A little while ago, in a post reflecting on the role of academics in the anticorruption movement, I noted the distinction between anticorruption classes that focus on “teaching of skills” (helping students become effective lawyers, policy analysts, critical thinkers, etc.) and “teaching of values” (using education to inculcate anticorruption norms and reduce cultural tolerance for corrupt activities). In this post I want to pick up on that latter theme, which has become increasingly important to anticorruption activists and policymakers. Fighting the “culture of corruption,” many have persuasively argued, requires not just changing incentives and formal institutions, but also changing norms and values. And one way to change values may be through education–not only formal classroom education at all levels, but other forms of educational campaigns. For example, many attribute the success of Hong Kong’s Independent Commission Against Corruption not only to its law enforcement efforts, but to its broad-based educational campaigns to change the attitude of the Hong Kong population. Many countries have tried to emulate some version of this broad-based “anticorruption advertisement” campaign, and there are at least anecdotal examples of such programs making a difference (though not, to my knowledge, and rigorous assessment through something like a randomized controlled trial).
But these sorts of education efforts, if not carefully designed, can prove not only ineffective, but counterproductive. I recently came across a very nice analysis by the political scientist Frederic Charles Schaffer making this point, drawing on a detailed case study of anti-vote-buying campaigns in the Philippines, and to a lesser extent in Thailand. (I haven’t yet had a chance to read Professor Schaffer’s 2008 book, The Hidden Costs of Clean Election Reform, but I gather it goes into much more depth and discusses a range of other issues and countries as well.) The paper is from 2005, so it’s possible some of the specific examples and criticisms might no longer be apt, but my sense is that the larger points are still highly relevant, and quite important to anticorruption reformers who want to use mass education/advertisement campaigns to change citizen attitudes and behaviors toward corrupt practices. I won’t try to summarize all of Professor Schaffer’s nuanced account, but here’s what I take to be the essential argument:
In the Philippines–a country with particularly pronounced class divisions–vote-buying (or, more broadly, the provision of particularistic material benefits by politicians or party operatives, in the expectation–implicitly or explicitly–that these benefits will be rewarded with votes) is a pervasive problem, particularly in poor communities. (That this is indeed a widespread problem, and that this problem results in worse public policy outcomes, particularly for those same poor communities, is reasonably well documented.) To address this, a range of Filipino NGOs, along with the Catholic Church, launched a series of initiatives designed to discourage vote-buying. These included, for example, magazine, newspaper, and TV adds, as well as fliers, saying things like “In this election, don’t sell your character, don’t sell your vote”, “Do you love the country or the money?”, “Don’t be blinded by money. Vote with your conscience”, and so forth.
But it didn’t work (or at least there’s no evidence that it did). To find out why not, Professor Schaffer and his collaborators interviewed a random sample of poor Filipinos. They learned a few things.
- First, many poor Filipinos feel that their more affluent compatriots don’t respect them, and look on them as filthy, ignorant, greedy, or even subhuman. They experience countless small humiliations, and they deeply resent this.
- Second, although much straight-up vote buying likely does occur in the Philippines, a great deal of what goes on is more like politicians or party operatives providing benefits (like paying for wedding or funeral expenses, or helping out with health care) that many poor Filipinos appreciate as a sign of respect and concern with their well-being–sentiments they don’t often feel from the wealthy and powerful.
- Third, when middle and upper-class activists campaign in poor communities with slogans like: “Don’t be blinded by money,” “Do you love your country or the money?”, “Your vote is important, don’t sell it for a bit of pocket money,” this doesn’t change their minds–rather, it triggers even more resentment and alienation. They experience these messages as condescending, and they reacted to the ads with comments like, “[T]his ad []is irritating; they think when I vote it’s because of money”, “Automatically this ad impresses upon you that you can be bribed. I find that insulting,” and “The words hurt; they see [poor] voters as greedy.”
This matters not only because voters are less likely to respond to educational campaigns that they view as condescending and out of touch, but also because this alienation can actually make things worse. Professor Schaffer doesn’t have the evidence to rigorously document the alleged adverse effects, but he offers anecdotal evidence at least suggesting that when poor voters feel more alienated, they are less likely to volunteer to assist in the NGOs that are trying to build support for anticorruption efforts, and less likely to volunteer for other crucial roles, like poll-watching in poor election districts. More generally, a feeling of alienation–a feeling that more affluent Filipinos don’t understand them and don’t respect them–can actually fuel corrupt norms by making anticorruption seem like an upper-class value invoked to criticize the poor and the candidates they support.
