 Welcome to this SOAS ACE webinar, titled Making Anti-Corruption Messaging Effective, the Critical Importance of Visibility and Targeting. Today's speakers are Karen Pfeiffer of the University of Bristol, Lena Coney Hoffman of Chatham House, and Heather Marquette and Nick Cheeseman, both of the University of Birmingham. Today's session will be moderated by Pallavi Roy, who is the Joint Research Director for the SOAS Anti-Corruption Evidence Programme. And with that, I'll hand over to Pallavi now. Now there seems to be some consensus on understanding the causes and costs of corruption, but little exists on how to effectively address and reduce it. So the SOAS ACE programme, SOAS University of London, an Anti-Corruption Evidence Programme, is part of a much larger consortium. There is global integrity. There is a senior and organized crime bit of ACE, which is headed by Heather Marquette and about to launch. So we're a big family, a rather large umbrella. This webinar is organized by researchers who are part of the SOAS ACE programme, the Anti-Corruption Evidence Programme. It is an innovative anti-corruption programme that adds to this discussion of what makes for effective anti-corruption policy design. And it's funded by the UK government, works across Asia and Africa. And the starting point of the SOAS ACE programme is that we don't take as a given that enforcement starts from above, that someone is watching and as soon as a rule is broken, the relevant authority will just sweep in and we'll take corrective measures. And we know that one of the most important reasons why corruption persists is that there is an enforcement deficit, even where very elegantly designed rules exist set to some kind of normative benchmark, desired normative benchmark, enforcement can leave much to be desired, particularly if many or most in that community do not want enforcement. So rules are broken because in the vast majority of cases, it's not in the interest of the community to stick to the rules. It's not just about that big, bad, politically connected guy, big corporation, big political party, big man who wants to break rules just because he or she can. In many cases, the community or the members of community break rules simply because they don't have an option. That's the only way they have to solve some problems that they are facing and rule breaking just becomes necessary behaviour because they don't have another choice. And more importantly, being rule following is of no help to them. You know, sticking to rules isn't usually going to help if that emergency hospital admission requires one to pay a bribe. That makes it what we call a very adverse context. The incentives for community to be rule following is just not there. It's not only a few bad apples who are breaking the rules. There is a societal issue over here. So as we describe it, this is the adverse context of designing in which one has to design anti-corruption policy. If there's little incentive for most people in society to uphold rules, what then makes for effective anti-corruption? And this is what we call where enforcement or horizontal enforcement matters, not just enforcement, but horizontal enforcement matters and that rules are upheld in self-interest. It's not top down, it's not vertical, but it's horizontal where I'll uphold a rule and my peer will uphold the rule and I can check that the peer is upholding rules because it's in my self-interest to do so. But millions have been spent by private and public donors to change, you know, private and public donors to change community opinion about corruption. When the reality is that unless incentives are changed, shaping or shifting, rather shifting opinions, not so much shaping and shape opinions, but you need to shift opinions, will actually do very little. And one of the most popular tools in this effort is to generalize messaging about corruption. That corruption is bad and that corruption can hurt society. And that's essentially what this conversation seminar is about today. Because research conducted in Lagos by SOASAS and the authors of that paper are very much part of this panel, has reinforced an emerging and very, very important strand of research and anti-corruption that if targeted messaging addresses specific problems, the chances of a positive outcome are higher, even more important bit of messaging we think. And that significant contribution is that generalized messaging can actually prime or even sort of motivate citizens to overemphasize the role of corruption in their lives and can result in what is called corruption fatigue. So our other panelists do have been working in parallel on very, very similar issues dealing with corruption that has become socially entrenched in some context, whether in the context of functional corruption or looking at it through the lens of social norms or in issues of enforcement and anti-government, enforcement of anti-corruption across governance agencies. Now this work also luckily fits very nicely with the SOASAS framework that states that there must be targeted players within the specific sector who want to implement relevant changes in policies, and this is important, and have the capacity to do so if the appropriate policy and messaging isn't taken, that's very important. Players who want to enforce change and players who have the capacity to enforce change, these two coming together is extremely important for us. To put it another way, our framework states that anti-corruption is more likely where actors are interested in following rules, specific rules, have the power to monitor peers and take them to task if they are actually breaking the rules. So the difference in this approach is that we are also devising strategies for horizontal enforcement while accepting and agreeing that vertical enforcement is as important. So we are also parallelly devising strategies for horizontal approach. So we need to talk, we need to ask about who will support the policy, who will enforce it, who will implement it. If there are no clear answers to who will enforce, messaging is unlikely to help. So we are therefore looking at a number of interdependent factors like effective messaging, like what is feasible messaging, feasible policy, what does that mean for anti-corruption policymaking? What are the incentives that we are looking to restructure? And in the context of this very adverse context where there seems to be a corruption equilibrium. So the conversation today is essentially about how these two will interact. So what we hope to explore today are answers to two questions that begs everyone in the policy design world the so what and the how to. So given who our panelists are, they are best placed to have this conversation and I have absolutely no doubt we will walk away with some very popular looking answers to these questions in the end. And it's time I introduce them. We have Karen Pfeiffer, Dr. Karen Pfeiffer who will begin the conversation. Dr. Karen Pfeiffer is senior lecturer of international public policy and governance at the School for Policy Studies University in Bristol, UK. Our research mostly focuses on the causes and consequences of corruption and the unintended impacts of anti-corruption policy. Karen has conducted research in several different countries and most recently in Papua New Guinea, Indonesia, Uganda, South Africa and Nigeria. She's one of the co-authors of the paper that I was just referring to. Her research has been published in esteemed journals like the British Journal of Political Science, Governance and World Development. She regularly advises transparency international and has worked with the World Bank, the UNDP, OECD, FCDO, the Development Leadership Program, the Australian Department of Foreign Affairs and Trade and the German Development Agency, GIZ on anti-corruption policy and analysis. Our next panelist is Dr. Lina Coney-Hosman. Lina is an associate fellow of the African Program at Chapman House and lead researcher for the Program Social Laws and Accountable Governance Project. She's an honorary senior fellow of the Global Evergreen Alliance. Lina researchers and writes on informal institutions, politics, governments, corruption, food security and regional trade in West Africa. Some of her work is very neatly with the work we're doing in this because we're all looking at how incentives can be changed in context of very embedded corruption. She was a technical advisor to the Permanent Interstate Committee for Drug Control in the Sahel from 2016 to 2020. She was also a Marie Curie research fellow at the Luxembourg Institute of Social Economic Research and an anti-corruption investigator in Nigeria's independent commission for our practices, more popularly known as the ICPC. We then have with us Heather Market. Heather is Professor of Development Politics at the University of Birmingham, UK and is currently seconded part-time to FCO's Research and Evidence Division as senior research fellow, Governance and Conflict. Heather's research, which has been funded by the British Academy Global Challenges Research Fund, DFID, Australian DFAT and the EU focuses on corruption and anti-corruption interventions, developing politics, aid and foreign policy and increasingly transnational organized crime. Heather is leading a new component of ACE, I referred to this in the beginning, focused on serious organized crime, transnational corruption, illicit finance and teleprosy known as SOCA, SOC ACE and with a website launching soon. So do keep a lookout for that one. Our other panelist, final panelist is Nick Chiesman. Nick is Professor of Democracy at the University of Birmingham, UK and was formerly the director of the African Studies Centre at the University of Oxford. He mainly works in democracy, elections and development and has published research in Ghana, Kenya, Malawi, Nigeria, Uganda, Tanzania, Zambia and Zimbabwe. Nick is the author or editor of more than 10 books on democracy and political institutions in Africa. He was awarded the Joanie Ladnowski Prize for Outstanding Professional Achievement by a mid-career scholar in 2019 by the Politics Political Studies Association of the United Kingdom. In the same year, his efforts to promote a better understanding of democracy and how it can be protected and strengthened won the Celebrating Impact Prize of the Economic and Social Research Council for Outstanding International Impact. He is a frequent commentator on African and global events and Nick's analysis regular features across leading global media platforms. We had another panelist with us, Idhayat Hasan. Unfortunately, Idhayat, who is the director of the Centre for Democracy and Development in Nigeria and one of the partners of SOASAS and has also worked extensively with the other panelists, she's known for her excellent work on corruption, anti-corruption and governance, especially of Nigeria's anti-corruption agencies. She would have been a fitting fifth person to complete this panel and be a part of this conversation. Unfortunately, she's in Sokoto with very, very bad internet coverage so she really can't make it. I mean, I can't even get my WhatsApp messages through to her. So that won't be joining us that she was supposed to be the fifth panel over to the panel now, starting with Karen. Thank you. Karen, over to you. Thanks very much to everyone who organized this and for putting this together. I was thinking about, you know, looking at the panelists and thinking about what could be, what I would be the best contribution I could make to this discussion, especially at the beginning. And I thought without presenting research or anything like that, it might do as well to talk first about sort of what the evidence says. So the kind of motivating question I have here is when can messaging change incentives sufficiently on its own? And when are parallel policy changes required to link messaging with feasible strategies? So really complex question. And I think it requires us knowing what, what the evidence has said about messaging so far. So I thought maybe I could, first give a really quick whistle to stop toward through what's been done. So the motivation behind these studies is that, as you said, these awareness raising campaigns are already occurring around the world. With the various partners, I've conducted three messaging studies so far. And in those, I've tested 14 different types of messages in three different