1 00:00:05,430 --> 00:00:13,570 Thank you to the organisers for inviting me to this conference. It's a great pleasure to be here in particular because many decades ago, 2 00:00:13,570 --> 00:00:18,670 Patrick O'Brien got me and Jeremy started on this path of looking at big questions. 3 00:00:18,670 --> 00:00:21,690 And so, you know, it's really nice to be here. 4 00:00:21,690 --> 00:00:28,710 And Patrick may not agree with a lot of things I'm going to say today, but you know, it's still a good exercise. 5 00:00:28,710 --> 00:00:32,910 So, uh, so I'm taking over from Leandro, 6 00:00:32,910 --> 00:00:42,750 who's tried painted a very nice picture of what was driving the great divergence and kind of reinstating Palmer on this timeline. 7 00:00:42,750 --> 00:00:54,420 I'm going to show you a much more disaggregated picture of Asia and in particular, look at three different countries India, China and Japan. 8 00:00:54,420 --> 00:01:00,630 Just to recap where we are, you know where I'm coming from. 9 00:01:00,630 --> 00:01:13,500 The big question that was posed by Pomerantz and the California school is when did Europe begin to diverge from Asia as a high growth economy? 10 00:01:13,500 --> 00:01:25,080 And they argued that the conventional view is Eurocentric. If you look at the young C delta, it was as developed as England in the 18th century. 11 00:01:25,080 --> 00:01:30,810 A similar argument was put forward by President and Parthasarathy on India. 12 00:01:30,810 --> 00:01:34,530 Looking at the wages of weavers in South India, 13 00:01:34,530 --> 00:01:45,720 Parthasarathy argued that the same amount of grain could be purchased by weavers in India as the weavers in in England in 1750. 14 00:01:45,720 --> 00:01:59,820 So basically a similar type of view that the Great Divergence actually dates from colonisation, coal and other kind of advantageous things for Europe. 15 00:01:59,820 --> 00:02:08,850 So what? What was the basis of this comparison? The basis of this comparison was quantitative and qualitative, but mainly qualitative. 16 00:02:08,850 --> 00:02:15,810 There was some evidence of wages and the questions that Steve Bradbury and I started with is that how 17 00:02:15,810 --> 00:02:22,770 representative were these numbers are the quantitative estimates that we found in this literature. 18 00:02:22,770 --> 00:02:29,010 So we came up with this conclusion that we need to document an average value, 19 00:02:29,010 --> 00:02:39,390 which is much more widespread and has many more wages that that than was there in Pomeranz book or in Parthasarathy Societies analysis. 20 00:02:39,390 --> 00:02:43,410 So this is an exercise we did for a fairly long period of time, 21 00:02:43,410 --> 00:02:50,940 collecting hundreds and hundreds of wage and price data on India in particular and some in China. 22 00:02:50,940 --> 00:02:57,900 And the question was that, you know, what was the timeline of this decline? 23 00:02:57,900 --> 00:03:08,970 So I'm going to show you what we found, but the two indicators that we used were single wages and grain wages. 24 00:03:08,970 --> 00:03:13,590 So silver wages were comparable money wages across different countries, 25 00:03:13,590 --> 00:03:21,990 and the grain wages were simply by dividing the silver wage by the look at the price of the local grain. 26 00:03:21,990 --> 00:03:33,480 What we kind of argued in our 2006 paper was that high silver wages reflected high productivity in the tradable goods sector, 27 00:03:33,480 --> 00:03:39,120 whereas a high grain raise reflected high productivity in the agricultural sector. 28 00:03:39,120 --> 00:03:48,600 So we can find fairly low silver wages, but high grain wages, and that's the indicator of living standard the grain wage. 29 00:03:48,600 --> 00:03:53,340 So the difference between Europe and Asia was not that great to start with. 30 00:03:53,340 --> 00:03:58,710 In 6500, we also looked at the wages of skilled and unskilled workers, 31 00:03:58,710 --> 00:04:05,370 and we found much higher skill premium in in Asia compared to Europe and which kind of, 32 00:04:05,370 --> 00:04:14,460 you know, suggested that skill premium was lower in Europe because the the supply of skilled workers was much higher. 33 00:04:14,460 --> 00:04:20,490 So this is this is an illustration from what we found if you look at the highlighted column there. 