An ongoing debate in many countries deals with whether allowing more immigration will result in increased levels of crime and increased political pressure for government regulation. In Denmark and other European countries this especially pertains to the immigration of muslims.
Among libertarians there are those who argue that monoculture keeps a society stable and peaceful and inhabitants will see less reason to use government power against people different than themselves. Because of this, monoculture is seen as desirable.
On the other hand, there are those who argue that in a polycultural society, government will tend to experience smaller growth because different cultural groups do not want to share the same institutions.
While I haven’t followed the debate meticulously it always seems very emotional and suffering from a lack of data. This is an attempt to bring more empirical knowledge into the discussion by seeking correlations between the religious composition of populations and freedom and crime rates.
To the extent that correlation is found, causation remains possible. To the extent that no correlation is found, a causal relationship is less likely.
I’m no expert on statistics and the following has been produced with simple approaches and tools. I welcome comments that point out if you do not think the data can support my conclusions. And I’ll make the data publicly available in case anyone thinks it deserves more attention.
A measure of monoculture
There’s probably no perfect way to measure the degree of monoculture. It was chosen to measure monoculture as the size of the largest religious group in a given country. I.e. if country A has 90% Buddhists while country B has 60% Christians, then A is more monocultural than B.
This definition definitely has flaws. The most obvious being that there is more to culture than religion though most will probably agree that religion is one of the most important aspects. The advantage is that data on the sizes of religious groups of different countries is readily available.
The data used for religion was grouped into the following: Christian, Muslim, Jewish, Buddhist, Hindu, and Other. Source.
Monoculture vs. freedom
Several studies are made yearly that measure different aspects of freedom for many countries. FreeExistence.org maintains a tool that calculates a meta-score based on nine different measures: Property rights, freedom from taxes, freedom of expression, limited government, gun rights, drug rights, freedom from corruption, freedom from inflation, and business freedom.
Sources of the tool include the Heritage Foundation Index of Economic Freedom, the Freedom House Freedom In the World index, the Transparency International Corruption Perceptions Index, and the Free Existence Gun Rights Index and Drug Freedom Index.
The tool allows you to weigh the different measures relative to each other. For simplicity and neutrality it was chosen to weigh them all equally.
This resulted in the following plot of monoculturalism versus freedom:
With this data from 194 countries, no correlation is seen between share of the largest religion and freedom.
Monoculture vs. reported crimes
Amount of reported crimes were taken from this source. The amount was divided by the national population to get crimes per capita.
No correlation is seen between share of the largest religion and reported crimes in 82 countries.
Monoculture vs. crime victimization
The above has the problem that it is not just a measure of committed crimes but also influenced by the willingness of victims to report crimes to authorities. To make up for this, studies are also made based on surveying population samples in different countries whether they have been subject to a crime in the past year.
Data on this was only found for 30 different western countries (Scotland was excluded because of missing data on religion).
No correlation is seen between share of largest religion and crime victimization.
Muslims vs. freedom
It is sometimes argued that immigration and/or polyculture in general is not the problem but that specifically muslim culture poses a problem. The same data on freedom, reported crimes, and crime victimization is therefore compared to the share of muslims in populations:
A slight correlation is seen between the share of muslims in a population and the freedom score. If a country has a higher share of muslims it tends to have less freedom.
It is also seen that most of this seems to be due to the countries with a share of muslims higher than 0.8 which of course are mostly countries with native muslims, not immigrants. It would be interesting to see if the same correlation shows in countries where muslims are not the largest group – i.e. where they tend to be immigrants:
In muslim-minority countries, no significant correlation is seen between share of muslims and freedom score.
Of course, these two results could also be interpreted in the way that the muslim share does not contribute negatively to freedom until muslims are in majority. It could be the case, but it could also be the case that muslims who leave their home country and join muslim minority countries are generally more liberty oriented than muslims who stay in their native country.
Muslims vs. reported crimes
A very slight correlation is seen between the muslim share and reported crimes.
Muslims vs. crime victimization
No significant correlation is seen between muslim share and crime victimization, i.e. the result of crime surveys.
Christians vs. freedom
A slight correlation is seen: The more christian the country, the more free it tends to be. The same statistic only for Christian majority countries:
No correlation is seen between share of christians and freedom in christian-majority countries.
Christians vs. reported crimes
A slight correlation is seen: The more christian a population, the more crimes tend to be reported.
