1 00:00:00,410 --> 00:00:23,370 Okay. Okay. 2 00:00:23,370 --> 00:00:29,100 I think with that, Reverend, hush, we should go ahead and make a start. Welcome, everyone, to this week's session. 3 00:00:29,610 --> 00:00:32,830 In the Political Economy Financial Markets Seminar. 4 00:00:32,850 --> 00:00:37,860 It's my great pleasure to introduce two very distinguished speakers for tonight's session, 5 00:00:38,460 --> 00:00:44,460 both of whom are going to be talking about the interface of economic and political crisis in interwar Europe. 6 00:00:45,060 --> 00:00:50,640 Our first speaker tonight is Kevin O'Rourke. He is professor of economic history at All Souls. 7 00:00:53,040 --> 00:00:59,520 He received his education from Trinity College, Dublin and his master's and Ph.D. from Harvard University. 8 00:01:00,000 --> 00:01:12,540 He has taught at University College, Dublin and Trinity College, Dublin and Columbia University in New York City before moving to All Souls in 2011. 9 00:01:12,870 --> 00:01:18,479 He's the author of a number of articles and books on economic history and international economics. 10 00:01:18,480 --> 00:01:22,650 I'm just gonna go ahead and read out a few titles, get a sense of his range and some of the things he's published lately. 11 00:01:24,720 --> 00:01:28,770 He published a book with a colleague called Globalisation History The Evolution 12 00:01:28,770 --> 00:01:34,200 of 19th Century Atlantic Economy that was published in 1999 by MIT Press. 13 00:01:34,620 --> 00:01:43,709 He followed that with a book entitled Power and Plenty Trade War and the World Economy in the Second Millennium, published by Princeton in 2007. 14 00:01:43,710 --> 00:01:52,440 And most recently, he is the co-editor of the Cambridge Economic History of Modern Europe, two volumes published by Cambridge in 2010. 15 00:01:53,220 --> 00:01:59,010 He's going to speak for about 45 to 50 minutes, and then we're going to have our discussion often. 16 00:01:59,010 --> 00:02:03,360 And a stock is going to deliver an intervention for us. I think it's quite well known to everybody. 17 00:02:03,360 --> 00:02:08,580 Here is kind of Lord of the Manor. He's the director of the European Studies Centre, director, of course, 18 00:02:08,970 --> 00:02:14,040 also senior research fellow in the Department of Politics and International Relations. 19 00:02:14,040 --> 00:02:22,409 And he too has published widely on a range of different subjects, including Balkan history, affairs of the European Union, 20 00:02:22,410 --> 00:02:28,920 and most recently, the rise of the radical right in Greece in particular and Europe in general. 21 00:02:28,920 --> 00:02:33,570 I just want to mention a couple of things that he's just published recently. 22 00:02:34,200 --> 00:02:38,040 Reforming Greece is a big task or Herculean challenge. 23 00:02:38,040 --> 00:02:44,040 What he did with Dorian seeing and also from crisis to recovery, sustainable growth in Southeast Europe. 24 00:02:44,040 --> 00:02:48,149 So I think we're going to then turn it over to Kevin for 45 to 50 minutes and then often we'll 25 00:02:48,150 --> 00:02:53,280 give a ten minute response and then we'll give Kevin a chance to respond to the response, 26 00:02:53,280 --> 00:02:59,519 if you'd like, and then we'll just open it up for general questions. So please refrain from any questions until we actually get to that point. 27 00:02:59,520 --> 00:03:04,889 But that moment will turn the podcast off and then have a general discussion probably somewhere around 6:00. 28 00:03:04,890 --> 00:03:08,730 Okay, please. Kind of thanks. Is it okay, podcaster, if I stand up? 29 00:03:09,720 --> 00:03:13,890 Yeah. As an Irish Catholic, I've noticed Protestants Protestant better than we do. 30 00:03:13,930 --> 00:03:19,290 I think it's because I stand up for him. I'm going to stand up while I sing. 31 00:03:19,470 --> 00:03:25,890 So I feel that I've been invited along today under false premises in a couple of ways. 32 00:03:26,320 --> 00:03:27,330 I mean, for sure, 33 00:03:27,660 --> 00:03:36,239 looking at the relationship to political extremism and the business cycle is an issue where history does seem to be rhyming quite a lot these days. 34 00:03:36,240 --> 00:03:41,250 And yet I think it's also a lot easier to think of ways in which today's Europe differs 35 00:03:41,490 --> 00:03:45,569 from the Europe of the interwar period and to think about ways in which they are similar. 36 00:03:45,570 --> 00:03:49,920 So it isn't actually clear how many lessons can be drawn from this interwar experience. 37 00:03:50,160 --> 00:03:54,360 I'm sure we'll be talking about that in the discussion as we go along. 38 00:03:54,960 --> 00:03:58,770 And secondly, as you'll see when I discuss the work that I've been doing on this, 39 00:03:59,220 --> 00:04:03,959 you'll see that the nature of the work is such that a lot of the most interesting questions that you might actually want to ask 40 00:04:03,960 --> 00:04:12,420 about the economic roots of interwar political extremism are questions that we can't handle very well given our methodology, 41 00:04:12,480 --> 00:04:19,440 right. So this is a very simple paper, actually. It's a very typical paper as written by quantitative economic historians. 42 00:04:19,440 --> 00:04:23,520 We're interested in average correlation. We're quantifying relationships. 43 00:04:24,300 --> 00:04:27,300 And there are at least two problems with that. We want to draw lessons. 44 00:04:27,300 --> 00:04:35,970 Firstly, if you're talking about something like the development of fascism or authoritarianism or whatever it is in a particular country in the 1920s, 45 00:04:35,970 --> 00:04:43,170 authorities, this is a case where the individual data points are as interesting as the average correlation and probably a lot more interesting. 46 00:04:43,170 --> 00:04:47,190 So that's a big caveat. And secondly, as you'll see, 47 00:04:47,190 --> 00:04:54,089 the nature of the exercise is such that I can't really say very much about the mechanisms that might be underlying the correlations, 48 00:04:54,090 --> 00:04:55,410 although I will be talking about them. 49 00:04:55,620 --> 00:05:01,200 But when I talk about them, I'd just be drawing from bits of the historical literature that Paul Hare knows a lot more about. 50 00:05:01,440 --> 00:05:06,330 I do, and we'll see if he thinks that I've been drawing, whether I've been drawn on the right bits of literature or not. 51 00:05:07,200 --> 00:05:19,360 All right. I started getting interested in the interwar period back in 2008, 2009 in the winter of 28. 52 00:05:20,050 --> 00:05:25,660 I noticed Barry Eichengreen noticed that output and trade were falling every bit 53 00:05:25,660 --> 00:05:29,380 as quickly as they had been doing in the first 12 months of the Great Depression. 54 00:05:29,390 --> 00:05:38,110 So the blue line is an index of world industrial output indexed to the pre Great Depression peak in June 1929. 55 00:05:38,350 --> 00:05:45,430 On the red line is world industrial output indexed to the pre Great Recession peak in April 28. 56 00:05:45,610 --> 00:05:50,139 And when we looked at this first and drove through these graphs, it looked pretty scary. 57 00:05:50,140 --> 00:05:56,560 It looked like we were basically experiencing something that was on a Great Depression type of scale. 58 00:05:56,560 --> 00:06:00,490 And immediately, you know, the first thing you thought was we got a rest this right, 59 00:06:00,490 --> 00:06:05,200 because if it goes on for year after year after year, I mean, that's what did the damage ultimately. 60 00:06:05,200 --> 00:06:08,889 It wasn't one or two years of bad growth. It was it was three or four or five years. 61 00:06:08,890 --> 00:06:13,000 But that's what does the political damage that's what that's how we were thinking. 62 00:06:13,450 --> 00:06:19,629 Now, no sooner had we started to worry about this then then the world economy starts to recover. 63 00:06:19,630 --> 00:06:27,670 I mean, economists are known for their sense of timing. So we, we we published an article that had it up to here, and no sooner had we done that, 64 00:06:27,670 --> 00:06:31,389 then the thing turned around, which is good news, but actually it was okay for that. 65 00:06:31,390 --> 00:06:36,560 We were okay with this intellectually also, because the point of that first article was to say, you know, 66 00:06:36,580 --> 00:06:41,440 we need to avoid the perverse policy response that followed the start of the Great Depression, 67 00:06:41,440 --> 00:06:48,490 where they engaged in pro-cyclical fiscal policy, where they on several occasions raised interest rates trying to keep countries on gold and so on. 68 00:06:48,640 --> 00:06:53,920 And the good news is, at the global level, we did actually do a lot better after 28 nine. 69 00:06:53,920 --> 00:06:57,010 There was the London meeting in April 2009. 70 00:06:57,010 --> 00:07:01,540 They agreed on well, it wasn't a very big coordinated reflation, 71 00:07:01,540 --> 00:07:06,549 but there was a bit of coordinated reflation and at the least they allowed automatic 72 00:07:06,550 --> 00:07:12,190 stabilisers to operate and there was a big concerted central bank response as well. 73 00:07:12,190 --> 00:07:17,890 So this is sort of more or less what we expected to see and we were happy to see it, however. 