As part of Charity Entrepreneurship's 2023 Top Ideas contest, will we select "Informing voters about political candidates’ criminal pasts" as a top Mass Media intervention?
Many LMICs suffer from having a high proportion of politicians with corrupt and/or criminal histories. These candidates often get elected because voters may not be fully informed, or because they don’t put enough weight on criminality when casting their vote. However, evidence suggests that electing such candidates leads to detrimental impacts down the line, including more corruption, less investment, and lower GDP growth. This organization would aim to reduce the vote shares for criminal politicians by sending text messages to voters in priority districts, informing them about which politicians are “criminal” vs “clean” (based on public records).
Mass media interventions
By ‘mass media’ intervention we refer to social and behavior change communication campaigns delivered through mass media, aiming to improve human well-being. We intend to select 2-4 ideas out of the 10 presented to recommend to entrepreneurs who enter our incubation program. This market resolves YES if this idea is chosen; NO otherwise.
About the contest
In partnership with Charity Entrepreneurship, Manifold is sponsoring a $2000 forecasting tournament to inform which ideas end up selected
You can win part of a $1000 prize pool as a forecaster, for best predicting which interventions we choose.
You can win one of ten $100 prizes for posting an informative comment on Manifold that most influences our decision.
For contest details and all markets, see the group CE 2023 Top Ideas.
I think this market is undervalued.
The utility of this intervention wouldn't come from the "journalism" part, it would come from the "texting people" part. Putting the information in front of people where they wouldn't have otherwise have seen it is very different than it just being a "different platform".
Where this would likely matter most is local and regional elections where people would otherwise just vote along party or ethnic lines. This text might be the only time they see the candidate's name. I think people tend to imagine the main impact of electoral interventions being on big, country-wide elections (like Mandela in South Africa). These likely already get a lot of media, I don't think this intervention would shift them much. Local positions drive a huge portion of corruption, I think there is big potential for impact here.
It's fine (probably even good) if having a criminal background doesn't preclude a candidate from being voted for.
The evidence for impact in electing fewer politicians with criminal backgrounds at the local level seems very solid, see:
22% difference in (proxied) economic activity from election of criminally accused politician
19% difference in consumption in districts where criminal politician narrowly won v. distrcts in which one narrowly lost
I looked into if text campaigns have evidence behind shifting voter behavior, I did find one study in which texting campaigns failed to produce changes in voting behavior. Not sure how much to update on this, couldn't find much other info (although I didn't look too hard).
Concretely responding to @RobertCousineau:
I am highly confident that at the local and regional level, people do not know of their leader's misdoings. I can try to find sources for this if it's relevant but I suspect your impression of people knowing might be based on solely thinking about larger elections and public figures in the spotlight.
I don't know how much impact a text drive from a charity (with its own name), funded by Charity Entrepreneurship, which is in turn associated with EA, would be able to create adversarial relations. The connection seems distant to me.
Investigative journalism in LMICs is already a thing. This is pretty much just putting it on a different platform. Everybody seems down on this and for good reason; it’s not clear why a new group would be particularly good at doing something that people are already doing, it’s probably going to be very politically controversial, and it’s not entirely clear how predictive criminal records are of bad governance in dictatorships or places where the most corrupt people are also less likely to get caught.
Aren't politicians in corrupt countries always trying to throw their opponents in jail, defeating the point of this? Like, public records would show Nelson Mandela as having a prolific "criminal" "history."
My low confidence intuition from spending some years in India is
all politicians have a corrupt background
People more ok than you’d think about their politicians being somewhat corrupt.
I would refrain from supporting this charity as I believe that its net impact on the world is likely to be negative.
The act of highlighting negative information about political candidates, particularly in LMICs, seem very likely to place the Effective Altruism (EA) movement at odds with the prevailing (and future!) power structures in these regions. This adversarial relationship will lead to increased difficulties in delivering aid, ranging from bureaucratic obstacles to direct resistance from these authorities.
Moreover, such intervention in foreign elections, even with noble intentions, is likely to perpetuate the notion that EA is a WEIRD neo-imperialistic endeavor. This perception could, in turn, undermine the overall EV of the EA movement by damaging its reputation and credibility.
The tractability of this initiative also appears questionable. It's likely to me that the majority of individuals living under corrupt regimes are already aware of a large number of their leaders' misdoings. This knowledge is probably overshadowed by feelings of resignation driven by electoral manipulations, the scarcity of untainted political candidates, and a history of poor governance.
In conclusion, while the intention to reduce political corruption is commendable, the proposed method appears to overlook the complex dynamics at play in LMICs and the potential benefits seem outweighed b y the potential harm to the overall EA movement.
@RobertCousineau If you think this response is valuable, please comment which parts and I can look further into validating my suspicions.