None of this is to say that anticorruption education campaigns are not useful, indeed vital, to inculcating ethical norms. And indeed, as noted at the outset, some such programs, like that in Hong Kong, appear to have been quite successful. But Schaffer’s evidence provides a number of reasons to be careful and conscientious in the design of those programs. A few suggestions, many drawn from Schaffer’s paper, with a few of my own thoughts thrown in:
- When designing these sorts of educational or advertising campaigns, it’s important to make sure one has an accurate understanding of how the (allegedly) corrupt activity actually manifests itself, and how it is experienced by the targets of the campaign, rather than relying on assumptions or subconscious stereotypes.
- Even when the diagnosis is accurate — for instance, even when it’s true that many people are crudely selling votes in exchange for cash — shaming approaches may not always work, because they provoke denial, resentment. and backlash. Simply telling people to behave might be less effective than weaving anticorruption messages into more general discussions about the problems and struggles of the target population.
- Where possible, members of the target population should be included in the design of the campaign and materials–indeed, when possible, members of those communities should be given a leading role.
- Any sort of media campaign should be pre-tested on focus groups. Amazingly, according to Schaffer’s account, this wasn’t done with the anti-vote-buying advertisements in the Philippines, even though such focus group testing would have been standard for marketing a commercial product.
- As I emphasized in an earlier post, in addition to surveys and focus groups, field tests using randomized trials (and, where possible, measurable behavioral outcomes) may be the best way to produce reliable information about what actually works, and what is counterproductive.
GAB | The Global Anticorruption Blog 
Very interesting topic. I will add that in most developing countries clergies and religious institutions play a crucial role in shaping the culture the majority of population and even schools curriculum include a class in religion. Since no religion will condone corruption, therefore I guess If governments manged to communicate with these institutions to engage them in action this might fast track the process. especially if they built on the same suggested standards of diagnosis and defining the targeted groups.
I generally agree, though one of the interesting things about Schaffer’s study is that the anti-vote-buying campaign in the Philippines that he studied _did_ try to work with the Catholic Church, but it still wasn’t so successful. Indeed, it’s possible that the campaign’s organizers may have overestimated the positive effect of involving the church.
But your more general point seems quite plausible to me. I haven’t seen very much on the integration of religious (or, to use the euphemism now popular in the US, “faith-based”) organizations can play in anticorruption campaigns. I know Roberto Laver and one or two others have written about this, but there’s surprisingly little given how, as you say, religious organizations might seem like natural partners in these efforts.
Of course, there’s also the concern about corruption _within_ religious organizations (see, for example, recent troubling stories about financial malfeasance in the Vatican), but that’s a topic for another day.
Matthew — A few thoughts. As a preliminary matter, I’m not sure I’m fully on board with the notion that these messaging campaigns can, as an empirical matter, be counterproductive. I think the principal problem is that we need to be precise about how we measure the effectiveness of these programs. If the measure of success is the ability of the program to inspire members of the viewing/receiving public to buy into or contribute to the broader educational mission, then there seems to be a straightforward causal theory for why a condescending messaging campaign could be counterproductive as measured on that spectrum. Absent empirical support — which you note is missing in Schaffer’s paper — however, I’m not prepared to buy into the idea that a poorly conceived and run messaging campaign like this will result in a meaningful increase in vote buying (or corruption more generally). It is harder to buy the causal theory that looks a little bit like: I’m offended by this message. Therefore, I will decide to sell my vote when I otherwise would not have done so.
That said, I think both Schaffer’s paper and your post make a very important point: if we’re going to devote resources to these types of value development campaigns, we should make sure we’re not wasting efforts or alienating the very people we’re targeting. And the reason for this is straightforward: if you alienate the target audience, they will grow numb to your message, and even if it doesn’t make the prevalence of vote buying spike in the short term, it may very well both worsen the problem and undercut any potential normative solutions in the long-term.
