 So, welcome everybody to our fire site chat today with Holly, Holly Sutherland, who for many years was director of the Euroma team at ISA and is also part of the board of the DRM program. So, we are very glad to have you here, gathering with us around our virtual fireplace, so you can hear the fire crackling in the back. And we have recorded the first part of this conversation, but there will be a part where you can ask questions interactively later on, so that much up front for the audience. So, Holly, great to have you. I will start in the very early days, and with a very simple question, how did you first come in touch with micro simulation? Oh, well, first of all, thank you very much for inviting me to join you. It's a lovely crackly fire. So, I think I first came into micro simulation through doing work on the UK back in the 80s, 1980s. And from that, the idea of using micro simulation to do comparisons across countries, comparisons of the effects of policies in different economic and social contexts to try and understand more about policy effectiveness, I suppose. I mean, that would be back in the mid 90s that we were doing that. And just for a selection of countries, based on the idea that we were a group of people working together anyway, and we could see that micro simulation would be a way of answering the questions that we had about different systems. So, the countries were a few EU countries, France, Ireland, UK, Italy, I think, were the first ones. And the idea, we all had our own national models, but the idea was that because they used different assumptions that were hardwired into these models and they all been built in different ways, we needed to build something designed for comparisons. So, that's what we did. So, the idea was not so much, it was really just to inform that research on inequality and poverty and the different effectiveness of policies in these different national contexts. And then, of course, that led almost naturally to the idea of having a European model, so being able to have a model that would not only compare the effects in different EU countries, but to look across the EU, to look at the EU as a whole. And in doing that, we were sort of way ahead of what the policy demands were at the time. This was very much an academic project. And so, we did that for the EU 15. And what that meant was that we had to build something that was very flexible and that could cope with different data, because it was before the EU Silk, for the EU, we were using different data for each country. And obviously, different policy systems. And it was almost as a byproduct of that, that we realized that we could make use of the framework that we built to easily, quickly build models for countries that didn't have models already. So, as the EU expanded, we did that for the new member states. And then, it became obvious that we could also use the same framework to build models for countries that didn't have them. And in fact, that were in quite different economic contexts. And the very first experiment was with South Africa. And where Michael Noble and Gemma Wright were, I think, they approached me. I can't remember. It was one of those real fireside chats, I think, where we happened to meet up and talk about it. And the idea of trying to build a model for South Africa came up. So, the idea of having a common framework for building models for different countries, which I suppose is what I ended up following my career, was that's what it was, was a sort of byproduct of the fact that we'd already had to develop something that was flexible in a way that a single national model for single set of purposes didn't have to be. Okay, so I have now, just out of curiosity, a question. What software did you use initially? Well, for the first UK models, we were using the very first model I wrote was for a mainframe, because we didn't have PCs then. And I used a combination of SPSS and Fortran. So those are the days. Then we rewrote the UK model using C++. But when we first wrote the first comparative multi-country model, we used Excel. So we didn't use the, I think the very first one, we actually used Excel calculation engine. Because that was a common, that was a software that was common and easily available and that everybody could use or easily learn to use. I mean, it was not ideal. So we moved on from that quite quickly. And so the software that we used was again C++ for the engine, the bit that nobody ever has to look at that's written by programmers. But the way in which Yoramard looks even now with those columns was based on that original idea from Excel. We could see how that sort of column structure, having a column for a policy system, that that could be made to work. And so, although it's a long way from Excel now, there are still some things that we sort of discovered along the way that turned out to be useful. Because in the beginning, we had no programmers as such working on the project. It was all, we were all economists at the time and we all picked up what we needed to pick up and went and tried to learn things. And it was very much not an expert programmer's program. Probably some programmers looking at it now would say the same thing. But it was fit for purpose because we could make it do what we wanted it to do. It has worked and still working. So when did you start to call it Yoramard? I think that was with the first, the EU 15. So when there was only 15 countries in the EU, it was when we first had, we built that model. We built a sort of prototype model, first of all, because that was the way that the funding went. We had funding from the EU scientific programs and they didn't really believe that what we were promising, which was an EU model, they didn't believe either that it was possible or that it was useful, which was a sign of the times, I think. And I think it's now very much recognized as a useful thing by the commission anyway. But at the time, they didn't see it. So instead of giving us a big grant to build a model, they gave us a small grant to prove that it was possible to build a prototype and to, if you'd like, try and sell it to explain why this was something that they were going to, or policymakers