34 00:04:20,490 --> 00:04:26,340 This shows the Indian grain wage as a proportion of the English grain wage. 35 00:04:26,340 --> 00:04:33,210 And clearly it was 83 percent of the English wage in in 6500. 36 00:04:33,210 --> 00:04:41,760 Ninety five percent in in the middle of the 17th century, but it began to fall soon after. 37 00:04:41,760 --> 00:04:46,500 And by the middle of the 18th century, it was only about 40 percent of the English wage. 38 00:04:46,500 --> 00:04:58,030 So we kind of argued in our paper that the timeline of the decline was indeed 7500 rather than 1800, as Pomerantz had argued. 39 00:04:58,030 --> 00:05:08,310 So you can see here that by the. Middle of the of the 19th century, it's less than 30 percent of the English wage. 40 00:05:08,310 --> 00:05:13,650 China looked pretty similar. I'm not. The Chinese data was worse than the Indian data, 41 00:05:13,650 --> 00:05:25,260 but it was a fall from 87 percent of the English wage to about 40 percent of the English feeds within a similar time span. 42 00:05:25,260 --> 00:05:29,310 So there were there are clearly some serious problems with the grain rates. 43 00:05:29,310 --> 00:05:33,450 One is that it's a very simplistic notion of living standards. 44 00:05:33,450 --> 00:05:43,860 People consume lots of different things. So how do we go from, you know, having just grain, a grain wage to a consumption basket? 45 00:05:43,860 --> 00:05:49,860 And this was done really nicely by a paper by Bob Allan and co-authors where they tried 46 00:05:49,860 --> 00:06:01,030 to construct a consumption basket using the basic grain and other consumption goods. 47 00:06:01,030 --> 00:06:07,750 But then there is yet another problem with the whole idea of using wages as an indicator of living standards, 48 00:06:07,750 --> 00:06:13,030 the most of the wage data that we used and others have used come from the urban sectors. 49 00:06:13,030 --> 00:06:20,620 And we know that China, India and Japan until even today are mainly agricultural. 50 00:06:20,620 --> 00:06:24,580 At the early 20th century were mainly agricultural economies. 51 00:06:24,580 --> 00:06:31,720 So, so, so using the wage data really actually distorts the the living standards. 52 00:06:31,720 --> 00:06:40,660 I'm going to come back to that in a minute. So just wanted to show you how Bob Allen and others constructed the wage basket. 53 00:06:40,660 --> 00:06:44,950 They used two types of wage baskets the bare bone based bread basket, 54 00:06:44,950 --> 00:06:55,960 which is the subsistence level basket and a and I'm a respectable basket which buys more than the subsistence level consumption. 55 00:06:55,960 --> 00:07:06,310 The welfare ratios are constructed by looking at the actual wage and seeing how many consumption baskets this wage would buy. 56 00:07:06,310 --> 00:07:15,520 And here the evidence was pretty unequivocal that the English worker was substantially above the barebones basket. 57 00:07:15,520 --> 00:07:21,970 The welfare ratio of the English workers was much higher compared to the Chinese or the Indian workers, 58 00:07:21,970 --> 00:07:27,070 who are mainly at the level of one bare-bones basket. 59 00:07:27,070 --> 00:07:30,790 So you can see the the lines here, the. 60 00:07:30,790 --> 00:07:36,520 So the top line, the the top that line is in is is London. 61 00:07:36,520 --> 00:07:42,790 The grey line is actually Oxford. And then you see the bunch of Asian countries bunched together, 62 00:07:42,790 --> 00:07:51,640 but they're clearly at a much lower level than the English living standards in the 18th century. 63 00:07:51,640 --> 00:07:56,890 So let me now kind of look at what has then followed, 64 00:07:56,890 --> 00:08:01,930 because this started this whole part of new research into the question of the great 65 00:08:01,930 --> 00:08:08,470 divergence and what happened in this context is really a move from wages to incomes. 66 00:08:08,470 --> 00:08:14,080 So how do we go from having just wages as an indicator of living standards 67 00:08:14,080 --> 00:08:20,320 to constructing some income estimates or GDP per capita estimates for India, 68 00:08:20,320 --> 00:08:25,090 China and Japan? So the question, of course, is is this wild speculation? 69 00:08:25,090 --> 00:08:31,630 We know that the data is not perfect as it is today or, you know, anywhere even close. 70 00:08:31,630 --> 00:08:39,490 But is it just wild speculation or can we claim some scientific basis of this construction? 