countries. So we've motivated, most recently worked together, Nick and I on Nigeria, but I've also done this in a similar study in Papua New Guinea and in Jakarta. And aside from that, there's also been a few other, other scholars who have led these types of messaging studies. So it's a growing kind of bubbling research agenda. If you take all the results together, it's only one message that's been tested so far that has seemingly had the intended impact. So one awareness raising message about corruption has seemingly influenced people to think about corruption or even anti-corruption in the way we would hope it would if we thought that awareness raising could be an effective tool in anti-corruption. And that one message was in Papua New Guinea and it suggested that when you discuss corruption as being a local issue that impacts local communities it might encourage people to think more positively about reporting corruption. So I think we just need to keep that in mind. Nick and I studied, it kind of goes further and has some worrying, but also some complex findings in there. So the worrying bit is that four out of the five messages we tested actually worked to encourage people to pay a bribe in a game that we designed on a tablet. So it's a bribery game. But if we split the sample up a bit, it seemed that it's actually that kind of negative impact is most pronounced in those who were already worried about the size of the problem. Those who were less pessimistic about corruption being extremely widespread in the society, we found there that there was one message, so a message about linking corruption to taxes and tax payment that that discouraged bribery in our game. So there's kind of glimmers of hope is what I want to emphasize. But lots of things have been tested. So this isn't just about raising awareness to the issue of corruption being widespread. But it's also about bringing in some of these messages that have been tested, which have had negative impacts in research, have taken other tones. One, for example, encouraged, only encouraged people to get involved in this anti-corruption civic response. It's now more than ever, it's easy to join anti-corruption organizations or to fight against corruption. Here's a phone number where you can report corruption that was tested which emphasized the success that the government had in fighting corruption. So the idea there is maybe if we put positive spin on it, it won't have such a negative impact. But both of those sort of neutral or positively tone messages did have the same backfiring effect. And what we think is happening with respect to this backfire, some messages are, even when we don't intend them to, but they're reminding people that corruption is widespread. And so it's triggering that corruption fatigue that Pilavi talked about. It's making potentially people think that the problem is too big to fight and so instead of resisting it, I mean, in what Nick and I found, it may be encouraging people to go with the brain. So it's really challenging to think about the question posed. You know, whether or not when can messaging change incentives sufficiently based on that evidence base? I mean, quite frankly, the evidence so far doesn't really let us ask the question or answer the question. We don't know about how different policy environments, for example, might impact on the efficacy of messaging strategies because the literature is still really small. It hasn't been tested in that many countries or in different types of contexts. There are some hypotheses that would be great to test going forward. For example, you know, to really examine whether or not people believe that the government has any sort of credibility in fighting corruption, maybe in that context, certain types of messages might work. But I think for me, the bottom line at any time I've talked about this body of work is the sort of like take home point is always that at this point, what we know is that messaging strategies need to be tested before they're deployed in any sort of big scale. And that's our best chance of, you know, avoiding doing more harm than good with them. I'll pass it over. It's okay. Fascinating. So I think that's also what one gets from your paper carrying with me, which is that messaging strategies need to be tested. We've had a lot of donor money going in. We need to do some small scale testing, you know, the policy background matters. And I think one important fact that I picked up from what you said is that the policy background matters because we need to connect this insight to creating sectoral incentives for horizontal support. If that's the policy background, well, how can messaging dovetail? You've made two very important points, which I think we can pick up on during the second half of the conversation. So I'm going to take that as, you know, conclusions from your intervention here. Next up with us is Lina. Lina, you're on next. Thank you, Pallavi. And thank you, Karen, for, thank you, Pallavi, first of all, for setting the scene. Thank you, Karen, for just ending on that note about testing messaging strategies. And I think it's really great that that's kind of the line. I would like to thread through what I'll be speaking about just responding to that first question of when can messaging change incentives sufficiently on its, on their own or its own. And when our parallel policy changes are required to link messaging with feasibility strategies. So I'll speak a bit about the work that we have done on the social norms and accountable governance project with Chatham House. I think it's, I think our inception all the time that we've been doing this work is around about the same, around about the same five to six years period of ace. So we've kind of been in the same environment and looking at the same phenomena from different sites. The specific site that we've been working, we've been looking at the phenomena of corruption on the snag project is looking at corruption as this collective practice and really trying to understand the social expectations, the social dynamics of this practice. What are the context or the conditions that allow for corruption to be normalized? What are the social influences that people are subject to when they are in an environment where they have to make a decision to either, for example, give a bribe or not. So I think the, the simple answer to, if messaging strategies can work on their own, as Karen said is no, messages have to take into consideration the environment in which they're introduced, the kinds of forms of corruption that are normalized or people are desensitized to in that environment, the kinds of corruption that people are apathetic to and the ones that, and I think I'll speak, I'll introduce kind of this point and then thread it to when I, when we come back the next round, the kinds of corruption that people are mistaken about the beliefs of others about. And that's really one of the very interesting findings that we have uncovered in our work. We've done two rounds of a national household survey in Nigeria looking at different corrupt behaviors. You could group these behaviors under bribery, embezzlement and electoral fraud. And we really wanted to test the kinds of assumptions. First of all, the assumptions people had about the behaviors and we also wanted to test people's expectations about what people do in those kinds of circumstances or those kinds, those kinds of, yeah, interactions. So we looked at law enforcement, healthcare. And we've also looked at vote selling. We've looked at education as well. And what we found across these behaviors that we found a couple of things quite interesting and to do with messaging across these behaviors is that actually in terms of talking about social norms of corruption and social norms being understood at the, as the behavioral rules of how people engage in society. And there is like the operational definition of a social norm that first of all, you hold onto an empirical expectation. That's your expectation of how people behave. And your normative expectation, the expectation, the beliefs you have about how other people think you should behave. And then you have a conditional preference to behave in that kind of way when you encounter that situation. The example that I think maybe the Nigerian audience or in similar audiences might connect to really well that we use in talking about social norms is the example of giving something with your left hand versus your right hand. I'm from Nigeria. I always, and I'm left handed. And I always adjust my practice of giving people things based on where I am. From the airport when you arrive, you know, I'm immediately told that your left handed behavior is not acceptable here. And regardless of whichever way I appear or I seem, there is a social norm that when you're giving someone something, particularly someone older than you, you have to give with your right hand. And now that I haven't been in Nigeria for such a long time, that is going to be like a real sharp adjustment that I will have to make. So in terms of talking about social norms around corruption, even though corruption might be pervasive in environments such as Nigeria, what we actually found across the behaviors and the context that we have explored is that social norms themselves are not actually as widespread as the actual phenomena is. And we think that's a very, very important finding to uncover and a very, very important finding to bear in mind when we're talking about corruption, particularly in these adverse contexts that you spoke about where there is a lot of functional corruption, where the options, for example, like the one you gave about if you're trying to get a hospital bed for your sick child to give 50 Naira or 100 Naira, it's not even, you know, there is no debate about the fact that a parent would engage in this kind of corruption to ensure that their child has a bed in the hospital. But what we actually found and bringing it to the point of messaging is that a lot of the corruption, the corrupt practices or at least the collective practices that we examined that are termed to be corrupt, what we actually found as driving those practices are descriptive norms. So people are really just engaging in this behavior. Of course, the context is very, very important for people engaging in these behaviors because they see others engage in this behavior. So it's really empirical expectation driven. And I think there's a lot of research from Christiana Bikari and I think there's a new paper out by, I'll get their names wrong, called Kershia or Keisha and it's around social norms of cooperation. And I think the general gist of that kind of research and I think Niles Kobles has actually done this on bribery is that when people's empirical norms and empirical expectations and normative expectations diverge, people tend to go with what other people are doing. So even though you might have a personal belief that this behavior is wrong, this behavior is unacceptable and morally opposed to this behavior, people engage in the behavior if they believe others would do the same. So you don't want to be the one that is frozen out or you miss an opportunity because you didn't give a bribe or you didn't engage in a choral practice. And one of the key points that we try to make in our work is that these descriptive norms are very, very important and they often and across the different types of bribery and embezzlement we've looked at, they drive behavior. So it's not social norms, it's descriptive norms. And another thing we found out that ties into messaging and I'll just end my points here is that people are often quite mistaken about what other people believe about corruption and or certain forms of corruption. And we also think that's really, really important in the literature is called pluralistic ignorance and informs a lot of the collective action problems we have around anti-corruption because for most people they find these behaviors unacceptable. They would rather live in a more honest and low corruption society but nobody wants to take on the individual cost of anti-corruption. Corruption avoidance is simply too costly to risky so no one wants to take on that cost on a day-to-day basis. So it's really in terms of messaging one of the things that we kind of put into one of the recommendations or intervention and recommendations that we make is that where