This post notes that, while international comparisons of governmental corruption draw attention to the federal government, much corruption can take place on state and local levels. Scholars have developed several ways of conceptualizing state-level corruption, and have arrived at somewhat divergent lists of most- and least-corrupt states. It seems the most stringent state-level definition of corruption might combine an analysis of state laws (to see whether various safeguards are in place) and corresponding state outcomes (to verify that those safeguards are working). It does not appear that any American states have distinguished themselves with that degree of commitment to clean government.
The following discussion notes that certain factors seem to be associated with corrupt state and local government. Such factors can include latitude, health outcomes, and population. None of these seems to deliver a simple and direct indicator of corruption. For example, the link between population and corruption can depend on population density, the distance between the state capital and population centers, the growth of deviant subcultures, neighborhood networks, and the existence of journalists or other watchdogs taking an active interest in government at state and local levels.
Due to such variations, two counties next or near to each other can achieve significantly divergent health outcomes not predictable from population or income alone. Corruption, as a possible cause of such divergence, might not be easy to verify, as many local jurisdictions seem to generate complaints of corruption, not all of which necessarily have merit. The following discussion illustrates some of these points with a look at the state corruption ranking of Kansas, and with health outcome comparisons among several contiguous Kansas counties.
In a previous post, I reviewed competing theories and concluded that, by international standards, the U.S. appears to be fairly corrupt. That study did not distinguish federal from state or local governmental corruption. It seemed, though, that American states might vary considerably in this regard. To verify that, I ran a search. That led to several studies:
- Using numbers of convictions of public authorities under federal corruption laws, Liu and Mikesell (2014, p. 351), cited by multiple sources (e.g., Fortune; Huffington Post), found that the least corrupt state was Oregon, followed by Washington, Minnesota, Nebraska, Iowa, Vermont, Utah, New Hampshire, Colorado, and Kansas, while the most corrupt state in the U.S. during the 1976-2008 period was Mississippi, followed by Louisiana, Tennessee, Illinois, Pennsylvania, Alabama, Alaska, South Dakota, Kentucky, and Florida.
- A Harvard study (Dincer & Johnston, 2014) likewise examined public corruption convictions as well as a survey of state political reporters, separated out among the three branches of government, to conclude that the most corrupt states include Alabama, Georgia, Illinois, Kentucky, New Mexico, New Jersey, and Pennsylvania, while the least corrupt states include the Dakotas, Idaho, Massachusetts, Michigan, Oregon, Vermont, and Wyoming.
- A State Integrity Investigation conducted by the Center for Public Integrity (Kusnetz, 2015) found that 47 of the 50 states scored D+ or worse, and 11 of those 47 flunked (in the traditional American grading system in which A+, A, and A- are best, followed by B, C, D, and finally F for Failing). This study looked for laws and procedures that could help to prevent corruption; the study did not examine the link between laws on the books and actual corrupt behaviors in any state (see Enten, 2015). The three states scoring relatively well in this study were Alaska (with a C), California (C-), and Connecticut (C-). States graded F were Delaware, Kansas, Louisiana, Maine, Michigan, Nevada, Oklahoma, Oregon, Pennsylvania, South Dakota, and Wyoming.
I decided to start with the lists compiled by the first two sources (i.e., by Liu & Mikesell, and by Dincer & Johnston). The combined list of most corrupt states, named by at least one of those two sources, included Alabama, Alaska, Florida, Georgia, Illinois, Kentucky, Louisiana, Mississippi, New Jersey, New Mexico, Pennsylvania, South Dakota, and Tennessee. The combined list of least corrupt states included Colorado, Idaho, Iowa, Kansas, Massachusetts, Michigan, Minnesota, Nebraska, New Hampshire, North Dakota, Oregon, South Dakota, Utah, Vermont, Washington, and Wyoming. The states boldfaced in those lists were mentioned by both of those sources.
The following discussion refers repeatedly to those combined lists of most and least corrupt states. Yet those lists are not without problems. The fact that the two scholarly sources agreed on so few states, and directly disagreed in the case of South Dakota, suggested that the sources were using somewhat divergent measures of corruption. It seemed the sources were trying, but were achieving only mixed success, in pinning down a multifacted and difficult-to-track phenomenon.
