Generosity and Political Preferences

December 26, 2012 4:52 PM

Generosity and Political Preferences [.pdf]

We test whether generosity is related to political preferences and partisanship in Canada, Sweden, the United Kingdom and the United States using incentivised dictator games. We document that support for social spending and redistribution is positively correlated with generosity in all four countries. Further, we show that donors are more generous towards co-partisans in all countries, and that this effect is stronger among supporters of left-wing political parties. All results are robust to the inclusion to an extensive set of control variables, including income and education.

Why are some people left-wing and others right-wing? One popular idea is that left-wing people are more other-regarding than right-wing people. The "bleeding hearts" of liberals and "heartless and cold"conservatives are typical stereotypes in U.S. politics (see e.g., Farwell & Weiner 2000). In contrast, social scientists often focus on economic self-interest as the determinant of political preferences and most economic-political models assume that voters are self-interested. Yet the explanatory power of self-interest in empirical studies is usually quite low. This discrepancy between theoretical assumptions and empirical …findings highlights the importance of …finding other determinants of political preferences. For example, political preferences might be related to "deep"preference parameters that also determinechoices in other domains, e.g. occupation.

In this paper, we focus on the relationship between social and political preferences. Spec…ifically, we investigate whether generosity in the dictator game is related to political preferences and partisanship. Our focus on generosity is motivated by the fact the there are two distinct reasons to support social spending and redistribution. On the one hand, people might support social spending because they directly bene…fit from it. This "demand side-view" of social policy is the standard perspective in economic-political models. However, there are also reasons to consider the "supply side"of redistribution: voters might prefer high levels of spending and taxation because they want to help people they think deserve government support. This latter view suggests that preferences for social spending could be related to generosity. The difference between the "demand"and "supply side"view is important for a wide range of questions in political science, not least the sources of public opinion and policy preferences, the basis of electoral competition, and the ideological positioning of political parties.

posted by wilful (35 comments total)

Summary

In the dictator game respondents are asked to split ~$100 between themselves and an anonymous person. In this study they also asked to split the money after being informed the anonymous person was liberal/conservative or a member of X political party. A small fraction of the respondent pairs then receive the money, split accordingly.

~70% of Swedes split the pot evenly, ~15% kept it all for themselves.
~20% of Brits split the pot evenly, ~50% kept it all.
Americans are slightly more generous than Brits, Canadians a bit more than that (but still far below the Swedes).

People who identify with conservative positions (but not necessarily parties) are less generous. People are less generous when giving to a member of the opposing party, but there's a pretty big spike if the other person has the same political views (liberal or conservative).
posted by Orange Pamplemousse at 5:16 PM on December 26, 2012 [2 favorites]

I want to believe this, and thus it demands extra scrutiny.
posted by Pope Guilty at 5:46 PM on December 26, 2012 [30 favorites]

Australian's pre John Howard

20% kept 80% split

after...

fuck you
posted by the noob at 6:11 PM on December 26, 2012

Noob brings up a good point. These sorts of studies seem to aim to explain why people ascribe to different political views based on character differences. But what if the causation (if there is any) goes the other direction?
posted by eviemath at 6:52 PM on December 26, 2012

I'll see your abstracted social science tomfoolery and raise you actual charitable giving statistics. Take a look at charitable giving by state. Top ten most generous states? All Red except for Maryland at # 10. Expand to the top twenty and you pick up New York and Oregon, plus Florida and Virginia as swing states. Top ten least generous states? All Blue except for North Dakota at # 43. Expand to the bottom twenty and you do get up to seven Red, but that's still a pretty poor showing.

Conclusion: there's a shit ton more going on here than an "experiment" this hokey could possibly hope to capture.
posted by valkyryn at 7:18 PM on December 26, 2012 [8 favorites]

I'll see your abstracted social science tomfoolery and raise you actual charitable giving statistics.

