The H-O theory has been being confronted with two main problems (i) problem of specialization and (ii) problem of reciprocal demand.

True but if the poor countries have nothing to trade with i.e. goods service or money there will be little trade between poor and rich countries. For example India has a commodity of people which in turn is services so they have something to trade with.

Of course India is a human resource endowment country but what she needs is specialisation and planning in human resource market according to needs of domestic and global growing requirements.

I view the HOS as a way of thinking about openness and specialization with closeness and diversification. Openness has been correlated with strong institutions while closed economies have weak institutions. Openness shas been accompanied with sustained growth and protectionism with poor human development progress. The main message of HOS is that moving away from comparative advantage through high protectionist barriers will worsen resource allocation. In fact, protectionism has shown to entail much higher costs than those predicted by the HOS model.

There are many confusions concerning international trade theory. Many people understand that HO theory is a natural generalization of Ricardian theory. But this is a grave error and misunderstanding.

I am afraid that Dr. Arjun Yallappa Pangannavar does not make this distinction. What  does he mean by the problem of reciprocal demand? Reciprocal demand is normally connected to John Stuart Mill's treatment of terms of trade. Marshall and Edgeworth followed him, but it  has little to do with Hechsher and Ohlin's formulation.  

Julio J. Nogues argues that HOS theory is a way of thinking about openess and specialization contrasted with closedness and diversification. Why does he ignore Ricardian theory of international trade? It also gives certain views on the same point but the implications of two theories are different. Does he thinking that HOS theory and Ricardian theory are identical or at least similar? If not, he should have argued the specific point of view that HOS theory gives us. 

I understand that Dr. Tahsen Alqatawni has asked a more specific question that HO model poses.  

Dr. Tahsen Alqatawni,

I have given a quick look read on Wade's paper on Poverty and Inequality. Although the "paper questions the empirical basis of the neoliberal argument," (last sentence of the summery) there is no mention on HO theory. What do you want to say by referring to HO theory? Please explain your intention.

Dear  all, 

I made a mistake. The paper that Dr. Tahsen Alqatawni cited was not the paper Wade (2004). He cited Bergha and Nilsson (2010). See the full information below.

Rober Hunter Wade (2004) Is Globalization Reducing Poverty and Inequality? World Development 32(4): 564-589.

Andreas Bergha and Therese Nilsson (2010) Do liberalization and globalization increase income inequality? European Journal of Political Economy 26(4): 488–505.

Although I made a mistake, my question above remains valid, because there is no mention on HO theory in Berga and Nilsson (2010), either. If I make a brief comparison, Wade (2004) is more critical to the mainstream understanding whereas Bergha and Nilsson (2010)is more concerend with the empirical fact findings.

Did Dr. Tahsen Alqatawni want to represent the mainstream economic thinking by HO model? If so, he is making an error, because Ricardian trade theory can have a very different implications from HOS theory. 

There are some sharp differences between Ricardian trade theory and HOS theory.

The latter is based on factor proportions. It predicts that labor abundant  country (when compared to capital) will specialize in the production of labor intensive products. In addition to the fact that this prediction is too rough and rude, the theory misunderstands that it is the country which trade internationally. This is of course simply false, because the actual subjects  of international trade are firms.

(On this point, see Faccarello's encyclopedic article Comparative Advantage and my comments on it. You can see the links below.)

For example, imagine a firm in India that works in a high tech industry. It is not the capital/labor ratio's that determines the competitiveness of the firm's product. If the production cost is  lower than that of any country's firms producing the same product, the Indian firm is most competitive in the products and can export it to the world. This is possible, because India holds a huge army of highly educated engineers and  the wage rate of those engineers is relatively low than that of more advanced countries. This is what the recent version of Ricardian trade theory tells. HOS theory has nothing to say on it.

China's case gives another example. China is now capital exporting country. Firms in China started  as village and town firms and then family firms and accumulated capital rapidly. At the beginning (in 1980's and 90's), China was labor abundant and capital rare country. Does the fact that China is now exporting capital mean that China is capital abundant?   