Reading through information provided by the State Integrity Investigation, I came to the general impression that corrupt influence pervaded state governments in the United States. It was not that every state office or agency was completely corrupt; it was that, with variations from one state to another, a person might not have great confidence that a given state office or agency was clean. In that sense, the low grades issued by the State Integrity Investigation were perhaps the most credible general indicator of state-level corruption. It seemed that a considerable share of America’s mediocre standing in international comparisons (above) could be due to state- as distinct from federal-level corruption.
Factors Contributing to Corruption
The foregoing studies identified the presence of corruption by looking at corruption convictions of public officials, reporters’ ratings of state governments, and safeguards written into law and practice. But these were not the only ways in which one might measure corruption.
One alternative way of detecting corruption came to mind as I reviewed an international study linking a nation’s level of corruption to the health of its people (“Sick Regimes and Sick People,” Witvliet et al., 2013, pp. 1245-1246). It appeared that corruption could result in lower governmental expenditures on public goods (see Campante and Do, 2012), meaning that not only health care but also other factors contributing to good health (e.g., education) could suffer. The money that might have improved health care would instead go to preferred projects and/or to the pockets of certain well-placed individuals.
There seemed to be some potential in the approach of using state-by-state health comparisons as a proxy for governmental corruption. For example, UnitedHealth Foundation (2015, p. 8) provided a ranking of U.S. states in which four of the most corrupt states listed above appeared in the bottom ten states healthwise, whereas none of the least corrupt states listed above were in that bottom ten; and among the ten healthiest states, I found none of the most corrupt states, but seven of the least corrupt. If corruption did not directly cause poor health outcomes, it appeared it might at least share some of the same roots.
There were other factors that also seemed to have some linkage to state corruption. One, mentioned in passing here, is that of latitude. For some reason (due, perhaps, to different cultures, or to the influence of a colder climate), a majority of the most corrupt states were in the South, while most of the least corrupt states were in the North. Alaska, as an exception, might owe much of its corruption to the concentration of mineral (especially oil) wealth.
Size also seemed to matter. Of the ten largest states by population, six appeared in the foregoing list of most corrupt states, while only two of the ten states with the smallest populations appeared on that list. Conversely, only one of the ten most populous states appeared on the list of least corrupt states, while four of the least populous states were on that list, and several other least corrupt states were only slightly higher on the population list. Population density (i.e., persons per square mile) might also make a difference: five of the ten least dense states were on the list (above) of least corrupt states and, again, several others were just a bit higher on the population density list, whereas only two of the ten least densely populated states appeared in the list of most corrupt states.
The link between population and corruption did not appear to be simple. For instance, Campante and Do (2012, Table 1) found that states whose capitals were relatively distant from their population centers tended to have higher levels of corruption. To cite several examples from the foregoing lists of most and least corrupt states, there is considerable distance between the capitals and largest cities of Illinois (Springfield vs. Chicago) and Florida (Tallahassee vs. Miami), but no distance in Massachusetts (Boston is both the capital and the largest city) and Colorado (Denver). Campante and Do suggested that a population centered around the capital tended to have higher voter turnout in state elections, and that the newspapers in such states tended to pay more attention to state politics. There seemed to be less corruption, on the state level, when there were more people watching.
Corruption in Cities and Counties
As described in another post, I had lately become aware of the trust that certain American militia gangs placed in the county sheriff, as distinct from state government. That trust was evidently frustrated when county sheriffs failed to support militia gang activities. But such trust — from militia gangsters or from anyone else in the community — would also be undeserved where city and county governments abused their power.
Corruption and other abuses would be unsurprising where economics and politics of a corrupt state are dominated by a corrupt city. For example, Illinois stands out, in the foregoing list, as the home of Chicago, notorious for its presence in current and historical accounts of American corruption (e.g., Fitzgerald, 2015; Economist, 2015; Huffington Post, 2012). Other major cities that might unsurprisingly appear on a rogues’ list (not all of which dominate their home states) include Detroit, Newark, Philadelphia, New Orleans, New York, Miami, and Las Vegas (e.g., Simpson et al., 2015; Junkins, 2011).
Size seemed to be a risk factor for corruption, not only in big states and major metropolitan areas, but also as an environmental concern for smaller cities. For example, City Mayors (2014) offered a list of mayors who had been indicted and convicted of various crimes while in office. Most of the mayors on that list were from smaller cities located near major population centers (e.g., Cicero, IL; Hoboken, NJ). The same seemed to be true of other cities and towns that generated headlines due to corruption scandals. Examples included Hampton, FL; Bell, CA; and Ferguson, MO (O’Neill, 2014; Nolan, 2010; Gordon, 2015).
