Military Deaths and Competence

I figured I’d take a break from tax issues and talk about the issue of the day… namely the war(s). These days, many are questioning the competence of the civilian leadership, and perhaps even of the military personnel they have elevated to do the day to day running of the war. Its always tough to evaluate military campaigns for a number of reasons, including the fact that much of the information needed to do so is classified, and most of us lack even rudimentary military knowledge.

“The amateur talks about strategy, the professional talks about logistics.”

Recently, I stumbled across a table on “US Active Duty Military Deaths per 100,000 Serving” for the years 1980 through 2004, and I started thinking about Omar Bradlye’s old dictum: “The amateur talks about strategy, the professional talks about logistics.” I started to think – perhaps one difference between amateurs and professionals is that professionals might pay attention to the kinds of things that lead to increased non-combat deaths, whereas amateurs would be unaware of those factors at all.

Let us consider four types of non-combat deaths shown on the table: accident, homicide, illness, and self-inflicted. Arguably, accident and illness are the two forms of death that those in charge could influence most easily. Investing in equipment and training can reduce accidents, and investing in quality medical care and medical training can reduce death by illness. Death by homicide, by contrast, can be due directly to the behavior of non-military personnel. Self-inflicted deaths can also, sometimes, be attributed to stresses of life outside the (military) job.

Below are the death rates per 100,000 for active duty military personnel for these four causes for several non-random years.


It seems that from 1980 to 2000, the accidental death rate dropped precipitously (it would have been nice to see data from before 1980) from 1980 to 2000. Death by illness also shows a similar, although less precipitous drop, perhaps because it was never as high as the accidental death rate to begin with. Homicides tended to drop over time, although not hugely, and self-inflicted deaths didn’t seem to show much of a pattern.

The obvious conclusion is that during the Reagan, GHW Bush, and Clinton administrations, there tended to be competent people in charge of watching out for the men and women in uniform.

The table above is repeated for four additionally non-randomly chosen years.


The accident rate is no longer dropping, and seems to be climbing slightly, as are deaths by illness. Deaths by non-combat homicide continue to drop – perhaps by moving so many military personnel out of the country, their likelihood of being murdered falls. (Also, some percentage of murdered military personnel are the victims of other military personnel – perhaps those few that are potential perpetrators, in a combat situation, find ways of satisfying whatever is in them that do not push up homicide statistics.) Self-inflicted deaths show no pattern. Death by illness appears to be on the rise – by 2003 it was at 1988 levels and climbing.

So what is going on? Perhaps its merely happenstance. Or perhaps accident and illness rates had somehow gone below their “natural” level by the late 1990s, and that the figures are reverting to these natural rates. Or maybe (as is often the case to hear some tell it) this is Clinton’s fault – he underinvested in the military, and somehow the pigeons held off on roosting until he left office. Or it is possible that this just has to do with growing pains – GW has increased the number of military personnel since taking office, but then, so did Reagan, and accident and illness rates dropped under Reagan.

Each of these explanations seems to rely on coincidence. A more likely explanation is that some deaths that might otherwise be attributed to combat are now being attributed to other (currently less unpalatable) causes. Another is that the folks making the decisions are not as competent as their predecessors in the three most recent administrations. Are there any others?