International Trade, 31E00500

Transcription

1 International Trade, 31E00500 Lecture 2 Pertti Aalto University School of Business

2 Trade and Technology 1 The oldest modern theory is based on the assumption that production technologies in the form of (labor) productivity differ across countries, the theory is known (for good reasons) as the Ricardian theory of international trade. Despite being the oldest theory it has experienced a revival as international and interfirm differences in productivity have become transparent with new data. The sources of productivity differentials are usually left unexplained, but there many possible explanations: slow international dissemination of information on new technologies (also due e.g. to institution like patent protection), economic and social history (learning).

3 Trade and Technology 2 One striking and tragic example of the impacts of history on economic and social development is analyzed in Grosfeld, Rodnyansky and Zhuravskaya Persistent Antimarket Culture: A Legacy of the Pale of Settlement after the Holocaust, American Economic Journal: Economic Policy 2013, the working paper version is here sol3/papers.cfm?abstract_id= In the Russian Empire Jews were forced to live in the Pale Settlement (which included e.g. Byelorussia, Ukraine), those living outside of it were forced to move in the Pale. Grosfeld et. al. argue that this created among non-jews in Pale antimarket and anti-entrepreneurship attitudes. They test whether these have persisted to this day, and it turns they do, although there was a dramatic change in the composition of population in the Pale area due to the Holocaust:

4 SugarSync/Luennot/International Trade 2017/Lecture 2/IntTrL1F1.pdf History and attitudes

5 Trade and Technology 3 Let us start with the simplest model: two countries (EU, K), two goods (f,c), one factor of production, labor. Given this and assuming that markets (exchange) are perfectly competitive the production technologies are fully characterized by labor productivities, or equivalently, by unit labor requirements. Assume labor to be perfectly mobile between sectors. Let af EU denote the unit labor requirement (e.g. in terms hours worked) in EU for production of f = the hours of work required to produce one unit of f. Obviously one hour of work produces then 1 units of good f af EU in EU = EU labor productivity in f.

6 Trade and Technology 4 The existence of international productivity differences means that unit labor requirements cannot be the same in all countries in all industries. What kind of differences matter for trade? The sine qua non for international trade trade are international differences in relative labor requirements/labor productivities, the comparative advantage. This is because they determine the production possibilities and thereby the trading opportunities as they determine the domestic opportunity costs of expanding production of one good in terms of the reduction of the other. If the opportunity costs differ internationally mutual trade is beneficial for both.

7 Trade and Technology 5 To understand this, assume that the EU labor supply is L EU. If all of it is allocated to production of c production of c is LEU. ac EU If the production of c is cut by one unit, the amount of labor released for production is a EU c. With this amount of labor aeu c af EU units of f can be produced in EU. This is the opportunity cost of expanding the production of f in EU. At the same time it also equals the (minus) of the slope of the production possibility frontier, see the next figure. Obviously, if all labor eventually is in production of the other good, its production is LEU a EU f Why can we assume that production is on the production possibility frontier, not inside it?

8 SugarSync/Luennot/International Trade 2017/Lecture 2/IntTrL1F2.pdf f Production possibilities slope = - c

9 Trade and Technology 6 Intuitively, in perfectly competitive markets wages and prices adjust to equilibrate all markets. Were labor supply to exceed labor demand wage rate would equal 0. More formally. it can be shown that with perfect competition factors of production are allocated to sectors so that the value of production is maximized for given prices. Let the prices of the goods be p c, p f and let L be the labor input allocated to production of c. Then the value of production is Y = Y = p c [ p c L a EU c L a EU c p f + p f LEU L L a EU f af EU ] + p f LEU a EU f (1)

10 Trade and Technology 7 So, what allocation of labor, what value of L, maximizes the value of production Y? Look at (1). If p c L a EU c p f L a EU f > 0 then clearly all labor should be allocated to production of c, L = L EU. But this is equivalent to having p c p f > aeu c af EU The opportunity cost of expanding production of good c is smaller than the opportunity cost the markets give. (2)