The central example  of the HOS theory is factor price equalization theorem. It contends that wage rates (price for labor) of trading countries are equal. Although Chinese wage rate is going up rapidly, it remains still about 1 fifth of the more advanced countries' wages. If we apply HOS model literally, if China is capital abundant,   its wage rate must be higher than that of advanced countries. These simple facts shows how HOS theory (or factor proportion theory) is irrelevant to the understanding  of what is happening in China.

Ricardian theory, generalized to many-country, many-products case, does not predict such a nonsense.

Readers can consult my recent talk in WIIW (The Vienna Institute for International Economic Studies). You can see what kind of topics are now discussed. At the end of my presentation, I have given references to some relevant papers. You can  read and download all of my papers from my Contribution page of RS.          

NB: Do not confuse with the presentation I gave in Hausdorff School in Bonn. It is purely mathematical and addressed to mathematicians. 

Chapter Comparative Advantage

Conference Paper The New Interpretation of Ricardo's Four Magic Numbers and t...

Presentation Recent Development of International Trade Theory and Some of...

First and foremost, the inherent problematic associated with the model is the naive assumption of comparative advantage, which is empirically untenable.

 Dear David,

I suppose you are talking about HO model. How do you think of recent development of Ricardian trade theory? Both of them talk about comparative advantage but assumptions and implications are quite different.

It is also necessary to distinguish theory and policy. Hajoon Chan rarely makes this distinction and blame Ricardian theory of international trade (including its modern developments) because of its policy arguments for free trade. 

You are right in emphasizing sociological and political aspects of the globalization. However, without having a good economic theory, you cannot explain what is now happening under the name of globalization.

The main force that drives world to globalization is economic. The economic policies and politics play some parts, but they are secondary and driven by the economic motives. Some parts are also influenced by the ideology embedded in the mainstream economics. In order to  criticize these complex phenomena,  it is necessary to appeal to an effective critique of economics. You cannot simply discard economic theory. You should present or construct a new theory.           

The Heckscher-Ohlin theory focuses as the raison d'être of international trade the difference of factor proportions between countries. In the simplest form, it assumes labor and capital as two major factors of production. It is normally supposed that rich country has higher capital/labor ratio and poor country lower one. If this differences of capital/labor ratios are real and effective causes of trade, major part of international trade should be made between rich and poor countries. Trade between rich countries and that between poor countries should be small. Of course, this is in contradiction to the fact. Trade between rich countries occupies majority of international trade.

Another consequence of HO theory is that trade liberalization will increase wage of poor countries and decrease that of rich countries. If, for example, the income of capitalists are higher than those of laborers, this would deduce inequality in rich countries and increase inequality in poor countries. Empirical study on this regard is quite ambiguous. Inequality in the U.S.A is increasing but the same tendency can be observed in poor countries too.

Dr. Tahsen Alqatawni's question must be directed on this discrepancy between theoretical predictions and the empirical facts. The most simple solution would be to abandon HO theory, because it has accumulated many anomalies.

HO theory or its generalized version HOV (Hechsher-Ohlin-Vanek) theory had been the workforce for more than 30 years and it was revealed that the factor proportion theory had a very poor predictive powers. It could not predict the trade volume and directions of trade. See Trefler (1993; 1995) for the first point and Bowen, Leamer and Sveiskaus(1987) for the second point.

Even Weinstein and Davis (1996), powerfull defenders of factor proportion theory, admit that "the Heckscher-Ohlin theory of trade is the perfect example of a theory which has been statistically rejected time and again." What they could find as an example supporting the original form of HOV theory was inter-regional trade inside of Japan (Davis, Weinstein, Bradford and Shimpo 1997).

Leamer, Treffler and others tried to modify HOV model by introducing technology differences. They got better results but it was doubtful if this modification is simply a fitting of parameters to the data.

Leamer and Levinsohn (1995) recommended not to "take trade theory too seriously." They argued that "theorems are neither true of false. They are sometimes useful and sometimes misleading." "If we approach a data base with contrary attitude, hoping to determine the verity or falsity of a theorem, we may statistically `rejects` the theory, but leave it completely unharmed nontheless. After all, we already knew it wasn't literally true."

Many argued their contentions but I do not follow the subsequent arguments. One thing seems evident for me. It is not the time to make a series of modification to adjust the theory to facts. Factor proportion theory is in a cul-de-sac. It is the time to think of a paradigm change. HO theory or any other forms of factor proportion theory should be abandoned as a basic theory of international trade.