74 00:07:19,080 --> 00:07:24,180 As you probably know, this wireless recovery is really a story about the emerging world. 75 00:07:24,390 --> 00:07:27,590 So this is the industrial output chart for the emerging world. 76 00:07:27,590 --> 00:07:30,899 And what you see in the rich countries is that there's a recovery. 77 00:07:30,900 --> 00:07:38,850 And then sometime around 2011, it starts to it starts to flatten out and it gets worse. 78 00:07:38,850 --> 00:07:43,650 Because if you look within the rich countries, what you see is that the recovery is really driven by the United States. 79 00:07:44,640 --> 00:07:55,740 In the eurozone in particular. The thing started flattening out in 2010 and has been on a clear, albeit gentle downward slide ever since. 80 00:07:55,740 --> 00:07:59,580 And again, all the we were irritated and worried about this intellectually. 81 00:07:59,580 --> 00:08:05,670 We were okay with it because this is kind of what the basic model would have predicted, because 2010 is, of course, 82 00:08:05,670 --> 00:08:14,730 when everybody starts worrying about Greece and when reflation in the eurozone is replaced with austerity. 83 00:08:14,790 --> 00:08:21,540 And any simple undergraduate model would tell you that that will not immediately but eventually lead to output declines. 84 00:08:21,780 --> 00:08:25,919 And so, in fact, it turns out to be so. 85 00:08:25,920 --> 00:08:30,180 The origins of the paper I'm going to talk about today are round about here. 86 00:08:30,930 --> 00:08:35,670 This is when we learned what the policy response to the Greek crisis was going to be. 87 00:08:35,880 --> 00:08:45,630 And if you're somebody like myself or Barry Eichengreen, you're going to instinctively start to worry that we're tightening too quickly, too early. 88 00:08:45,630 --> 00:08:48,930 And what are the implications of this going to be? 89 00:08:49,050 --> 00:08:56,280 And we started immediately thinking about the intentional political consequences as well as the economic consequences, because, you know, 90 00:08:56,310 --> 00:09:04,950 once you start thinking about the age of war as an analogy, flawed or otherwise, it's the politics of that analogy that make it so, so striking. 91 00:09:05,790 --> 00:09:12,059 And that's what happened to industrial output. Of course, the thing that really could do the damage in the long run politically is unemployment. 92 00:09:12,060 --> 00:09:17,310 And again, it's the same sort of story. There's a concerted rise in unemployment during the Great Recession, 93 00:09:17,310 --> 00:09:24,210 and then it starts to fall in the United States and has been on a downward trajectory ever since I started to rise. 94 00:09:24,480 --> 00:09:28,950 It's the same timing. It's 2010, 2011 in the Eurozone. 95 00:09:28,950 --> 00:09:35,250 And we know how high unemployment is right now in countries like Greece and Spain. 96 00:09:35,730 --> 00:09:40,470 All right. So not just economic historians, policymakers, too. 97 00:09:40,800 --> 00:09:45,060 Around about 2010 started to worry that this might turn sour. 98 00:09:45,720 --> 00:09:55,410 Ultimately, here is a quote by somebody that many of you will know, I suppose, Adam Posen, who was still at the Bank of England in those days. 99 00:09:55,410 --> 00:09:59,990 And this is from a speech he gave at home. And it's a it's a nice quotations. 100 00:10:00,180 --> 00:10:03,690 Everybody was worried about the debt, the public sector debt in these days. 101 00:10:03,690 --> 00:10:10,499 And he's pointing out that there are other stocks of capital that you might legitimately worry about if you're a policymaker. 102 00:10:10,500 --> 00:10:15,840 For example, the stock of Democratic capital that has been built up since World War two in Europe, 103 00:10:15,840 --> 00:10:21,630 maybe we should worry about handing that on intact to our children and grandchildren as well. 104 00:10:21,840 --> 00:10:24,749 All right. So that's that's when we started thinking about this. 105 00:10:24,750 --> 00:10:32,480 I put a post on the Irish Economy blog and then Barry's that he wrote a paper about this and and that's how it all started. 106 00:10:32,490 --> 00:10:38,430 However, today is completely different to the 1930s in many ways. 107 00:10:39,390 --> 00:10:46,470 We haven't come out of World War One. It came out of 5060 years of unprecedented peace and prosperity and so on, so that, you know, 108 00:10:47,120 --> 00:10:52,229 the reparations and, you know, revanchism and frontiers and all of the aftermath of our side. 109 00:10:52,230 --> 00:10:57,870 I mean, that's that's that's that's not got anything to do with us economically. 110 00:10:57,870 --> 00:11:03,209 If you think about the welfare state, I mean, the consequences of being unemployed in Germany or somewhere in 1929, 111 00:11:03,210 --> 00:11:08,250 1930 were much more radical than the consequences of being unemployed in most of our states today. 112 00:11:08,250 --> 00:11:15,329 And that charity has got to have an important economic and political consequence. 113 00:11:15,330 --> 00:11:21,270 And we have bigger states, you know, so even if countries aren't actively replacing their economies, 114 00:11:21,270 --> 00:11:27,810 those bigger states are going to be a stabilising mechanism. The automatic stabilisers going to be working more effectively and so on. 115 00:11:27,960 --> 00:11:33,840 In a sense, the proof of the pudding there is in what happens if, if, if we haven't allowed the automatic stabilisers to work. 116 00:11:33,840 --> 00:11:42,299 So in a country like Ireland or Latvia or so on, you've seen these double digit declines in output of Greece since this thing starts that 117 00:11:42,300 --> 00:11:48,900 that tells you what could happen if we completely ignored the automatic stabilisers. 118 00:11:48,900 --> 00:11:52,890 The difference today is that it's only little countries that don't collectively matter very much, 119 00:11:53,400 --> 00:11:57,660 that have been engaged in that kind of very harsh pro-cyclical policy. 120 00:11:57,870 --> 00:12:02,730 In the twenties and thirties. It was all the big economies that were doing the same thing. 121 00:12:02,970 --> 00:12:06,959 So I do want to emphasise that today is completely different authorities in all sorts of ways. 122 00:12:06,960 --> 00:12:10,140 I'm sure we'll hear more about that in the discussion. 123 00:12:10,320 --> 00:12:16,890 And yes, so this was the point of this is the point of my blogpost that sparked this paper. 124 00:12:17,220 --> 00:12:21,810 Yet there were people saying, look, we don't. Not to worry about this at all, you know, because Europe is just very, very stable. 125 00:12:22,110 --> 00:12:28,050 And where we you know, famously Hitler only got 2.6% of the vote in 1928. 126 00:12:28,710 --> 00:12:31,950 It was a tiny, minuscule party at that time. 127 00:12:31,950 --> 00:12:36,120 And now I know that these people are not Nazis or anything close. 128 00:12:36,120 --> 00:12:41,490 I know that. All right. Thank you. I do know that. But, you know, some of these people are quite unpleasant. 129 00:12:41,520 --> 00:12:44,980 I mean, the obvious is really Jobbik is really very unpleasant writer. 130 00:12:45,320 --> 00:12:50,130 These are these are quite big numbers and these are quite big numbers before the crisis hit. 131 00:12:50,610 --> 00:12:58,860 You know, and so I said, you know, can we really assume that all of this nastiness is completely in the past and I mean, ever since? 132 00:12:58,860 --> 00:13:04,160 I mean, now we see the national numbers in France, of course, and that is genuinely alarming. 133 00:13:04,210 --> 00:13:08,160 That was the point of that initial post of mine. 134 00:13:08,160 --> 00:13:12,840 And so well that it morphed into our project, morphed into a historical paper. 135 00:13:13,050 --> 00:13:17,100 So it's me, it's Barry, and it's a guy called Alan Bromhead who I don't know. 136 00:13:17,100 --> 00:13:19,860 He says he is Dphil student here at Oxford. 137 00:13:19,920 --> 00:13:26,700 So it's a paper by the three of us, and we ended up being sufficiently worried about drawing lessons from the past for the present, 138 00:13:26,760 --> 00:13:32,340 from the past, with the present that we kept that out of the paper. This is just a purely historical paper. 139 00:13:32,700 --> 00:13:38,070 We're not really so much concerned with the question of can we generalise from the 1920s and thirties to today, 140 00:13:38,190 --> 00:13:42,809 but can we generalise from the German experience in this period to the experience of other countries? 141 00:13:42,810 --> 00:13:46,280 So Germany is different in lots of ways. 142 00:13:46,290 --> 00:13:50,369 I mean, it's famously the country where, you know, mass unemployment leads to Hitler and so on. 143 00:13:50,370 --> 00:13:54,569 But can you really generalise from this? It was distinctive in all sorts of ways. 144 00:13:54,570 --> 00:14:02,040 It had a particularly reactionary agrarian aristocracy before World War One, which sort of lingers into the interwar period. 145 00:14:02,310 --> 00:14:06,090 They had been defeated in World War One. Not everybody was defeated by World War One. 146 00:14:06,420 --> 00:14:13,920 Hard to pick people winning it, but not everybody was defeated. It was driven by all sorts of class, some logical and religious divides. 147 00:14:14,040 --> 00:14:19,679 And the Weimar electoral system was famously friendly towards small parties. 148 00:14:19,680 --> 00:14:29,370 So maybe the association that we think exists between unemployment and the Depression and Hitler in Germany is sui generis, 149 00:14:29,370 --> 00:14:33,210 and there's nothing more general that can be said about this phenomenon. 150 00:14:33,600 --> 00:14:41,790 So we decided that we would see if there was a general relationship between business cycle conditions and extremism in the interwar period. 151 00:14:41,790 --> 00:14:48,120 So there's two big questions. Do we find this link in general between economic hard times and voting for extremists in other countries? 152 00:14:48,450 --> 00:14:49,379 And then secondly, 153 00:14:49,380 --> 00:14:56,430 but this is a minor part of the paper actually do the structural features that have been advanced as explanations for German voting patterns, 154 00:14:56,430 --> 00:15:02,760 for example, they're their medieval aristocratic junker, just, you know, 155 00:15:03,000 --> 00:15:06,870 are there are those sorts of structural features also at work in other countries? 156 00:15:07,080 --> 00:15:13,260 As it turns out, there's not that much we can say about this, but we we do try to say something. 157 00:15:14,190 --> 00:15:18,270 So what we do, we have data for 171 elections in 28 countries. 158 00:15:18,660 --> 00:15:27,240 All right. And we're going to analyse the determinants of voting for extremism in these 171 elections we're interested in. 159 00:15:27,450 --> 00:15:30,900 Was there a link between the business cycle and voting for extremists? 160 00:15:31,110 --> 00:15:36,750 And we're interested about what conditions might have led to that link being particularly tight. 161 00:15:36,840 --> 00:15:39,389 Maybe it's not a one size fits all correlation. 162 00:15:39,390 --> 00:15:45,210 Maybe there are different sorts of countries, and this correlation is tighter in some types of countries than in other types of countries. 163 00:15:45,600 --> 00:15:52,450 What the paper is not about is the collapse of democracy. So an Oxford College ironic approach has written a big book about this. 164 00:15:52,450 --> 00:15:56,550 And well, there's a problem once you start thinking about this, because a lot of the democracies collapse before the Depression. 165 00:15:56,730 --> 00:16:02,230 So we're you know, we're just not going there. All right. This is a paper purely about voting behaviour. 166 00:16:03,240 --> 00:16:06,510 And then we started looking for literature. 167 00:16:06,510 --> 00:16:12,630 There is some literature on voting for right wing extremists today or over the last two or three decades. 168 00:16:12,780 --> 00:16:16,049 It tends to be interested in things like where you have more unemployment too. 169 00:16:16,050 --> 00:16:22,830 People are more likely to vote for a right wing populist party, and they find that probably yes. 170 00:16:23,010 --> 00:16:28,169 But then there are nuances. So this paper here says there is a link between unemployment and populism in 171 00:16:28,170 --> 00:16:31,320 countries where there is an immigrant class that can serve as a scapegoat, 172 00:16:31,620 --> 00:16:35,820 but not in countries where there are not so many immigrants. And then there's a debate about that. 173 00:16:36,030 --> 00:16:41,010 So there is a statistical debate in the political science literature about this. 174 00:16:41,020 --> 00:16:44,950 And anyway, this is the paper that first made us think maybe we could write a paper. 175 00:16:44,950 --> 00:16:48,360 It's just looking at not election results, but an opinion surveys, 176 00:16:48,690 --> 00:16:55,950 and it's finding that there is a link between between growth and self-declared support for extremism. 177 00:16:55,950 --> 00:17:02,820 But of course, self-declared support for extremism isn't the same as supported at the ballot box, which is what we're really interested in. 178 00:17:04,710 --> 00:17:12,570 And there's obviously huge literature, historical literature about all of these phenomena in the interwar period, 179 00:17:12,570 --> 00:17:19,709 but we couldn't find exactly what we were looking for. Right. There's a there's a big literature on the rise of the fascists and the Nazis. 180 00:17:19,710 --> 00:17:23,910 There is a quantitative literature on the determinants of the mood for Hitler within Germany. 181 00:17:23,910 --> 00:17:30,450 And I'm going to be referring to that briefly. But we couldn't find similar quantitative literatures that talked in general about 182 00:17:30,450 --> 00:17:35,160 links between right wing extremism and the business cycle across countries. 183 00:17:35,880 --> 00:17:39,750 And so we felt that there was a definite gap in the market there. 184 00:17:40,490 --> 00:17:45,780 Let me talk a little bit about the literature on Germany, because it's the best developed for obvious reasons. 185 00:17:45,990 --> 00:17:50,400 And if there are students here, I mean, it's the sort of literature where there are actually research questions that can be, 186 00:17:50,940 --> 00:17:54,340 I think, generalised to other countries potentially and so on. 187 00:17:54,360 --> 00:17:59,890 So in the in the German literature, there are two distinct questions that you could ask. 188 00:17:59,910 --> 00:18:04,080 Firstly, what's the impact of the Great Depression? And secondly, who voted for Hitler? 189 00:18:04,890 --> 00:18:11,550 This is a sort of a time serious question. It's about what drove the upward trend in the Nazi vote over time. 190 00:18:11,790 --> 00:18:16,839 And this is about, you know, in a given election, what groups in society voted for Hitler or not? 191 00:18:16,840 --> 00:18:22,740 And these two questions can end up being conflated sometimes because unemployment is what goes up during the Depression. 192 00:18:23,370 --> 00:18:27,230 It's not our best measure of the Depression, but that doesn't mean that the unemployed voted for Hitler. 193 00:18:27,240 --> 00:18:34,420 In fact, the unemployed probably some of them did, I suppose, but they're probably much more likely on average to vote for the communists. 194 00:18:34,440 --> 00:18:38,070 And so there's been a statistical literature about how you think about this. 195 00:18:38,280 --> 00:18:42,660 So this is the argument that says there's something going on here, folks. 196 00:18:42,900 --> 00:18:52,170 But so this is the Nazi vote. I cheated very slightly in that there are two elections in 1930 to support the November election as signed up to 1933. 197 00:18:52,290 --> 00:18:53,490 So the graph would look nice. 198 00:18:53,730 --> 00:19:00,210 And I assigned to the December 24 election just 25 so that, you know, it wasn't having to do with two elections in one year. 199 00:19:00,810 --> 00:19:01,250 But I mean, 200 00:19:01,290 --> 00:19:08,460 actually doing that makes sense because famously the Nazis do see a dip in the share of their votes and the economy is recovering by late 32. 201 00:19:09,210 --> 00:19:13,680 So that's the Time series correlation. But in the cross section, this is a stylised picture, 202 00:19:13,680 --> 00:19:20,280 but the the correlation coefficients are are drawn from real, real data exercises in the cross section. 203 00:19:20,520 --> 00:19:26,900 You know, the the communists are more likely to vote, you know, for for the car, 204 00:19:27,000 --> 00:19:29,850 the unemployed are more likely to vote for the communists than for the Nazis. 205 00:19:30,060 --> 00:19:36,270 So there's a negative relationship, say, across districts between the share of the Nazi voters and the unemployment rate. 206 00:19:36,480 --> 00:19:39,920 But what's happening over time is that negative correlation is shifting out. 207 00:19:40,380 --> 00:19:48,510 That's what that's what's happening. And so essentially the the relationship that you find regarding was a to unemployment and 208 00:19:48,630 --> 00:19:53,130 voting for Hitler depends on whether you're primarily picking up time series by indication, 209 00:19:53,280 --> 00:19:56,250 in which case you will find a link or cross variation, 210 00:19:56,400 --> 00:20:00,900 in which case you won't write into the statistical literature about how you tease all of this out. 211 00:20:01,350 --> 00:20:05,700 The more interesting literature really is who actually voted for him, right? 212 00:20:05,700 --> 00:20:11,910 So so this is Gary King, who's a political scientist at Harvard and a brilliant guy. 213 00:20:12,660 --> 00:20:18,510 So, you know, at the constituency level, we know who voted for the Nazis, the liberals, you know, and everybody else. 214 00:20:18,810 --> 00:20:25,800 And we also know from the census data, because we can match that to the constituencies now who are self-employed, blue collar, white collar and so on. 215 00:20:26,160 --> 00:20:31,890 And so what we'd really like to know is, you know, what share of the unemployed voted for the far left, the centre, you know, the Nazis and so on. 216 00:20:32,430 --> 00:20:38,190 Now you will notice that there are a lot more question marks here than there are numbers and this is a problem. 217 00:20:38,490 --> 00:20:43,049 But Gary King is a clever man and he's found a solution. So he has a maximum likelihood estimate. 218 00:20:43,050 --> 00:20:47,730 It allows you essentially to fill in all of the blanks. The proof of the pudding is in the eating. 219 00:20:47,730 --> 00:20:51,629 So what he does is he comes up with results that look sensible. 220 00:20:51,630 --> 00:20:59,550 So he finds that the areas where you are most likely to get an increase in the Nazi vote moving from 28 through 221 00:20:59,550 --> 00:21:06,660 to 32 are in areas where you have high proportions of self-employed people and domestically employed people, 222 00:21:06,720 --> 00:21:11,879 which includes agrarian peasants, but predominantly in Protestant areas like his. 223 00:21:11,880 --> 00:21:15,090 In the Catholic areas, they're still voting for Santorum and so on. All right. 224 00:21:15,090 --> 00:21:19,530 So that's sort of what we thought we knew. And so in a sense, that helps to validate the methods. 225 00:21:19,830 --> 00:21:24,570 I should point out that the same problem Head, who's a co-author with us, is doing his dphil here, 226 00:21:24,720 --> 00:21:32,340 and he's exploiting the fact that in about 20% of the constituencies that counted the women's votes on the men's votes separately, 227 00:21:32,580 --> 00:21:37,650 which is an extraordinary things, you can actually see how the women voted in constituencies compared with the male votes. 228 00:21:37,890 --> 00:21:42,150 And what you find is that the women were actually more moderate than the men and 229 00:21:42,150 --> 00:21:46,590 then they become more like the men in 32 and they become more like the men, 230 00:21:47,100 --> 00:21:49,950 especially in these groups. They're self-employed and domestics. 231 00:21:49,950 --> 00:21:54,660 So there's all sorts of things that you can do with this methodology, and you could do this for other countries too. 232 00:21:54,900 --> 00:22:00,270 I'm sure we're not doing anything quite so exciting, unfortunately. 233 00:22:00,930 --> 00:22:03,960 So we're interested in the impact of economic. 234 00:22:04,000 --> 00:22:08,170 Growth on right wing extremist voting. 235 00:22:08,440 --> 00:22:11,770 It would be better to do on employment and the right wing votes. 236 00:22:12,100 --> 00:22:17,110 But unemployment statistics for the interwar period are just not comparable. It's a complete mess, so we can't do that. 237 00:22:17,120 --> 00:22:23,540 So we have to look at GDP growth. All right. We have this for a bunch of countries over time. 238 00:22:23,560 --> 00:22:27,940 So we have several Norwegian elections. We have several German elections before. 239 00:22:27,940 --> 00:22:31,120 We do want to take account of the fact that Norway is not the same as Germany. 240 00:22:31,390 --> 00:22:34,900 This would seem sensible. And so we have fixed effects. So. 241 00:22:34,900 --> 00:22:41,200 So. So we have a dummy variable in our regression that is one for Norway and zero for everybody else. 242 00:22:41,200 --> 00:22:44,260 So you're allowing for the fact that Norway is different from Germany. 243 00:22:44,380 --> 00:22:49,150 But that means that if there are any other features of being Norway that are invariant across time, 244 00:22:49,300 --> 00:22:53,020 we can't tell you anything about how they influence the vote in Norway. 245 00:22:53,170 --> 00:22:59,920 They're all sucked up on that dummy variable. And that's that's a basic problem when it comes to thinking about mechanisms and structure and so on. 246 00:23:00,420 --> 00:23:08,110 All right. Most of the data are in Europe, but we have some non-European data also, and we're relying on the political scientists here. 247 00:23:08,290 --> 00:23:13,329 So we're particularly interested in votes for anti-system parties or anti-system parties defined 248 00:23:13,330 --> 00:23:18,190 as a party that would change the system of government rather than just the government itself. 249 00:23:19,180 --> 00:23:24,040 And we're particularly focusing on the extreme right wing anti-system parties and the the communists. 250 00:23:24,400 --> 00:23:29,320 And well, I'm going to start up. So this is your favourite, your country. 251 00:23:29,320 --> 00:23:34,250 You can see if it makes sense. And I'm going to quickly scroll down just in case it doesn't. 252 00:23:34,480 --> 00:23:38,950 There are some judgement calls here. So for example, the NDP is this anti system, 253 00:23:39,190 --> 00:23:45,410 so we're relying on somebody else's classification and he says they become anti-system after a certain point, you know. 254 00:23:45,430 --> 00:23:50,020 Well, we well we worry about as you'll see in a moment. 255 00:23:50,560 --> 00:23:54,910 So this is the relationship we're interested in relationship growth and extreme exposure. 256 00:23:55,120 --> 00:23:59,230 We have control variables of various sorts or there are other things that might matter. 257 00:24:00,700 --> 00:24:04,060 There are people who've argued that modernity is good for democracy. 258 00:24:04,070 --> 00:24:08,560 So we collected data on urbanisation. That's a variable that Lipsett thought was important. 259 00:24:08,770 --> 00:24:14,860 We know whether in your country you had a nasty, reactionary, pre-war agrarian elite. 260 00:24:15,130 --> 00:24:20,950 We have data quoted from other authors on whether they're just an for cleavage, as we know. 261 00:24:20,950 --> 00:24:27,520 Did you win or lose World War One? We know. Did you have a poor electoral system and so on? 262 00:24:27,780 --> 00:24:33,730 Here's the one that turns out to actually be important. We look out whether you were a democracy before World War One or not, 263 00:24:34,750 --> 00:24:41,530 because there is an argument that says that exposure to democracy helps inoculate you against non-democratic partisans. 264 00:24:41,530 --> 00:24:46,720 That's a theoretical claim that's been made, and we're going to talk about that later on. 265 00:24:46,990 --> 00:24:49,360 Okay. So I'm going to show you some regressions. 266 00:24:49,360 --> 00:24:57,210 But before I show you the regressions, let me just show you some sort of average voting data, if you like. 267 00:24:57,220 --> 00:25:02,710 So this is the votes for anti-system parties from the last election before the Great Depression, 268 00:25:03,340 --> 00:25:12,070 and this is the votes for the anti-system parties in the post 1929 election where those extremist votes were at their peak. 269 00:25:12,490 --> 00:25:16,360 Right. So it's before and then it's after. But it's, you know, the worst that ever goes. 270 00:25:16,450 --> 00:25:21,720 Right. And it's just divided up by countries. So I'm going to help you see what you ought to see. 271 00:25:21,730 --> 00:25:29,170 So, first of all, what you see is that there is an increase in voting for extremists from three 4% on average to 10%. 272 00:25:29,470 --> 00:25:35,950 We want to exclude Finland sometimes because they suppress their communists and there's lots of variation in the data. 273 00:25:35,950 --> 00:25:42,130 So there are countries like Germany and Czechoslovakia where there was a pretty big extremist vote before and it gets even bigger. 274 00:25:42,520 --> 00:25:49,990 There are countries like Romania and Bulgaria where there was not a big extremist vote before and then a big one after. 275 00:25:50,440 --> 00:25:56,410 And then there are moderate, tolerant countries like Ireland and Norway where there were no extremists before. 276 00:25:56,560 --> 00:25:59,590 I know extremists after. So this is nice. 277 00:25:59,590 --> 00:26:02,829 I mean, it means that there is variation in the data. It also means you can't generalise this. 278 00:26:02,830 --> 00:26:09,730 Most economists are always torn between wanting variation in the data and wanting to generalise, and these things don't always go together. 279 00:26:10,870 --> 00:26:13,870 Generally, it tends to be the extreme right that does well. 280 00:26:13,930 --> 00:26:21,100 So these countries here are both highlighted. We see a big increase in the extreme right wing vote, for example, Romania. 281 00:26:21,760 --> 00:26:26,500 But there are some countries where it's the communists that see a big increase in their vote. 282 00:26:26,510 --> 00:26:32,710 So again, it's hard to really generalise. There are average correlations, but hard to really generalise. 283 00:26:32,860 --> 00:26:37,629 Let me let me begin by just say something about these country characteristics that 284 00:26:37,630 --> 00:26:41,590 I'm not going to be able to deal with my regressions because they're fixed by time. 285 00:26:41,590 --> 00:26:43,360 So I'm interested in, for example, 286 00:26:43,540 --> 00:26:51,620 whether having a pre 1914 agrarian aristocracy meant that you had a bigger fascist those in the interwar period, for example. 287 00:26:51,640 --> 00:26:56,800 All right. So, so I can see whether it means that you have it. 288 00:26:56,830 --> 00:27:03,880 I can see whether having a pre-war agrarian least leads to a higher fascist vote. 289 00:27:03,940 --> 00:27:12,310 In general after 1990. And I can also look and see whether it leads to a bigger increase in the clashes both after and after 1929. 290 00:27:12,610 --> 00:27:16,120 And by the way, I'm going to say fascist, even though I know they're not all fascists, 291 00:27:16,120 --> 00:27:20,800 but pretty, you know, extreme right, anti-system party is a bit of a a mouthful. 292 00:27:22,270 --> 00:27:25,690 All right. So I'm going to show you pictures. 293 00:27:25,690 --> 00:27:28,360 Don't worry. All right. The pictures are ultimately based on regressions. 294 00:27:28,660 --> 00:27:34,239 And actually, let me just because time is can move on, basically, we find nothing for good. 295 00:27:34,240 --> 00:27:38,470 You have a pre-war agrarian least where you urbanised. Did you have a religious divide? 296 00:27:38,830 --> 00:27:43,900 Did you have an ethno linguistic divide? There's no average correlations there in the data whatsoever. 297 00:27:44,080 --> 00:27:47,890 All right. Where we get where we got correlations. 298 00:27:47,920 --> 00:27:51,250 First of all, it's the extreme right that an average does better. 299 00:27:51,340 --> 00:27:56,710 So this is how much did the fascist vote go up on average after 1929? 300 00:27:56,950 --> 00:28:00,739 And it usually goes up by quite a bit. And they're often statistically significant. 301 00:28:00,740 --> 00:28:05,120 And if you compare that with the communist. You know, on average, they're not doing that well. 302 00:28:05,140 --> 00:28:09,070 So it's this kind of paradox. It's a crisis of capitalism. 303 00:28:09,070 --> 00:28:12,400 And it's the it's not the communists who are reaping the benefits. 304 00:28:12,910 --> 00:28:14,020 All right. Pictures. 305 00:28:15,640 --> 00:28:27,160 So if we slice and dice the the country sample into different groups, what turns out to be statistically significant and quantitatively meaningful? 306 00:28:27,310 --> 00:28:32,290 So one way of chopping up the sample that seems to make a difference is where you were World War One, loser or not. 307 00:28:32,650 --> 00:28:39,100 So if you were a World War One loser, then after 1929, you have big fascist vote, 308 00:28:39,430 --> 00:28:42,780 much bigger than if you had not been on the losing side in World War One. 309 00:28:42,790 --> 00:28:46,060 So these are post 29 votes. These are 29 votes. 310 00:28:46,540 --> 00:28:51,040 There's no fascists. Most any place before a Great Depression afterwards there is. 311 00:28:51,040 --> 00:28:56,560 But especially if you are on the losing side. Communists, it looks different with communists. 312 00:28:56,860 --> 00:29:04,090 Again, being a loser matters. But the effect shows up immediately after World War One, whereas the fascists, it only showed up after 1929. 313 00:29:04,300 --> 00:29:10,480 Nothing causal here. But I mean, these are just interesting. These are interesting facts about the days, I suppose. 314 00:29:10,630 --> 00:29:16,720 And then the other one that that seems to matter is where you had democracy before World War One or not, 315 00:29:16,990 --> 00:29:23,020 which is obviously largely the same as where you on the losing side in World War One, which is a problem. 316 00:29:23,020 --> 00:29:26,160 But we'll get to that. It's the same it's the same story. 317 00:29:26,170 --> 00:29:33,640 Right. So the war democracies are largely immune from the siren call of fascism. 318 00:29:34,480 --> 00:29:41,860 Yeah, it's the same thing. Essentially, the only other kind of country characteristic that seems to matter is, is you have a PR system. 319 00:29:42,340 --> 00:29:47,050 But interestingly enough, it wasn't the fascists that seem to have systematically benefited from the communists. 320 00:29:47,560 --> 00:29:54,250 All right. So so this is saying that if you got PR, you had a much bigger communist post than if you didn't have PR. 321 00:29:54,790 --> 00:30:00,160 Yeah, it's it's it's much bigger, in fact. And then and then it gets even bigger still after the depression. 322 00:30:01,060 --> 00:30:08,440 Okay. So summing up thus far, I guess I have I guess I have I guess I have 15 minutes left. 323 00:30:08,560 --> 00:30:15,070 So I'm sure that the fascist vote increases by more after 29 than the Communists vote 324 00:30:15,100 --> 00:30:19,570 on average countries have better on average in countries that have lost World War One. 325 00:30:19,750 --> 00:30:23,980 And now you have a quote there to suggest that this makes sense. 326 00:30:24,550 --> 00:30:30,820 And in countries of of PR electoral systems. So so where, you know, these poor people who get in the trenches, I mean, you know, 327 00:30:30,990 --> 00:30:35,530 if they get mad and if they're on the front lines, they seem to have gotten louder on average. 328 00:30:35,530 --> 00:30:42,160 And this leads. Well, I like this guy liberated. You know, this guy who's a political scientist, I must say, I like his account of all of this. 329 00:30:42,520 --> 00:30:45,459 I think we can talk about that in the discussion. 330 00:30:45,460 --> 00:30:54,220 Make it fascist, do better in countries that lost World War One, but only after 1929, in new democracies, especially after 1929. 331 00:30:54,910 --> 00:30:59,110 And now what have we not accounted for? It's just been before and after. 332 00:30:59,110 --> 00:31:04,360 Before 29. After 29, we haven't accounted for the fact that the Depression starts at different times in different 333 00:31:04,360 --> 00:31:08,500 countries and at different times in different countries is worse in some countries than others. 334 00:31:08,770 --> 00:31:11,890 And this is where we want to do some regression analysis. 335 00:31:12,100 --> 00:31:17,739 So this is what the data look like. So this is the growth over three years. 336 00:31:17,740 --> 00:31:20,740 And here's the fascist vote share. You can see there's lots of zeros. 337 00:31:21,190 --> 00:31:24,489 There's lots of countries that never had a fascist party to vote for. All right. 338 00:31:24,490 --> 00:31:29,620 And that's going to influence the type of regression that we run. 339 00:31:29,650 --> 00:31:36,610 The type of statistical exercise that we engage in has got to take account of the fact that there are all of those zeros in the data. 340 00:31:36,910 --> 00:31:44,170 All right. And we're also going to want to see if any correlations we find a robust to excluding Germany for obvious reasons. 341 00:31:44,170 --> 00:31:50,680 Is this just a German story or is it something more than that? Um, yes, I've said that. 342 00:31:51,190 --> 00:31:55,150 So let me let me show you some regression results. Okay. So. 343 00:31:56,650 --> 00:32:04,510 These regression results. They show that they show the impact of various variables on or the statistical 344 00:32:04,510 --> 00:32:09,520 relationship between various variables and the extreme right wing anti-system votes. 345 00:32:10,180 --> 00:32:18,880 I'm looking at in particular the impact or the relationship between growth and fascist vote growth over the previous one year. 346 00:32:18,910 --> 00:32:23,170 Over the previous two years, over the previous three years. That's what I'm really interested in. 347 00:32:23,620 --> 00:32:25,959 All right. There are some control variables there. 348 00:32:25,960 --> 00:32:35,220 So we do want to allow for the fact that 1920 933 was different to 25 to 28 was different to 19 1923. 349 00:32:35,230 --> 00:32:41,370 So we are at least controlling for some things that are making the world worse and worse as you move forward in time. 350 00:32:42,100 --> 00:32:44,410 Nothing much there. If you look at urbanisation. 351 00:32:45,370 --> 00:32:51,190 But that's because a country is more or less urbanised and that's a sort of pretty country fixed effect. 352 00:32:51,580 --> 00:32:55,659 I mean, it will it will become a little bit more urbanised over the course of this period. 353 00:32:55,660 --> 00:32:58,600 But basically, there are some countries that are more urbanised, some that are less. 354 00:32:58,750 --> 00:33:01,550 And so this is all going to be swallowed up in these country fixed effects. 355 00:33:01,900 --> 00:33:06,070 I was talking about if you just forget about the fixed effects and exclude Germany, 356 00:33:06,580 --> 00:33:14,799 then indeed more urbanised countries have smaller fascist votes, which is, I guess, what Lipsitch would have predicted, you know. 357 00:33:14,800 --> 00:33:18,820 But Germany messes things up for him because Germany is very modern and has a very big fascist. 358 00:33:19,840 --> 00:33:27,010 All right. But it's fairly weak. And there's there's something there if you look for evidence of of electoral systems mattering. 359 00:33:27,490 --> 00:33:34,480 So for any political scientist, you know, in a PR system, you have an electoral threshold like 5% in Germany, we all know. 360 00:33:34,750 --> 00:33:40,600 But then you can also think about what's the effect of electoral thresholds just based on things like given that you have constituencies, 361 00:33:40,840 --> 00:33:44,470 how many seats are in the constituency, how many parties are vying for those seats? 362 00:33:44,710 --> 00:33:52,210 Then de facto, what share of the vote you need to gather in order to have a probabilistic chance of getting somebody into the parliament? 363 00:33:52,570 --> 00:33:54,280 All right. So it's a de facto kind of thing. 364 00:33:54,460 --> 00:34:00,610 And there's some weak evidence there that the higher the threshold you have to jump over, the lower is that fascist vote. 365 00:34:00,880 --> 00:34:05,080 But it's not terribly strong. But there's something there. This is what we're interested in. 366 00:34:06,170 --> 00:34:13,390 And what's interesting is that if you look at the impact or the Statistical Association for the growth in the fascist vote over one year, 367 00:34:13,540 --> 00:34:17,350 if you exclude not control variables, it goes away. It goes away. 368 00:34:17,620 --> 00:34:20,620 It goes away for two years. It's there for three years. It's there for four years. 369 00:34:20,770 --> 00:34:26,230 It's there for five years. So there's a statistical relationship between growth over three, four or five year horizon, 370 00:34:26,350 --> 00:34:30,520 but not over the one and two year horizon, which from introspection makes sense. 371 00:34:30,910 --> 00:34:35,050 But that's just from introspection. You know, you lose hope at a certain point. 372 00:34:35,320 --> 00:34:42,040 That's why it's important to cut these things off. Now, you might be worried that the you might be worried that if we're looking at growth 373 00:34:42,040 --> 00:34:45,310 over three years and if there are several elections in that three year period, 374 00:34:45,490 --> 00:34:48,070 that we're double counting bad experiences somehow. 375 00:34:48,370 --> 00:34:54,880 So if you do it again but exclude all elections falling within three years of each other, you get the same result. 376 00:34:55,960 --> 00:35:00,400 And if you exclude Germany, you get the same result. It's not just a German phenomenon. 377 00:35:00,600 --> 00:35:06,489 Okay, so there is a statistical relationship there, it seems, between growth over the longer than, 378 00:35:06,490 --> 00:35:11,680 say, for three, four or five year horizon and the fascist vote, but not shorter periods of time. 379 00:35:11,980 --> 00:35:23,230 I'm interested in how big these effects are. And let me actually I give you a better I give you a better slide on that in a moment. 380 00:35:23,950 --> 00:35:26,620 All right. I'm here. Here's where I'm coming from. 381 00:35:27,160 --> 00:35:34,090 We're interested in whether the association between growth and the fascist vote is bigger in some countries than in others. 382 00:35:34,780 --> 00:35:42,430 So what you can do is you can chop the country of the country, sample up into countries that were democracies before World War One, and that won't. 383 00:35:42,970 --> 00:35:51,460 And then you can basically see if the association between growth and the fascism fascist pose is higher in one group of countries than in the others. 384 00:35:51,850 --> 00:35:58,990 So what this says here, this is an interaction for the competition between growth and a pre-World War One dummy. 385 00:35:59,230 --> 00:36:02,920 What this is basically saying is if you are a pre-war democracy, 386 00:36:03,100 --> 00:36:09,160 the total association between growth and the fascist vote is this plus this, which is almost zero. 387 00:36:09,460 --> 00:36:18,610 It's very small. All right. So if you're a Briton or if you are Ireland or if you're another country that had been democratic before World War One, 388 00:36:18,910 --> 00:36:25,870 there's no real relationship there. The the total relationship that I found earlier is being driven statistically by the relationship 389 00:36:26,470 --> 00:36:32,110 that you find in the sample of countries that hadn't been democratic before World War One. 390 00:36:32,230 --> 00:36:37,300 And when you then look and see how big this statistical association is. 391 00:36:37,720 --> 00:36:43,960 Well, one way of doing it is to think about if you experienced the growth decline that they experienced in Germany, 392 00:36:44,470 --> 00:36:49,360 how big of an increase in the fascist vote would you expect given this average correlation? 393 00:36:49,600 --> 00:36:58,150 And the answer is you'd expect an increase in the fascist vote of eight percentage points, eight percentage points, as opposed to one out pence. 394 00:36:58,360 --> 00:37:04,630 If you take that the DMV role was fascist, then the increase in the vote is only 20 percentage points. 395 00:37:04,660 --> 00:37:10,990 Let me explain quite a lot to if you think that they flipped in an important matter, then we're explaining a smaller share. 396 00:37:11,260 --> 00:37:15,250 But we're so we're explaining quite a bit. But it's not the whole story. 397 00:37:15,670 --> 00:37:22,270 But there is something there. All right. So this is just a quote from Eric Hobsbawm. 398 00:37:22,270 --> 00:37:24,040 You know, there are you know, the notion, 399 00:37:24,040 --> 00:37:30,970 the notion that having been a democracy over a longer period of time might inoculate you against these kinds of political temptations. 400 00:37:31,150 --> 00:37:35,860 Isn't a new story all that we're doing here is we're finding statistical evidence for it. 401 00:37:36,550 --> 00:37:45,000 If there are economists in the audience, you're going to be worried about the possibility that it isn't the slump that is causing fascism, 402 00:37:45,010 --> 00:37:47,140 but the fascists that are causing the slump. 403 00:37:47,470 --> 00:37:53,830 So maybe I'm glad that you're smiling because I mean, I think the direction of causation is pretty clear in myself. 404 00:37:54,040 --> 00:38:00,400 But the argument would be that if you're an investor and you see Hitler doing better in 1930, maybe you panic and this has about it. 405 00:38:01,030 --> 00:38:04,210 And there should be was a bit of that going on. Right. 406 00:38:05,170 --> 00:38:10,510 So we we try to convince our readers that this isn't so much of a problem, 407 00:38:12,250 --> 00:38:16,899 but insofar as it is a problem, I think it strengthens our key political message, 408 00:38:16,900 --> 00:38:22,480 which is that if there had been better macroeconomic policy earlier on in the crisis, 409 00:38:22,720 --> 00:38:30,370 then this supposed a two way feedback mechanism between worsening economic conditions and worsening political conditions wouldn't have gotten going. 410 00:38:30,370 --> 00:38:36,750 So mean. That's ultimately what our agenda is. So we feel, you know, okay, about this, correlations are interesting. 411 00:38:36,760 --> 00:38:42,129 Sometimes we do a bunch of robustness checks. If you are worried about this kind of thing. 412 00:38:42,130 --> 00:38:49,180 So this is a base case estimation risk with something called fixed effects tobit you could do a round of effects total, 413 00:38:49,180 --> 00:38:52,809 which you can do with your dummies. You can do no one else with your dummies. 414 00:38:52,810 --> 00:38:58,120 It's pretty robust. You try different things and you got a pretty robust correlation. 415 00:38:58,360 --> 00:39:03,460 And there's some evidence this is looking at growth and growth squared. 416 00:39:04,210 --> 00:39:12,610 So that's to see it as a non-linear effects. So when growth gets worse, does that have a much bigger bout effect on the voting behaviour? 417 00:39:12,610 --> 00:39:19,030 And it does. You know, there's some evidence, some evidence there of nonlinear linearity which is maybe not so surprising. 418 00:39:20,380 --> 00:39:24,220 This is the next slide that I really want to talk about. 419 00:39:24,370 --> 00:39:27,340 Do the same thing for communists. And there's nothing there. There's nothing there. 420 00:39:27,550 --> 00:39:32,170 There's no relationship between growth in the communist voters, which is, you know, surprising. 421 00:39:32,170 --> 00:39:35,469 Maybe, but maybe not. Depends on whether you're a historian or not. 422 00:39:35,470 --> 00:39:43,840 Possibly. I show this to economists and they're very surprised. So they then say, well, why on earth are the fascists benefiting from this? 423 00:39:44,230 --> 00:39:49,900 And the communists aren't. And here I can say, look, I'm only I'm only an economist myself, you know, what do I know? 424 00:39:49,900 --> 00:39:54,580 Right. But so so this is the sorts of things that I think historians say, right? 425 00:39:54,580 --> 00:40:00,610 So there's an extent to which Social Democrats were discredited in some countries by their involvement in World War One. 426 00:40:01,930 --> 00:40:05,499 The left wing movement splinters, you know, the communists and the Social Democrats. 427 00:40:05,500 --> 00:40:12,070 There's civil warfare between the two sides, especially after this left turn in 1929. 428 00:40:12,250 --> 00:40:16,510 And then if we were looking for resonances, because I do, I guess, share. 429 00:40:16,770 --> 00:40:19,799 Cars. The cars view what useful history as you know. 430 00:40:19,800 --> 00:40:20,070 So. 431 00:40:20,220 --> 00:40:28,380 So if you're looking for resonances, I don't know what the real historians in the room think of this book, but I quite like this book by Shari Berman. 432 00:40:28,620 --> 00:40:33,510 So Shari Berman says, you know, the problem with a lot of the speed parties is they were Orthodox Marxists. 433 00:40:33,750 --> 00:40:40,530 They were waiting for the revolution to happen. They never really came up with programs that would have preserved private property, 434 00:40:40,530 --> 00:40:46,440 but banned the market so that it serves the interests of people and not the other way around. 435 00:40:46,620 --> 00:40:50,399 In countries like Sweden, where they did that, you know, or even Ireland, I suppose, 436 00:40:50,400 --> 00:40:54,030 as a kind of funny example of that, you know, where there was interventionism that seemed to work. 437 00:40:54,210 --> 00:41:01,030 People loved that. And the problem was the Orthodox left in many countries wasn't wasn't offering that. 438 00:41:01,080 --> 00:41:05,549 So that's what I read from the literature. And there might indeed be resonances there. 439 00:41:05,550 --> 00:41:11,250 But but maybe historians want to disagree with this interpretation of the data. 440 00:41:11,400 --> 00:41:15,240 All right. We also looked at what determines how many people you got into parliament. 441 00:41:15,780 --> 00:41:18,840 Well, the higher the vote share, the bigger the seat share. 442 00:41:19,020 --> 00:41:22,080 That's not very surprising, but less obvious. 443 00:41:22,430 --> 00:41:27,660 The higher the electoral threshold that you have to jump over to get into parliament, the lower the seat share. 444 00:41:28,020 --> 00:41:31,589 And so so so there is evidence. There's weak evidence. 445 00:41:31,590 --> 00:41:36,390 If you look at voting and there's very strong evidence. If you look at how many deputies you get in the parliament, 446 00:41:37,020 --> 00:41:44,160 that the electoral systems matters and of the people who after the war looked at Peter and said, you know, there's a problem here. 447 00:41:44,370 --> 00:41:49,530 They weren't necessarily completely wrong. So that's that's just that's just making that point. 448 00:41:50,190 --> 00:41:54,390 All right. A final slide and then I'll move on to conclusion common. 449 00:41:54,410 --> 00:42:03,140 So up till now, what I've been doing is looking at the relationship between the fascist vote and the communist vote and GDP growth. 450 00:42:03,830 --> 00:42:09,620 And I found that it only mattered over the three year horizon for about a half of our countries. 451 00:42:09,620 --> 00:42:16,939 In our sample we have consumption growth. This is from Barrow and one and you might think that consumption is something 452 00:42:16,940 --> 00:42:19,940 that's getting a bit closer to what households are actually experiencing. 453 00:42:20,390 --> 00:42:27,200 All right. And actually what you find is that you get massively bigger and more statistically significant effects, 454 00:42:27,200 --> 00:42:31,819 even though we've chopped our sample size in half. So I think that there may be, as you know, 455 00:42:31,820 --> 00:42:39,950 work that one could do potentially thinking about what aspects of macroeconomic economic activity are particularly politically significant. 456 00:42:39,950 --> 00:42:44,989 But but then part of me thinks that actually the more historical questions that are really interesting ones. 457 00:42:44,990 --> 00:42:52,580 So in conclusion, do Germans general like German lessons generalise for the scariest some do nice fascist in general. 458 00:42:52,580 --> 00:42:56,569 Is it better in countries that have been the losing side in World War One had limited democratic 459 00:42:56,570 --> 00:43:02,750 capital possessed electoral systems conducive to entry and experience in serious slumps, 460 00:43:02,900 --> 00:43:08,960 which was our major result. So these are all aspects of the German experience that do seem to carry over. 461 00:43:09,200 --> 00:43:15,019 As I say, it's not clear me that you would want to generalise today, but if you really wanted to take these results seriously, 462 00:43:15,020 --> 00:43:21,350 then you might, I suppose, expect to see bigger extremist right wing votes in recent democracies that have very severe slumps. 463 00:43:21,800 --> 00:43:25,640 So, you know, make of that what you will. 464 00:43:25,670 --> 00:43:28,670 And there is also a longer run political context. 465 00:43:28,670 --> 00:43:36,350 So let me just finish with a few more slides. And I think this is the longer run political context, because what we have now is we have a sharp, 466 00:43:36,350 --> 00:43:45,409 sharp shock that's being superimposed upon a much longer and slower shift that is having major political implications. 467 00:43:45,410 --> 00:43:48,710 And that shift has basically got to do with the changing division of labour globally. 468 00:43:48,950 --> 00:43:53,540 So back in the sixties, right until quite recently, actually, 469 00:43:54,680 --> 00:44:00,350 southern exports to northern countries were 90% primary products, only 10% manufactured goods. 470 00:44:00,620 --> 00:44:04,880 And that 10%, 90% ratio, you could send that back into the late 19th century. 471 00:44:05,120 --> 00:44:11,030 It was thus for forever, seemingly. And then things change and they change very quickly. 472 00:44:11,330 --> 00:44:16,340 And what that means is that now the composition of their exports is such that what they're 473 00:44:16,340 --> 00:44:20,960 exporting to us is potentially competing with stuff that we are or were producing domestically. 474 00:44:21,470 --> 00:44:25,880 And what that means is there are now losers, potential losers from this trade as well as winners. 475 00:44:26,180 --> 00:44:31,730 All right. So if I were to go back into history to find an analogy for this, I wouldn't go to the interwar period. 476 00:44:31,910 --> 00:44:33,380 I go to the late 19th century. 477 00:44:33,530 --> 00:44:40,490 I coached the invasion of cheap grain from the Americas and the Ukraine that had serious damage to farmers and landowners and so on. 478 00:44:40,640 --> 00:44:45,140 Right around Europe and where those farming interests were sufficiently powerful, 479 00:44:45,440 --> 00:44:50,809 they got protection and the backlash was sufficiently powerful that it completely overturned 480 00:44:50,810 --> 00:44:56,570 the road towards liberalism that characterised the continent until the mid 1870s or so. 481 00:44:56,570 --> 00:45:02,150 And if you read one, Rakowski, whatever they tell you very nice stories about how all the major political cleavages in many 482 00:45:02,180 --> 00:45:07,520 European countries were driven by this basic issue of how we responded to the great invasion, 483 00:45:07,790 --> 00:45:12,829 who were the equivalents of the 19th century? Farmers and landowners. 484 00:45:12,830 --> 00:45:20,780 But it's blue collar workers. Right. And what we know from survey evidence is that in rich countries, in rich countries, 485 00:45:20,990 --> 00:45:27,080 being high skilled is associated with being less protectionist, more in favour and free trade. 486 00:45:27,440 --> 00:45:33,500 So the number here is is basically the correlation, if you like, between being high skilled and being protectionist. 487 00:45:34,220 --> 00:45:41,840 So negative number means being high skilled is associated with being not protectionist or being low skilled is associated with being protectionist. 488 00:45:42,140 --> 00:45:45,290 And so in rich countries, it's poor people who are protectionist. 489 00:45:45,470 --> 00:45:49,820 It's rich are people who are cosmopolitan and liberal and all that good stuff. 490 00:45:49,820 --> 00:45:55,280 And I mean, very suddenly the correlation actually reverses in poorer countries. 491 00:45:55,400 --> 00:45:59,570 Now, this has been, in a way, a completely latent political force, I would say. 492 00:45:59,570 --> 00:46:04,100 You don't see you know, you got who is the giant sucking sound man in America. 493 00:46:04,970 --> 00:46:09,980 Yeah, Ross Perot. But it went nowhere. Right. And Pat Buchanan tried this in America and it basically went nowhere. 494 00:46:10,250 --> 00:46:16,190 Right. So maybe you're not going to vote for Ross Perot to be president United States just because you're ticked off with the Mexicans for 495 00:46:16,430 --> 00:46:23,090 sucking your jobs out and maybe you're not going to vote for the whole national that's hope just because you don't like globalisation. 496 00:46:23,240 --> 00:46:29,660 But when people are given a chance to vote on trade in general and the market in general, as in France in 20 or five, 497 00:46:29,960 --> 00:46:35,060 you saw exactly the correlation emerging in the ballot boxes that we see in the opinion surveys. 498 00:46:35,060 --> 00:46:39,320 It was it was quite blatant. And in Ireland for the Lisbon Treaty it was the same thing. 499 00:46:39,320 --> 00:46:46,130 All the blue collar worker districts voted against Lisbon, all of the bourgeois areas voted in favour. 500 00:46:46,220 --> 00:46:51,370 So you've got this potentially important class divide on which now you have this shock. 501 00:46:51,810 --> 00:46:59,250 Being superimposed on that is, I think what gives me a bit of a pause for concerns in other papers over I'm just editorialising, 502 00:46:59,550 --> 00:47:03,960 but I worry about a situation where in this kind of context, 503 00:47:05,280 --> 00:47:08,999 if all of the centrist parties in Europe just throw up their hands and say, 504 00:47:09,000 --> 00:47:13,559 We'd love to be able to help you, we really would, but there's nothing that we can do on the trade front. 505 00:47:13,560 --> 00:47:15,150 We're constrained by the WTO. 506 00:47:15,300 --> 00:47:23,160 On the macroeconomic front, we're constrained by the ECB and the six pack of a two pack and every other bloody pack you can imagine. 507 00:47:23,160 --> 00:47:31,080 And you know, and anyway, the Germans won't wear any change. Then at a certain point people maybe will start to look elsewhere because as we know, 508 00:47:31,080 --> 00:47:34,830 in the interwar period, you know, there were constraints on macroeconomic policymaking. 