Sentences to ponder (model this)

by on June 2, 2013 at 12:28 pm in Political Science, Religion | Permalink

Our main result is that an increase in per-capita Muslim expenditures generates a large and significant increase in future religious conflict.  An increase in Hindu expenditures has negative or no effect.

Here is more, from Mitra and Ray.  The pointer is from Andres Marroquin.

Context June 2, 2013 at 1:22 pm

The single quote seems somewhat misleading: “our preferred explanation for this strong and curious relationship rests on the theory outlined in Section 3. The fact that Muslim expenditures display a significant and positive connection with later conflict, while Hindu expenditures have a negative link, suggests that (statistically speaking) Hindu groups have been largely been responsible for Hindu-Muslim violence in India, or at least for violence driven by instrumental, specifically economic considerations.”

Ungated link here:

Bill June 2, 2013 at 6:15 pm

Misleading is par for the course.

Thaniks for doing your homework.

dan1111 June 3, 2013 at 8:55 am

Their “preferred explanation” is largely speculative. It is not shown by their data, but is based on one possible model of what could be happening, as the paper makes clear:

“it is entirely possible (especially with some stretching of the creative imagination) to offer up a variety of explanations that are compatible with this finding.”

They also appear to have priors which pre-emptively exclude certain kinds of explanations:

“we do not believe that a particular religious group is intrinsically more predisposed to the use of violence. Our personal opinion is that religious fundamentalists are of the same ilk everywhere.”

This is questionable, to say the least. Religions differ widely in their teachings about the appropriateness of using violence; it is inconceivable that these different beliefs actually have no effect on how predisposed adherents of a religion are to using violence. There are even fringe religious groups that advocate and directly carry out violence. Are they no more predisposed to violence than pacifist groups? Their statement is an absurdity that is indefensible on logical grounds. It only makes sense as a political/ideological statement attempting to shift blame away from a particular group–and this suggests they may not be able to conduct an unbiased analysis of what the evidence shows.

I am not saying their preferred explanation is wrong, and I am certainly not saying that their evidence proves Muslims are to blame. But given how speculative and politically shaped their analysis is, it is not necessarily misleading to report only the factual finding and omit their interpretation of it.

mulp June 3, 2013 at 2:31 pm

In the early 20th century, Jews were doing well economically…

The economic power of Jews was a major complaint of Christians.

Who was responsible for the violence?

dan1111 June 4, 2013 at 3:39 am

I’m not sure what this is supposed to show in response to my comment. My points were:

1) Their preferred model is a speculative explanation and not proven by the data (as the authors themselves admit).

2) They appear to have prior commitments which heavily influence what kinds of explanations they would consider.

It is certainly true that one group’s increased prosperity can cause another group to react violently. But this only shows that their model is plausible, not that it is correct.

dearieme June 2, 2013 at 2:33 pm

How can one be “statistically speaking” responsible? If they mean that Hindus are the aggressors, shouldn’t they say so? If they don’t mean that, shouldn’t they cut the BS and say what they do mean?

Zaphod Beeblebox June 2, 2013 at 2:42 pm

No, they are being cautious and that is a good thing. “statistically speaking” means they want to emphasize 1) the conclusion is empirical in nature rather than theoretical value based, 2) there is a degree of uncertainty involved in their conclusion. You might know all this, but this is not BS actually great. though their abstract seems to be misleading. High time people started studying such issues in a detached way.

ThomasH June 2, 2013 at 6:50 pm

The whole abstract is not misleading, but the sentence that Tyler excerpted is ambiguous.

dan1111 June 3, 2013 at 9:09 am

The excerpted sentence accurately shows their main finding. Note that their abstract (linked above) is substantively the same as what Tyler quoted; it does not include any of the speculation about what the cause is.

David Brown June 2, 2013 at 3:35 pm

It means they don’t have evidence of a causal link. As an academic (in the hard sciences) this strikes me as really standard. They’re being scientists not hacks.

In any case the conclusion is not obvious. Here are two opposite theories that the evidence cannot distinguish between:
1. When Muslim’s income rises, they become more aggressive and start violence.
2. When Muslim’s income rise, Hindus feel threatened and start violence.

dearieme June 2, 2013 at 5:36 pm

“It means they don’t have evidence of a causal link.” Then it’s dreadfully irresponsible to use “responsible” of what, I take it from your remark, is just a ruddy correlation. Shame on them.

Rahul June 3, 2013 at 12:54 am

Most discoveries of causal links start as ruddy correlations.

So Much For Subtlety June 2, 2013 at 7:06 pm

The evidence from India. But Muslims live in other parts of the world. The evidence from Lebanon, the rest of the Middle East, Nigeria, Thailand, Indonesia, and the Philippines would suggest it is not the Hindus. On the other hand, the evidence from Burma might suggest it is not the Muslims.