References:

Bowen, Harry P., Edward Leamer, and Leo Sveikauskas (1987) “Multicountry, Multifactor Tests of the Factor Abundance Theory,” American Economic Review 77(5):791-809.

Davis, D.R., and D.E. Weinstein 1996 Empirical Tests of the Factor Abundance Theory: What Do They Tell Us? Eastern Economics Review 22(4) 433-440.

Davis, D.R., D.E. Weinstein, S.C. Bradford and K. Shimpo 1997 Using International and Japanese Regional Data to Determine When the Factor Abundance Theory of Trade Works. American Economic Review 87: 421-446

Leamer E.E. & J. Levinsohn 1995 "International Trade Theory: The Evidence," in The Handbook of International Economics: Vol. III , edited by G. Grossman and K. Rogoff, Elsevier Science B.V, pp. 1339-1394.

Trefler, D. 1993 International Factor Price Differences: Leontief Was Right! Journal of Political Economy 101: 961-987.

Trefler, D. 1995 The Case of the Missing Trade and Other HOV Mysteries, The American Economic Review 85(5): 1029-1046.

Apart from its poor predictive power, HO theory is conceptually defective. Followings are some naive questions. Can somebody explain? Why does nobody raise these simple questions?

(1) Factor proportion of a country is not a crucial condition for firms in the country.

Suppose a country that has low capital/labor ratio. Even in such a country, many firms can capital intensive. What counts is not the capital/labor ratio, but the low unit cost of production. Equipped with highly efficient machines, a firm can produce at a very competitive cost.

(2) Factor proportion is a result, not the cause.

When a firm plan to choose the location of a new factory, it may choose a location where it can hire good workers. When it constructs the factory, machines would be bought from elsewhere and be installed in a suitable combinations. A factory is not planed by the existence of capital goods. If the plan is good, banks will offer to lend necessary capital money. Capital/labor ratios are determined without any reference to capital/labor ratio of the regions. When new factory come to a rural area where there were no big, this new factory determines the capital/labor ratio. Factor proportions is a result of investment for a factory and not the reason of investment.

(3) Capital goods are importable. Why not import them?

Capital goods (machine-tools, materials, parts, fuels, etc.) can be imported. Why not import them? What is the difference between consumer goods and input goods?

(4) Is capital money lacking in low capital/labor ratio country?

By capital money, all these goods can be imported. What difference are there between consumer goods and capital goods? In a low capital/labor ratio country, is wage rate low and is profit rate high? If then, why the high capital/labor ratio country invest their money into low capital/labor ratio country?

(5) Is capital endowment just like labor?

All machines and equipment are constructed or imported. They are not a natural endowment. They are result of investment.

(6) Why are capital accumulation neglected?

Capital accumulation was the most important theme for classical economists? Why do neoclassical economists in international trade theory ignore this important question?  

Se the following article on the subject:

J. Hellier (2012), ‘The North-South HOS Model, Inequality and Globalization’, in: J. Hellier & N. Chusseau Eds, Growing Income Inequalities, Economic Analyses, Chapter 4, Palgrave MacMillan, 107-146.

I have read Joel Hellier's  paper that is uploaded in RG (the first of the two below).

• The North-South HOS Model, Inequality and Globalization
• The New Interpretation of Ricardo's Four Magic Numbers and the New Theory of International Values / A Comment on Faccarello's " Comparative advantage "

This gives a good example in some special point of view. I will explain later days more in detail, but this paper shows the actual state of theory that HOS theory faces with regards to globalization.

Joel Hellier knows well where and how HOS theory applied to North-South trade situation fails to predict various stylized facts. Table 1 lists 11 stylized facts among which NS-HOS model predicts only 2 facts correctly. Its prediction is irrelevant for 2 other facts and contradicts all other facts, i.e. 7 facts among 11.

Normal attitude would be to doubt the relevance of HOS theory to North-South trade and actually proceeding globalization. Heller knows this very well. He lists 3 possible attitudes. He explains the second  attitude is to admit "the inability of the HOS model to capture the key characteristics  of globalization and NST" [Shiozawa's comment NST=North South Trade] . However, Hellier prefers or choose the third attitude. It is to accept "the HOS framework as a starting point and to significantly amend it so as to model certain key mechanisms that is  its traditional version cannot portray because of over restrictive assumptions."

