With support from the University of Richmond

History News Network

History News Network puts current events into historical perspective. Subscribe to our newsletter for new perspectives on the ways history continues to resonate in the present. Explore our archive of thousands of original op-eds and curated stories from around the web. Join us to learn more about the past, now.

Why Voting for the GOP is Bad for Your Health: Interview with Dr. James F. Gilligan


Photo Credit: Flickr/Edward Allen L. Lim.

Dr. James F. Gilligan is the author of Violence: Reflections on a National Epidemic and PreventingViolence: An Agenda for the Coming Century. Having formerly headed the Institute of Law and Psychiatry in addition to directing mental health programs for the Massachusetts prison system while on the faculty of the Harvard Medical Scholl of Law and Psychiatry, Dr. Gilligan is currently a Collegiate Professor in the College of Arts and Science, Clinical Professor of Psychiatry in the School of Medicine, and Adjunct Professor in the School of Law at New York University.

His latest book is Why Some Politicians Are More Dangerous Than Others, which argued that there is a direct statistical correlation between murder and suicide rates and the party of the incumbent president. A Republican president, he writes, "increases the rates of suicide and homicide."

This is a condensed version of an email interview I conducted with Dr. Gilligan.

Critics of your research will say that you had a predetermined outcome and cherry picked your numbers to fit. How do you respond to this type of criticism?

Asking a scholar that question is a bit like asking someone "When did you stop beating your wife?," which I understand used to be a popular question in the old vaudeville circuit. Nevertheless, I am glad that you asked it, because it will give me an opportunity to explain the mistake made by those critics who are applying the well-known formula, "if you don’t like the facts but can’t refute them, attack the methodology."

"Cherry-picking" presumably means selecting only a certain subset of data (which gives you the result you want) from a larger data-base (which would give you a different result, which you do not want). Thus I could be accused of "cherry-picking" if I had not selected the largest data-base possible with which to answer the question I was asking, namely: has there been a statistically significant difference in the increases and decreases in age-adjusted annual rates of violent death during the presidential administrations of the two political parties? But the data-base on annual violent deaths that I selected was in fact the largest one possible -- namely, the entire record of age-adjusted deaths per 100,000 population from the earliest year in which these were tabulated on a yearly basis (1900) to the last year for which such data had been reported as of the time in 2011 when my book manuscript was published (2007).

But I went even further than that in my effort to include all possible relevant data in my examination of the records of the two political parties. Appendix A in my book explains in detail how I tabulated not just one but all four of the possible sets of data on the yearly increases and decreases in rates of death from homicide, suicide, and total violent deaths (meaning the sum of the suicide and homicide rates) from 1900 to 2007, during the administrations of each party, in order to avoid the possibility of bias in any of them, and to allow readers to compare the results obtained from all of them.

How did you decide what you would count as a violent crime in your research?

I did not count anything as a violent crime, because I was not studying crime, I was studying violence. That is, I was approaching violent death as a problem in public health and preventive medicine, not crime, which is a legal category. Violence and crime are two different (and only partially overlapping) subjects. Most crime is not violent (property crimes and violations of the drug laws are much more common than violent crimes), and most violence is not a crime (e.g., suicide, war, capital punishment, most police shootings). In fact, suicide is considerably more common than homicide in the U.S. and other developed countries, and while it is a violent death, it is not a crime (though it used to be). Only a minority of violent deaths are considered to be the result of violent crimes, and labeled "murder" or "manslaughter". "Homicide," as opposed to "murder," "manslaughter" or "execution," is a category in medicine and vital statistics. From a medical standpoint, it is merely one among the many possible causes of death, to be documented by a physician on a death certificate, whether it is inflicted by a "criminal," a prison warden, a police officer, a soldier, or anyone else. Thus, only some homicides count as violent crimes. To have studied only criminal homicides would itself have been a form of "cherry-picking," if one’s purpose is (as mine was) to determine how violent or non-violent our society is at any given time.

Of the twenty-eight presidential elections since 1900, fifteen have gone to Republicans and thirteen to Democrats. Why do Americans vote the way they do?