Conclusion: Completely different data set seems to imply things which are not implied by the first data set. Because they're completely different and don't even begin to address the same issues.
posted by pompomtom at 7:24 PM on December 26, 2012 [5 favorites]

But a 'red' state tends to be, like, at most 60% Republican, 40% Democrat, and vice-versa. It's not like Alabama is 100% conservative.
posted by shakespeherian at 7:32 PM on December 26, 2012 [2 favorites]

In leftie circles there is a social benefit to being generous, in the same way as being generous is of social benefit in religious circles. It's behaviour that is approved of within the group.

You give money to the person on the street who asks you for money because, partly, it makes you feel better about yourself (at a very reasonable price). It's beneficial for rich people to be seen to be generous, but not to actually give enough to cause themselves inconvenience. And so on…generosity has a transactional element in most cases (hence the debate about private versus public donations).

I'm not surprised that people who see themselves as left-leaning and socially aware are more generous, because they identify themselves as being generous types. So I don't really see the point, in the end.
posted by chrisgregory at 7:36 PM on December 26, 2012 [1 favorite]

Conclusion: Completely different data set seems to imply things which are not implied by the first data set. Because they're completely different and don't even begin to address the same issues.

Yes, certainly. On the other hand, you wouldn't want this study of a highly artificial environment to give you the impression that, as stated in the summary "People who identify with conservative positions (but not necessarily parties) are less generous" in any sort of broad way. In fact, there's empirical evidence that in some situations (like philanthropic giving, liklihood to give blood or to volunteer (all the the U.S.)) there is evidence to the contrary.
posted by TheShadowKnows at 7:36 PM on December 26, 2012 [2 favorites]

For reasons too tedious to elaborate on (short version: some of my drinking buddies were poli sci types, they needed technical help staging some experiments), the only published academic research my name is on boils down to this sort of thing. In retrospect, I more or less concur with "tomfoolery".

It was kind of fun, though.
posted by brennen at 7:43 PM on December 26, 2012 [1 favorite]

I'll see your abstracted social science tomfoolery and raise you actual charitable giving statistics.

I'd like to know if "charitable giving" here means any donation "qualifies for itemized deduction on tax returns" and includes religious tithes. Because if it is, bringing it up doesn't provide any kind of counter to the study. Giving to a church organization one belongs to would be giving to a tribe/in-group one identifies with, and one might expect a spike there in the same way you would with a similar political group.
posted by weston at 7:56 PM on December 26, 2012 [20 favorites]

valkyryn: perhaps if you included tax payments into the equation the determination of exactly how charitable certain states were might change. Some of us consider taxes our patriotic civic duty. I imagine it might line up slightly differently.
posted by Freen at 7:57 PM on December 26, 2012 [4 favorites]

"The US study was conducted as a part of a computerized laboratory experiment at the University of California-San Diego between November 2010 and March 2011. 262 students participated in the experiment. The advantage of laboratory experiments is that they provide a higher degree of control. The obvious drawback is that a sample of students at a prestigious university is not representative of the general US population. The results from the US study should therefore be interpreted with caution."
posted by John Cohen at 8:32 PM on December 26, 2012 [4 favorites]

Two differences between this experiment and charitable giving:

- The money is won by chance instead of earned. I'm not sure if this would make people more or less likely to share. If you found $100 in the street, would you feel obliged to share it with a random person? OTOH, people might feel more entitled to keep money they earned and more inclined to smooth out the unevenness of chance.
- Giving to an anonymous person (who might be richer than you for all you know) is different to giving to a homeless person or some other deserving recipient. Maybe the people who didn't share rationalized it by telling themselves they could share it later with someone they thought deserved it. Perhaps sharing with an anonymous person is done out of an abstract sense of fairness, while giving to a specific person in need is done more out of a feeling of empathy.
posted by L.P. Hatecraft at 9:46 PM on December 26, 2012 [5 favorites]

If you found $100 in the street, would you feel obliged to share it with a random person?

No, if I couldn't tell who it belonged to, I would think the only rational thing to do is keep it for myself. So this study tells me that conservatives are more rational than liberals.

Maybe the people who didn't share rationalized it by telling themselves they could share it later with someone they thought deserved it.