71 00:08:39,490 --> 00:08:47,830 So there is Madison's data, which is which was widely used before the new data has become available. 72 00:08:47,830 --> 00:08:50,560 And I'm afraid that I'll show you the comparison. 73 00:08:50,560 --> 00:09:00,880 These were really based on guesswork, and they gave a picture of of stagnant Asian economies and a more growth, 74 00:09:00,880 --> 00:09:07,930 more buoyant European economies until the 1980s. 75 00:09:07,930 --> 00:09:15,850 The new research is more has much better data. It's much more careful in its construction of the GDP numbers. 76 00:09:15,850 --> 00:09:28,300 So for India, we have good data. After 1870, we have reasonably reliable data, patchy but reliable data for 15 ninety and later on. 77 00:09:28,300 --> 00:09:37,180 For Japan, we have pretty good data from 17 from seven hundred and thirty and four China from nine hundred and eighty. 78 00:09:37,180 --> 00:09:42,340 So this work owes a lot to Steve Bradbury, who got us all in in, you know, 79 00:09:42,340 --> 00:09:49,990 frenzy to get different data from different parts of the world and construct these estimates of GDP per capita. 80 00:09:49,990 --> 00:09:53,320 I just want to say how we did it. 81 00:09:53,320 --> 00:10:02,920 This is done on the basis of a demand model, which means that let's look at the population of a country at a particular point of time. 82 00:10:02,920 --> 00:10:12,190 How much? How much agricultural output would need to be produced to feed this population at a bare minimum. 83 00:10:12,190 --> 00:10:19,990 Similarly, how much industrial commodities would be needed, such as cloth to close this population at a given point in time? 84 00:10:19,990 --> 00:10:30,370 So the bare minimum estimates of GDP are constructed using population and the demand for the basic consumption goods. 85 00:10:30,370 --> 00:10:37,420 Then we use the baby wage data that we already have to build some flexibility into this estimate. 86 00:10:37,420 --> 00:10:46,360 Some countries were better off than the others, and therefore we do get some variation in terms of the consumption. 87 00:10:46,360 --> 00:10:53,230 So the demand model gave us a set of numbers on GDP per capita, but clearly we can't be sure about it, right? 88 00:10:53,230 --> 00:10:59,960 There is a lot of guesswork. There are some assumptions. So the main cross check is to find. 89 00:10:59,960 --> 00:11:06,560 Supply side estimates, so output data for some benchmark years, and that was hard, 90 00:11:06,560 --> 00:11:15,500 but I think we did it pretty carefully and we got some good supply side estimates based on output of agricultural products, 91 00:11:15,500 --> 00:11:25,790 industrial products, etc. for certain when benchmark years and try to see whether it matched with the demand side estimates that we constructed. 92 00:11:25,790 --> 00:11:28,220 And it did it. It matched pretty well. 93 00:11:28,220 --> 00:11:36,740 So I can I think I can say with some conviction that our GDP numbers are not just they are not really wild guess guesstimates, 94 00:11:36,740 --> 00:11:40,520 they are not absolutely accurate. 95 00:11:40,520 --> 00:11:51,950 I have to say that, but they are pretty close and we have GDP per capita converted into 1990 dollars, which Leandro has already shown up. 96 00:11:51,950 --> 00:11:57,950 So here you can see our estimates on the top. I don't know how well you can see that. 97 00:11:57,950 --> 00:12:06,890 You can see our estimates on the on the on the top of the table and at the bottom of the table and Madison's estimate. 98 00:12:06,890 --> 00:12:14,730 And if you focus on Japan, China and India, they don't change very much until much later, up to eighteen hundred, 99 00:12:14,730 --> 00:12:25,460 they kind of stay pretty stagnant, whereas our estimates on the top of the graph show a trend for China, Japan and India. 100 00:12:25,460 --> 00:12:33,380 India's data starts months later, but we still see a declining trend from 6500. 101 00:12:33,380 --> 00:12:37,430 So when did the when did the great divergence begin? 102 00:12:37,430 --> 00:12:44,330 And here I'm going to show you some different patterns for India, China and Japan. 103 00:12:44,330 --> 00:12:51,330 So let me start with India. This is based on my work with Steve Rothbury and Johan custodians. 104 00:12:51,330 --> 00:12:56,060 And this is a little bit of fancy graphics. 