The state-of-the-art of trade theory represented by HOS theory reminds me the state of the astronomy just before Nicolaus Copernicus (or the first half of the 15 th century). At that time, Ptolemaic system accumulated amendments and modifications for more than 1000 years and had become a complex but very refined system. Prediction accuracy was quite high. Tycho Brahe, who was the most talented observer in the latter half of the 15th century, continued to believe in Ptolemaic geocentric system against Copernican heliocentric system.

It is true that Hellier (and his collaborators) can amend the HOS model infinitely when a new stylized fact appears. However, we can also argue that HOS model has accumulated so much contradictions (in side of the theory and with observed facts) and it is time to consider a change of theory itself.

If he has no other available theory, he may be obliged to use the old theory until a new theory appears. It seems he does not know the new development of Ricardian trade theory. 

For a quick overview, see for example my recent presentation at WIIW. (The presentation data t appears in my Contribution page in RG). Readers can also read my conference paper (the second of the two below).

For details please wait several days. I am writing a draft paper and planning to upload it as discussion paper in RG.

These are not destined to trade theory specialists. I am trying to keep them readable for all economists who are interested international trade and economic development.  

Chapter The North-South HOS Model, Inequality and Globalization

Conference Paper The New Interpretation of Ricardo's Four Magic Numbers and t...

Just a short answer.

I am not a defender of the HOS approach in any case. I even have huge critiques on its hypotheses. I just think that the most important element of NST is the factoral divergence between the North and the South in terms of skilled vs unskilled labour. There is obviously a very large initial difference in terms of technology, but capital and technology are very mobile because of firms' multinationalization,  whereas labour mobility is much smaller. Then, starting from this difference is to me rather appropriate. What I mean by 'starting point' is just to assume that the key difference between the North and the South is the difference in  relative skill endowment, because the other differences can be lessened by capital and technological mobility. Note that differences in technology are inserted in the framework I proposed in my paper, but those are shrinking with time because of catching up through adjustment to the  technological transfers.

Finally, I thing that we do not have to choose between the factoral and technological (or other) approaches to international trade, both having an impact. But this impact differs depending on the type of trade which is analysed. E.g., as regards North-North trade, I do think that the Melitz' s type models as well as the quality models are the most appropriate because of a rather limited factoral divergence  between advanced countries.

Dear Joel,

I do not contend that you are defending a HOS approach. All right. If your position is as such, why did you not mentioned a word on it? In the last paragraph of your conclusion, I read

• The analysis presented here shows that the NS-HOS framework remain an essential tool to model the observed developments of globalization and inequality under the condition of modifications and extensions that correct its most simplifying and controversial assumptions.

You know that HOS framework does not work. It has too many anomalies and discrepancies to defend it. Do you claim that it deserves to be examined because it is the unique acceptable framework for many economists? That is fine. I can still use your paper as a good example to prove the necessity of a paradigm change. (I want to discuss this point in near future.)

As for your "starting point," it may be a good one although I do note see what kind of theory framework you are imagining. Does it come to HO theory in the final account? You may not bother in this questions, because you are more oriented to understand about the inequality than to worry about theory. However, even in this point, what you suppose as background theory orients your investigation very much.

I admit that Ricardian trade theory (or more widely classical theory of value) does not (yet) have a good theory by which to analyze in-homogeneous labor market. If you want to analyse wage disparity between unskilled labor and skilled labor, you are obliged to appeal to HO type analysis. However, you are confined do argue the inequality within a country. HO theory (even if it is modified as NS-HOS with factor intensity differences) does not give any causal explanation on why wage inequality sometimes increases and sometime decreases.

HOS theory may explain why inequality decreases but not why inequality increases. It did, before 1990's. We may compare several time periods: before and after industrial revolution, from 1860's to 1960's, or even between 1950's and 1990's. This increase of inequality is only explained by Ricardian trade theory (in a form of new theory of international values) and not by HOS theory. Then, the real option is this. Are you more concerned with internal inequality, or with international (or North-South) inequality?

