see in the one-year data. Further analysis using the 2020 ACS 5-year estimate when it becomes available will be necessary to see how growth trend reflects across the state. 

Methodology

A 5-year estimate is an average of Census surveys over a 5-year period. The 2010 estimate includes data collected in 2006, 2007, 2008, 2009 and 2010. A single estimate is provided by the Census Bureau along with a margin of error indicating there is 90% confidence that the true value lies in a range equal to the estimated value plus or minus the given margin of error. For this analysis, the indicated margin of error was added and subtracted from the estimate to determine the 90% interval for each county’s median housing value as well as a statewide estimate. The 2006-2010 estimate range was adjusted to 2019 dollars using an index of the average of the 2006-2010 and 2015-2019 Personal Consumptions Expenditures Price Index from the U.S. Bureau of Economic Analysis.

Note that the 2020 ACS 5-year estimates are not released yet. The 2019 data is the latest data available for this analysis.

Median home value estimates were assigned one of three labels: lower, higher, or steady. A steady label was given to counties with no statistical significance between the estimates over that period. If an estimate was labeled as lower, the 90% confidence interval for the 2019 estimate was lower than the 90% confidence interval for the 2010 estimate. If the estimate value was labeled as higher, the 90% confidence interval for the 2019 estimate was higher than the 90% confidence interval for the 2010 estimate.