Go to HHH home page.
Smart Politics
 


Will Violent Crime in Minnesota Increase Along With the Jobless Rate?

Bookmark and Share

The news released on Thursday by the Minnesota Department of Employment and Economic Development that Minnesota’s unemployment rate had reached 6.4 percent, prompted Smart Politics to examine how the current jobless trend, and specifically the large rate of job loss, is unprecedented, going back decades in the state.

The problems of increased unemployment, of course, don’t end at the unemployment line. Job loss is perhaps but the tip of an iceberg of other subsequent social problems.

For example, with more and more people out of work, one might logically expect the state to be more vulnerable to higher incidents of violent crime as a result. (The reason being that a loss of jobs in an economic downturn creates more financial hardships, so more people will turn to, say, robbery as a solution for these hardships: robbery, one of the four violent crime measures compiled by the F.B.I. for its violent crime index, has historically comprised about one-third to one-half of all violent crimes committed each year in the state. The other three measures are murder and nonnegligent manslaughter, forcible rape, and aggravated assault).

Such has not been the case in the Gopher State, however. In fact, the numbers are quite startling.

A Smart Politics analysis of trends in unemployment and violent crime rates since 1976 finds there is no positive correlation between the two social ills. In fact, a bivariate correlation conducted of these variables finds they are negatively related (-.553, significant at the .01 level). In other words, increases in the unemployment rate are associated with decreases in the violent crime rate in Minnesota, and vice-versa. Even after a 1-year time lag was added to the model (as crime would not necessarily occur in time with job layoffs), the results were virtually identical (-.564, significant at the .01 level).

If this sounds topsy-turvy, there is some data here that makes sense: unemployment data does, in fact, have a more intuitive association with trends in Minnesota’s gross domestic product (i.e. economic growth). Looking back at yearly trends in both measures to 1976 finds the variables are negatively correlated (-.541, significant at the .01 level). In other words, increases in the rate of GDP are correlated with decreases in unemployment, and vice-versa. Which makes sense: more growth, more jobs, less people out of work. (GDP was curiouslypositively associated with violent crime in Minnesota: .508, significant at the .01 level).

So what happens when these three variables interact together?

Smart Politics conducted a linear regression analysis, using Minnesota’s unemployment and Gross Domestic Product rates as independent variables and violent crime as the dependent variable. The results, surprising as they may be, find that since 1976 an increase of one percentage point in the jobless rate causes a drop in the violent crime rate of 15 incidents per 100,000 people, holding Minnesota’s GDP constant. Thrity-seven percent (R Square = .367) of the variation in violent crime in this model is explained by changes in unemployment and GDP (only inserting unemployment into the regression model explains thirty-one percent of the variation in crime). The model as a whole is highly significant (at the .001 level).

One would expect that if there was not a causal, positive relationship between these variables then there simply might be no relationship at all; many violent crimes (except robbery) are not borne out of problems due to economics and the loss of one's job, but deep-rooted social and psychological ills. But a negative relationship between unemployment and violent crime – that is something altogether unexpected.

Why would violent crime rates drop during periods of economic hardships? Is Minnesota’s social services system that strong to catch people in its net while they are falling? Or is it a case of Minnesota neighbors helping neighbors in their time of greatest need?

Stay tuned for further research at Smart Politics to examine these issues. Coming next: an analysis of property crime and unemployment data.

Minnesota Unemployment, Violent Crime, and GDP, 1976-2007

Year
Unemployment rate
Violent Crime rate
Gross Domestic Product
2007
4.6
288.7
254,970
2006
4.1
312.0
242,095
2005
4.2
297.3
232,001
2004
4.6
269.8
223,454
2003
4.9
262.9
208,179
2002
4.5
267.2
198,558
2001
3.9
263.7
190,231
2000
3.1
280.8
185,093
1999
2.8
274.0
172,874
1998
2.7
310.2
164,897
1997
3.3
337.8
155,938
1996
3.9
338.8
141,664
1995
3.7
356.1
131,357
1994
4.1
359.0
124,733
1993
4.9
327.2
114,946
1992
5.1
338.0
111,919
1991
5.2
316.0
103,791
1990
4.8
306.1
100,327
1989
4.3
288.3
96,165
1988
4.3
290.1
89,991
1987
5.1
285.4
83,947
1986
5.6
284.6
78,194
1985
6.0
256.4
74,808
1984
6.3
211.5
70,296
1983
8.0
190.9
61,029
1982
8.1
219.3
56,952
1981
5.7
228.5
54,966
1980
5.9
227.8
49,717
1979
4.3
221.0
46,353
1978
4.0
189.6
41,126
1977
5.3
193.8
36,380
1976
5.9
189.0
32,610

Minnesota Unemployment Rate: seasonally adjusted (Source: data compiled from the Minnesota Department of Employment and Economic Development).
Minnesota Violent Crime Rate: per 100,000 inhabitants (Source: FBI Uniform Crime Reports).
Minnesota Gross Domestic Product: in millions of current dollars, all industries total (Source: Bureau of Economic Analysis, U.S. Department of Commerce).



Previous post: Minnesota Unemployment Rate Reaches Highest Level in Nearly A Quarter Century
Next post: Getting to 15 Percent: Two Different Paths for Dean Barkley (’08) and Tim Penny (’02)

6 Comments


  • I actually did a post on this subject recently here. The SAFE officers I spoke with mentioned that robberies drop because there's less cash available, making the crime less worthwhile since the criminal is less likely to score a significant amount.