and academics working on policy issues, we're going to find useful. So we did those things and then we had funding to build the whole thing. But I think the term Yoramard was first invented when we first put in the big proposal. We've always had a bit of trouble with it because it's a kind of obvious name to use, no, for a European model. But it's actually copyrighted by the fashion industry. So there are various Yoramard and Yoramard sort of trademarks that are registered. So we've never been able to, we never were able to sort of own the word, but we carried on regardless since we felt that we weren't really in competition with the fashion industry. That's interesting. Yeah, that you consider yourself not to have been in competition with the fashion industry. You mentioned the South African model and actually Michael Noble will be another guest in this series in the upcoming, I think for the next day that he will be speaking with my colleague Rodrigo Oliveira. So we can actually check whether that idea was or happening at a real fireplace with Michael Noble. And it brings me also back to the next question, basically. So obviously the South African model was in a way the front runner for a model that's not from a developed or European country specifically. But like, when you first heard that at Union United is the idea of building such models, like, what did you first think? Like, oh, these people are crazy. Oh, great that finally someone comes around with it or like your initial reaction. Well, there were sort of two initial reactions, because in fact, that there had been talk of doing some micro simulation for Africa, for Africa, before sometime before South Mott. So I think the idea had been around for a while, but didn't actually gel in a way that South Mott has gelled. So it was not a single moment. I have always been keen for the work that we've done and the ideas that we've had and the approach that we took for EU countries to be picked up and adapted, if necessary, and carried out anywhere in the world. And in fact, one of the things I'm most proud of that that happened and South Mott, first of all, the SA model, the South Africa experience was such a great sort of trail blazer. But then the idea of doing something for several countries and really with a rather open agenda about where it might lead was, of course, really exciting. I was, it was exciting, but I was also concerned that people wouldn't think it was easier than it was. Or, because again, this is my experience going back with the EU and with your mod and even with the UK, is that when people hear the word model, they think that this thing is going to be able to do anything. I had the experience a long time ago with a very clever person working for the European Commission. She wasn't, she was by no means stupid or ill-prepared, but she asked me if Euromod could predict the effect, the gendered effects of increased R&D funding to the aerospace industry. Well, I mean, the quick answer is no, not really. I mean, the long answer could be, well, it depends what else you have there. Possibly it could be part of some kind of analysis, but it wouldn't be the first place you'd look. And so I do have, I mean, I'm very aware that people often think that this model will answer all the questions that they might have. And so one always had to be careful not to raise expectations too high and to be quite practical. And I'd always been keen on the idea of having quite simple and quite appropriate for the model questions to look at first as a way of demonstrating the kind of core things that the model could do. I don't know, you know, the distribution effects of a new simple child benefit, that kind of thing, as a way of trying to help people understand what it can do. And then of course, there are all the steps after that about explaining how that's not the end of it, that one can be, you know, there are a lot, it's basically, we can't imagine all the questions that we can answer. They will come along and people will have new ideas based on work that's been done already and so on. It's a kind of dynamic organic sort of process. I think I can say from inside that both your initial reactions to certain extent have helped through that. It has, surely sometimes a challenge and kind of navigating the expectations towards the models is a constant challenge, obviously. On the other hand, I think we can say that they are there and they are coming out of the very infant stage. We're going into the toddler stage, I think at this point. And actually, then let me tie in my third question in that way as well. So now if you think forward, what would you, where do you see Southmond going? Or where would you want to see it going? Oh, it's not up to me to want. All I would want is that it's used and that it makes a difference and that the people who work with it love working with it. I mean, I think there are, this is a bit of an aside, but there are people who really love doing this kind of work. And if they do, they should be grabbed and given the opportunity. It's always been such a pleasure for me to work with teams of people who have enthusiasm for this work. But I think there's another thing to bring in here, which is the question of data. So the models are not anything much unless they have good data. You can make up for data problems to some extent, but in the end, if you have no data, you have no model. And I think in a way that's the future direction for those sort of technical sides of the model is for the models is to improve and in a sense diversify the data that they run off. So thinking about administrative data of various sorts, but also forging strong links with data providers so that the data that is available is improved and that the data providers understand the needs of the model. So I think that's the future. It's also the past. I mean, it's always been necessary to try and build good relationships with data providers, not just about access and availability, but also about improving what's there. Obviously, that's difficult because there are always many demands on particular data. But I think it's an essential part of carrying the things forward. The other thing, of course, is