there are mistaken beliefs and particularly around behaviors associated with bribery and education and healthcare around bribe solicitation from law enforcement we actually found in and it's we did this study across Nigeria so not just Lagos like this particular one but in Sokoto and the like we found in specific states people are mistaken about what other people in their community believe about that behavior so most people thought I think in Enugu around law enforcement four out of ten people thought everybody else in my community is okay with this so when you have that kind of a situation where you are going through the world or going through your society environment and you think everybody else is okay with this form of corruption collective action is really really challenged in that particular context so on what kind of messaging works messaging that indicates you know what descriptive norms are and messaging the highlights people's mistaken beliefs about what other people believe because that's how you galvanize collective action if you realize that most people in your community really really don't like this practice and you're able to provide the kind of evidence that supports that position or that or that shared belief that is a powerful powerful component for collective action so just end on that and pass it on. Thanks Lena fascinating I mean I've always found your work on social norms in Nigeria extremely extremely fascinating and you make this very important distinction between what is a descriptive norm and what's a social norm and it's important to identify or make the distinction between the two because in this in this adverse context of you know corruption equilibrium we need to look for those opportunities and I think often anti-corruption you know people who work on anti-corruption tend to be sort of deceived by gloom and doom but I think across the board across what our panelists are speaking to what Karen spoke to what our work also speaks to is that there are that there are opportunities but I think the trick is that you know policy should allow most people within the sector to operate without corruption that's the policy benchmark that we want to use. Policy has to be such that it takes into account those those incentives for horizontal enforcement so the policy needs to be such that it allows whether it's in education or health care or in the electricity sector allows people to operate in such a way that essentially what is happening that you can isolate what we call the genuine thieves from the free riders the ones who have to who don't want to pay the cost of everyday corruption when that happens that's when you can start organizing collective action and that's what the evidence gathering really has to be about so what evidence do we have that this is the kind of policy which will allow us to create that condition that self-enforcement will happen and once self-enforcement happens I can monitor my peers and collective action that then comes in so I think we're already beginning to talk in very common terms and our next panelist Heather she's also spoken a lot and worked a lot on exactly these aspects of anti-corruption because one of one of the main points that she makes is corruption has a function it's functional corruption and therefore if you're just trying to kind of just clean everything out with a broom there are these unintended consequences so we need to understand that particular context and then design a feasible anti-corruption policy so Heather on to you now. Great thank you so much Palavi and lovely to hear from Karen and Lena so far I'm very pleased to be here as much as I would love to talk about functionality and could do I don't know I really wanted to focus as you say on unintended consequences and what this research Lena's research and others is telling us about this I've been very lucky to have followed this research from its inception and it just goes from strength to strength you know I'm really really pleased to see this this study coming out it struck me straight away when the original pilot project which was Karen had done in Jakarta when the Phoenix came in and they were so surprising because they didn't fit that common sense narrative around what we would expect to see with anti-corruption interventions as well and the issue of unintended consequences was one of the first things that we talked about Karen and I have a book chapter coming out very soon in a routledge handbook on transnational organized crime and it's a chapter on corruption and transnational organized crime and in it we look at five lessons coming out of anti-corruption evidence research and evidence that are relevant to the organized crime space the first one is you know we need to be problem led not solution driven the second is we need to get more specific the third unsurprisingly we need to think about functionality the fifth is we need to stop using political will as an excuse but the fourth one I think is really relevant here which is we need to test what works including our sacred cows and including things that are common sense as well and I think all of that aligns really nicely with Aces work with unintended consequences the research that Karen and Nick have done here really highlights why we need to be paying much more attention to potential unintended consequences in anti-corruption interventions as well you see the concept of unintended consequences in all sorts of policy analysis literature but it rarely features in researcher policy on anti-corruption as well and Ozreki writing back in 2015 pointed out you know and I'm quoting here although corruption and anti-corruption have a long history as research topics in the social science little is known about the unintended side effects of anti-corruption measures and it seems to be a really major gap in the literature given that anti-corruption is at its heart about forcing those with the power to abuse their positions for their own private gain to give up this power and to be accountable for their actions to imagine that that won't have potential unintended consequences just seems a bit unimaginable for more unintended consequences are about errors that come about when policy actors haven't considered all of the likely impacts of their policies or their interventions either because they lack the necessary information or because they refuse to consider information as well and that's the problem with sacred cows it's actually quite hard to tell which one it is on the ground another thing that makes it really challenging with corruption in particular is that any discussion of corruption or anti-corruption often comes with it moral outrage as well and that moral outrage the need to do to be seen to do something and to not question whether or not you should actually be doing something carries its own unintended consequences too and it's really why investing in research like this is so important and why testing matters because it's the way to move past our current belief about sacred cows in order to learn more about what works and importantly what doesn't work even if it's emotionally or politically difficult to do so one of the potential unintended consequences that the research could look at going forward if the team would like as well that really strikes me as important is when you think about anti-corruption messaging on the ground well extremists and authoritarians platform on anti-corruption messaging so their anti-state anti whoever is an authority as well and they use that messaging very effectively to attract followers who may themselves have picked up on messages whether it's formal anti-corruption messaging like Karen and Nick have tested whether it's journalism you know gossip anything like that as well you know this is a hypothesis I'm not saying that this is a finding you can claim from the research or anything like that but what if all of those anti-corruption billboards all of those lessons in school all of those things over the past 20-30 years have been feeding into those narratives and I think Karen and Nick's follow-up paper on what this might mean for democracy make some unbelievably important points and I assume and hope Nick's going to talk about those as well you know is there a possibility that we've been undermining democracy and feeding extremist narratives by trying to do the common-sense thing and we just don't know which is why you test and why we should be testing more rigorously the second point and final point I wanted to make about this thinking about messaging as well and looking at it in the other direction is the importance for researchers of getting policy messaging right as well especially when you have tricky research findings like this one as well. Researchers are not always well equipped or open to considering how those messages might land on the other side of the table so to speak and especially if those might be quite politically challenging keeping in the sort of vein of the counter extremism I've put on Twitter today and I think Duncan put up our Twitter accounts with a fantastic article by Lydia Wilson about counter violence and extremist policy and I'm just going to quote a little story she says very quickly where she was presenting to the UK government on what works in counter extremist work which is basically in her findings here pretty much nothing and she says here I'm quoting I suggested that we quit the disastrous CVE approach to focus on making our societies more inclusive and hopeful thereby addressing many social problems at once and simultaneously avoiding the stigma of accusing communities of potential terrorism this didn't go down well as policy advice do you disagree with anything I've yet said I asked my interlocutor and he repeats I cannot go back and tell my government this I lied in the question to the room does anyone disagree with what I've said and an advisor to the then prime minister said he would never publicly say what she'd said she forestalled my outrage I know she said but to save time in this meeting you need to know that I won't be passing that on so let's not get into it that's one of the things I think when you're presenting research like this to a policy audience in particular that it's important to understand what that actually means and how to impact that I think Nick and Karen are incredibly experienced researchers when it comes to engaging with policy and they do this it's much better than pretty much everybody has seen as is the wider SOAS team I do think there's always more that researchers can do to think about no but recommendations to help our friends and colleagues who have to make the case for why our research that's telling them to do things differently to politicians and decision makers who have the authority to do that and I really think this is a great project to watch for that as well because what I've seen and what they've written as well as what I've heard them say is they do that process and they start to go through so they help make that transition between what you're doing is absolutely wrong potentially quite dangerous and we need to stop doing it to okay here's what you do to get changes done and I think as they go forward in the next phase which I'm really pleased that through SOCASE it's going to enable them to take this research and adapt it to counter SOC messaging too it gives them a whole new audience and a whole new challenge as well there and it's going to be great to see what with that research aligns with the SOAS research on anti-corruption what's similar and what's different as well and how to continue to boost that consistent messaging about unintended consequences in doing things differently and I think I'll end it there thanks Heather as always fascinating intervention there especially the point you make about the unintended consequence of moral outrage and whether we are getting that right I'm frankly sitting here in India and to go a little off piece I certainly seem to think there's no great research that has gone into it but this whole anti-corruption moral outrage very well can provide the fig leaf that an authoritarian fundamentalist extremist government can actually clothe themselves with so and it's not just India we've seen that across it all begins with anti-corruption you feed into a certain sense of moral outrage and exclusion and then you just fall off a cliff but anyway we shan't discuss that here