509 00:47:35,250 --> 00:47:39,770 There were some people willing to think outside the box and those people eventually got the votes. 510 00:47:39,790 --> 00:47:44,550 So that's my editorialising. With no scientific basis for it whatsoever. 511 00:47:44,760 --> 00:47:48,480 All right. So so over you. 512 00:47:48,540 --> 00:47:54,090 All right. Thank you very much. Oh, thank you for Thanksgiving. 513 00:47:54,110 --> 00:47:57,569 I thought that was particularly interesting. And also for me, 514 00:47:57,570 --> 00:48:03,389 because I'm not very familiar with the quantitative research and because I've done a lot of my work 515 00:48:03,390 --> 00:48:08,070 and reading on extreme right to see all of this quantified that for me was really challenging. 516 00:48:08,070 --> 00:48:13,379 And one thing that I particularly appreciated was the richness of all these variables, 517 00:48:13,380 --> 00:48:19,590 because all of them are actually very, very relevant when you study the the notion of of extremism. 518 00:48:20,700 --> 00:48:29,879 And then I started thinking first of your title because, you know, in your in your chapter paper, do German lessons generalise. 519 00:48:29,880 --> 00:48:34,980 And this, I think, is such a relevant question also for the interwar period, 520 00:48:34,980 --> 00:48:44,520 because it's always a question whether Nazi Germany is comparable to any other case in European fascist, 521 00:48:44,520 --> 00:48:49,860 fascist trade or whatever, you know, other authoritarian regimes that exist. 522 00:48:50,310 --> 00:48:55,450 I think the only comparative would be which makes it not comparable to have the, 523 00:48:55,500 --> 00:49:00,480 you know, the Soviet Union communist regime compared with other East European countries. 524 00:49:00,990 --> 00:49:10,060 But what I what I think is correct about your approach is that you take the the example of Nazi Germany and you kind of, you know, 525 00:49:10,080 --> 00:49:15,840 deconstructed and you see all those because it was a a very rich configuration of 526 00:49:16,380 --> 00:49:21,960 circumstances that led to this kind of very extreme phenomenon in Nazi Germany. 527 00:49:22,170 --> 00:49:32,280 And that was how I read it. I mean, by looking at your variables and how you deconstructed the reasons why we have this rise of of extremism. 528 00:49:32,880 --> 00:49:37,740 And then the other thing is that whether it makes it relevant for today, 529 00:49:38,100 --> 00:49:46,320 because we do see many references as to, you know, what's happening today with with the economic crisis. 530 00:49:47,100 --> 00:49:50,309 Most of them are actually from an interesting perspective. 531 00:49:50,310 --> 00:49:55,380 I mean, we we were discussing the other day there was an article in the in the Financial Times 532 00:49:56,640 --> 00:50:01,380 that was quite provocative by saying whether Greece reminds us of Weimar Republic. 533 00:50:01,680 --> 00:50:08,730 And that is actually, you know, is there a truth to it or is it just for sinister reasons that we have this discussion? 534 00:50:09,180 --> 00:50:12,780 Well, I'm among those who believe that, you know, you can take lessons. 535 00:50:12,840 --> 00:50:15,149 I mean, you know, everything is unique, 536 00:50:15,150 --> 00:50:26,140 but you certainly can take lessons from from Nazi Germany and especially that connection between cumulative growth, the connection with that, 537 00:50:26,160 --> 00:50:32,969 with other variables that you mentioned, the democratic culture or the the loss, you know, 538 00:50:32,970 --> 00:50:37,049 the big loss from from World War One, although that's not relevant at all today. 539 00:50:37,050 --> 00:50:45,210 One could see a you know, a comparison between Greece feeling internationally defeated and that kind of Greek creates a certain resentment. 540 00:50:46,590 --> 00:50:56,459 One thing that I would like to to ask and talk about, I mean, voting behaviour is particularly important, but in the context of the interwar period, 541 00:50:56,460 --> 00:51:04,830 there's also this lead and in some cases it did lead to authoritarianism and dictatorship, but in others it did not. 542 00:51:05,760 --> 00:51:11,250 And Cappadocia has this very interesting article about the Belgium, 543 00:51:11,790 --> 00:51:19,740 Czechoslovakia and Finland who did not succumb to authoritarianism actually, and they preserved their democracy. 544 00:51:20,520 --> 00:51:24,809 And that for me is, you know, as a political scientist, it's also a very interesting question. 545 00:51:24,810 --> 00:51:27,030 So why why did this happen? 546 00:51:27,330 --> 00:51:35,740 But it also relates to what you said about democratic culture and democratic tradition, because one thing, of course, is to see democratic transition, 547 00:51:35,880 --> 00:51:42,540 tradition and culture with the longevity and having these countries, you know, 548 00:51:42,540 --> 00:51:47,130 preparing what Hobsbawm said, actually preparing better defences against authoritarians. 549 00:51:47,150 --> 00:51:51,240 But the other thing which is very important is also what are the short term kind of. 550 00:51:51,290 --> 00:51:54,650 Responses towards the rise of extremism. 551 00:51:54,890 --> 00:51:59,210 And that may not be in your analysis, but I think it's particularly important. 552 00:52:00,110 --> 00:52:05,810 And the reason why those some parties did better than others and actually some outcomes were 553 00:52:06,950 --> 00:52:13,040 towards dictatorship in that how the other parties responded to it was particularly significant. 554 00:52:14,300 --> 00:52:21,440 And this kind of strategy and that is also relevant today because there's a lot of discussion, especially now in view of the European elections, 555 00:52:21,680 --> 00:52:29,930 as to how the mainstream parties respond to the rise of extremism in in many European countries. 556 00:52:30,620 --> 00:52:34,699 And then it confuses us also with this argument of democratic culture, 557 00:52:34,700 --> 00:52:40,640 because we see the rise of of extremism in countries that have a very, very solid democratic culture. 558 00:52:41,030 --> 00:52:44,510 And, you know, you see it in the Netherlands or in France. 559 00:52:45,620 --> 00:52:53,029 And then it's interesting also that in other countries where you've got similar circumstances, like the south of Europe, you see it in Greece, 560 00:52:53,030 --> 00:52:59,510 but then you don't see it in Spain or you don't see it in Portugal or Ireland where you've got similar kinds of pressures. 561 00:52:59,870 --> 00:53:06,530 So that violence actually is particularly significant to look at. 562 00:53:06,980 --> 00:53:12,810 And then when you talk about the issue of growth and you mentioned, you know, 563 00:53:12,830 --> 00:53:17,299 this related to unemployment, but I was just wondering whether it's also the issue of inequality. 564 00:53:17,300 --> 00:53:28,040 I mean, inequality. So one should definitely see how a growth or at times of peak of a big recession or depression actually where, 565 00:53:28,040 --> 00:53:34,370 you know, the whole output is how the output is distributed, whether it loses for the winners. 566 00:53:34,700 --> 00:53:38,299 And that sense of inequality in particular, I think is very, 567 00:53:38,300 --> 00:53:50,060 very important because that leads the the voting the extremist vote is on the right because there is this there is actually a reality of inequality, 568 00:53:50,060 --> 00:53:56,450 but also the perception and imagined inequality in many ways that that leads. 569 00:53:56,450 --> 00:54:07,489 And that is also connected with the fact that there are many variables that cannot be measured quantitatively, especially in phenomena like this, 570 00:54:07,490 --> 00:54:16,010 because there's a lot of of emotion when voters actually go and vote for a party which is anti-system, 571 00:54:16,010 --> 00:54:21,620 which is out of the of the mainstream kind of space. 572 00:54:22,040 --> 00:54:26,479 And that is something that also affects voters at the time of voting. 573 00:54:26,480 --> 00:54:33,260 I mean, how they feel about that's why it may be it's important that the the issue of the losing of 574 00:54:33,260 --> 00:54:37,040 World War One is particularly important because this is how it plays into the minds of people. 575 00:54:37,430 --> 00:54:40,399 And to this day, I was I you know, 576 00:54:40,400 --> 00:54:50,090 I also believe that one of the most important factors in the German case and the German the rise of Nazism and also going into the war was that, 577 00:54:50,420 --> 00:54:55,340 you know, sense of injustice that, you know, that was from the so that they could never I mean, 578 00:54:55,340 --> 00:55:00,860 they had to to set the record straight with possibly with a victory in the Second World War. 579 00:55:02,180 --> 00:55:14,870 So this is you know, what I have to say. One final question that I wanted to ask is why don't you include Turkey in this in your group of countries? 580 00:55:14,870 --> 00:55:24,409 Because there are many, you know, comparable kind of variables there in terms of loss of one with one non-democratic tradition. 581 00:55:24,410 --> 00:55:32,840 And yet this is a country that didn't experience an extreme right and rising extremism, but a different kind of political reaction. 582 00:55:32,840 --> 00:55:33,050 Yeah.