I read many papers which discuss wage inequalities. Most of them are concerned with the internal inequality (between unskilled and skilled labor forces). I see few paper which discuss or analyse international inequality (i.e. wages between workers in the South and in the North). Does this state of interest only reflect the fact that more people are interested in the internal inequality questions? I doubt it. This may well reflect the present state of economic theory. HOS approach is still dominant in economics and slaving our thinking.

If you are more interested in the international inequality (and I hope you are), you should choose a right theory.

This is related to the last point you raised. You put that "we do not have to choose between the factoral and technological (or other) approaches to international trade, both having an impact." You may be thinking to measure relative impacts of factor proportion factor and technological factor, but what is really required is the choice between approaches, i.e. theories or paradigms.

It is not a wise custom to mix two different and conflicting theories (even if it is widely spread). You are contemplating the necessity of generality and universality of a basic theory. This is in fact the very basis of modern science. Milton Friedman neglected this important characteristic of a science. If you are thinking that you can employ any theory as you please, you are then a slave of Milton Friedman type methodology.

Yoshinori

).  In 2006, Kremer and Maskin noted there are at least two pragmatic problems with this model as follows:

1. It predicts that bilateral trade will be largest when factor funding is most different.  There is little trade between developing countries such as the U.S. and Developing countries such as Chad.

2. the clear indication from examination of particular developing countries following trade liberalization, and from cross country research does not recommend that trade liberalization generally decreases inequality in Developing countries. In fact, repeatedly recommends that trade liberalization can jack up inequality.

Reference

Kremer, M., & Maskin, E. (2006). Globalization and inequality. Retrieved from http://www.economics.harvard.edu

We can read Kremer and Maskin's paper at 

http://www00.unibg.it/dati/corsi/8997/25752-GlobalizationInequality_Oct06.pdf

I will try to read it.

I quickly read Kremer and Maskin's paper. So the following is a rough impression.

As Dr. Tahsen Alqatawni mentioned, the authors pick up two problems of HO-theory.  Namely

1. It predicts that bilateral trade will be largest when factor funding is most different.

2. the cross country research does not show that trade liberalization generally decreases inequality in Developing countries. 

These two point are not in accord with the standard consequences of HO theory. Kremer and Maskin show these two propositions are logical consequences of standard HO theory. Then, they examine the empirical aspect and they admit that cases of two  propositions are empirically observed.

After this argument, they pick up the second problem and they starts to ask if these two anomalous phenomena can be explained by  an HO model. The model they constructed  is the one "in which there are two countries and just one consumption goods".

Each country have two kind of workers: highly skilled worker and low skilled workers. Rich country has high skilled A and low skilled B workers and poor country has high skilled C and low skilled D workers. 

Innovative point in the model is that they think of a production which is composed  two tasks: mangers task and workers task. There are 4 matching patterns and  In the self matched case the product per unit of work hour will be

             H2  and L2.

If workers are cross matched, i.e. a high skilled worker is assigned to manager's task and a low skilled worker to the workers' task, the product per unit of work  hour will be 

          H2 L.

If the production can be divided just like offshoring fragmentation, it turns out that the both consequences of trade liberalization (increase and decrease of inequality) are possible. More precisely, if the global inequality increases depends on the relative numbers of A, B, C, D workers.

It seems that two authors are satisfied by this result. Except pointing that these results are possible in a HO model, they do not produce any judgement on the validity or usefulness of HO models.

What we have to ask is what this result means? Kremer and Maskin's model is so different form standard model. The fact that their model can tell two opportunities means that HO model is useful? I doubt it.

Let us consider an imaginary situation. We have a  theory (or theory framework that may implements many models)  and two phenomena A and B. The classical model can explain only A. The new models can explain A and B. Does this mean a development of the theory?

A theory framework comprises infinitely many models of almost infinite variety. Then an existence of a model which is adjusted to some of the phenomena  does not signify that a theory is true or right. We cannot even say it is healthy. Such a theory lacks something very essential to a science. A science has to have a criterion when or at which situation a theory is applicable. It must have a theory that tells how to choose a model. If a model is applicable in such and such situations and not applicable in such and such situations, we should have a criterion that tells how to do in such choice.      

Have we all forgotten even a least criterion of a science?  Is this a consequence of predictionism? Something in economic methodology seems to be seriously ill. 