but I think that would strike a chord across all our participants and the other point that you made about researchers and how to land their research it's the messaging by the messengers that you want in a sense that how is it that we as researchers package I think that's not something we talk about and I'm really glad you brought it out and I think that's something that across the three sort of branches of ACE we've been working on how to make it practical so that we don't say it in a way that the policy implementer says this is not something I can take to my minister and that's a very critical issue to learn and I think the other point is that coming back to some of the points that we make in SOAS ACE and where I think the similarity lies in these contexts where the common sense approach can lead to the solidifying of a particular authoritarian power base I think our approach is that when power is important because you need to create very countervailing sources of power and that is really most likely to happen by splitting the powerful into those who will then contest that authoritarian regime in their own interest and I think that is the other puzzle that we in SOAS ACE are also trying to work on so thanks Heather for that really interesting intervention and Nick I think it's over to you to round off this particular session, over to you Nick. Hi everyone yes so if we haven't depressed you thoroughly enough I'm going to start by throwing in a couple of additional problems and then hopefully Chi will up a little bit by rounding up some of the things that have been said so far in terms of potential solutions. The first thing to say is I completely endorse everything that the previous speakers have said lots of great insights from people who are working at the cutting edge of the field so I don't think I need to add too much to that. The one thing I was going to add to start with that I think is another problem and Palava you kind of touched on it there with the comments about India. Is it another of the things that I think we haven't thought about enough as an international community I'm talking about now as anti-corruption activists is the ways in which anti-corruption campaigns can be localized and instrumentalized by political leaders for their own purposes. You can see this a classic example would perhaps be Tanzania under Maga Fuli where on the one hand anti-corruption yes was a great and much needed initiative to sweep a clean broom and remove corruption from Tanzania and Tanzania was a corrupt country where far too much money was being lost but if you look at the work of my brilliant former PhD student Dan Padgett you see that a lot of that process was also manipulated in a way that enabled Maga Fuli to both increase his control over the ruling party CCM and also increase the costs of being outside of his personal networks and therefore to essentially make it easier for him to establish control and of course we know that that control was then used to create a political system which was much more oppressive and controlling. We could look at India elements of the same we could look right now at you know some of the anti-corruption campaigns we see in sub Saharan Africa where there's the same risk and I mentioned this not just because it demonstrates how international efforts to combat corruption can be subverted or just because it reminds us that often we could end up celebrating and applauding something that we don't actually fully understand turns out to have unintended consequences but also for its impact on domestic populations. I think one of the things that I've been saying for example about Zambia where we've just seen a transfer of power is it is great that the new government wants to do a big anti-corruption drive. It's really important to demonstrate to Zambians that they genuinely elected a new government that will not be as corrupt as the government of the past led by Egilungu but it's also important that that does not start to look like a witch hunt that that doesn't start to look like rounding up lots of opposition now opposition, former ruling party individuals and prosecuting them because if you do that in countries like Zambia you start to get the sense that actually one party which might draw support from certain parts of the country is now prosecuting systematically another party which will support from other parts of the country and you can ethnicize and politicize the anti-corruption campaign in ways that actually lose it legitimacy. And so one of the risks right now for President Egilungu and Zambia is that depending on how we prosecute the anti-corruption campaign it could actually lose it legitimacy in certain parts of the country. We need to make sure that the anti-corruption campaigns we see target leaders of different parties and different ethnicities and different regions equally to demonstrate that there's a principle being applied there. All too often over the last 20 years anti-corruption efforts in sub-Saharan Africa in particular have not followed that rule and therefore lost sort of broader public support and legitimacy. And of course if we then add that to everything we've just been saying about the unintended consequences of anti-corruption messaging i.e. that sometimes simply by telling people about corruption you can actually exacerbate the extent to which they think it's a problem and make them defeatist and therefore make them go with the grain and not fight corruption you can see how actually you know big the problem is. And Karen and I are actually working on a second paper now which takes it even further and what we can show in the second paper which she touched a little bit on but she didn't say so much about is that the same anti-corruption messages that in our first paper are shown to make people more willing to actually pay a bribe in the game we play with them also make people less likely to be willing to pay tax. So there's actual consequence there on the social contract there's a consequence there for government revenue potentially of anti-corruption messaging and of course it makes perfect sense if you're telling people constantly about the dangers of corruption they think well the dangers of corruption mean that tax revenues are being misused if tax revenues are being misused why should I pay my taxes. So there could potentially be a relationship here between international anti-corruption efforts and the difficulty of raising tax which again you know brings to mind that the kind of unintended consequences and the inconsistent interrelationship between different international priorities and interventions because of course tackling corruption and raising taxation have been two of the big priorities of the international financial institutions over the last 20 30 years. So that just brings I think you know to a head quite how difficult and challenging this environment is to end then I just wanted to throw out you know some of the things that we know about you know what works and what makes things better and here Lina's done a great job and I'll echo some of the things she said and maybe add a few more a little bit more kind of flesh to the bone the skeleton that she kind of identified of how this works and what we know and I'll draw on some of the work that's coming out of social psychology and other people like Alice Evans who've been who've been looking at this over the last sort of five ten years and as Lina said what we know is that people don't change their minds when you simply give them information they tend to change their minds when they think people like them are already changing their minds in other words the way to get people to shift is to persuade them that people in their peer group have already shifted you can shift attitudes on gender homosexuality corruption if you persuade people that people like them are already opting into the belief systems you want you can't do it simply by telling them that their views are wrong and immoral and they should actually opt into these other views because they're the right ones so we know from that research it's much more effective if you can get that sense that people are already moving into the behaviors that you want to encourage the challenge with anti-corruption is how do we do that in a context where there is not a lot of evidence we could design adverts that tell people people like you you know educated people, business people people you aspire to are not corrupt they're leading the anti-corruption fight and we could design lovely posters with lovely images but unless they ring true for ordinary citizens they're not going to be effective and would those things ring true? Probably not and the reason for that I think is something that none of us have actually perhaps pinpointed as explicitly as we should which is that all of this is underpinned by fairly high levels of public distrust in government and public belief that there are high levels of corruption so for example in the paper that Karen and I look at we look at assumptions about corruption assumptions about the state and as Lena has already said they're often much worse than the reality there's often people who believe corruption and bribery is even more prevalent than it really is and in most of the countries we've been talking about but also if you look at the research in the United States in Denmark in Australia there's a public perception of the public sector which is often even more problematic than the reality so a recent paper on Denmark showed that simply by labeling an institution a public institution rather than a private one people's assumptions about its efficiency went down when all the information provided to them was the same all you changed was private to public so we can see that a lot of what we're talking about here is is so strong the reason this triggering mechanism is so strong is partly because there's a prior set of understandings and assumptions about the nature of the state and about corruption that this is then playing into and in our paper Karen and I actually show that people who have a pre-existing kind of pessimism we call them pessimistic perceivers people who already believe that corruption is really pervasive and endemic are much more likely to be negatively affected by the problem than people who do not have that existing position so one of the key questions that we actually have to ask if we really want to do the revolutionary job of trading a situation where anti-corruption messaging works is how do we actually get to that underlying perception how do we deal with those pessimistic perceivers and turn them into something you know into people who don't have those negative perceptions of government and of corruption and there has to be actions we can't do all of that through messaging and through persuasion there has to be a change in reality that we can do that with and in a sense I'm therefore bringing us back completely full circle to what Karen said at the very beginning which is we need to look at the interaction between messages and actual government policy and changes because it needs to be the connection between the two we do see in some messaging in particular some of the messages that have been tried around the effect of corruption on local communities appear to have less negative impacts and in some cases more positive impacts we also know as I've said from the social norms literature that positive reinforcement of social norms can work but both of those things are only going to be really effective if we also see governments taking action and operating in ways that reinforce what we actually need is a combination of two things happening at the same time we need to be much more careful in our messaging we need to target our messaging and not just aiming at everybody we need to be very careful to try and avoid sending the worst messages to the most vulnerable groups in terms of the unintended consequences but we also need to make sure that when we design those messages they're related to actual government programs that provide people with significant resonance within the populations that we're targeting in that. If we can do those two things we can perhaps start to break through this deep-seated set of perceptions which I think is underpinning a lot of the problems that we're talking about that is a massively difficult challenge it's a difficult challenge because it requires both good