    However, burglaries are NOT included in the violent crime index, and those are on the rise right now. Prostitution is another crime that tends to experience upswings in hard economic times.

    Lastly, I think the exercise is interesting, but there's some oversimplification going on here. Take the '01-'03 recession, for example. Crime was highest during the recovery period -- '05-'06. However, crime tends to lag recessions, especially in modern times when state and municipal budgets get cut as a result of deficits. In '05-'06, the crime rate skyrocketing was a direct result of public safety budgets getting slashed. And there's a much more direct correlation there than with unemployment.

    I'd be curious to see if that holds true in other years as well, or if it's something that has only come into play in the last decade or two.

  • > However, crime tends to lag recessions, especially in
    > modern times when state and municipal budgets get cut as > a result of deficits.

    Right - that's why I also ran the #s with a 1-year time lag as well. Moreover, even without the lag, these numbers are aggregated yearly, which should give us more of a chance to see the impact of joblessness and the economy on crime, as opposed to using the month-to-month data which is also available.


  • I'm thinking the 1 year lag may not be enough, however. The recession officially ended around late '03/early '04, for example, but '06 had the highest violent crime rate (and property crime rate) for that era. And I tend to believe that it's less tied to unemployment dropping than it is a shortage of cops/enforcement/prevention programs over the two or three year period combined. Budget allocations for public safety and human services really only started catching up in '06/'07.

  • > And I tend to believe that it's less tied to unemployment
    > dropping than it is a shortage of cops/enforcement/prevention
    > programs over the two or three year period combined.

    Ah, I see. There is no doubt this initial model i put forth is far from explaining the multi-varied reasons for the ebbs and flows of crime. Many long reports, dissertations, and books have been written about that. And that is why i raised the issue (3rd to last paragraph) that some would feel there should be little relationship of unemployment to crime. What this blog highlights is a counterintuitive negative relationship; and i guess what you are suggesting is that this result is either a random by-product of some odd data, or odd use of data, that simply the data hasn't been lagged properly. In no way was this entry meant, however, to provide a definitive model as to the causes of the fluctuation in crime rates.

  • No question, and I didn't think that's how you presented it. I just find it to be an interesting discussion. Obviously, since I took the time to write about it (albeit in a less mathematical fashion) over at Secrets of the City.

  • MN saw a large population increase in the 1990s due to people coming here for social services. As a result it is only natural that you will get more individuals who might do crime based on population growth alone. The violent crime rate for much of the 1990s somewhat backs this idea up.
    It is possible the reason the crime rate isn't so high the last eight years is because people arrested for crime in the 90s are serving their prison sentences. This could also explain a surge of crime one year and a drop a few years later.
    Any increase in future crime should be compared to incarceration and/or prison release trends and not solely economic changes.

  • Leave a comment


    Remains of the Data

    Gender Equality in the US House: A State-by State Quarter-Century Report Card (1989-2014)

    A study of 5,325 congressional elections finds the number of female U.S. Representatives has more than tripled over the last 25 years, but the rate at which women are elected to the chamber still varies greatly between the states.

    Political Crumbs

    Small Club in St. Paul

    Mark Dayton is one of just three Minnesotans ever elected to three different statewide offices. Dayton, of course, had previously served as State Auditor (1991-1995) and U.S. Senator (2001-2007) before winning the governorship in 2010. At that time, he joined Republicans Edward Thye and J.A.A. Burnquist on this very short list. Burnquist was elected governor in 1914 but then became governor after the death of Democrat Winfield Hammond in 1915. He then won the gubernatorial elections of 1916 and 1918 and eight terms as attorney general two decades later (1939-1955). Thye was similarly first elected lieutenant governor of the Gopher State and became governor after the resignation of fellow GOPer Harold Stasson in 1943. Thye won one additional full term as governor in 1944 and then two terms to the U.S. Senate (1947-1959). Twenty Minnesotans have been elected to two different statewide offices.


    Respect Your Elders?

    With retirement announcements this year by veteran U.S. Representatives such as 30-term Democrat John Dingell of Michigan, 20-term Democrat George Miller of California, and 18-term Republican Tom Petri of Wisconsin, it is no surprise that retirees from the 113th Congress are one of the most experienced cohorts in recent decades. Overall, these 24 exiting members of the House have served an average of 11.0 terms - the second longest tenure among retirees across the last 18 cycles since 1980. Only the U.S. Representatives retiring in 2006 had more experience, averaging 11.9 terms. (In that cycle, 10 of the 11 retiring members served at least 10 terms, with GOPer Bill Jenkins of Tennessee the lone exception at just five). Even without the aforementioned Dingell, the average length of service in the chamber of the remaining 23 retirees in 2014 is 10.2 terms - which would still be the third highest since 1980 behind the 2006 and 2012 (10.5 terms) cycles.


    more POLITICAL CRUMBS

    Humphrey School Sites
    CSPG
    Humphrey New Media Hub

    Issues />

<div id=
    Abortion
    Afghanistan
    Budget and taxes
    Campaign finances
    Crime and punishment
    Economy and jobs
    Education
    Energy
    Environment
    Foreign affairs
    Gender
    Health
    Housing
    Ideology
    Immigration
    Iraq
    Media
    Military
    Partisanship
    Race and ethnicity
    Reapportionment
    Redistricting
    Religion
    Sexuality
    Sports
    Terrorism
    Third parties
    Transportation
    Voting