making sure that the models are used and that everyone who has a use for them has an opportunity to use them, that they're not just hidden away in a single person's office, that they're disseminated well. And of course, I know you've been running training courses, but thinking about how to do that in the most effective way is always there, but it's something that depends on the context. It depends on who's being trained, why they want to, or I've been asked to learn how to use the model and what they will do with it. I mean, I see every use of the modeling, even any training course, is an investment. So it has its immediate effect, but it also has potentially a kind of longer term dynamic effect. Thank you for this. Now we'll close the recording here and go over to taking questions to follow up what we just started to discuss about. So here we are. And because I already had the opportunity to ask Holly so many questions, we can have anyone else ask follow up questions of what we just started discussing. You can either put them in the chat, but I'm also happy if you just speak up if that's technically possible, Anna. Yeah, I think it's possible so that people can ask to share video and audio and I will admit them in and then they can speak. So we can do that very well. Okay, so we have a few more minutes so you can basically just shoot away and just ask then Anna or just try to speak and then Anna will let you in. So actually, I think Holly, we can follow up one of the questions we had because I think Michael is actually here in this room. And we can actually check with Michael. Was it so Michael that actually Holly made the match between at the time Jukka and you guys at SASPRI in the very early stage of Southmont? Do you remember Michael? Michael, are you able to ask her to be let in? We need to admit you in first. It's the blue button you should see on your screen. I think it says ask to share, Michael, if you want to. Ask to share video, audio, if you see a button like that. I don't know if he's written something. Ah, he is answering by writing. Yes, Holly was the matchmaker. Yes, Holly was the matchmaker. So now we know Holly. So may I ask something, actually? Yeah, go for it, Anna. Yeah, so I'm just interested because I don't know so much about this topic still. So is there some examples like how the EuroMod has been used? Like, have you collected some kind of impact stories or like, or is it very widely used nowadays to like in everyday decision making in the EU? I hope I can be heard now. Yes, yes, we can hear you. So there are plenty of impact stories and also academic papers that have used EuroMod. I would direct you both to the website at Essex, the ISO website, to look at the the EuroMod and the micro simulation pages there. But I would also direct you to the EuroMod website at the JLC at the European Commission, because they also put material on the web about how the model is being used currently at the Commission. There is a working paper series where when Essex was in charge of EuroMod, we used to ask that all analysis that all sort of substantial analysis was published through that as a way of collecting together a kind of library of ideas and analysis that other people could improve on or build on or whatever. But certainly there are now two web places to look at impact type studies too. I don't know if somebody wants to put the web address there. Yeah, Matteo Arribi, thank you. Yeah, thanks Matteo. Thank you. And I think actually, yeah, we have to, I think, close this soon Anna, right? Because we need to be super much on time in these online conferences. I think so. Maybe we can have just one quick question. Can we have one more question? Yes, so if there's one more question, please, you can type it in the chat. It seems it's a bit hard to get let in. Yeah, maybe. So we see a lot of hellos in the chat. It's just nice too. So that's the part that we are missing from a real live conference. You would now go for a beer probably. Oh, I think that will come and ask something. Yeah. Rodrigo. Hi. Sorry, I'm without the video because I was not prepared for that. But I have one question. Tomorrow we will listen a lot from Michael about this somewhat. But I have this curiosity about how was the process for other countries to decide to develop their models for other developing countries? So was the countries needed to approach you? Or was the in remote team and the university and all the partners that went to the country and help to then to think and propose to develop the model? Well, I think it's a it's a mixture. And I think it's almost pure that should be answering this question in relation to the countries that are part of South Mod. It was very much initiatives taken by people at UNU wider. It was a little bit initiatives taken when people contacted me and I wanted I passed on the ideas that there were people in countries that wanted to build models themselves. So there were also, Gemma and Michael had lots of contacts in other Southern African countries. So there was all sorts of ways in which people were taking the initiative. And from the Euromard side at the time, it was both the more the merrier. But also we were aware that it's we that the effort needed not to be spread too thin. And that there were some fundamental necessary conditions for building a good model, data being one of them, and people who are prepared to commit themselves and who are in a position to commit themselves being another. And I'll stop there because I know we're short of time, but I think it's something that that the PL may well have a very long answer to how these countries get chosen and what happens. And maybe you can continue with the Michael fireside. Oh, by the way, I think Matthew is missed in some other session. If I'm understanding right. So I think all a lot of us will right move over to Yeah, okay. Well, in that case, so I think we're gonna like run out of this room and run into the next room. But before that, I want to thank you, Holly, for making the time. I know you have very important commitment now coming up with your grandson. Thank you for taking the time and look forward to next time. Thank you so much. Thank you. Bye.