messaging and actual government policy change but I think we have to face up to the depth of the problem before we can start to break through the problem is that at times we're also trying anti-corruption messaging in countries where the messages aren't actually fully true because we're encouraging people to adopt forms of behavior that aren't actually rational for them given the governments they have and the context in which they work and so unless we deal with those two things at the same time the actual context on the ground as well as the messaging we won't be able to design anti-corruption messages that have just both added a little bit of nuance on maybe how much the great the challenge is but also a few more concrete points on ways that we might be able to move forward and look forward to the questions and answers coming up. Thanks a lot. Thanks Nick absolutely I mean I don't think you ended up being as pessimistic as you thought you would be or indeed wanted to be and I think we were taking away a lot of very very good points here. I just wanted to make a few points again going back to the sort of framing questions that we started with and the work that we do on is absolutely dark fields of what you said Nick about you start identifying people who are already moving into behaviors that we want them to move into and that's what the evidence base has to unearth and that's where anti-corruption really has to pitch for and that's completely relevant to the work that we do because when we do deep dive we're working in electricity we're working in healthcare sector absenteeism we're working in the skills training sector what we do when we do our deep dive looking for evidence is real instances of behavior that's different from the conventional sectoral corrupt behavior not everybody's out to make huge amounts of rents from investments in the electricity sector not every skills provider you know provides data so that they can make their claims from the government not every doctor in the public healthcare sector actually starts out wanting to be absent we in fact find a large majority and we've done this in the healthcare sector with discrete choice experiments they've been extremely successful thrown up some very fascinating data that in each of these cases we actually find a large body of actors who are behaving in the way that we want them to we've done this in the healthcare sector you know power plants or power companies in Bangladesh which can set up power plants at very very low costs and are actually supplying power at very very inflated costs it's the same company just doing two kinds of behavior we see doctors who come into work at a time when a lot of other doctors are absent we see skills providers actually providing skills training and actually upskilling justifying what they are invoicing the government so when we look at these kinds of behavior we then try and investigate what is it that's making their behavior to say what is the incentive structure that they are facing collect this evidence base and this is really that optimistic evidence base that we look for in SOASIS so I think the way that you are approaching messaging and then the way that we are approaching messaging at SOASIS is in that sense extremely important to all of you you've talked about you know how we need to message we've talked about instrumentalizing it the unintended consequences I think we can start looking somewhat at how we gather this evidence how do we know that this is where behavior has actually moved towards Lena talked about it this is the behavior that we need to capture and Nick you talked about it you know so I think my next question here and please also now that we've laid the foundation what would be really interesting is if you you know started questioning each other I can jump in and ask a few questions if you Nick you just raised a few points about Lena's work feel free to go ahead and do that and we can have a sort of moderated ground but really how do we build up this evidence base how do we do that nudging that you're talking about and what really would be feasible and effective in these context so you know over to you and maybe we can start in reverse order so it can be Nick it can be Heather it can be Lena and it can be Karen just to make sure that we are fair all around Nick would you like to start Yes I mean I sort of said everything I was thinking of saying on that particular issue I think you know one of the things that I think we might want to sort of talk about as a group is also whether or not we can start to think about different kinds of contexts you know Karen has done great work in a number of different areas not just Nigeria as she talked about and we've seen similar effects around the world so one of the things that we're initially thinking is that at least some of the unintended consequences we're seeing might be fairly universal in shall we say what a traditionally called developing countries in urban areas that are you know diverse and have high levels of corruption it seems like what we're finding is fairly common in those places and in a number of different areas like that in different continents people have found roughly the same things that anti-corruption messaging doesn't help or is counterproductive but I think we could perhaps go further than that are there particular contexts in which different types of messages are more likely to be successful and unsuccessful so for example are we also looking to some extent at the hardest cases you know by looking at for example Papua New Guinea, Lagos etc are they some of the toughest cases because of public understandings and expectations of corruption is it possible that the messages that we've been testing would have a slightly different effect if we applied them in a country with lower levels of corruption for example so one sort of first cut might be to try and think about not necessarily kind of another redundant typology but starting to think through whether or not we need to say more about the kind of context in which different campaigns are more likely to be problematic and not problematic one of the things we haven't really seen so much so far is comparative research so Karen has done some great work in a number of different countries but most people who do this because you need quite a lot of resources and you do it as an experimental design by showing people treatments you tend to do it in one country or two countries max we haven't seen big research programs that have tried to do this in 10-15 countries varying the levels of development varying levels of education varying levels of corruption so going back to Heather's point about needing to test things I think one constructive way might be to start by doing that we as Heather was saying as part of the extension of this research an extension of ACE we'll be looking to do that in Albania and hopefully trying to find out some interesting things there obviously significant differences as well as similarities to the cases we've looked at in the past and I think that's one aspect of this that might be really interesting trying to do more comparative research as we've heard you know most of the people on the panel most of the others in this area are focused on case studies so leveraging those comparisons and trying to get some insights into what might work in different contexts I think would be something that we might take forwards but I'll look forward to the answers of the rest of the panels to these quite difficult and intractive points of curiosity here I'll throw in as moderator and you know all of you feel free to jump in here but you know your research identifies a non-pessimist based on based on you know very robust survey data what I'm curious about is how do we know what types of messages non-pessimist would respond to positively that's what I meant by how do we generate the evidence based I'm looking at it in terms of you know really methods because that's the method that you use you and Karen use in the paper is extremely interesting and could be a guide for us in doing this kind of anti-corruption research and designing the messages so in this case it happened to be that tax was what the non-pessimist responded to positively how do we design it that's what I meant by generating that evidence based to me it would seem that what a particular you know sort of section of non-pessimist would respond to is based on let's say the material interest and their perceived feasibility of their actions having an effect on their you know own welfare how do we sort of build this these are difficult questions but it's something that you make and Karen have set yourself up for because these are the kinds of things that you know one one is curious about so I was wondering if the two of you had any response to that that's really it well I'll let Karen answer some of that I think one thing that we do need to do in the future is move beyond the kind of lab so you know one thing Karen and I have talked about is would it be good to shift from doing this under a more controlled setting or we did the last project in Lagos on household interviews but would it be interesting you know to move from that into actually looking at you know real world posters in a real world setting so not showing someone something on a tablet in a formal interview but having posters you know every one part of a town having not posters in another part of town and working yet how that operates I also think well I agree with your points about you know the sorts of things you might think about that might drive and trigger more positive responses I think the other thing to add to that is of course you know local context and local resonance and one of the things that you know most of us have tended to do at least in our say Lagos papers to use English predominantly and you know we often the sorts of messages we see pushed by the international community tend to be more generic and there's perhaps more to be done in terms of tackling into local norms and cultures that can then leverage some of this so if you look for example at John Lonsdale's you know work on moral ethnicity in Kenya he talks about a very rich set of moral debates within the Kikuyu community of Kenya in particular in which there are certain ideas about how wealth and leadership should operate now those don't necessarily contain kind of democratic ideals of accountability nor do they include socialist ideals of redistribution but they do contain norms and values that can be used to hold leaders to account when they breach those legitimate distribution and wealth and so you could then use that as a way of trying to develop a set of messaging that would capture to much greater extent how the community is historically thought about and talked about the distribution of resources to try and make sure that things are seen to be as it were more illegitimate when you talk about them and also then to perhaps remind people that there was a point in time when things weren't as they were today where these ideas evolved out of so that might be one way of trying to tap into that collective memory on a more collective basis than the individual basis the other option is unity of course which you know your comment about individual incentives kind of draws us to is to use social media messaging and targeting to get messages to people so we kind of reverse the evils of Cambridge Analytica to use that kind of big data analysis to target people with individual messages that would not be general would not be blanket and would be specifically targeted at the right community in order to achieve the maximum positive effect without doing the harm that's I think something we do need to really look at for the future the challenge at the minute of course is that big data is most available right now on developed economies where people are on Amazon and by their shopping on Tescos and therefore you can build that complex data profile of every individual it's less easy to do that in a lot of the countries that we've been talking about where that data doesn't exist in that kind of way and so that kind of targeted messaging can't be quite as specific just yet but I think that is the next frontier and that's something that you know ace keeps going and people hopefully give us some more funding maybe that can be the project in four or five years time but over to Karen. Yeah just