Does Blue America Have A Violent Crime Problem?
Political pundits have often claimed that democrat run areas have more violent crime. Is this true? We dig into the numbers.
Introduction
Information has been circulated on the internet to suggest this like this graphic. On the campaign trail in 2020 then President Donald Trump tweeted the following:
But is this really the case? We dig into the data to find out if this is true, and to find out what it can tell us.
The Data
We collected violent crime data for over 10,000 cities from a variety of government and private sources, not just the FBI data as other’s have previously used. Our dataset details down to the exact number of incidents for the violent crimes of murder, rape, robbery, and assaults. The data covers not just one year’s worth of data but the last 10 years. In addition, we also have the presidential election county level voting results from 2020, not just who won but why how large of a margin. This data will come in useful as you’ll see.
Results
Looking at this chart, an average of crime over the last 10 years, it appears they do, and by a large margin. However, if we stopped here in our analysis we would be borderline negligent. This graphic does not account or the varying number of each party’s underlying population. Once we do this below, on a 100,000-person basis, we have a much better comparison.
We can see that, yes, democrat areas do have a slightly more crime, but not anything significant. What is interesting is the both the trend and makeup of crime.
Crime Trends
Here we can see that both Democrat and Republican areas have declining crime rates with Republican crime rates decreasing slightly more. And always remaining slightly lower.
One explanation for this difference is often that democratic areas tend to have more people. So we tested this below by breaking cities into 10 equally sized deciles based on population. Interestingly enough, below we do find this is true on average and for Republican areas. However, this is not true for Democrat areas. Based on this graphic, democrat areas actually tend to decrease in crime as they grow larger.
We were not fully convinced that this trend was true considering that 1) on average there is an increase and 2) there was a large amount of crime in 2009. We reran the same analysis without 2009 and found that yes, there was some uptick when population increased for Democrats, but not nearly as pronounced as Republicans. This suggests that this cannot be the majority of the cause for this difference in crime rates.
We also did this same analysis based on density rather than population. We found that as density increased crime tended to fall. While the reasoning for this is outside the scope of this analysis one hypothesis might be that policing is more scalable and efficient in a smaller area on a per person basis.
Differences in Crime
While gross levels of crime were similar the underlying crime make up was different between Democrat and Republican held areas. Republican areas tended to have more assaults and rapes per year, with assaults decreasing for Republicans but increasing for Democrats; rapes increasing for both Democrat areas and Republican areas. Democrats tended to have more murders and robberies per year with murders increasing and robberies decreasing for both Democrat areas and Republican areas.
But to What Degree?
We took our analysis even further and reviewed crime on a more granular scale. In the above analysis cities were assigned the party regardless of how large the margin between Democrat and Republican were.
We wanted to look at crime based on the percent of the vote won by each party. We did not expect to find that the Democrat/Republican ratio was not directly inverse considering the two-party nature of U.S. politics but wanted to ensure that the analysis was complete and looked at both. Here we were not surprised. In all cases the trends for each were opposite.
When we looked at total crime and percent of party vote we found that the larger the percent that voted democrat the larger the crime rate became.
In this more detailed and granular analysis murders and robberies increased as the percent of voters that voted democrat increased. This matches our earlier analysis as well.
With Republicans, we also found that rapes increased as the percent of voters that voted Republican increased. This also matches our earlier analysis.
The earlier and less granular analysis showed more and increasing assaults for Democrats, and more but decreasing assaults led by Republican areas. However, as the share of Democrats that won the vote increased so did assaults.
What About the Mix?
Until now we classified each as Republican or Democrat based on the which party won the area. We felt we should also consider how strongly the difference in margin won plays into crime rates. Here we found the most interesting finding of our analysis.
We found that as the homogeneous rating increased, crime decreased. Not exactly riveting, until you break it out by party. Here we found that Republican areas, as they increased in homogeneity or increased in diversity of politics crime rates remained the same. Democrat areas with low homogeneity or high diversity of politics experience higher crime rates and fall as homogeneity increases. Democrat areas with 70% of voters voting Democrat tended to have similar crime rates as Republican areas regardless of Republican homogeneity. Areas where democrats won less then 70% of the vote tended to have more crime. Democrat areas that won more than 70% of the vote tended to have less crime.
Overall, Democratic cities do not have significantly more crime, but there are some interesting underlying findings, both in the structure of crime between Democrat and Republican areas and in the relationship between crime rates and the homogeneity of the political vote. Splitting all areas into two categories, one for each party, did not show a difference in crime. However, the larger the share of a vote overall that was democratic, the more crime that area experienced.