And in fact, this would make the most sense for the most generous people.
posted by John Cohen at 10:09 PM on December 26, 2012

I'll see your abstracted social science tomfoolery and raise you actual charitable giving statistics.

I personally see American-style "charity" as a partisan subset of generosity (and one with an ugly underbelly). If you did the same study on a different partisan subset of generosity, such as "I would vote to raise my taxes so government can help more people", you'll see blue and red swap places.

It doesn't seem significant to me that the US right wing's partisan "alternative" to government is more attractive to the right wing than to the left. Well Duh. :)
posted by anonymisc at 10:21 PM on December 26, 2012 [4 favorites]

You know, I was thinking about it, and I were presented with the choice of splitting money or keeping it, I believe I would split it evenly in most cases where I knew what the other person's political views were.

Interestingly, though, it would be for different reasons. If I knew the other person was conservative, at least in the normal spectrum of being conservative, I would want to split evenly to indicate fair dealing. I don't really understand the motive properly, but I think it's meant as a signal, not as true generosity. If the other person was liberal, I'd want to do an even split to make them feel good. I might even split 60:40 in their favor, because I know they'll go home and tell their family, and maybe they'll be a bit more generous with someone else.

But if the other person was batshit-crazy conservative, I think I'd still split evenly, but I would feel dirty. I'd feel even worse if I kept all the money, so I'd still split it fairly, but I have so little respect for that position that it would piss me off that they had the money, after.

It's really interesting, seeing the different motives. The same basic outcome is likely in all three cases, but the process for getting there is different. It's a signal to conservatives, generosity to liberals, and not thinking of myself as greedy with a whackjob.
posted by Malor at 10:24 PM on December 26, 2012

Take a look at charitable giving by state.

But the generosity ranking changes when religion is taken out of the picture. People in the Northeast give the most, providing 1.4 percent of their discretionary income to secular charities, compared with those in the South, who give 0.9 percent.

It looks like the 'Chronicle of Philanthropy' study counts all church donations as charity. It seems to me if secular communities had social organizations to substitute for church (or maybe just more liberal churches?) you might see more community involvement where it's needed.

Other Findings:

The rich aren’t the most generous. Middle-class Amer­i­cans give a far bigger share of their discretionary income to charities than the rich. [...]

The 1 percent really are different. Rich people who live in neighborhoods with many other wealthy people give a smaller share of their incomes to charity than rich people who live in more economically diverse communities. [...]
posted by Golden Eternity at 11:42 PM on December 26, 2012 [4 favorites]

This may be an example of Simpson's paradox . It's possible that

1. Religious liberals are more charitable than religious conservatives.
2. Secular liberals are more charitable than secular conservatives.
3. Conservatives are more charitable than liberals.

So, in data where religion is controlled, the result is that liberals are more charitable. In data where religion is not controlled (e.g. donations by state), the result is that conservatives are more charitable.
posted by ILuvMath at 11:57 PM on December 26, 2012 [6 favorites]

Swedes favor consensus in politics; Americans prefer majorities and the results derive directly from this. Generosity is not even a variable.

Of course the study doesn't reveal that the Americans arrived at their result in a matter of minutes and in Sweden it took weeks.
posted by three blind mice at 12:02 AM on December 27, 2012

This type of experiment doesn't tell a lot. But I like the part of their theory where they assume people may act out of other reasons than self-interest. If I were their advisor, I'd ask them to focus on that, and try to figure out ways to study that.
It always bothers me when economists focus on (economic) self-interest as the sole driver of human activity. Then why would smart people ever be school teachers, or philosophers? Sometimes, I have the feeling some of my economist friends conclude that smart people are not school teachers, because their definition of "smart" is "the ability to grab money".

Apart from the problem of who is stupid and who isn't, the focus on self-interest is one of several reasons their models are flawed, and consistently show that tax-breaks are good for growth, in spite of very strong empirical evidence to the contrary.
posted by mumimor at 1:28 AM on December 27, 2012 [2 favorites]

So, in data where religion is controlled, the result is that liberals are more charitable. In data where religion is not controlled (e.g. donations by state), the result is that conservatives are more charitable.