Consistent with the observation (above) that researchers disagreed on which states were most corrupt, I gathered that corruption could be defined in different ways, to include different kinds of behavior. In this sense, corruption seemed to be like other kinds of crime. That raised the question of what one could learn about corruption from crime research. My findings, on brief review:
- Chamlin and Cochran (2004) cited several theories that could account for a link between crime and population. One theory was that population growth increases the frequency of negative interpersonal contacts that could result in crime, while decreasing the frequency of positive contacts that could inhibit crime. Another theory held that larger population means more, and greater tolerance of, deviant subcultures.
- Chang et al. (2013, p. 7) indicated that many studies have found a direct link between population size and crime rates, though a few have found a negative relationship between population density and crime rates.
- Hipp and Roussell (2013) and Hipp et al. (2013) found that network ties among neighborhood residents did predict crime levels within neighborhoods, and that the effects of population varied according to the type of crime. For example, robberies (typically involving strangers) did vary according to the population of the surrounding region, while homicides (often occurring among non-strangers) did not.
It seemed, then, that types of corruption, like other forms of crime, might vary according to the absolute population size, the relative population density, and neighborhood social capital. It seemed there could be less corruption in parts of a larger city and more corruption in a more remote or less networked place, if the latter lacked watchdogs. The City Mayors (2014) webpage did list some recent indictments and convictions taking place well outside of major metropolitan areas.
If health outcomes could be a proxy for corruption on the state and national levels (above), it seemed that one could also suspect corruption in some counties with poor health. So, for example, despite its size, California was not on the foregoing list of most corrupt states, but California’s relatively rural central district did rank high in corruption convictions (Lee, 2012). Similarly, I noticed that the boundaries of two California counties — Lake County and Marin County — are only 30 miles apart, and yet the two were near opposite extremes in terms of health outcomes, according to the Population Health Institute at the University of Wisconsin (2015). Lake County had far more premature deaths, twice the rate of poor or fair health, and about 50% more days of poor physical or mental health. Corruption did appear to be a concern in Lake County.
Generally, a search suggested that many county and small town governments across the country had encountered various forms of corruption. As worst-case scenarios, Jensen (2013) told stories of a county welfare agent who created fictitious welfare recipients and scammed the government for hundreds of thousands of dollars in a central Montana county with a total population of only 2,168; a county sheriff in southeastern Kentucky who confiscated and resold guns and drugs for a $350,000 profit; and a treasurer for a town of 16,000 people in western Illinois who embezzled $53 million. It appeared that, in some such cases, people were able to get away with these schemes, for as long as they did, precisely because the place was small and there was hardly anyone in a position to document and do anything about them.
A Case Study: Kansas
As noted in another post, I spent some time in Kansas. For those who are not sure, I can confirm that this was actually rather different from what I had experienced during my years in New York and Los Angeles. Compared to life in those places, small-town Kansas was clean, calm, and considerate. Nonetheless, Kansas had mixed results in the state-level rankings of corruption cited above:
- Liu and Mikesell (2014, p. 348) placed Kansas in their top ten list of least corrupt states.
- Dincer and Johnston (2014) distinguished legal from illegal corruption: the difference was whether the politician’s vote or influence was being bought with campaign contributions (or other legal benefits to the politician) rather than with illegal bribes. Illegal corruption was rated as only slightly common in all three branches of the Kansas state government (but note that a “slightly common” rating in the judicial branch put Kansas among the 16 most corrupt states in that regard). Legal corruption was rated as very common in the executive and legislative branches, and slightly common in the judicial branch.
- Kansas was among the eleven states receiving an F grade in the State Integrity Investigation conducted by the Center for Public Integrity (Kusnetz, 2015). This F grade included failing grades in a number of components. For example, Kansas was ranked 43rd among the states in the area of Judicial Accountability because, among other things, Kansas state judges are not required to give reasons for their decisions; Kansas was ranked 46th in State Budget Processes because, for example, the state does not publish a citizen budget providing budget information in terms that lay readers can understand; Kansas ranked 49th in Procurement because, among other things, the law does not require competitive bidding, companies guilty of bribery are not barred from participating in future bids, and the results of major procurement bids are not reviewable in detail by the public; and Kansas ranked 50th in Internal Auditing because, among other things, the state’s internal auditing agency does not independently initiate its own investigations.