to add to what Nick was saying I think I mean Nick and I have sort of lived in this world of how do we make what would the what would the dream study be with all the resources in the world and how could we address the limitations and I think what's important to acknowledge at this point in the conversation is that the evidence space we have is they're based off of limited you know studies that have serious limitations and some of those are as Nick said already that they've only been tested in certain contexts and really only if I I'm thinking right really only in urban areas certainly my studies have been but aside from that there's several and I think what that does is it lets us think about what could the next the next sort of studies or the dream study do so I think what Nick was saying about the comparative study that would be ideal to really look at is this universal a universal impact or is it context specific I think also maybe what's missing so far from the discussion is is to look at who is the messenger and I think that this relates to SOAS ACEs sort of perspective so you know if the government's lost its popular credibility on really effectively fighting corruption so is there another group that has more credibility or legitimacy around this and so it does speak to that SOAS ACE approach about recognizing and identifying who are the potential influential people and groups that can shift things and this would be really intimately dependent on context I would think you know who has legitimacy whose messages might work how can you test that in the field I think it would really depend on context and then legitimacy for who so what we know about the research on legitimacy of course is that this there isn't you know one group that holds legitimacy for everybody and so then it gets us into all sorts of potentially really like dividing the population and kind of hairy ground about who do we put our resources into thinking about who can carry legitimate messages something that Heather was talking about earlier that I think also relates to this point is so Heather and I have done research on a range of other things and one of the messages that we've had in a different in a corruption functionality framework is to look first at what is the aim of anti corruption and I think it's important to recognize that messaging usually has more than one aim or if it has one aim it's not explicit so is it is the messaging supposed to get people to report corruption is it to get people to think differently about their votes about how they can mobilize our votes to create more accountability is it to get people to avoid engaging in corruption is it you know what is what is the it might be the case that some messages from certain messengers might be better achieving certain aims than others so that's another thing to think about and then I guess before I hand it over the bigger kind of I don't want to cause too much more pessimism but the bigger question I think is we've only talked about messaging but what about anti corruption in general so the policy energy around anti corruption to the extent that the public is made aware of it may very well have the same impacts on what the public thinks about corruption and anti-corruption and so while it's nice and neat for us to talk about awareness raising specifically and its efficacy what I wonder is whether or not what we find and what we think about the impacts of awareness raising if there's a more generalized impact that we might be finding with other anti-corruption policy energy that's being heavily invested in all over the place impact yeah yeah we always come back to impact and maybe Heather you might want to come in and respond or speak to both what Nick and Karen have said actually I wanted to go to one of the questions slash comments in the chat because it was fascinating and I started thinking about something with SOASACE in general and actually it's connected to something coming out soon in SOCACE which I can't believe I've never thought of and so I'm really grateful for that and we'll see Karen and Nick might go actually that's rubbish but so Benam Zoki and I apologize if I've not pronounced that right but you know it's not about messaging strategy in Iran and how in Iran the public have very high expectations from anti-corruption efforts so you know the opposite of what Nick and Karen described in that kind of sense and how it's really challenging because that goes with a lack of tolerance for cooperation and coalition making because of a polarized politic what I will reply to that is to say that actually one of the challenges I think that SOASACE's research overall really points to is the importance of feasibility so high expectations for anti-corruption efforts that aren't realistic may actually be extremely harmful and so whether there's an anti-corruption messaging element in there if people have expectations that can't possibly be met whether it's because the political environment is too polarized whether it's because the resource levels are too low or because forms of corruption fill functions like Karen and I have written about people don't have they've not been paid they can't get by without it whatever it might be then that's going to end up feeding that sense of despair and so on and I think this is such an important contribution from SOASACE about the need to really think about feasible reforms which almost inevitably as Lena and others those of us working on collective action and corruption have talked about it's almost always going to involve some kind of cooperation and formal or informal coalitions and so on but I was thinking about two of the first outputs that will come out hopefully mid-October from SOASACE there's going to be an evidence note on what the evidence tells us about serious organized crime and political will so like SOASACE we share a similar approach looking at politically feasible interventions and there's also going to be a short conceptual piece that goes with it and one of the things that we're writing about is a kind of older definition of political will as the sum of political want and political must so we often focus on the political want but we don't think enough about the can and the must and can thinking about Derek Brinkerhoff's groundbreaking paper on political will and anti-corruption about 10 plus years ago can is about capacity so if there isn't capacity whatever the messaging is telling you it's just not going to succeed so there has to be a degree of feasibility that goes into messaging and I don't have never seen anything like that I don't know if Nick or Karen have seen that the must sort of Derek talks about how that really links to public pressure and citizen engagement as well and so what anti-corruption messaging could do is to create that momentum that sense of must to go with it and then the want is really about things like personal sense of civic duty of personal sense of integrity and so on but the point being made in that is that you don't get political will whatever that may be without all three of those so you can want it and there could be public pressure but if you don't have the resources if it's not feasible and there isn't the capacity there then it's not going to happen and then that could have blowback so I'm not sure if that answers your question or not but it's definitely made me think that we all need to have a conversation afterwards what to do with that because I'm now really excited about that so thanks Thank you Heather maybe Lina did you want to come in and respond to that a lot of issues have been thrown up how to gather evidence there's big data there's the can must want how do we use the social norms you know what the social norms research is telling us to actually generate evidence you know my question to you could well be how do we we know there are descriptive norms and there are social norms how do we you know sort of sift the two apart and then design a messaging or policy accordingly take your pick you really have I have a lot and while everyone it was talking and carrying and Heather's like just my brain was going off with every tangent because there's lots of really really interesting points and we could be talking about a whole bunch of things for a very very long time but I think we've kind of covered a lot of ground on the on the feasibility side of things and I think while and everyone was talking about this feasibility point I wanted to actually I didn't say this in the beginning but I should have also mentioned that we should also bear in mind the sociability aspect of corruption and this is something that we looked at in our work on the role of religion in in the social in the social acceptance of corruption this is work that we put out I think in March of this year and I'm going to talk about evidence gathering after this but in this particular part of our survey when we're looking at the role of religion in how sociable or people's judgments of a particular corrupt act we actually found that when you turned the purpose of say someone embezzling government funds or taking government funds for their personal use when you turned that question on its head that it wasn't just it wasn't about personal use it was about a person taking government money for supporting the religious community or building a mosque or a church or something like that or giving to charity in through the religious community people's judgments of that behavior changed even though on the whole most people thought this particular practice was unacceptable when we did the randomization it was really really cool to be able to finally actually get something statistically significant when you randomize this type of behavior when we randomized and changed the purpose or the reason why somebody would engage in you know what would be a typical kind of public sector corruption regardless of the amounts or whatever if it was for your community people generally had a lower threshold negative judgment of that kind of behavior of course for the most part most people thought it was wrong but when they thought okay if it was actually for your religious community we can actually justify that so one of the reasons why we did that kind of investigation was that at this to the point of funding and evidence gathering that we're all talking about and all the anti-corruption programming that's going on everywhere is that a lot of funding is going through faith-based organizations or at least faith-based anti-corruption messaging and a lot of the and I think Heather's done this work for years and years and years and we actually even reference her work in our paper in our project this aspect of our project is that a lot of assumptions contribute to moralizing the corruption problem and we already know that that is hugely a problem particularly in these contexts that we speak about when people actually have specific negative personal beliefs about corruption you don't need to moralize a problem that is functionally okay for most people so we found it really really important to understand the institutional context and I think Nick mentioned that institutional context matters and there was a study I wish I don't know if there's an update of it but there was a study in 2012 I think it's summer now I can't remember his first name there was summer and a whole bunch of other people who did this study when they were looking at the relationship between religion and corruption to the point about institutional context mattering and they found that the only context where religion had a positive role or reducing corruption or a role in anti-corruption was where political institutions were democratic and societies had strong democratic values so I think this also ties it to the point and the concerns that everybody has raised about political will and political action and that political actions can be targeted against political opponents and the like so what we recommend in that context because we also notice a lot of work being done around behavioral change supporting the work of a lot of faith-based organizations and this particularly in Nigeria what we recommended in that context was that you can have faith-based organizations because of their social influence their collective action convening power faith-based organizations working in the space of amplifying citizens voices for community monitoring and this also ties into