11 Trade and Technology 8 You can convince yourself that when pc p f be allocated to production of f. What happens if pc p f = aeu c af EU? < aeu c af EU all labor will Obviously, income is maximized by any allocation of labor. In this case allocation of labor will be determined by demand. At the same time prices would be determined by technology,the previous formula tells that relative prices pc p f determined by the unit labor requirements/labor productivities. This is the labor theory of value developed and used by classical economists. are

12 Trade and Technology 9 The labor theory of value holds if a) all goods are always demanded (as then complete specialization = allocation of all labor to one sector only) would not be an equilibrium, there would be excess supply of one commodity and excess demand of the other), and b) the economy is closed. Those interested in history of economic thought would benefit from this account of the labor theory of value by Branko Milanovic: labor-theory-of-value-primer.html. It is more about Karl Marx s version of it, who along with Ricardo realized that the classical theory does not hold exactly, in Ricardo s case because of the possibility to international trade, in Marx case because of the heterogeneity of labor.

13 Trade and Technology 10 How can we understand the results on labor allocation from the perspective of individual agents, through their actions and choices? In this simple model with just one factor of production, labor, resources would with given prices obviously flow to the sector with highest return = highest wage (people are self-employed). As with perfect competition no pure profits can exist prices have to equal unit costs. We have already implicitly assumed that labor is fully mobile across sectors. With this, wages have to be equal in both sectors if both goods are produced, otherwise the economy will be completely specialized.

14 Trade and Technology 11 The unit costs of production = costs of producing one unit of the commodity are a EU c w c, a EU f w f (3) Here w c (w f ) = the wage that can be earned in producing good c (f ). Since unit costs must equal the price the wage rates are p c = ac EU w c (4) p f = af EU w f (5) w c = pc a EU c w f = pf af EU (6) (7)

15 Trade and Technology 12 People switch producing in sector c if the returns there are higher than in sector f, w c > w f. This happens, using (6) and (7), when p c p f > aeu c af EU This is exactly the same condition as derived above (see (2)) for the complete specialization of production on c. You can go yourself through the other cases. Thus, with both goods produced the labor theory of value holds also from the point of view of maximizing the returns to labor. (8)

16 Trade and Technology 13 The final item to notice before going to international trade case is that, as shown by equation (2) and our discussion after this, the resource allocation depends only on relative prices, p c p f, and not on absolute price levels, the exact values of p c and p f, but only their ratios. This is a reflection of the fact that we do not have any money in the model. But even with money we would reach the same result by assuming that there is not money illusion in the economy (or alternatively, money neutrality). Another thing worth noticing is that the result on the competitive economy maximizing the value of production for given prices is used actually in large computable general equilibrium models in solving the model. These models are extensively used in trade policy analysis.

17 Trade and Technology 14 We are now ready to introduce the second country K. As in EU resource allocation depends on the ratio of market prices relative to unit labor requirements there. Thus, if pc p f > ak c a K f, K produces only c, if pc p f = ak c a K f < ak c a K f only good f and if pc p f determined by demand, it can produce both goods. it produces its resource allocation will be Notice that I have assumed the goods prices to be equal in both countries. This is the standard assumption, I come later to issues which arise if there are trading costs (trade policies are among them).

18 Trade and Technology 15 Now we know that if ak c a K f = aeu c af EU, there will not be any trade between the two countries (or more precisely, trade would be indeterminate, the pattern of trade, which would exporting what, would be indeterminate, and it would not have any impact). Why? Assume now that ak c = aeu af K c af EU Let us define comparative advantage, CA, as follows: EU has a comparative advantage in producing good c if and only if aeu c < ak a f EU c a f K Obviously K has comparative advantage in producing f.