What I want to see is how much of church money actually goes to charity. With all of the prosperity preaching and megachurches in the south, I sincerely doubt that much of it does anything but line the pockets of the organization and buy influence for the person "tithing." Getting together on Sundays to sing in a giant auditorium with a live band through a multi-million dollar sound system does diddly squat for someone who needs a meal around the corner.

Here's what I did find:

Do religions engage in charitable work that addresses the physical needs of the poor? Many do, but that is not their primary focus. Religions are quick to trumpet when they do charitable work—ironically for Christians, since the Bible explicitly says not to (Mathew 6:2). But they don’t do as much charitable work as a lot of people think, and they spend a relatively small percentage of their overall revenue on such work. For instance, the Church of Jesus Christ of Latter-day Saints (the LDS or Mormon Church), which regularly trumpets its charitable donations, gave about $1 billion to charitable causes between 1985 and 2008. That may seem like a lot until you divide it by the twenty-three-year time span and realize this church is donating only about 0.7 percent of its annual income. Other religions are more charitable. For instance, the United Methodist Church allocated about 29 percent of its revenues to charitable causes in 2010 (about $62 million of $214 million received). One calculation of the resources expended by 271 U.S. congregations found that, on average, “operating expenses” totaled 71 percent of all the expenditures of religions, much of that going to pay ministers’ salaries. Financial contributions addressing the physical needs of the poor fall within the remaining 29 percent of expenditures. While these numbers may be higher as a percentage of income than typical charitable giving by corporations, they are not hugely higher (depending on the religion) and are substantially lower in absolute terms. Wal-Mart, for instance, gives about $1.75 billion in food aid to charities each year, or twenty-eight times all of the money allotted for charity by the United Methodist Church and almost double what the LDS Church has given in the last twenty-five years.

Based on that evidence, I would say donations to churches shouldn't count as a charity, because most of it goes in the pockets of the people running the organization and not to the needy. If you want your money to go to the poor, donate to a charity like Oxfam or Action Against Hunger where 80-85% goes directly towards assistance instead of rituals and sound systems.
posted by tripping daisy at 5:11 AM on December 27, 2012 [4 favorites]

I'd like to know if "charitable giving" here means any donation "qualifies for itemized deduction on tax returns" and includes religious tithes. Because if it is, bringing it up doesn't provide any kind of counter to the study. Giving to a church organization one belongs to would be giving to a tribe/in-group one identifies with, and one might expect a spike there in the same way you would with a similar political group.

Of course people are going to give more to causes that they happen to like. This is as true of liberals as conservatives. Doesn't a liberal giving to Greenpeace, the ACLU, or NPR count as a "tribe/in-group one identifies with"? The kind you and others here want to discount? Because if it doesn't, "controlling for religion" is essentially "controlling for causes that secular liberals don't like." Fuck that shit.
posted by valkyryn at 9:02 AM on December 27, 2012 [3 favorites]

It always bothers me when economists focus on (economic) self-interest as the sole driver of human activity. Then why would smart people ever be school teachers, or philosophers?

Why wouldn't they? The basis of liberal economics is maximising utility and a lot of teachers get a bigger kick out of teaching compared to investment banking or what have you. One issue is that utility is conflated with money because it's easier to measure the latter than measure the utility from teaching successfully a concept to a student after a few unsuccessful tries.
posted by ersatz at 9:02 AM on December 27, 2012

Because if it doesn't, "controlling for religion" is essentially "controlling for causes that secular liberals don't like." Fuck that shit.

Nope, it's 'controlling for organizations that actually give a large portion of their income towards the cause they are ostensibly championing, versus frittering it away on personal expenditures'. Because religions are, quantifiably, pretty crap at the whole 'charitable organization' game. (see links above, and all data about church tithing ever).
posted by FatherDagon at 9:20 AM on December 27, 2012 [1 favorite]

I'll see your abstracted social science tomfoolery and raise you actual charitable giving statistics. Take a look at charitable giving by state.