Another point: It seems that the designers of these campaigns could learn a simple lesson from general parenting or basic social interactions — shaming is often not an effective mechanism for teaching values and changing associated behavior. This campaign associated positive values (anti-corruption messages) with negative emotions (the feelings of condescension and class-bias), eroding the power of the former. Perhaps the lesson to be learned is that it is better going forward to frame the lesson (don’t sell your vote) in ways more likely to evoke positive emotional reactions (e.g., “Vote with your conscience.”; “Your vote is your own.”; “Speak Your Mind. Vote With Your Conscience.”).
Jordan, I’d like to push back on your last point a bit, because I don’t actually think it’s as simple a lesson as you suggest. Class divisions probably play a significant role in how people perceive different messages, and I think Matthew is right to point out that people differently situated might miss how certain messages are received. For example, I’m not entirely convinced that a slogan like “vote with your conscience” actually resolves the problem in this case. It could be received in a similar way, i.e., “what else do they think is informing my voting?” or “do they think I wasn’t using my conscience before?”
I think this is particularly sticky in scenarios where the targeted vote-buying behavior is less explicit than a cash-for-ballot exchange. I don’t think it’s hard for someone to rationalize that voting for someone who provides their family with important benefits like weddings or funerals (or, in the Philippine Vice President’s case, birthday cake) is voting with their conscience.
This is why I like Matthew’s point about members of the targeted population taking a lead role. It’s important for those people to have a say in what is empowering for them, and what they perceive as problematic in the current system.
Jordan,
Your point about the absence of strong evidence of counterproductive effects is of course well-taken. We always need to be cautious about not over-reading the evidence or jumping too quickly to unwarranted conclusions.
That said, I think I was perhaps not clear enough about the possible “counter-productive” effects that Schaffer (as I read him) is suggesting might occur. He’s not arguing that saying “don’t sell your vote” will lead to an increase (certainly not a short-term spike) in vote-buying. Rather, the main counterproductive effects may occur through two channels: First, exactly as you suggest in your comment, is alienation of the target audience, causing them to “tune out” anticorruption messages, or associate them with upper-class condescension. The second, more short-term effect is not a spike in vote-buying as such, but rather a decrease in the likelihood that people in the target population will join local NGOs, volunteer their time to monitor polling stations, etc. — again because they are turned off by the condescension coming from those groups.
I think your ideas about positive messaging are helpful, though as Bea points out in her comment, it may not always be straightforward for outsiders to design a message that is perceived in the right way. That’s why involving members of the targeted communities, or at least doing some focus-group studies, may be so important.
I think Mohamed makes an excellent point. Although it is superficially unrelated to the topic at hand, there is some very interesting literature on the critical role Shiite mosques played in the dissemination of information and crystallization of norms in post-Saddam Hussein Iraq. As David Patel, a political scientist who has studied the topic, writes “[t]he influence of religious authorities is not just due to the message of Islam; it is due to the ability of Islam to deliver messages.” The same could likely be said of Christianity, or any other religion that involves periodic gathering in large groups to listen to community-wide messages. I’m not sure I share Mohamed’s optimism that religious figures will inherently condemn corruption. But an additional suggestion, similar to the ones you’ve already made, Matthew, is that campaigns should build upon pre-existing, widely recognized methods of communication (which, in many cases, will include religious networks).
On a different note, related to your first suggestion, I think a successful campaign against vote-buying must do much more than preach values. A necessary component is a genuine understanding of the people’s conception of democracy. In countries where elected officials (and other wealthy elites) act more as patrons than as representatives, the vote itself has little inherent value. If I live in a society where I can get more from selling my vote than from voting, then whether an anti-vote-buying campaign is framed in a positive (esteem-building) or negative (shaming) light doesn’t really matter all that much. Of course, if everyone voted for their candidate of choice, the political environment and the role of the government might well change. The trick of a campaign is to get people to forego short-term rewards for long-term benefits (and to hope that those benefits materialize).
Building a bit on Liz’s point, I think it’s important to disaggregate two different kinds of corruption when looking at educational initiatives. On one hand you have vote-buying–I trade my vote for a specific benefit from [you, politician]. On the other is extortion–I should rightly receive a certain government benefit, but must pay a bribe to get it. Both cases are examples of governance failures, and both are examples of official corruption. But the position of the (vote/bribe) giver is very different, and this matters for how you choose to address the problem.