19 Trade and Technology 16 The other way to put this is to say that country has a CA in producing c if its relative unit costs of producing it are lower than in the other country. CA implies a precise pattern of trade: If there is any trade between the countries the country will be exporting the good in the production of which it has comparative advantage. ak c a K f Take the case in which aeu c pc. af EU p f The prices here are now the world market prices. By the analysis above, in this case at least one of the countries would be completely specialized. EU would produce always c, K always f. Thus, if in both countries there is some demand for the good produced in the other, EU would be exporting c and importing f, K exporting f and importing c.

20 Trade and Technology 17 With all other price ratios both countries would be specialized in producing the same good, and there would not be any trade. But this cannot be an equilibrium if there is demand for both of the goods in the global economy at all prices. So we assume that holds. a EU c a EU f pc p f ak c a K f Using (9) one can show that there are gains from international trade. (9)

21 Trade and Technology 18 One way of doing this is to look at real wages (wages deflated by prices), if they increase then welfare increases. Let us look at EU. We know then that EU produces (and exports) good c. In this case ac EU w EU = p c which implies that the real wage in terms of good c is w EU p = 1. c ac EU This is the same as without trade, so the purchasing power of domestic wages over the export good does not change. So, look at the other real wage: w EU p f = w EU /p c p f /p c w EU /p c af EU /ac EU = 1 a EU f

22 Trade and Technology 19 The inequality comes from (9). Now 1 is the real wage prevailing in EU without trade. Thus, af EU this real wage increases with trade and does strictly so when EU is completely specialized (strict inequality holds in (9)). What happens is that with trade based on comparative advantage citizens possible choices are not bound by the production possibility set but by the market value of production. The EU-citizens budget constraint (assume EU is completely specialized) is p c C EU c + p f C EU f C EU f p c Q EU c = p c LEU a EU c pc L EU p f ac EU pc p f C c EU

23 Trade and Technology 20 Here the C s and Q s are consumption and production volumes. Next figure shows the difference between production possibilities and consumption volumes:

24 SugarSync/Luennot/International Trade 2017/Lecture 2/IntTrL1F3.pdf Trade and welfare Consumption possibilities slope = slope = - Production possib.

25 Trade and Technology 21 But if in both countries have larger consumption opportunities than production opportunities, then international trade must increase global production of both goods. And this happens because of gains from specialization. The logic: Start by assuming that all labor in the world is allocated to the production of c. Assume then that the production of c is reduced by one unit. Will this reduction take place in EU or in K? Here it must take place in K as it has comparative advantage in producing f, it is relatively more productive in production of f. This continues until all labor in K has been transferred to production of f. Further increases in production of f must come at the expense of production in EU. Next Figure repeats this and shows the resulting global production possibilities.

26 SugarSync/Luennot/International Trade 2017/Lecture 2/IntTrL1F4.pdf International division of labor and production possibilities slope = - slope = -

27 Trade and Technology 22 With trade changes in global structures of production can always be made by using technology with the smallest opportunity costs even when the most efficient technologies cannot be used in individual countries. This is equivalent to saying that trade increases global productivity even with the given technologies. These gains are one source of growth impacts trade creates. Empirics of trade and growth? The results give quite a rosy picture from gains of trade. But in a sense this is a lousy model of trade. Why? We ll get back to this.

28 Trade and Technology 23 Up until now we have not specified the demand side at all except for assuming demand structure where all goods are demanded and assuming that consumer welfare is increasing in her real income (real wage). The backside of this is that we have not been able to pin down the equilibrium prices but in the closed economies. By augmenting the model with an explicit demand side it becomes a rudimentary example of a general equilibrium model of the world economy. These type of models are among the main tools used in analyzing the impacts of changes in trade (and other) policies. The are usually hard to solve analytically, that is why they are called C(omputable)G(eneral)E(quilibrium) models.