A globalization paradox?

Hellier's paper The North-South HOS Model, Inequality and Globalization comprises too many topics. I confine myself in this post to discuss only one point that he refers to. 

In his Introduction, Hellier wrote:

• If the NS-HOS framework is well tailored to model North-South trade (NST), and if most of its predictions are not in line with observed facts, the NST is not a major explanation for the changes in inequality.

In 1980's and 1990's, many economists argued that NST is not the cause of increased inequality of wages in USA, because the weight of trade is too small. Globalization changed the situation much. Now the weight of emerging countries in world trade is critically increasing. So Hellier argues that

• With the advent of China and India, most of the HOS prediction should be verified if the HOS model provides a reliable picture of the main mechanism of North-South openness.

However, the American economy is experiencing totally different trends that HO theory predicts. Such paradoxical facts were discovered by Robert Z. Lawrence for the period after the last 1990's. See

Lawrence, R.Z. 2007 The Globalization Paradox: more trade less inequality. VOX, 4, September 2007

http://www.voxeu.org/article/globalisation-paradox-more-trade-less-inequality

Lawrence, R.Z. 2008 Blue-Collar Blues: Is Trade to Blame for Rising US Income Inequality? Peterson Institute for International Economics. A digest at https://piie.com/publications/papers/lawrence0208.pdf

 Lawrence shows in a series of many figures that 

• The US economy has also become increasingly open to trade.
• Big volume shock occurred. Ratio of trade balance to GDP declined rapidly from the end of 1990's.
•  The ratio of white- to blue-collar compensation stopped rising!
•  Relative earnings of high-school dropouts stopped falling.
• Ratio of earnings of advanced degrees relative to high school diploma rests stable.

As a conclusion Lawrence defines his concept of paradox: more trade yet little increased wage inequality in developed country such as USA.

Here are my comments. HS model predicts that wage inequality between unskilled and skilled labor increases in developed (capital or skilled labor intensive) country when the trade is liberalized and its volume and weight increase. What happened is the opposite of this prediction.

Lawrence and Hellier produce their interpretation. In my opinion, Lawrence's paradox (globalization paradox) is not a paradox at all. From my standpoint (see my post #13 in this question page),

(1) Factor proportion of a country is not a crucial condition which determines the volume and specializations.

(2) Factor proportion is a result, not the cause. They have no predictive power on what happens in international trade.

Lawrence's findings simply mean that HO theory (and all other factor proportion theories) made a bad emphasis on the role of factor proportions. They are not important factors that determine wage inequality between unskilled- and high-skilled labor. HO theory and models has posed a bad focus on the cause of international trade. Argument based on this irrelevant theory (including Stolper-Samuelson theorem) only revealed how it is totally mistaken.

The history of North-South HOS models shows a typical science philosophy of majority of economists:

I made a promise to discuss the Hellier’s methodological stance on North-South HOS models. In his consciousness, what he is doing is a historical survey and assessment. With this regards, he is a good worker and I learned much by reading his paper (The North-South HOS Model, Inequality and Globalization; cited in posts #13 and 14.). It is a full scale survey on how North-South HOS theory and its modifications or extensions tried to explain many stylized facts which are in variance with the theory’s predictions. If one wants to try any related investigation, Hellier’s paper must be the shortest way to get an overview on the problem. However, as I have written (in post #14), this paper gives us a good example which show the disastrous state of our science methodology, or methodological thinking in economics.

I am not criticizing Hellier himself. He is an able economist equipped with a standard science philosophy. The trouble is this science philosophy that lies in the deep mind of almost all economists. Hellier’s paper made clear what economists are doing when an anomaly appears. First, some economists contest the validity of the new discovery. When the knowledge became a stylized fact, other economists present a modified model and explain how it can occur. If this happens only once, this might be admissible. However, if anomalies appear in groups in different aspects, the normal attitude would be to doubt the general validity of the core theory. Hellier’s paper shows (in a beautiful way) how economists continues to accumulate modifications. The system becomes more and more complex with no unifying principles.      