In the first case, the voter holds something of value (their vote) that they trade for a benefit for themselves or their group. There’s obviously a sliding scale here–we might think it’s permissible for a union to vote for a candidate who promises to increase benefits to public sector workers, but find it distasteful when a politician directly pays for a family’s healthcare to secure their vote. Nevertheless, the voter in that circumstance could very well be deciding to maximize the benefit they can receive from the government. The root problem here is generally related to unequal distribution of government benefits. And as the paper finds, for those who don’t receive much from their governments, it sounds a bit condescending to call them greedy for taking the few perks on offer. Addressing this problem probably requires a higher-level initiative to ensure that government services are provided (more) equally.
The second situation is, in my view, much worse. There, the cards are stacked against the putative recipient, and they *must* cough up funds to get some kind of benefit or service. I’d imagine some educational initiatives, particularly training on how to take corrupt local officials to task, could be much more useful here.
Even there, I’m broadly skeptical of value-based initiatives, because it’s not clear to me that corruption is generally driven by lack of values. Even in schools where bribing teachers for high grades is rampant, that behavior is publically disavowed and students and teachers alike maintain the fiction of meritocracy. Why? Because folks aren’t confused about whether it’s wrong to pay for an A.
It seems more often driven by a combination of unchecked (or poorly checked) power in the hands of government officials (or some other person in authority), the bribe-payer’s inability to function effectively absent bribery, and the fact that some folks with money can now pay for something that ought to be earned rather than bought.
I think what’s really interesting is the effect the type of vote-buying has on how these campaigns are perceived. The challenges involved with condescension and shaming are particularly acute in situations where actual vote-buying is not taking place, and instead individuals are responding to benefits received from politicians. It’s actually difficult to conceive of a way that this could be dealt with through catchy slogans. In fact, I think that this may be where more comprehensive educational efforts are most needed.
In local elections, where this more indirect form of voter-buying is taking place, it’s possible that individuals perceive this exchange as a legitimate practice. Votes are given to politicians who directly serve the needs of the people through the provision of benefits. Educational efforts that disassemble the culture of corruption are really needed in these circumstances, to point out why these types of indirect buy-offs are inappropriate. I agree with you, Matthew, that it’s very important that any such efforts take into account the real struggles of the target population. In the end, an approach which highlights that population’s stake in anti corruption efforts will be more successful than broad appeals to nationalism or civic duty– particularly for populations that feel alienated from those in power.
I’m very much in agreement, but maybe it’s worth noting that we’ve got a short-term/long-term problem here. More comprehensive educational efforts, as part of a concerted effort to alter social norms and expectations, may indeed have significant long-term positive effects (though it may be worth pausing to note that we do not, to my knowledge, yet have any strong systematic evidence of this). But in the short term, if there’s indeed a problem — with vote buying, patronage, or similar — it may not seem entirely satisfactory to say that we need a comprehensive educational effort that disassembles the culture of corruption. I don’t really have any good answers, and I suppose any successful strategy may need to have multiple components.
I agree with Bea that it’s possible/probable that the locals don’t perceive that they are doing anything wrong. Especially if there is an effective secret ballot (don’t know enough about the Philippines to know whether that’s the case), people would be voting not in what they see as a quid-pro-quo buyoff but rather for the party they think best supports their needs, as seen through the direct support they provide. The issue may not be condescension in general, but rather the fact that locals felt as though they were being accused of a corrupt activity that they didn’t believe they were doing.
I’ve seen some interesting research on how to ask people about sensitive questions like whether they think they are selling votes — basically the idea is you ask one group ‘how many of these things have you done’ and include a list of 8 harmless things. Then you ask another group the same question, but give them a list of 9 things, including vote selling. You never ask anyone to confess directly to vote selling, but by comparing the numbers of the two groups, you can figure out the overall presence.
Perhaps a poll like that before this kind of campaign might have revealed that people don’t believe they are selling their votes. This could have convinced the anti-corruption workers to focus on showing people why their current actions left them worse off, rather than trying to convince them to stop doing something that they didn’t think they did.
Pingback: Conflicting Philippine Identities and the Fight Against Corruption | GAB | The Global Anticorruption Blog