When voters vote for candidates who are elected but then do not support the voters’ own interests, I can only conclude that the voters have made a mistake (in their own terms). However, what voters consider their own interests varies so widely that to answer this question would require a wide-ranging discussion beyond the space limitations we have here. Possibly some voters place a greater value on issues like abortion or gay marriage, for example, than they do on becoming employed or having health insurance. In that case, even if the person they elected leaves them unemployed and uninsured, they might consider that to be a price worth paying, for the sake of achieving their larger goals, if the candidate, once elected, succeeds in banning abortion and gay marriage. In that case, I would not presume to say that they were mistaken, but rather, that they had decided that there were other things that were more important to them than money and health. And if that happens to them and they then change their minds, and decide that money and health are in fact essential, they may vote for a different candidate, or one from a different party, the next time.

However, when the voters themselves feel that they have made a mistake, we may well wonder why they made it. Sometimes they may decide that they have been "duped" by the candidate’s political rhetoric or advertising.

Lyndon Johnson described what he called the "Bourbon strategy" in the South ("Bourbon" referring to the white ruling class, not the whiskey). He said that it was to the political and economic advantage of the Bourbons for racial discrimination to continue to exist, because as long as it did, the poor whites would have people even poorer and of lower status in the nation’s social hierarchy than they were (i.e., poor blacks) whom they could look down on, and to whom they could feel superior – the implication being that that boost to their self-esteem would distract them from noticing that they too were being discriminated against (by the Bourbons, who would benefit financially and politically as long as that strategy worked). Whether that means the poor whites were voting "against their own interests" would depend on what they considered to be the interests they cared most about. But clearly they were voting in a way that worked to the benefit of the economic and political interests of the Bourbons.

There appears to be a significant drop in violent deaths per 100,000 people per year during the first four years of the Reagan administration (followed by a sharp rise during his second term on into the first Bush administration). What do you think accounted for this drop?

I take it for granted that violence is multi-determined. I believe it can only be understood as a product of the interaction between multiple biological, psychological and social forces, including, among the latter, sociological, cultural/anthropological, historical, economic and political forces. No one of those forces alone predetermines whether or not any given person will commit a violent act. So I also take it for granted that, even if lethal violence is statistically more frequent when one party is in power than when another is, and that we can detect that difference if we study a long enough time-period, the resulting death rates will show considerable random variation from one year to the next, due to the fact that they are the result of far more variables than we can possibly control for during any one year. In fact, what I find most astonishing about the correlations between political parties and violent death rates that I have discovered is how loudly and clearly the "signal" comes through, given all the "noise" and "static" there is surrounding it. Despite the uncountable multiplicity of variables that are unavoidably involved, there are some patterns that keep recurring.

Thus, to me the most important fact about those rates during the years of the Reagan/Bush Sr. administration was that it remained at epidemic levels (i.e., it never dropped below the epidemic "floor" of 19.9 per 100,000 per year, which was the median violent death rate for the century as a whole) throughout the entire twelve years they were in power. During that time it ranged rather randomly from 19.9 to 22.4 and averaged 20.9, but never dropped below 19.9. To be more specific, the age-adjusted figures (with 1940 as the standard year) indicate that the total violent death rate (homicide plus suicide) was 21.9 during Reagan’s first year in office, and dropped to 20.1 by his fourth year; it then increased slightly to 20.5 by his eighth year, and increased to 21.7 by the end of Bush’s term in office -- almost exactly where it had been at the beginning of the entire twelve-year Republican administration.

What is most astounding is what happened after Clinton replaced them: the violent death rate dropped during every single year he was in office, and finally ended at 16, some 5.7 points lower than the rate of 21.7 that he inherited from the Reagan/Bush administration when he first took office in 1993.

The last available statistics published in your book are from 2007. Taking into account the slow growth in the U.S. economy and high unemployment levels during the Obama administration, do you see any reason why the trends you document might reverse?