29 Trade and Technology 24 An particularly easy way to model the demand side is to assume that spending on each good is a constant proportion of income. This type of structure comes from assuming Cobb-Douglas-preferences U = C α c C 1 α f, 0 < α < 1 (10) The consumer maximizes this by choosing consumption levels from her budget set p c C c + p f C f Y (11)

30 Trade and Technology 25 Here Y is the nominal income. The optimal choices are C c = α Y p c (12) C f = (1 α) Y p f (13) Assuming both countries are completely specialized Y EU = p c LEU and Y K = p f LK. ac EU af K

31 Trade and Technology 26 For EU the demands are C EU c = α LEU ac EU, Cf EU = (1 α) pc L EU p f ac EU (14) Analogous expressions hold for K. Notice that demands depend, in both countries, only on the relative price pc p f. Thus, there is only one variable for which the model can be solved, the relative price. But there are two equilibrium conditions, one for each good, is not this a problem, as we need then two variables (prices) to get a solution?

32 Trade and Technology 27 No, the equilibrium conditions are not independent, the Walras Law holds: If one of the markets is in equilibrium then the other market is also in equilibrium. The budget constraints are p c C EU c + p f C EU f = p c LEU a EU c p c C K c + p f C K f = p f LK a K f Sum these side by side and reorganize the expression to get the equation [ ] [ ] p c Cc EU + Cc K LEU L ac EU = p f K af K Cf EU Cf K (15)

33 Trade and Technology 28 The equilibrium condition for the market of c is C EU c + C K c = LEU a EU c If this holds then also, by (15), also the market for f is in equilibrium: C EU f + C K f = LK a K f Thus, we have only one equilibrium condition and it can be used to solve for the equilibrium price-ratio.

34 Trade and Technology 29 Since only relative prices can be solved, a standard practice is to choose some good (or basket of goods, or some nominal aggregate) as the numeraire in terms of which all other prices are expressed. In our case one could e.g. choose good f as the numeraire and set p f = 1. All other prices and incomes would then be measured in terms of good f. The standard practice in deflating GDP s or aggregate expenditures or wages by CPI or GDP deflator are examples of this. We are now ready to solve the model just built.

35 Trade and Technology 30 The equilibrium condition for the market of good c is α LEU a EU c + α pf p c L K a K f = LEU a EU c (16) This can be solved for pf p c : giving p f p c = 1 α α p c p f = α 1 α L EU a EU c L K a K f L K a K f L EU a EU c (17) (18)

36 Trade and Technology 31 Recall that we have assumed, in calculating this equilibrium, that both countries are completely specialized, that a EU c a EU f < pc p f < ak c a K f (19) holds. We must then look the conditions under which this is true: a EU c a EU f < α 1 α L K a K f L EU a EU c < ak c a K f Thus: differences in expenditure shares cannot be two different, countries size differences (in terms of labor force) cannot be too large, and countries productivity differentials cannot be too large. (20)

37 Trade and Technology 32 The last point is important: This is the first time in this lecture when absolute advantage, the absolute level of productivity is mentioned. Before Ricardo (and unfortunately also after him) absolute advantage was thought to be crucial for trade, Ricardo pointed out that also a country with absolute disadvantage (with lowest productivity in all activities) can successfully enter in beneficial trade with other countries. But: absolute (dis)advantage can play a role in global structural change.

38 Technological Change and Global Restructuring of Production 1 One drawback of the basic Ricardian model is that it is completely static. One possible element of dynamics is to allow (some of the) productivities/unit labor requirements to change. One way to do this is to assume productivity to change through learning-by-doing = experience in producing, discovered during the 2nd WW and first analyzed theoretically by Kenneth Arrow (1962). The idea is that practice leads to routines which make labor more efficient.

