***This is a guest post from MAPPhd***
An earlier discussion posed on this blog questioned which parts of the country have the highest poverty levels, thereby indicating some remediation by some outside entity (I personally believe that people should take care of each other, not governments. People have hearts, not governments. Besides, God told us to care of each other, but did not tell governments that.) It is widely believed that places like Mississippi, Tennessee, West Virginia, and rural areas have the highest poverty levels. There are some fascinating data on the US government web site that pertains to this question.
In particular, I have been analyzing data from www.census.gov/compendia/statab/ using the 2012 Statistical Abstract. Most poverty level maps are displayed based on the absolute value of income or on absolute median income data. Those displays do not take into account the cost of living in various parts of the country. I have lived in six states, lived in two countries (US and UK), visited all 50 US states, visited 36 countries, and have good friends from another two dozen countries around the world. Cost of living does count. It costs more to live in San Francisco than in Houston, more in the urban center of Houston than in the suburbs, and more in the Houston suburbs than in the Hill Country of Texas.
In this article, I have used the 2009 median income data by state and have used the 2009 urban center cost of living data. These are the latest data available for both. My analysis leaves the median income data untouched. However, to obtain a cost of living index for each state, I grouped the urban center data by states. Within each state, I calculated the mean, standard deviation, and standard error of the cost of living index (an index of 100 indicates an average cost of living). This calculated mean cost of living index by state was used to normalize the median income by state across the US. The standard deviation and standard error were used to determine the variance in the data. Well over half of the standard errors were less than 4%. However, the standard error for New York was 30%, due largely because it costs a lot of money to live in Manhattan (index of 217), but Buffalo is relatively cheap at 96. I will have more on this observation at the end of the article.
I then split the US into three regions, which turns out to be closely correlated to the latest presidential election results. Table I shows this split.
Table II shows the 2009 average median income as reported in the 2012 Statistical Abstract and displayed on most maps compared with the average median income normalized for cost of living index (COLI) by state by the method described above.
If one simply looks at the average median income, it appears as if Region #1 (West Coast) and Region #3 (East Coast and north central) have median incomes more than $8,000 dollars higher than Region #2 (i.e., rural Midwest and southern states). An obvious conclusion is that the rural and southern states areas (#2) are much poorer. However, once the COLI is accounted for, then it is pretty much of a wash between the regions. In addition, the data in the 2009 urban center cost of living index do not take into account state and local income and sales taxes. Note that Regions #1 and #3 are dominated by states with high state and local income taxes, while this is not common in Region #2. Hence, if taxes were accounted for, Regions #1 and #3 would be reduced even further in the Normalized Median Income relative to Region #2.
Another way of looking at the data is to compare the change in ranking for a state based on the Median Income versus its ranking based on the Med Inc Normalized by COLI. To do this exercise, I ranked each state by its Median Income, then its Normalized value, and subtracted the Median Income from the Normalized value. Hence, negative numbers mean that the normalization moved the state up higher in the ranking after accounting for the COLI, while positive numbers mean that the ranking dropped. Plot #1 below shows the change in ranking for each region. Most of the states in Region #2 tend to move higher, and demonstrate a clear tendency to outshine those in the other two regions. The key exception is Alaska, which has a very high cost of living. An interesting observation is that Hawaii ranks number 5 in Median Income at $64098 and ranks number 51 (since DC is included) in Normalized Med Inc by COLI at $38674. Therefore, it plots in Region #1 and a Change in Ranking of 46. Based solely on Median Income, living in Hawaii is desirable, as it is in the Top 10% of all of the states. However, it is the lowest when normalizing by COLI. In effect, Hawaii could be considered the poorest state because their cost of living is so very high. The average change in ranking (negative numbers indicate movement upward in rank when normalized) for each Region was +3 for #3, -7 for #2, and +10 for #1.
Plot #2 shows the basic data for Median Income and for Normalized Median Income by COLI by state. The three tranches from left to right are for Region #3, Region #2, and Region #1. If anything, Regions #1 and #3 have a greater disparity in Median Incomes between the richest and poorest states than does Region #2. For Region #3, the highest and lowest Median Income states drop significantly when they are normalized. Nearly all of the Region #2 states increase their Normalized values relative to the Median Incomes. For Region #1, nearly half of the states have their Normalized values drop significantly, while the other states see little change relative to the Median Incomes. Clearly, the Normalized values for each region are relatively uniformly distributed across common trend lines, showing that on a normalized basis each region is roughly equal.
Finally, Plot #3 shows the relative ranking for each state for the Median Income, the Normalized Median Income by COLI, and the COLI using a similar format to Plot #2. Whereas Region #2 contains 9 of the 10 poorest states based on Median Income, all Regions are generally evenly distributed once the COLI is used for the normalization.
As a final comment, it would be interesting to split the US into regions based on the counties rather than states. For example, much of the western half of New York is dominantly rural and generally supported Romney. As we saw above, living in Buffalo is much less costly than living in Manhattan. My suspicion is that the results above would be even more strongly confirmed, and the standard errors in my analysis would be much less. However, those data appear not to be available.
***This is a guest post from MAPPhd***