I'll see your charitable giving statistics and raise you an ecological fallacy. You're making inferences about individuals based on aggregate (i.e., state-level) data.
posted by jonp72 at 9:57 AM on December 27, 2012

Based on that evidence, I would say donations to churches shouldn't count as a charity.

Fair point. Let's exclude donations to religious organizations:

If donations to all religious organizations are excluded, liberals give slightly more to charity than conservatives do. But if measuring by the percentage of income given, conservatives are more generous than liberals even to secular causes.

Source.
posted by ILuvMath at 10:41 AM on December 27, 2012

"I say there, Liberal/Conservative Subject Person, what would you do with this fake money? "

Pretty unreal premise, I'd say.

What they really should have done was attach a stranger to electrodes, put Subject Person in a remote control room with the generator, and then show SP the stranger leafing through Mother Jones/National Review.

religions are, quantifiably, pretty crap at the whole 'charitable organization' game. (see links above, and all data about church tithing ever).

Not in the mind of the donor, which for the sake of this study is all that really matters. By the same token, giving to Planned Parenthood or giving to Pro-life America each qualifies as charity in the mind of the donor, and a pernicious misuse of money in the minds of their opponents.
posted by IndigoJones at 10:53 AM on December 27, 2012

Of course people are going to give more to causes that they happen to like. This is as true of liberals as conservatives. Doesn't a liberal giving to Greenpeace, the ACLU, or NPR count as a "tribe/in-group one identifies with"?

Not unless the liberal giving to Greenpeace or the ACLU is receiving direct services from them in the same way that the churchgoer donating to their own church does.

This does mean, though, that liberals would have to admit that donations to NPR/PBS arguably aren't really charitable either.

Because if it doesn't, "controlling for religion" is essentially "controlling for causes that secular liberals don't like."

No, controlling for religion is an attempt to get at how much people donate to do actual charitable work in the world as opposed to what amounts to unenforced payment for services. Hiring a minister for your own church or putting in new stained glass into your own church isn't what most people would consider charity.
posted by ROU_Xenophobe at 11:23 AM on December 27, 2012 [2 favorites]

I like stuff like this--probably because it's self-affirming, regardless of the problems (I mean, I've seen many studies that conservatives or religious donate more than non-religious, but then, I don't know if I trust those "studies" as well).

Terror Management Theory (TMT) has shown over and over again, IF these participants had had to think about their own death, they would have been more likely to share with those of similar political/religious/etc beliefs and much less likely to share with those of different political/religious/etc beliefs.
There's a lot of academic stuff on TMT, but only the documentary "Flight from Death" seems geared toward the mainstream. (Available on youtube and hulu for free.)
posted by whatgorilla at 12:04 PM on December 27, 2012

I think they should have tested Swedes vs Swedish Americans. As someone who lives in a city that's about a third Swedish, I can vouch that they are generous people.
posted by michaelh at 12:16 PM on December 27, 2012

Because religions are, quantifiably, pretty crap at the whole 'charitable organization' game.

Oh, so you know how many aid orgs are religious vs secular, and what the breakdown of costs and processes is for both? That's impressive. Heck, that's a PHD.

I am as atheist as they come, but blanket generalisations like this cannot pass uncommented. Many of the world's biggest - and, yes, best - aid organisations are religious or religious-founded. They do amazing work without care for faith, creed, or nationality. Blackening them like this without a shred of evidence beyond anti-religion dogma is uncharitable (har har) in the extreme.

What is it about aid that brings out the innner Generalisation Fascist in people? Generalisations about aid are such a waste of time, but they are so popular. Everyone like a just-so story about starving Africans that confirms their racial, political, social or religious prejudice. Aid is complicated; it defies generalisation.
posted by smoke at 3:40 PM on December 27, 2012 [1 favorite]

Hiring a minister for your own church or putting in new stained glass into your own church isn't what most people would consider charity.

Actually, it kind of is. You may not like it, but a majority of the country--indeed of the world--subscribes to one sort of religious faith or another.