Btu still – Where Muslims live there is violence. Where Buddhists and Hindus live, no so much.

Ankur June 2, 2013 at 11:39 pm

“Where Muslims live there is violence. Where Buddhists and Hindus live, no so much”

Srilanka? Though one can argue that it’s not as much Hindu-Buddhist but Tamil-Sinhala conflict

So Much For Subtlety June 3, 2013 at 6:40 am

Well I didn’t say none!

I would say that was part of the wash up from the Cold War. The Soviet Union sponsored, armed and trained terrorist groups all over. Most places had quasi-Communist groups of one sort or another. Some places they were more violent than others.

The Tamil Tigers come out of that world without actually being Marxist as such. But since the collapse of Communism we have had a great decrease in violence across the world.

Adrian Ratnapala June 4, 2013 at 8:50 am

As far as I know, only one country both a big Hindu and big Buddhist popluation: Sri Lanka. And the conflict there is defined by ethnicity, not religion. If anything about Sri Lanka is relevant to this discussion, it might be the following points:

(1) Sri Lanka also has a large Muslim population; which has stayed out of the violence – except as a victim: these people do not consider themselves to be either Sinhalese or Tamils.

(2) Most Christians identify as either Sinhalese or Tamils; and have joined their respective sides in the conflict.

(3) The discrimination which fed Tiger propaganda was partly driven the image of Tamils as being a successful elite minority. (E.g. some say the British favoured them in order to “divide and rule”).

Rahul June 2, 2013 at 3:13 pm

Which expenditures are these? Too bad its gated. Sounds interesting.

ad*m June 2, 2013 at 3:49 pm

Muslims conquered India most likely through jihad, killing up to 60-80 million Indians in the process (see
Where Islam does not succeed in becoming the majority, as in Pakistan and Bangladesh, or is eliminated, as in Spain in 1492, the resulting diversity results in recurring violence that is still ongoing. Indonesia, which has been majority Muslim since the conquest of the Majahabit empire by the Muslims around 1527, is another example. Only those areas in Indonesia where jihad is not yet complete, such as Sulawesi, does violence recur as in India.

Economics may be the proximate cause in Muslim-Hindu violence in India, but the root cause is Islam.

Adrian Ratnapala June 4, 2013 at 8:52 am

Are you saying the reconquista is a root cause of Basque seperatism?

Matt D. June 2, 2013 at 6:39 pm

The British intentionally starved tens of millions of Indians to death during several famines during the 19th century, in very well-documented cases (both in terms of British culpability and the number of victims). I’d be willing to bet there’s a pretty strong statistical correlation between white people and genocide. Does this mean that the root cause is enlightenment philosophy?

Chip June 2, 2013 at 9:04 pm


Famines were frequent before British rule. The East India Co lost money during famines. The first famine under British rule led to much criticism in Britain of British failures, this famine led to a British insurance scheme to alleviate future famines and the construction of infrastructure like rail and communication networks eventually helped diminish famine significantly.

Incompetence and even indifference surely played a part. But intentionally. That’s just silly.

Rahul June 3, 2013 at 12:52 am

“Intentionally” could have more indirect shades. e.g. What’s the relative value of letting a farmer grow indigo / cotton versus a food crop? Does a dead farmer in itself matter outside of his utility as a raw material producing input?

Such trade-offs may conceivably tweak the calculus of famine relief. Basically, it’s complex.

So Much For Subtlety June 3, 2013 at 6:24 am

It is irrelevant which one the peasant grows except in so far as a peasant who grows a cash crop for the market is less likely to suffer famine than a peasant who grows for the market.

Why is left as an exercise for the reader.

It is not complex at all.

Adrian Ratnapala June 4, 2013 at 8:56 am

Matt D’s is talking about genocide.

To sustain that “intentionally” needs to come in a strong shade: e.g “Haha! We shall starve those black buggers to death so that we can replace them with [Englishmen? Chinese coolies? robots?]“

So Much For Subtlety June 3, 2013 at 6:20 am

Matt D. that is a rather sad and pathetic lie. If you go to the India Office department of the British Library you will find a massive section of government documents on famine prevention in India. Once the East India Company noticed that its maladministration caused a famine in Bengal in the 18th century, they and the British government spent more and more time concentrating on the prevention of famine.

And they were highly successful at it. Between 1880 or so and 1940 their record was about as good as India’s post-independence government’s from 1947 onwards. Even though the Indians are richer, have American aid and modern transportation. Oh, and the Green Revolution.

Rahul June 3, 2013 at 7:37 am

…….and a lot more people to feed!

Adrian Ratnapala June 4, 2013 at 8:59 am

Which is in large part a consequence of the better famine prevention in the post-independence and (apparently) late-colonial periods.

Comments on this entry are closed.

Previous post:

Next post: