WEBVTT NOTE Weak Links in the Theory of Comparative Advantage: Missing Markets 1 00:00:00.000 --> 00:00:11.000 First of all, I'm very happy to be here, this is my first time here, I'm grateful to Professor Salerno for giving me this opportunity to present some of my work. 2 00:00:11.000 --> 00:00:27.000 This paper is actually an extension of one of my chapters from my recently completed PhD thesis that I did under the guidance of Professor Glenn Fox at the University of Guelph 3 00:00:27.000 --> 00:00:32.000 in Ontario, Canada. 4 00:00:32.880 --> 00:00:36.200 I'll first start with a little bit of context 5 00:00:36.200 --> 00:00:40.560 how I ended up dealing with this question. 6 00:00:40.560 --> 00:00:45.560 And sometime in 2004, the province of Saskatchewan, 7 00:00:47.920 --> 00:00:50.800 political authorities of the province of Saskatchewan 8 00:00:50.800 --> 00:00:54.920 demands a reassessment of the national allocation 9 00:00:54.920 --> 00:00:57.500 of, in that case, poultry production, 10 00:00:59.020 --> 00:01:01.620 according to the principle of comparative advantage. 11 00:01:01.620 --> 00:01:06.240 Now, you may ask, why would provincial political authorities 12 00:01:06.240 --> 00:01:07.940 ask for such a thing? 13 00:01:07.940 --> 00:01:12.940 Well, because that's the only way it can be done, 14 00:01:13.580 --> 00:01:17.780 because some of the agricultural commodities in Canada, 15 00:01:17.780 --> 00:01:22.740 like poultry, milk and eggs, 16 00:01:22.740 --> 00:01:26.900 are under a specific system called Supply Management. 17 00:01:26.900 --> 00:01:30.020 And what that basically means, it's a quota system. 18 00:01:30.020 --> 00:01:32.780 If you're a producer, you have a license 19 00:01:32.780 --> 00:01:36.500 to produce a certain quantity of output. 20 00:01:36.500 --> 00:01:38.560 If you want to expand your output, 21 00:01:38.560 --> 00:01:42.160 you have to obtain a license from another producer 22 00:01:42.160 --> 00:01:43.780 so you can buy it, but it goes 23 00:01:43.780 --> 00:01:48.080 through the provincial marketing boards. 24 00:01:48.080 --> 00:02:10.080 Another thing here is that you can buy from producers in your own province, but not from the producers in other provinces, so you can see that the total quantity of quota within a province can only increase if the national or in some cases provincial agencies allow that. 25 00:02:10.080 --> 00:02:14.080 That's why they go through the political means. 26 00:02:14.080 --> 00:02:16.760 Now, why comparative advantage? 27 00:02:16.760 --> 00:02:18.920 Why would they want this? 28 00:02:18.920 --> 00:02:23.100 Well, at the time of the forming of this policy 29 00:02:23.100 --> 00:02:27.380 in the 1970s, this term comparative advantage 30 00:02:27.380 --> 00:02:32.380 was put in the policy and it said in one of the articles 31 00:02:34.000 --> 00:02:38.760 in the law that the national supply management agencies 32 00:02:38.760 --> 00:02:41.040 will take into account the principle 33 00:02:41.040 --> 00:02:43.280 of comparative advantage when allocating 34 00:02:43.280 --> 00:02:45.640 production to different provinces. 35 00:02:47.060 --> 00:02:51.080 So that's where the question comes from. 36 00:02:51.080 --> 00:02:56.080 So the Canadian government officials ask economists 37 00:02:57.080 --> 00:03:01.360 how to allocate this production according to this principle. 38 00:03:01.360 --> 00:03:06.360 So economists start working and they started working in 2005 39 00:03:06.860 --> 00:03:11.860 and they're still working and they give some advice 40 00:03:11.860 --> 00:03:21.860 and it's generally contradictory advice and it's different statistics, different aggregate statistics, you should use this number or you should use that number. 41 00:03:21.860 --> 00:03:40.860 So when I kind of started with this literature I started wondering is there something in our theory that sort of points us in this general direction in looking for these statistics but doesn't really answer the question which one, so we have lots of them. 42 00:03:40.860 --> 00:03:47.660 So, then I went back and looked at the theory and I started with Adam Smith. 43 00:03:47.660 --> 00:03:56.260 There are some earlier references to division of labor and that would come from the principle 44 00:03:56.260 --> 00:04:04.300 of comparative advantage, but since Adam Smith is the most well known, I started there and 45 00:04:04.300 --> 00:04:07.620 onward in the history of economic thought. 46 00:04:07.620 --> 00:04:14.860 And what I found, and this is my thesis, is that while in the 1920s and the 1930s the 47 00:04:14.860 --> 00:04:20.180 focus of international trade theory, which is based on this principle of comparative 48 00:04:20.180 --> 00:04:28.140 advantage, theory was slightly sidetracked by attempts to cut a perceived tie to Ricardo's 49 00:04:28.140 --> 00:04:34.380 labor theory of value, but maintaining the emphasis on modeling using aggregate variables 50 00:04:34.380 --> 00:04:41.580 from 1940s onward has left out important aspects of the principle of comparative advantage 51 00:04:41.580 --> 00:04:47.620 and what are those aspects, heterogeneity of capital including human. 52 00:04:47.620 --> 00:04:53.980 Austrian theory maintains that heterogeneity of capital is one of the major factors that 53 00:04:53.980 --> 00:04:58.140 leads to the coordination and calculation problems. 54 00:04:58.140 --> 00:05:07.260 So in the calculation problem we have questions like how much of what we should produce and 55 00:05:07.260 --> 00:05:09.660 consume and in what way. 56 00:05:09.660 --> 00:05:15.140 In the coordination problems we have questions like who does what, where and when. 57 00:05:15.140 --> 00:05:23.580 And these questions, when we assume away heterogeneity of capital, become of tertiary importance. 58 00:05:23.580 --> 00:05:33.740 Now, the implication of this is that this has blurred the limits of economic policy 59 00:05:33.740 --> 00:05:38.860 with the meaning of who and how can implement this principle of comparative advantage. 60 00:05:38.860 --> 00:05:45.620 To the people in this room it's probably more clear to begin with why this would blur the 61 00:05:45.620 --> 00:05:47.860 limits of economic policy. 62 00:05:47.860 --> 00:05:59.540 But if we start with a model that begins with homogeneity, that may not be as obvious. 63 00:05:59.540 --> 00:06:08.860 Now I decided to start with Adam Smith and he was one of the first economists who points 64 00:06:08.860 --> 00:06:15.280 out the benefits of division of labor and he says, division of labor, however, so far 65 00:06:15.280 --> 00:06:18.680 are as it can be introduced, occasions in every art 66 00:06:18.680 --> 00:06:20.240 are a proportionable increase 67 00:06:20.240 --> 00:06:22.320 in the productive powers of labor. 68 00:06:22.320 --> 00:06:25.000 The separation of different trades and employments 69 00:06:25.000 --> 00:06:27.680 from one another seems to have taken place 70 00:06:27.680 --> 00:06:29.660 in consequence of this advantage. 71 00:06:29.660 --> 00:06:32.640 This separation too is generally called furthest 72 00:06:32.640 --> 00:06:35.360 in those countries which enjoy the highest degree 73 00:06:35.360 --> 00:06:37.000 of industry and improvement. 74 00:06:37.000 --> 00:06:40.880 So I wanna highlight two things that come out of this 75 00:06:40.880 --> 00:06:43.560 is that he talks about individuals and says, 76 00:06:43.560 --> 00:06:55.560 Individuals within a country specialize, and as a result, there's an increase in total productivity, so some benefits of individual specialization. 77 00:06:55.560 --> 00:07:05.560 Now, the next major figure that talks about division of labor at the national level is David Ricardo. 78 00:07:05.560 --> 00:07:20.560 And I highlighted one of his quotes that says because it would be advantageous to her, meaning Portugal, he's talking about Portugal and England deciding whether to trade or not. 79 00:07:20.560 --> 00:07:27.560 Rather to employ her capital in the production of wine for which she would obtain more cloth from England, 80 00:07:27.560 --> 00:07:31.000 then she could produce by diverting a portion of her capital 81 00:07:31.000 --> 00:07:35.320 from the cultivation of vines to the manufacture of cloth. 82 00:07:35.320 --> 00:07:38.240 And I want to highlight a few things here. 83 00:07:38.240 --> 00:07:42.280 First, he uses a metaphor, an anthropomorphism, 84 00:07:42.280 --> 00:07:45.640 where countries sort of look like they decide 85 00:07:45.640 --> 00:07:50.440 to specialize in exchange as single entities. 86 00:07:50.440 --> 00:07:55.720 But the language that he uses had a specific purpose, 87 00:07:55.720 --> 00:08:03.080 to provide a rational argument to political authorities of that time in favor of opening borders, 88 00:08:03.080 --> 00:08:07.720 not necessarily a watertight description of reality. 89 00:08:07.720 --> 00:08:22.200 Now, it seems that this literary metaphor, it seems that it has initiated a process of translating this broad literary metaphor 90 00:08:22.200 --> 00:08:27.160 into the Language of Exact Aggregate Mathematical Variables. 91 00:08:27.160 --> 00:08:36.560 And the first major figure in this process is Bertil Olin. 92 00:08:36.560 --> 00:08:42.520 Actually, he uses some of the earlier findings of Heckscher, 93 00:08:42.520 --> 00:08:49.200 who was working on the factor endowment theory of comparative advantage. 94 00:08:49.200 --> 00:09:19.200 But Bertil Olin also thought that Ricardo's theory of comparative advantage was based on wrong theory of value and the labor theory of value and he says the classical theory of value has long ceased to be a satisfactory solution to the problems of price determination, nevertheless it continues to form the basis for the theory of international trade even in contemporary exposition, even in contemporary exposition. 95 00:09:19.200 --> 00:09:21.400 This is clearly an anomaly. 96 00:09:21.400 --> 00:09:22.520 So he does two things. 97 00:09:22.520 --> 00:09:29.040 He attributes Ricardo's labor theory of value to the theory of comparative advantage. 98 00:09:29.040 --> 00:09:37.000 And in his subsequent work, he thinks he can solve this by replacing Ricardo's fixed proportion 99 00:09:37.000 --> 00:09:47.380 technology with variable proportion technology, but maintains Ricardo's country person metaphor. 100 00:09:47.380 --> 00:09:51.380 has a reasonable abstraction in, but now a description of reality. 101 00:09:51.380 --> 00:09:58.620 His attempt is not just to give a reasonable argument, but to provide a description of 102 00:09:58.620 --> 00:10:01.580 reality that is reasonable. 103 00:10:01.580 --> 00:10:10.140 Now Gottfried Haberler, one of the Austrians, thought that the theory of competitive adage 104 00:10:10.140 --> 00:10:15.180 is not necessarily logically based on the labor theory of value, but it may appear to 105 00:10:15.180 --> 00:10:18.660 to be, and he was more correct. 106 00:10:18.660 --> 00:10:24.700 He says the theory of comparative costs, that's the synonym that some theorists use, was developed 107 00:10:24.700 --> 00:10:29.300 on the basis of labor theory of value and all theorists who have accepted have indeed 108 00:10:29.300 --> 00:10:33.840 assumed that it rests also logically on the labor theory of value. 109 00:10:33.840 --> 00:10:39.940 So he seeks ways to separate the theory of comparative advantage from the perceived link 110 00:10:39.940 --> 00:10:43.180 to the labor theory of value. 111 00:10:43.180 --> 00:10:59.180 He said, fortunately, however, it is possible to reformulate theory in such a way that its analytical value and all conclusions drawn from it are preserved, rendering it at the same time entirely independent of the labor theory of value. 112 00:10:59.180 --> 00:11:04.180 This may be most readily be shown in a diagrammatic representation of our theorem. 113 00:11:04.180 --> 00:11:10.180 And this is the introduction of the production possibilities frontier in economics. 114 00:11:10.180 --> 00:11:21.180 Now, Mises didn't use diagrams, but he clarified that the issue is not actually about value, but it's about productivity. 115 00:11:21.180 --> 00:11:31.180 Mises said, the law of comparative cost is as independent of the classical theory of value as is the law of returns, which its reasoning resembles. 116 00:11:31.180 --> 00:11:37.180 In both cases, we can content ourselves with comparing only physical input and physical output. 117 00:11:37.180 --> 00:12:02.180 He clarifies the nature of the concept and says that not everything in classical economics was based on the labor theory of value, but adds an important distinction, to whom do we primarily attribute productivity differences? 118 00:12:02.180 --> 00:12:08.180 He says individuals associate with each other to take advantage of their productivity differences. 119 00:12:08.180 --> 00:12:18.180 And he says Ricardo was fully aware of the fact that his law of comparative cost, which he expounded mainly in order to deal with a special problem of international trade, 120 00:12:18.180 --> 00:12:27.180 is a particular instance of the more general case of the law of association that's in the line below. 121 00:12:27.180 --> 00:12:30.220 Now Rothbard, Murray Rothbard maintains 122 00:12:30.220 --> 00:12:34.700 that individuals specialize and exchange goods 123 00:12:34.700 --> 00:12:38.940 to take advantage of their productivity differences 124 00:12:38.940 --> 00:12:40.700 to their mutual benefit. 125 00:12:40.700 --> 00:12:45.380 So he maintains this individual focused approach. 126 00:12:45.380 --> 00:12:49.580 But this is not the case in the traditional 127 00:12:49.580 --> 00:12:51.900 and currently widely used, 128 00:12:51.900 --> 00:12:54.300 actually all in Samuelson-types models. 129 00:12:54.300 --> 00:12:56.940 We have two production possibility frontiers 130 00:12:56.940 --> 00:13:01.460 and they are actually all in variable proportions, 131 00:13:01.460 --> 00:13:04.460 technology put in a general equilibrium framework 132 00:13:05.980 --> 00:13:08.580 and expressed using Haberler's production 133 00:13:08.580 --> 00:13:12.860 possibility frontiers, so these are country variables, 134 00:13:12.860 --> 00:13:14.960 country production possibility frontiers 135 00:13:14.960 --> 00:13:19.740 and what Samuelson does, he still maintains Ricardo's 136 00:13:19.740 --> 00:13:23.220 country person metaphor by homogeneous inputs, 137 00:13:23.220 --> 00:13:26.420 capital and labor within an economy of each country, 138 00:13:26.420 --> 00:13:31.540 Single National Production Function and one optimization problem with all starting variables 139 00:13:31.540 --> 00:13:33.900 known and given. 140 00:13:33.900 --> 00:13:41.900 Now there were some model extensions and modifications making it more complex, but in the essence 141 00:13:41.900 --> 00:13:46.420 the core country person metaphor was preserved. 142 00:13:46.420 --> 00:13:51.300 Even when there is some firm heterogeneity within an economy, it is not necessary for 143 00:13:51.300 --> 00:13:53.020 the existence of prices. 144 00:13:53.020 --> 00:13:57.620 You could assume a way heterogeneity and the model structure the way it's mechanically 145 00:13:57.620 --> 00:14:04.420 built would still give you prices, market prices presumably. 146 00:14:04.420 --> 00:14:09.860 Heterogeneity is a secondary wrinkle in a general model, not a general feature of actual 147 00:14:09.860 --> 00:14:10.860 markets. 148 00:14:10.860 --> 00:14:12.980 So the emphasis is different. 149 00:14:12.980 --> 00:14:18.380 So why is not ignoring heterogeneity important? 150 00:14:18.380 --> 00:14:22.740 A number of speakers today and in previous days talked about that. 151 00:14:22.740 --> 00:14:26.300 In this instance of comparative advantage, why is it important? 152 00:14:26.300 --> 00:14:29.780 Well, first, it provides a logical basis for markets. 153 00:14:29.780 --> 00:14:35.560 For example, we would maybe simplify a little bit, but we could ask a question. 154 00:14:35.560 --> 00:14:40.700 Why do individuals continually specialize in exchange if everything is perfectly homogeneous 155 00:14:40.700 --> 00:14:47.380 and divisible everywhere as it is implied by the continuous functions? 156 00:14:47.380 --> 00:14:51.260 Second, if everything is the same everywhere, 157 00:14:51.260 --> 00:14:53.260 is there a knowledge problem? 158 00:14:53.260 --> 00:14:56.940 Any need for market prices as a source of information 159 00:14:56.940 --> 00:14:58.980 or a coordinating mechanism? 160 00:14:58.980 --> 00:15:03.420 And second, this comes from the actual comp of people 161 00:15:03.420 --> 00:15:06.580 who were trying to figure out whether you can 162 00:15:06.580 --> 00:15:11.540 mathematically aggregate heterogeneous production functions 163 00:15:11.540 --> 00:15:15.140 into an aggregate production function. 164 00:15:15.140 --> 00:15:17.820 and the conclusion was by Philippe and Fischer 165 00:15:17.820 --> 00:15:20.780 and Fischer did a lot of work in the 70s and 80s 166 00:15:22.260 --> 00:15:27.260 after the Cambridge Capital Theory controversy 167 00:15:29.420 --> 00:15:31.620 to see whether this actually can be done 168 00:15:31.620 --> 00:15:33.460 and he says, concludes, 169 00:15:33.460 --> 00:15:37.580 it is not theoretically sound to assume that. 170 00:15:37.580 --> 00:15:42.120 Now, what are the messages that we get from our models? 171 00:15:42.120 --> 00:15:48.600 Well, if we omit individual heterogeneity, the message that our model tells us is how 172 00:15:48.600 --> 00:15:54.120 do economists measure comparative advantage in a world where all the initial variables 173 00:15:54.120 --> 00:15:55.880 are given and knowable. 174 00:15:55.880 --> 00:16:00.840 And this is more predominant message that we get. 175 00:16:00.840 --> 00:16:07.040 The other message is if we don't omit individual heterogeneity, how do individuals determine 176 00:16:07.040 --> 00:16:11.880 their own comparative advantage relative to others in the world where most variables are 177 00:16:11.880 --> 00:16:15.880 that are not given and not knowable in the actual world. 178 00:16:15.880 --> 00:16:18.880 Implications for economic research. 179 00:16:18.880 --> 00:16:21.880 In the first case, when heterogeneity is omitted, 180 00:16:21.880 --> 00:16:24.880 we look for aggregate statistics with little attention 181 00:16:24.880 --> 00:16:27.880 on how markets give rise to these statistics. 182 00:16:27.880 --> 00:16:34.880 For example, try calculating capital-to-labour ratio in a country X without market prices. 183 00:16:34.880 --> 00:16:37.880 When individual heterogeneity is not omitted, 184 00:16:37.880 --> 00:16:40.880 we look for institutions that are best suited 185 00:16:40.880 --> 00:16:48.120 Discovering Individuals Comparative Advantage in a World of Uncertainty and Ignorance. 186 00:16:48.120 --> 00:16:54.080 Some conclusions from this, so the major shift in the theory of comparative advantage was 187 00:16:54.080 --> 00:17:01.480 from the Ricardian to the actual Olin-Samuelson aggregate models, so from fixed proportion 188 00:17:01.480 --> 00:17:08.680 technology to a variable proportion technology, however maintaining the country person metaphor 189 00:17:08.680 --> 00:17:13.240 has weakened the logical basis for the existence of markets in these models 190 00:17:13.320 --> 00:17:17.720 and extended the perceived limits of central economic planning 191 00:17:17.800 --> 00:17:21.360 beyond what may not be realistically feasible.