39 Technological Change and Global Restructuring of Production 2 Very recent evidence is provided in Haggag, McManus and Paci Learning by Driving: Productivity Improvements by New York Taxi Drivers, American Economic Journal: Applied Economics, vol. 9, No. 1, 2017, New, inexperienced taxi drivers learn by accumulating neighborhood specific experience and improve their productivity. The following figure, from Haggag et. al. shows the taxi-divers learning from accumulated driving :

40 SugarSync/Luennot/International Trade 2017/Lecture 2/InTrL2F1.pdf Productivity of taxi-drivers

41 Technological Change and Global Restructuring of Production 3 This learning curve is typical: Experience increases productivity initially very much, the effect becomes much smaller with further experience and may wither away altogether. We assume this to be the case in the analysis to follow (based on Brezis et. al). Let us now use this mechanism in analyzing structural change in the model developed above. Assume that the experience relevant for LBD is accumulated production. To make things easier (e.g. consistent with perfect competition and allowing the neglect of strategic considerations) assume that learning is industry-wide and completely external to the agents.

42 Technological Change and Global Restructuring of Production 4 Assume that LBD is faster in c-production than in f-production, and simply assume that there is no LBD in f-production, and also assume that af K = af EU = 1. Assume also that countries are of same size: L K = L EU. Assume that at the beginning of this narrative both countries are completely specialized with EU producing c, K producing f. ( In this case the unit labor requirement ac EU Q EU), Q = accumulated production, declines with new production, a c EU < 0. Assuming that there are diminishing returns to LBD is equivalent to assuming that a EU c > 0

43 Technological Change and Global Restructuring of Production 5 These imply that the labor productivity 1 experience but at diminishing rate. ac EU grows with Indeed, assume that we are initially in the situation where the productivity effects of further experience in EU are small. Complete specialization implies, using assumptions just made and remembering the wage equations that (using (18)) that w EU w K = α 1 α (21) We take EU to be an industrial country and K a developing country, and assume α > 1 2, EU has higher wages than K.

44 Technological Change and Global Restructuring of Production 6 Assume now that a new c-production technology a c emerges, vaialble for use in both countries. Assume it to more productive at a given level of LBD than the old technology, a c ( Q) < ac EU ( Q). ( What happens, if a c (0) > ac EU Q EU)?

45 Technological Change and Global Restructuring of Production 6 Assume now that a new c-production technology a c emerges, vaialble for use in both countries. Assume it to more productive at a given level of LBD than the old technology, a c ( Q) < ac EU ( Q). ( What happens, if a c (0) > ac EU Q EU)? Obviously EU will not adopt the new technology. This happens if LBD has improved productivity with the old technology very much. What else?

46 Technological Change and Global Restructuring of Production 7 Now, it is profitable for K to adopt the new technology if, at current wages the unit cost of producing c in K with the new technology is lower than the unit cost of production in EU a c (0) w K < a EU c ( ( Q EU) w EU aeu c Q EU) a c (0) > 1 α α (22) In this situation K would initially start to produce both of the goods while EU would continue only in c-production? Why does not K specialize completely in c-production?

47 Technological Change and Global Restructuring of Production 8 Were this to happen EU would have to produce both of the goods. In this case the relative price of c would be higher than in the original situation as the new technology would be less productive than the old. This would make also EU to concentrate in production of c, contradiction. Check! (Hint: Which country has CA in c-production?). Assuming that the returns to LBD in EU very small while returns in K to LBD are large, the global economy would after some time experience a change in production structures, K would concentrate in c-production, EU in f-production. During this process the EU-wage relative to K-wage declines, check it yourself! Examples? At industry-level car production, USA-Japan? The fate of mobile phone production in Finland?

48 Technological Change and Global Restructuring of Production 9 The advantage of a less-developing (low wage) country in technology adoption is one example of the benefits of backwardness brought in our consciousness by Alexander Gerschenkron long time ago. The model is also consistent with Gerschenkron s claim that the benefits from backwardness show up very fast if they are utilized. This was one side of his criticism of the view that development would/could be even and linear.