Kansas was also noteworthy, at this writing, because its governor, Sam Brownback, had instituted a relatively radical (some said ideologically driven) campaign to cut taxes, and so far this campaign had produced poor results for the state. In light of the factors just listed, this development raised the question of whether Campante and Do (2012) would consider Topeka, the Kansas state capital, to be relatively isolated from population centers, and thus somewhat insulated from critical scrutiny. Wichita, the largest city in Kansas, was 140 miles away; Kansas City, KS was an hour to the east; and perhaps neither was large and feisty enough to support aggressive journalism that might draw substantial voter attention to state government.
I looked into county-by-county variations in corruption, using as a possible proxy the map of health factors and outcomes provided by the Population Health Institute (2015). Among other things, I was puzzled to see that Barton County (county seat: Great Bend, 25 miles from where I lived) was ranked 86th among the state’s 105 counties, while Ellsworth County on its east side and Ellis County at its northwest corner were ranked 24th and 11th, respectively. Compared to Ellis County, Barton County had a 64% higher rate of premature death, 23% higher violent crime, twice the rate of teen births, 55% more poor mental health days, and 31% more poor physical health days — despite having somewhat more primary care physicians per capita.
It appeared that something was not going well in Barton County. It, and some other counties in the Population Health Institute’s map, seemed to be at least partially surrounded by other counties that were doing better. No doubt some of those contrasts were due to differences in economic opportunity. But that did not seem to be a factor in this case. The U.S. Census Bureau webpages indicated that median household incomes in recent years were fairly similar: $44,981 in Barton County, $43,085 in Ellis County, and $45,865 in Ellsworth County. Ellsworth was smaller, but the populations of the other two were nearly the same. Nonetheless, there were some demographic differences. At 6.7% of total population, Barton had nearly three times as many foreign-born persons (mostly Hispanic, I suspected) as the two other counties. At about 20%, Ellsworth and Barton had a noticeably lower rate of bachelor’s degree holders than did Ellis. Wikipedia notes that “Ellis County is an anomaly in western Kansas, having voted several times for Democratic presidential candidates, even when the vast majority of the state’s 105 counties went for the Republican nominee.”
It was possible that the inferior health outcomes experienced in Barton County were due to politics, demographics, or something other than corruption. In addition, I was reminded of Hanlon’s Razor, which urges that “one should never attribute to malice that which is adequately explained by stupidity.” I thought that perhaps there could be a similar aphorism involving corruption and ineptitude or just plain bad luck. It seemed, that is, that the postulated link between corruption and expenditures on public goods (above) might not always provide the best explanation of poor county health outcomes.
But this is not to deny corruption in Barton County. A search led to reports of a Great Bend banker indicted on charges of embezzlement and money laundering; a Great Bend mayor‘s home being purchased by a corporation to which the city had given “hundreds of thousands of dollars in tax incentives”; a corruption complaint against a judge hearing a case in Great Bend; and a Facebook page objecting to corruption in the Great Bend family law court. There would surely have been more if I had dug further. I did not investigate the merits of any of these materials. I had begun to gather that most communities tended to generate some such complaints and materials. I doubted all were true; I doubted all were false.
Over time, as I traveled and lived in various spots around the state, I came to see that governmental corruption (and, for that matter, ineptitude), in a place like small-town Kansas, is not what I might have expected. From national news reports and dramatizations, a person could get the impression that corruption triggers lawsuits, protests, and arrests. But there were not going to be any police confrontations or mass demonstrations out in these little Kansas towns, nor were there going to be many sensational news reports uncovering massive fraud. I met some surprisingly positive and proactive law enforcement people out there. But I also met one or two bad cops, and they scared me. Because if that kind of cop decided to go after me, nobody was going to be there to protect me. The public trust was more responsibility than some people could handle, and that did not seem likely to change.
The public response to corruption in many places seemed to reside primarily, not in news reports and prosecutions, but rather in gossip and memory. People could trot out all kinds of stories about things they had seen and heard. Some of the stories were shocking — about ongoing shakedown operations, about knowingly false prosecutions. For the most part, there would be no justice in these cases. The social fabric and professional pride did often appear to instill expectations of appropriate behavior, but even in settled times those consumed with self-advancement or personal preference could sometimes find ways to circumvent the ordinary controls.