105 00:12:56,060 --> 00:13:06,770 So this is Indian GDP decline from 6500 to 1870, followed by a relative decline with Britain. 106 00:13:06,770 --> 00:13:11,060 The lower graph is much steeper than the graph above. 107 00:13:11,060 --> 00:13:22,130 So the relative decline of India is not only driven by the slight that the slow decline of Indian GDP per capita, 108 00:13:22,130 --> 00:13:28,340 but also the rise of British incomes or English incomes very, very fast. 109 00:13:28,340 --> 00:13:34,350 After 17 50. 110 00:13:34,350 --> 00:13:44,490 This is the data which Steve Bradbury, Agron and Lee constructed, this is should be at twenty eighteen for China. 111 00:13:44,490 --> 00:13:55,530 The Dark, but the dark line gives you the high estimates for China, and the broken line gives you the low estimates of China. 112 00:13:55,530 --> 00:14:04,320 So there are two estimates one based on data from the Yangzi Delta and the other is full of, you know, based on the rest of China. 113 00:14:04,320 --> 00:14:13,680 And you can see, as in the Indian case, these are not that different fora for a long time, up to 7500. 114 00:14:13,680 --> 00:14:16,860 They look pretty similar with the grey line for Europe. 115 00:14:16,860 --> 00:14:28,260 But then the divergence began from 17 40, where Europe moved ahead and China began to decline, both according to the high and the low estimates. 116 00:14:28,260 --> 00:14:37,680 So both India and China do show a decline in GDP per capita from 7500, you know, 117 00:14:37,680 --> 00:14:47,910 which is very different from what I'm going to show you for for Japan in in in the next slide. 118 00:14:47,910 --> 00:15:01,410 So why is it is Japan a special case? And here I think one, you know, there's a comment on Leandro estimates that once you add Japan to the story, 119 00:15:01,410 --> 00:15:12,360 we lose the decline that we see in 17:00 because Japan looks very different from from India and China for a long time. 120 00:15:12,360 --> 00:15:22,770 It looks similar to Britain in a particular way, which is that it doesn't have a sharp decline in GDP per capita at at any stage. 121 00:15:22,770 --> 00:15:33,090 So the big difference is it did not decline. The GDP for Japan did not decline in absolute terms, as it did for India and China. 122 00:15:33,090 --> 00:15:37,920 And this is this is my work with Steve Bradbury. 123 00:15:37,920 --> 00:15:43,080 Jean Pascal, No George Fuko and Masanori Takeshima. 124 00:15:43,080 --> 00:15:57,360 And you can see that as in the case of India and China, the Japanese GDP per capita declined from a fairly high share of the English GDP per capita, 125 00:15:57,360 --> 00:16:04,110 but it did decline, but it did not decline at all in absolute terms. 126 00:16:04,110 --> 00:16:12,720 So if you look at the highlighted section, a Japanese GDP was six hundred and sixty seven in 600, 127 00:16:12,720 --> 00:16:19,650 rising to six hundred and seventy five then eight hundred and twenty eight, nine or three and 2011. 128 00:16:19,650 --> 00:16:31,200 Whereas for China, it began to decline from 1100 to 1950 to seven hundred and twenty seven all the way down to 16:18 for India. 129 00:16:31,200 --> 00:16:37,860 Again, there is a decline from six eighty two to six thirty eight all the way to five twenty six. 130 00:16:37,860 --> 00:16:43,410 So the decline that we see in India and China never happened in Japan, 131 00:16:43,410 --> 00:16:50,970 which kind of puts Japan in a very different category amongst the Asian countries and therefore, 132 00:16:50,970 --> 00:17:00,090 you know, quite similar to to to Britain in in in in in in terms of the trend in GDP per capita. 133 00:17:00,090 --> 00:17:07,140 So my concluding comments are the estimated historical GDP provides us with 134 00:17:07,140 --> 00:17:14,790 some evidence that Europe did have higher living standards compared to Asia, 135 00:17:14,790 --> 00:17:21,240 not in 6500 us as much as by 18 7400. 136 00:17:21,240 --> 00:17:31,710 And we can think of the beginning of the Great Divergence as from 17 from the from the 17th century. 137 00:17:31,710 --> 00:17:44,190 The little divergence in Asia explains why Japan stands out as an economic unit, and it never declined in absolute terms, 138 00:17:44,190 --> 00:17:54,240 and it may explain why Japan was able to take off from 1868, and that's the topic which Karl Rove is going to pick up. 139 00:17:54,240 --> 00:18:03,920 I think in the next talk, OK, I'm going to stop there.