I wrote in my previous post (post #14) that state of HOS-related theories and models reminds me the state of the astronomy just before Nicolaus Copernicus. By that time, Ptolemaic system had become very complicated by a series of modifications. The movement of each planet is adjusted by using deferent and epicycles. In later years, two or three epicycles were used to obtain sufficient accuracy of the movement. What economists are doing is quite similar to those astronomers who added new epicycles when they observed deviation from predicted orbits.

In his paper, Hellier cites 11 stylized facts which are invariance with the “traditional” North-South HOS model. A traditional North-South HOS model is a 2×2×2 HOS variant which replaces two classical factors capital and labor by skilled and unskilled labor. The only difference between North and South is that North has higher proportion of skilled labor than that of South.

The biggest trouble with is model is that Factor Price Equalization (FPE) theorem holds in a standard situation, or when the two countries produce both goods (in Hellier’s terminology, the case where both countries lien in the D-cone or Diversification cone). In this case, FPE theorem tells that wages of both counties are equal and there is no big difference between South and North. Hellier does not pay enough attention to exclude this situation (his Figure 4b-1 case). This is a careless mistake perhaps because he is concentrating his mind in wage inequality between skilled and unskilled workers. How ever, this is not an important point.  (To be continued)

How Hellier estimates the traditional North-South HOS model.

Hellier derives 9 predictions (listed in Table 3) of the “traditional” model and compares them with 11 stylized facts. If we count yes/no by stylized fact, North-South HOS model’s predictions are correct only for two facts. Excepting two “irrelevant” cases, predictions are at variance at 7 stylized facts. If we count yes/no by prediction side, predictions are suitable for 2 stylized facts and not for 7 other facts.

If there are so many discordances between facts and theory predictions, the normal reaction would be to abandon the theory. It would be better to restart from searching a new theory or framework. Maybe, Hellier’s real intention is to make people acknowledge these facts and Incite them to change their traditional orientation, but this point is not very clear. In his “general assessment and conclusion” (section 9), as I have cited in a previous post, Hellier observes:

• The analysis presented here shows that the NS-HOS framework remain an essential tool to model the observed developments of globalization and inequality under the condition of modifications and extensions that correct its most simplifying and controversial assumptions.

He remarks after this sentence that this may not mean that “the globalization is the essential explanation of these changes.” He concedes that globalization alone may explain these changes and it may be necessary to combine studies on globalization, institutional change and technological progress. To admit weak predictive power of NS-HOS is not to discharge the theory of its logical consequences. If NS-HOS theory is a good theory, it must have a partial theory within the theory that makes clear which it can explain and which it cannot. The weak predictive power may indicate that NS-HOS theory is excessively simplifying the economic mechanism and disfiguring its workings. It is a pity that Hellier does not think about these questions. This must also reflect the actual state of science philosophy that many economists have spontaneously and unconsciously. (To be continued)

How economist added modifications after modifications:

Hellier’s paper does not stop by making the two yes/no tables. He makes an extensive survey on how economists modified the model in order to explain each of many anomalous facts. As an example of this history, Hellier’s paper is indeed a good survey.

I do not intend to explain all troubles and the “solutions” they gave. Hellier picks up 4 major problems and explains how economists modified the models and reshaped it to be in accordance with some stylized facts. Each time, new variables were introduced and some extra hypotheses were added.

Section 4 deals with the transition form one stage of diversification to another. In a “traditional” NS-HOS model, factors are given amounts. Hellier introduces the size factor to this model. This implicitly assumes the change of existing factors (or growth of factors). Therefore he endogenizes actor endowments but he does not give any account how these factors are changed or accumulated. He simply changes the size and proportions in such a way he can make a story how South catches up North and the skill premium (in South and North) changes.

Section 5 introduces wage rigidity. Just like neo-Keynesians, Hellier points that NS-HOS model can deals with unemployment. He introduces Davis’s model and efficiency wage arguments. In the latter case, which it seems Hellier supports personally, the model introduces effort of workers. By assuming appropriate levels of efforts, it is easy to make a story as we please.

Section 6 treats technological differences and technical change. New variables such as technology gaps are introduced. Then it is easy to predict whatever happens if we manipulates the change of these gaps. 