Christians vs. crime victimization
No significant correlation is seen.
Thoughts
In the data used, no significant correlations were found between the degree of polyculture and freedom from crime and government.
Slight correlations were found between the share of muslims and christians in a country and freedom: More muslim countries tended to be less free, and more christian countries tended to be more free. However, if one looked only at countries with muslim minorities and christian majorities, i.e. typically countries of immigration and not countries originally muslim/non-christian, the correlation disappeared.
If there was a significant causal effect between religion and freedom or crime, we should expect it to result in correlation given significantly large samples of countries.
It can be argued that there is too much noise and too many variables to tell from the above if monoculturalism or size of muslim population has any effect at all, but in that case at least the above shows that other variables are apparently more important than monoculturalism or size of muslim (minority) population. That monoculturalism and muslim (minority) population plays a seemingly insignificant role in how free and peaceful a society is.
It would also be interesting to analyze the development in crime rates and freedom scores through a period of years with a growing group of immigrants for different countries. However, such historical data is hard to find.
Unfortunately, I don’t recall seeing any other attempt to empirically find a relationship between the degree of monoculture and muslim population and freedom from crime and government. If you know of any, please post them in the comments.
Data sheet
Are monocultural societies more free and peaceful?
Sejt Lasse! Mere af det!
Lidt data om monokultur. Multikultur er næppe kilden til kriminalitet, hvilket Singapore og Hong Kong naturligvis viser, men her er så det store billede.
The analyses presented are fundamentally flawed and it is
impossible to say from the plots presented whether there is a causal
connection from monoculture to freedom, crime, victimization etc.
The problem is the following:
If you want to estimate the empirical effect of a (possibly)
causal factor, X (monoculture), on a dependent variable, Y (say, freedom), then
you have to formulate a complete statistical model where all the factors
determining freedom are included as explanatory variables. Those explanatory
factors may be both qualitative and quantitative, and the relationship can be
linear or non-linear in a number of ways.
Then you estimate your statistical model on a cross-section
of countries or possibly on a mixed time series and cross section of countries,
and then you test whether the marginal contribution from your factor, X, is
significant or not. Of course you have to test your model for possible
multicollinearity and subject it to other forms of specification and misspecification
tests.
The analyses presented here fail
to make corrections for the causal effect of all other determinants of freedom,
crime, victimization etc. The statistical model is misspecified and suffers
seriously from an omitted variable bias. It is therefore not surprising that
you don’t get much information out of the “models” presented here. In general,
the more “structure” you impose on a model, the more information can you get
out of it. It is possible that a fully specified statistical model would show
that monoculture is highly significant for freedom, and the opposite is
possible as well. We just can’t draw any conclusions at all from the above
analyses.
Similar problems occur when one tries to find a simple
relationship between the tax level and economic growth. We know from economic
theory that there should be a negative effect of high income- and capital taxes
on the level of GDP per capita and on the rate of growth of GDP, but simple
plots of tax rates against GDP per capita reveal no such relationship. In
detailed statistical models, however, tax rates and the size of the public
sector have a significant negative effect on economic growth.
Torben
I briefly discuss the influence of other variables in the article:
“It can be argued that there is too much noise and too many variables to tell from the above if monoculturalism or size of muslim population has any effect at all, but in that case at least the above shows that other variables are apparently more important than monoculturalism or size of muslim (minority) population. That monoculturalism and muslim (minority) population plays a seemingly insignificant role in how free and peaceful a society is.”
So the article remains open to the fact that other variables can be important. Actually I only conclude that IF multiculturalism has an effect, then it appears that other variables have even greater effects. Otherwise correlation would have been found, right?
Also, don’t you agree that given a significantly large sample of data, for instance if one million countries existed and we had this data for all of them, the effect from other variables would on average be cancelled out?Of course we don’t have a million countries to measure, but then the question is if about 200 is sufficient for this.
Lasse
Regarding your question of whether the effect from the other variables would cancel out on average for a large sample, I am pretty sure the answer is no. At least if there is a correlation between the share of the largest religion and the other variables that explain freedom (or crime). I.e. if monocultural societies also tend to be those with the largest geographical area (due to historical development), and the longer distances between people reduce crime, the high correlation between multiculturalism and crime would capture the effect of area. It is just an example, many better (more realistic) might be available.