In honor of today’s celebration of the Day of the Dead, this post ponders why some countries practice cremation more than others. The Cremation Society of Great Britain estimates the cremation rates (i.e. cremations divided by total deaths) for about 40 countries around the world. The rates vary dramatically from country to country, ranging from 99.9% in Japan to 0.3% in Romania. For comparison purposes, the world and US cremation rates are 49.5% and 37.6% respectively. What accounts for this substantial variation across countries?
Several factors come to mind. First, greater population density (i.e. more people per unit of land) might be correlated with greater cremation rates because land for burials is scarce and expensive. One way to capture the effects of land scarcity is via the number of people per square mile.
Second, people likely base their cremation preferences on religious beliefs. Shintoism in Japan, for example, prescribes cremation within its funeral rites. Christianity, on the other hand, is more associated with burials.
Third, a country’s GDP per capita might correlate with cremation rates because one form of last rites might be more expensive. Burial is probably more expensive relative to cremation in developed countries, while the cost order is likely reversed in the developing world. By comparison, burial is a relatively labor-intensive task, while cremation is more capital-intensive. Crematoriums aren’t high tech, but they usually require greater physical capital than equipment required for burials. In the developed world, unskilled labor is scarce and consequently the labor-intensive products expensive. Conversely, capital is relatively abundant in the developed world, making capital-intensive products cheaper.
To test which of these factors is associated with greater cremation rates, I employ simple ordinary least squares models. I use robust standard errors to deal with the heteroskedasticity that arises due to the dependent variable’s truncated range (0 to 100%). Additionally, some of the models include regional fixed effects that accounts for observable and unobservable differences that vary by geographic region (e.g. East Asia, Latin America). Lastly, in some models I exclude Hong Kong and Singapore because they are outliers in terms of population density.
In every model, Christianity and GDP per capita are statistically significant at the 0.05 level. As hypothesized, when a country’s dominant religion is Christianity, the cremation rate is 15 to 30% lower. A 10% increase in GDP per capita is associated with 2% increase in cremation rate on average. This marginal effect of wealth is much smaller than Christianity, but nevertheless is robust to various model specifications. Curiously, land scarcity as measured by population density is not associated with cremation once the above factors are controlled.