Section 7 discusses “international outsourcing”. Hellier presents many trials which aimed to explain rapid growth of outsourcing (or offshoring, fragmentation, and unbundling). It is evident that “traditional” framework of NS-HOS model cannot treat these phenomena, because it has no input goods. The classical HOS model assumed capital but in this case neither it cannot treat outsourcing appropriately because thee are no traded goods. Facing this serous defect of the theory, majority of economists (you can find many eminent names in the second paragraph of section 7) appealed to a chip trick to introduce plural segments in the production process. Even that trick did not work well if they want to “integrate” each firm’ behavior to a single production function with aggregated arguments. They have to assume a peculiar case where the segmented functions are again Cobb-Douglas functions. With segmentation new variables are introduced. In a simple 2 stage segmentation, a new variable is introduced.

(To be continued)

Final overall assessment

 In the last section before conclusion, Hellier adds four more extensions: (1) human capital accumulation, (2) skill upgrading, (3) capital-skill complementarities, and finally (4) multiple numbers of goods and factors. The story proceeds as before. New assumptions and new variables are introduced.

In the end, we have Table 4: The extensions of NS-HOS model facing Stylized facts. We have 9 modified models in addition to the initial “traditional” model. To my astonishment, there are no models that satisfy all 11 stylized facts. The model that has minimal number of “no”s (Factor technological bias) counts only 1 “no” but also counts only two “yes”. All other checks are “not relevant” except one “yes/no.” Sector technological bias model give the similar results with 2 “yes”, 2 “no”, and 7 “not relevant.” All other modified models counts at lest 4 to 6 “no”. Although the number of “no” is decreased, they have at least four stylized facts that they fail to predict correctly.

In view of this result, Hellier is quite optimistic. He emphasizes that “different type of extensions are adapted to different shortcomings of the traditional NS-HOS approach”. He is logically right in this remark. But does this mean any improvement was obtained? We have 11 types of variants and yet no model can explain all the established stylized facts. Logically this means that no one passes the screening that Popper asked for.

The real question starts here. Is it wise that we continue to search a new model, or is it better to change our research program? As I have repeatedly stated, state of HOS-related theories and models is very similar to the state of the astronomy just before Nicolaus Copernicus. Adding new epicycles does not resolve the problem. A new insight is required.         

Some pessimists would say. Yes you are right, but we have no new theory in hand. We may have to wait another one hundred years (reducing by the factor of 10 the time passed from Ptolemy to Copernicus). I do not think so. A new theory is already prepared. See my essay below:

https://www.researchgate.net/publication/304717720_New_Theory_of_International_Values_A_General_Introduction?ev=prf_pub

Working Paper The New Theory of International Values: An Overview

amazing...

I have posted this new question:

Trump and Sanders are opposed to the TPP. How the International Trade theory and International Political Economy explain this?

https://www.researchgate.net/post/Trump_and_Sanders_are_opposed_to_the_TPP_How_the_International_Trade_theory_and_International_Political_Economy_explain_this?_tpcectx=profile_questions

Although this is not the particular defect of the HO-model,  the inability to explain why these strong opposition emerges despite that HO-theory and other GET based trade theory preaches the merits of free trade. I believe that this is really what International Trade Theory explain. How do you think? Please argue freely in the new question page.

Chisato, 

are you talking about  

Arja H. Turunen-Red and Alan D. Woodland (1991) Strict Pareto-Improving Multilateral Reforms of Tariffs. Econometrica 59(4): 1127-1152.

If so, theirs is not a complete model of international trade. They consider

• a perfectly competitive general equilibrium model of the world, consisting of K (> 2) nations trading in N (> 2) internationally tradeable commodities. Each nation consists of many producers, a single consumer, and a government.

In this sense, theirs is a multi-country, multi-commodity  model. As they assume as input factors any number of non-traded goods, we can also say it is multi-factor model, if we assume that non-traded goods are factors.

Even though, it is not a complete model, because they assume a peculiar GPN function whose arguments are exogenously given prices vectors. Turunen-Red and Woodland want to consider welfare effects of tariffs and they assume this,but this kind of production function has no relevance to the real economy and moreover has no micro foundations.

Basically H-O theory insists on Factor-endowment as determinant of size of gain from trade, internal or external; in present context considering total cost factor richness still holds good, for instance, labour intensive goods of labour rich countries are still comparatively cheaper to labour poor countries, from production point of view. Provided demand and supply elasticities are ignored.