I think this is where we say "deal with it."
posted by valkyryn at 6:45 AM on December 28, 2012

And, of course, there is the added complicating factor that many conservatives seem to believe in charity as the only moral solution to certain social problems, whereas liberals tend to view those same problems as more reliably and thoroughly solved by government action (and in a manner more respectful to the victims - much better to receive a payment on a regular basis from the state than to have to simper like a slave in the hopes of receiving money on the whim of a rich man).

What is interesting about this study is that (in a very small and limited way) it goes against a common conservative talking point - that conservatives are more generous than liberals.

For some reason, conservatives seem desperate to cling to the idea that they are somehow "better" people than liberals. This is the subtext of so much of what you can see pumped out every day by conservative news sources - an image of liberals as hypocritical, ridiculous and above all greedy - wanting to take from YOU to give to THEM.

By contrast, when liberals look at conservatives, it always seems to me to be in a spirit of bafflement. Why do these people vote against what seem to be their own best interests? Why do these people hate us so much? How can they not see that they are being duped? It can be insulting to be treated as an object of study, or for someone to assume that you are being tricked because you are disagreeing with them. But it is far, far more insulting to be actually directly insulted - to be subjected to an endless harangue from people who have no right, no qualifications, no track record of achievement that would justify their extraordinary moralising high-handedness.

The idea that liberals might not be degraded untermensch seems to be an anathema to the conservative mindset. Could it be that liberals are intelligent and honest people who simply disagree with conservatives? Could they have looked at the same evidence and just come to different conclusions? Hell, could they even have looked more closely, thought more deeply and come out with ideas that are right? Impossible! the conservative mind replies.

And yet.

If this little study, whatever its limitations, whatever its flaws, could serve as even a tiny crack in that vast edifice of conservative myth-making about liberals - another tiny refutation of the puffed-up and unpleasant way that conservatives talk about anyone who disagrees with them - well, then, whatever its limitations as a piece of research, it is a good thing.

Also, including church giving in a study of generosity does, in fact, undermine the value of the study. It is a little difficult to explain why, but the essential point is that by doing so you are controlling the results of the study: you are turning it from a genuine experiment into something that just tells you what you already want to hear.

Look at it this way: an experiment like this has no sense or validity unless it compares like with like. The boundaries of the experiment must be applied consistently to all the participants. You are trying to establish which group is more generous. But not everyone involved in this study regards giving to a church as "generosity".

If you have a group of people who regard giving to a church as charity and another group of people who regard the same ends as better served by giving to the state; and then you purport to measure generosity by including the church donations of the former under the heading of charity, but not the tax donations of the latter; well, then, you have put your finger very firmly on the scales in favour of the first, church-tithing group. They will come out as more "generous", but only because you have arranged things to make it seem that way by carefully choosing a loaded measure in the first place.

You need to select a category for comparison that does not immediately and arbitrarily bias your experiment in favour of one group or the other. Including churches makes your experiment biased, because you are valuing one society's expression or idea of generosity over another's. At that moment, your experiment ceases to be an experiment. It becomes instead a kind of slippery, rhetorical "quasi-argument" that looks a bit like a proof but is actually nothing of the kind. Focussing on giving to secular charities, however, suffers from exactly the same problem.

This is why a real-world example like the one Valkyryn cites above is interesting, but does little to refute the study - it obviously exists on a completely different intellectual level and has, as others have pointed out, limitations of its own.

For my own part, I am vaguely disgusted with the whole business. It would be better, I think, to change society for the better to fix social ills than to waste time arguing over whether this group over here has fundamentally better morals than that group over there.
posted by lucien_reeve at 7:29 AM on December 29, 2012 [1 favorite]

Doesn't a liberal giving to Greenpeace, the ACLU, or NPR count as a "tribe/in-group one identifies with"?

I can't tell from the Chronicle of Philanthropy site whether Greenpeace and ACLU donations and the like (501(c)(4), not tax-deductible) are counted in their study. Anyone?
posted by naoko at 8:54 PM on January 1, 2013