The most informative source on mass government murder – for which RJ Rummel coined the term “democide” – is the University of Hawaii professor’s Freedom, Democide, War site. It’s definitely worth doing some poking around in.
The statistics that I cited are by no means precise, but they are reasonable. The 4x estimate is the easiest, as it was derived by dividing the number of total 20th century victims of democide (169 million) by the number of people believed to have been killed in all the wars and civil wars taking place during that same time (38.5 million).
The comparison between murder and democide is tougher, as it can be calculated in a variety of ways. One significant challenge is that estimated murder rates often include government killings, another is that relatively few countries keep record of such statistics. I derived the number as follows:
– The United States has 19,000 murders in a population of 290 million. This is 6.55 per 100,000, or 0.000065517.
– This is considered to be high by US historical standards (1991 US Senate report) and by world standards as well. For example, Japan’s murder rate is 11 times lower, the UK 4 times lower, Ireland 9 times lower. Since the current US rate is high by both world and historical standards, estimating the 20th century global murder rate by dividing the US rate by 4 seemed a reasonable educated guess, which gives 1.64 per 100,000, or 0.000016379.
It’s hard to know how precisely how many people lived in the 20th century, but looking at a number of demographic sites indicated that the number was somewhere around 18 billion people on the high end. Dividing the 169 million victims of democide by 18 billion, we get a 20th century democide rate of 0.009388889, or 938.88 per 100,000, which is 570 times higher than the 1.64 per 100,000 estimated annual global murder rate. (Update – of course, multiplying the latter number by 100 gives a direct comparison of 938.88 per 100,00 with 164 per 100,000, or 5.72x.)
This seems too high, especially since Rummel himself estimates the average annual domestic democide rate at only .00235 per year for the countries where democides took place – in a little over one-third of the countries belonging to the UN.
How to rectify these differences? The problem likely stems from not comparing apples to apples, by comparing a percentage calculated from a total to an annual rate. But we can get a ceiling on the global 20th century murder total by multiplying the number of annual US murders by 100, then multiplying again by the global population divided by the US population. (19,000 x 100) x (6 billion / 290 million) = a maximum 39,310,394 murders in the 20th century. Since as was mentioned before, the US rate is high on two accounts, plus the global population is much higher now than it was at the beginning of the 20th century. The divider could therefore be increased somewhat, but since I prefer to be conservative here, using the previous divider of 4 gives us a best estimate of 9.83 million criminal murders in the 20th century, compared to 169 million legal democides.
Thus, I conclude that one’s chances of falling prey to a homicidal government in the 20th century ranged from 4.3 (low estimate) to 54.5 (high estimate) times that of being murdered by an individual acting on his own, with the most reasonable estimate being 17.3 times.