Actually, since my book was published, the National Center for Health Statistics has published violent death rates for 2008 and 2009, and partial ones (95% complete) for 2010. What they indicate is that those rates increased from 16 to 17.2 during George W. Bush’s eight years in office (an average increase of 0.15 per year), and have decreased by 0.3 during Obama’s first two years as president (also an average of 0.15 per year, but of course in the opposite direction – down, not up). What is also relevant is the fact that, as we all know, the nation was in a deep and deepening recession during the last 14 or so months of the Bush administration, which did not end until Obama had been in office for six months (that is when the economy reversed from one of contraction to expansion, according to the National Bureau of Economic Statistics). However, the unemployment rate, which had escalated to 8.5% by January 2009, the month in which Obama took office, continued to increase to greater than 10%, following which it has been going down quite regularly year by year, to a point that by now is lower than it was when he first took office. According to Michael Grunwald’s new book The New New Deal, Obama’s economic stimulus measures have saved and created 2.5 million jobs, helped the economy grow by 3.8%, kept the unemployment rate from growing to 12%, provided $1.5 billion in rent subsidies and emergency housing that kept 1.2 million people from becoming homeless, increased the availability of food stamps, unemployment benefits and Medicaid, and has kept 7 million Americans from falling below the poverty line. Those are among the reasons why I think the Obama administration has begun diminishing the lethal violence rate, and is likely to continue to do so.

Is this a strictly U.S. phenomenon, or does this hold true for liberal and conservative governments worldwide?

In both the United Kingdom and Australia, both suicide and unemployment rates since 1900 have increased when conservative governments were in power and both have declined when liberal ones were. I do not know whether that applies to homicide rates as well, but at least as far as suicidal violence is concerned, their pattern seems to replicate ours.

Do your findings hold true on the state level as well? By that I mean, do the so-called "red" states have a higher rate of violent crime than the "blue" states? And does the violent crime rate decrease if a state switches from blue to red during an election cycle?

Yes, the red states do have higher rates of both homicide and suicide than the blue states (in 2000, 2004 and 2008). What was most striking to me was the finding that the red states have commited 20 times as many legal homicides (executions) as the blue states have, since capital punishment became legal again in 1976 -- more than 1000, compared with a little over 50.

I have not had time to investigate whether that changes in any one state when it changes from one color (party preference) to another -- I have only summarized statistics on groups of states.

One tends to equate violent crime with large urban areas. Yet, most cities tend to elect liberal candidates to local offices. Do the statistics hold true on an urban/rural level?

I have not studied the differences (if any) between Democratic and Republican mayors of large cities, but it is true that the homicide rates in all 25 of the largest cities in the country declined dramatically after Clinton was elected president (not just New York under Giuliani, who was in office during almost exactly the same years as Clinton). Urban homicide rates in general have tended to be higher than those in rural areas -- but as I said, I just do not know whether the partisan affiliation of the mayor makes any difference. I have focused on the nation as a whole, and on the groups of states that have qualified as red or blue. What I hope is that my book may stimulate a much larger research program than I have been able to undertake individually, and that others may investigate important questions such as this one.

What effect does Congress have on violent crime statistics?

I have not been able to detect any statistically significant effect of Congress on violent death rates. In my book I discuss why the president may be more influential than Congress in determining the morale of the nation’s population as a whole. The twentieth (and now twenty-first) centuries have been, after all, what Arthur Schlesinger Jr. called the age of the "imperial presidency."

In the final chapter of your book you discuss the criteria by which the International Agency for Research on Cancer suggests to "establish beyond reasonable doubt whether a given agent (e.g. cigarettes) could be regarded as causing a given outcome (e.g. lung cancer)." I have to ask, this being an election year, what would be worse for me on November 6th, starting a cigarette habit, or voting for Mitt Romney?

Cough, cough! If Romney were elected and brought about a recession/depression like so many of his Republican predecessors, and millions faced the humiliation and despair of suddenly becoming unemployed, homeless, hungry and unable to get adequate medical care, then I think their deaths might well happen even more quickly than they would if Romney did not get elected and they started smoking. Smoking, after all, usually takes years to kill people, whereas unemployment, depressions and sudden loss of one’s social and economic status can kill people -- or provoke them to kill themselves and/or others -- much more rapidly than that.