Kevin's point about localized messaging so you can have faith-based actors working in that dimension of anti-corruption not directly on anti-corruption in the sense of moralizing the problem because it seems to be the angle that as a faith-based organization it makes sense that's kind of the message and their messaging to take the approach of moralizing the problem and appealing to people's individual ethical standards and their belief systems and so on but again what is the case in a lot of these adverse context is that when you load these individual costs people are unable to meet these individual standards because the individual cost is too great but faith-based organizations can intervene in this context where they distribute the cost of anti-corruption of resisting either a shakedown etc etc faith-based leaders faith-based institutions, faith-based actors can play a powerful role in that and another aspect that I wanted to speak about is around and I think this is really if we were all fantasizing if we had a whole bunch of money to do more research my kind of fantasy around our social norms work is if we had money to integrate some social norms and social network analysis because we were talking about who the messenger is and the legitimacy of the messenger what I really would love for a lot of work on behavioral interventions to move into is this kind of is identifying norm violators identifying the kinds of characteristics of a first mover as somebody who would abandon a behavior and take on a new behavior there of course in social psychology there is their metrics and the different measures for measurement tools for identifying these kinds of specific people within a network identifying their norm sensitivity, their risk perception their risk sensitivity and when we did our work on law enforcement that was kind of one of the things that I hoped that we could move forward on doing identifying individuals within a context where social norms are actually present and drive a behavior identifying individuals who are norm violators and you need another layer to be able to do that people that you can target messaging to in context where you don't have these kinds of individuals there's a lot of work around entertainment but you can create fictional characters but this is mostly being tested around risky sex behavior health and the like but we haven't had a lot of that in anti-corruption that would be great to see if you can model and project images of norm violators who go against the grain who behave differently and if you can have kind of collective behavior change off the back of that but in terms of evidence gathering what would really be great for us to see a bit more around behavior change is identifying norm violators who the messengers are the kinds of characteristics they have and targeting messaging or different types of messaging for all of these different components of a reference network just end there I think there are a couple of questions I think a ton of them Lina thanks I was going to invite the panelists to answer Claudia's question but I can see Heather typing away so that's absolutely great Heather for picking it up because I was going sequentially Claudia was the first one to get a question and Heather do you want to keep typing or do you want to say it up you're on mute Heather line of the century actually if I say it because that way it actually what I was going to say links to Kathy's really interesting question as well and so hi to both of you by the way but I was going to say to Claudia if research is definitely out there on this I think it's worth looking into but I wonder if there's research around service delivery that's not linked to to corruption specifically that could be worth looking into particularly around any interventions that are aimed at improving citizens understanding about what they can expect from service delivery and provision so kind of rights based approaches to civic education or something like that that are focused very much on what you're entitled to and what the quality should be and so on because it would be interesting to look to see if there's any evidence already or if maybe there could be future research to see does that indirectly affect their attitudes towards corruption as a positive unintended consequence of those sorts of activities and just like we don't pay enough attention unintended consequences on anti-corruption but that could be positive as well as negative and Karen and I have papers out on that as well having said that I'm not convinced then that even if the evidence is there to say that happens I'm not convinced that we either can or should try to design for those kinds of positive unintended indirect consequences I don't think we're that clever and I think the sort of more we we kind of add layers and layers of complexity onto what we do that we end up not doing anything well but also if we kind of go to the point about you know the research showing that problem-driven approaches rather than solution-led approaches are important then trying to combine two different problems in one thing could actually it's not very problem-driven and it could lead to problems but I think that we don't do well enough at identifying those indirect effects which is I think a whole new area of potential research as well so I'm not sure if that answers and I'd love to know if there was research out there on this because I think it would be fascinating and important but thanks Karen can I yes yes yes so Claudia's question in case anyone didn't see it is basically is there evidence on whether or not messaging that doesn't use the term corruption like what impact has that had on attitudes about corruption I think it's fantastic it could be a great focus for one of a treatment in the future but so all of the messages I've personally tested and all that I've seen do explicitly mention corruption but one thing Nick and I found is that for the non-pessimistic procedure so the people who aren't already convinced that corruption is extremely widespread that this message about taxes linking corruption to taxes so it did explicitly mention taxes but this message about taxes didn't have such a negative impact instead it actually discouraged those people who weren't pessimistic perceivers to not bribe in our game so it was like our one positive or optimistic finding from the study and I think part of the way that Nick remind me if I'm getting it wrong but part of the way we interpret that is that that message aside from the rest actually don't implicate people it's not mentioning the specific government but instead it's mentioning taxes and corruption and so we thought maybe what might be the case is you can talk about corruption but if you're not pointing the finger directly perhaps that's when you aren't necessarily priming this issue of the government being involved so I think there might be something to that is the moral of the story Right I'll just quickly go to the other question that Rosemary Ventura had I'm just going by time so this is not favoritism she says with his excellent example from Zambia Nick touched on concerns that corruption efforts may be perceived as partisan if prosecution seem unfairly targeted towards one party or group can this exacerbate conflict dividers existing tensions grievances in fragile context examples thanks I think Nick that one's for you Yes absolutely I think that it can I think we've seen that in countries like Kenya for example where within the power sharing government and president Kenyatta and opposition leader opposition in the kind of use of corruption allegations was kind of a form of politics and played into ethnic tensions that were left over from the political violence after the 2007 elections so yes I think that kind of use of corruption as a kind of divisive issue as a way of persecuting people or removing people of particular communities particular parties can then exacerbate and play into all of the kinds of divisive politics that we think of sometimes in those highly ethnicized or highly salient community identity countries in which politics may be conflictual and of course the tricky thing is that it's also in some sense is just I mean it may just be that these people committed terrible crimes it may be that they were corrupt and so yes they should be prosecuted in that sense you know having these people jailed and recouping the money is you know just the problem is you know that as we've been talking about public perceptions are critical if your aim is to shift public attitudes and public behavior and so even if something is you know formally just you then have to consider the appearance of it so yes I think that's a very real concern. Great thanks Nick I think the next question both Karen and Heather have answered your question Florencia you had a number of very interesting points and I'm just going to read that through but I think there was one question in chat which just got picked in so here we are there is Brunilda and she talks about how do we identify the right champions that could serve as examples for other peers so to start this horizontal monitoring and collective action how to do that in countries such as Albania where finding such people are difficult and if I may since Brunilda you've talked about horizontal monitoring we are not really talking about picking specific champions so to say for messaging since this is a much more sectoral approach electricity or fertilizer or skills training we're actually looking for sets of actors who behave differently so they're not champions in that conventional sense of the word that you pick up somebody who becomes an anti-corruption champion but these are people who signal different behavior they are in the same sort of adverse political economy if you want but they are actually their behavior is much more rule following and therefore more efficient but the sectoral outcome is more efficient they're producing maybe power at a cheaper rate or training people without capturing the sort of training fee and then we work backwards from that collect that kind of data sectoral and then see how we can devise policy in the sector broaden it out so that other players in the sector can also follow that policy but if others wanted to take Brunilda's point about identifying the right champions it's an important question anyone or okay I guess not silent means they're done with this question Kathy your points I've seen them all as answered Lena I don't think Kathy's question has been answered should I read it out to you fascinating Lena in terms of messaging to combat okay here again fascinating Lena in terms of messaging to combat descriptive norms do we have examples of how this has inspired collective action I'm finding it hard to think about messaging that might do this in a way that would be positive not make individuals feel guilty or upset and would spur collective action would you have any examples they would be helpful well yeah I've been I was thinking about that and I think we were thinking about that when we were doing our work and when we continually just uncovering that a lot of the corrupt behavior we're looking at was being driven by descriptive norms we're thinking okay so what do you do with this kind of evidence what do you do how do you design messages or just generally interventions that demonstrate that a lot of the corruption that people are engaging in they disapprove of it they would rather a different kind of equilibrium so the only example I don't short answer to that don't have a specific kind of tidy example of how this has inspired collective action but internally and just in my mind I think the collective action movement and I think I wrote that in my response to a cafe was the collective action movement or against police brutality and police corruption in Nigeria last at the end of summer into October as an example of people realizing that a a particular behavior or practice was unacceptable to most people and actually acting upon it of course we know how tragically the ensars the movement to end police brutality that was a hashtag ensars what happened with the killings in Leci in Lagos but I find it really a very inspiring example of increasing numbers of people realizing that a behavior is unacceptable to most of us and actually taking civic action against it or at least to demonstrate or to demonstrate protest resist highlight