49 Ricardian Comparative Advantage with Many Industries 1 To be of use in empirical research (and to be testable) it is necessary to develop a multi-sector extension of the model (and also multi-country version) Assume now that there are N industries, let a EU i, a K i be the unit labor requirements in industry i. Industry i production is in EU if ai EU w EU ai K w K w EU w K ak i a EU i EU thus produces and exports goods in which its unit labor requirements are relatively small (labor productivity high). (23)

50 Ricardian Comparative Advantage with Many Industries 2 Number the industries in such a way that industry 1 is the one in which EU has largest comparative advantage, 2 the second highest etc.: a K 1 a EU 1 ak 2 a EU 2... (24) This is called the chain of comparative advantage. Obviously there exists an industry ĩ such that EU will be exporter of that while in industry ĩ + 1 K will be the exporting country.

51 Ricardian Comparative Advantage with Many Industries 3 Empirically: One can form chains of comparative advantage for any country pairs by utilizing industry level labor productivity data. But how does one measure the (relative) intensity of trade? The answer to this are indices of Revealed Comparative Advantage, the most famous of which is the Balassa-index. The index for industry i is defined as the the ratio between the share of industry i exports in a country s (e.g. EU) total exports and the and the share of industry i exports in total exports of the reference country (e.g. K).

52 Ricardian Comparative Advantage with Many Industries 4 If the index has a value larger than unity for industry i the country is said to have a Revealed Comparative Advantage in industry i. Balassa indices are widely available. One can then simply calculate correlations between the relative labor productivities/unit labor requirements and Balassa indices to get a simple test of the Ricardian theory. In the real world one must also take into account the existence of all kinds of trade barriers. How do they fit in the present story?

53 Ricardian Comparative Advantage with Many Industries 5 This is easy. Think of EU imposing a uniform (ad valorem) tariff τ EU > 0 on all imported goods and at the same time K imposes a tariff of its own τ K > 0 Then EU will produce in all industries for which ai EU w EU (1 + τ EU) pi W = ai K w K w EU w K Here p W i ( 1 + τ EU) a K i a EU i (25) denotes the world market (free-on-board, fob) price. This shows that the tariff increases the range of goods produced in EU.

54 Ricardian Comparative Advantage with Many Industries 6 Analogous condition holds for K, it produces goods for which w EU w K a K i (1 + τ K ) a EU i (26) Again, the range of goods produced in K increases with tariffs. Altogether all the goods/industries for which ( 1 + τ EU) a K i a EU i w EU w K ( 1 + τ EU) a K i a EU i (27) are produced in both countries and not traded internationally at all: Trade barriers create closed sectors.

55 Ricardian Comparative Advantage with Many Industries 7 The implication is that at intermediate points in the chain of comparative advantage there should not be much mutual trade if there are trade barriers. The model should be completed by specifying the demand side to determine the world market prices and relative wages. You can find that in R. Dornbusch, S. Fischer and P. Samuelson: Comparative Advantage, Trade, and Payments in a Ricardian Model with a Continuum of Goods, American Economic Review, vol. 67, No. 5,

56 Gravity Equation in the Ricardian Model 1 One of the most important tools in empirical analysis of trade flows is the so called Gravity Equation. It explains the bilateral trade flows by the market sizes ( masses ). We know that the volume of imports by K from EU are (as K does not produce c) M K EU = α p f Q K f p c (28) Here p f Q K f = nominal income in K, with Q K f as the volume of good f produced = GDP volume.

57 Gravity Equation in the Ricardian Model 2 At the same time, by definition, EU income must equal the total global expenditure on the goods it produces: ( ) p c Qc EU = α p c Qc EU + p f Qf K (29) Solve this for the price of good c: Use this in (28): p c = α ( p c Qc EU + p f Qf K Q EU c ) (30)

58 Gravity Equation in the Ricardian Model 3 Gravity equation: Q M = EU Q K p c Qc EU + p f Qf K P w is a measure of the global price level. QEU Q K P World (31) It can be regarded as a measure of all kinds of trade barriers.