the fact that this behavior is unacceptable of course I believe we haven't heard the last of the movement against police brutality and police corruption in Nigeria I think there will be political ramifications for that moving forward but I think in terms of just the example an example of people acting on their true beliefs and shifting away from or realizing almost at the same time of course not coordinated by an NGO and that's another point necessarily just neatly coordinated by a group of NGOs who thought okay today we're going to start off this movement and XY numbers of people are going to get on board etc etc I think the fact that it was so organic the fact that it was almost as though there wasn't any hierarchy of leadership added legitimacy and authenticity to the movement and you had more ordinary people being able to feel confident feeling that they weren't taking on the individual cost of resisting this or at least indicating that this behavior was unacceptable knowing that there were more people within their community that found these types of practices or the behavior of the police unacceptable and feeling that they would have allyship with others in their community so for me that's kind of like the biggest really dynamic example there's a lot of lessons to be I hope I able to examine and really write about that particular example and as an example of people acting on their true beliefs in the context of descriptive norms versus injunctive norms so that is definitely an example I think we should look at what worked what made it happen what were the triggers, how was it sustained and how can you have that kind of collective action happen in a way that there were fewer risks and you don't have these kinds of unfortunate unintended consequences of people dying in civic for demonstrating their political position against police brutality I think Kathy has also brought up the point about the N-SARS movement and how norms shifted so that's interesting but again we were all as Nigeria observers watching the N-SARS movement some of it unfolded on twitter back in the day when twitter was very much allowed but it was interesting that there was a different take coming from many youth movements in northern Nigeria where the whole idea was you can't do away with the police because we need the police look at the kind of insecurity we are facing in this matter so again where and how messages are devised and again where it's landing in this case becomes very interesting there is no one size fits all even though N-SARS was extremely motive and did take a lot of inspiration from the Black Lives Matter movement but to move away I don't know Lina should I move on because I'm mindful of that I just want to say that different circumstances the protests play out differently in northern Nigeria so just social media platforms there are definitely contextual factors as the kinds of norms and just power distance and all of those types of you know how power is distributed in northern Nigeria and just how these conversations happen quite differently from the south country and urban centres so all of those should be taken into consideration absolutely sorry Florence are really great points there but I think the point that you make about disinformation you've provided some very helpful links to everyone who can't access it please look at your Q&A window she's provided some very helpful links that add to the conversation and she's picked up the point about disinformation unfortunately Dayat Hassan is not part of the panel today but they've been doing some great work in CDD on disinformation in Nigeria and I think we've managed to answer most of the questions here I'll just see if there is anything open left here in the Q&A I don't think we can see Claudia's links and questions once again that's from Kathy, Kathy says she can't see Claudia's links and questions I'm not sure Claudia has provided any links it's Florencia who's provided us the links Claudia's questions are all now in the tab called answer in your Q&A panel we've got open, answered and dismissed there are seven questions answered and four open if I've missed anything just drop something into the Q&A can we see her links are all there in the Q&A tab if you just look for the open tab that's I think exactly where it is if you can't see it we can paste it and put it on the Q&A for you but they're all there in the Q&A tab under open so she's got there you go Duncan's done it thank you so much multitasking is never a good thing when you're moderating I'm mindful of the fact that we are close to closing time and whether the panelists want to make very, very quick closing interventions and then I will wrap up at exactly some time left for 2pm panelists any closing interventions lots of good support none for me like you say lots and lots of things to think about lots of notes so thanks to everyone I'm going down the list Nick Karen Lina nothing from me just thank you so much for this conversation there are lots of things to take into consideration when we're messaging but I think just the top of the line point that Karen made about localised messages of corruption apply across different contexts and they apply when we take into consideration factors such as religion we didn't talk about gender which is something I should have perhaps mentioned that the types of messages you should deploy in a context where you have really, really complex gender norms and gender and quite high levels of gender inequality would be really different because of the kinds of social costs that women take on so just this localised and just context specific messaging approach is one that I think we are quite unanimous so thank you so much for all of your work thank you Lina Karen any closing thoughts just thank you for organising this it's clear and it's great to walk away from it with lots of sort of new energy around potential new research avenues so thanks Nick just to say on that point that we've all been making about comparability and increasing the cases I think collaboration is great and that can be formal collaboration and coalition building as some people have been pointing out in the Q&A thanks for those excellent comments and questions it can also be though making sure that we read each other's work and we try and replicate and build on each other's work so Karen and I have tried to use questions that are not too far off the question she's used in the past if other people come along and do similar studies and draw on some of those questions we can start to build up essentially comparative research without actually doing a whole comparative project by making sure that we're kind of building on and testing the phrases and the ideas and the messages that have worked in the past or haven't worked in the past and if we keep doing that intuitively we will hopefully hit upon things that work better but if we do it sort of collaboratively taking care to test each other's questions in different contexts and to build on what's been done before that's probably the most effective way to do that so in that regard thank you very much Balavi and Duncan and everyone at ACE for organizing this excellent search seminar because I think it will help us to do exactly that and to coordinate and just to encourage people who are listening and interested in collaborating with the speakers and to you to keep the conversation going thanks very much Thank you, I missed a question from Ambeka who says do you have do you have examples of who can be brought together maybe somebody can type that in as to how influencers were brought internationally into a campaign without feeling forced and manipulative maybe this was for Kathy but I just taking on some extra leverage would sort of like to round off today's discussion I have to say that Nick and Karen's research was one of the earliest bits of research that we began with in ACE, I think Nick and Karen this research was part of the absolutely initial sort of bit that we had made when ACE was just starting out and it is not the kind of research that SOAS does traditionally say is sectoral it's about what we would call the productive sectors it's much more to do with how does the electricity sector operate how do we design fertiliser subsidies so in a sense if I might lapse into technical talk for a second this is much more supply side which is how do we increase effectiveness it's almost developmental anti-corruption because if you get electricity producers to produce electricity better quality electricity and provide electricity to everyone of a standard and a cheaper tariff we've not only reduced corruption we've actually produced a better growth outcome so that if you have in a sense is sort of the business case for SOAS so what we are really looking for and maybe it kind of got mixed when I was beginning but we are really looking for actors within sectors and this is why we work with sectoral data this speaks about the fact that we don't have big data yet in developing countries so we actually find it easier to work with sectoral data and you can take care of the confounding variables that you don't find at national level data and we can experiment so some of our approaches have been very experimental whether it's been in the electricity sector in Bangladesh and Nigeria or in the skills training sector in Bangladesh so we also have these very experimental approaches to generate evidence and how do we do that and how do we go deep into what the sector looks like what the behaviour of various actors in the sector basically looks like and we try and find people who are corrupt who are not corrupt but in their self interest so these are people who are in the sector where you have corrupt people and these are people who are not corrupt because for them it doesn't pay to be corrupt they are profitable, they are delivering a service they are producing a good without being corrupt and for us well how does that happen how is it that they are behaving differently when everybody else is behaving differently and then we pose the feasibility question well they are doing something which makes it feasible to be not corrupt and then the experiment is in designing policy that way using that as a sort of template that can be replicated for others in the sector and then this becomes our suggestion for feasibility and I think where the complementarity is with the conversation that we were having with how to message effectively is to then target messaging and devise messaging which actually targets these policy measures to reinforce these policy measures and I think that comes across as a very unique anti-corruption developmental package that there are these sort of supply side issues where you are increasing productivity, you are increasing service delivery and you are reinforcing them with very targeted specific delivery of messages because these are your non-pessimists in a way, people who are behaving in a rule following manner in a sector which is actually corrupt well these are the non-pessimists that we have already identified so I think there are some very unique complementarities here so it would be absolutely great for the two bodies of research to talk together and I am really glad that we got this project in this really great work that Nick and Karin have done to talk to the other kind of work that SOS is doing and of course it also as Heather says adds to what SOC SOC is now doing so something to end here just to give you an idea of how the complementarities can actually end up in making research anti-corruption research even more effective this is one way of reducing that corruption fatigue you are actually designing research built on something that is already working you are replicating it and then you are reinforcing it with targeted messages and I think this is really one good way of reducing that sort of despair and fatigue and maybe we can have much more complementary work going forward thank you to all the panelists thank you to everyone for all the questions and the interaction that happened it's been a long session but it's been extremely rewarding thank you once again and have a very good evening morning afternoon wherever you are and do keep in touch Duncan has been tweeting out all the links all the Twitter handles all the websites are there Heather's website is going to be launched soon absolutely do access it and keep this conversation going thank you