EN_S4J2_critiques-test-leontief

In-depth exploration files

REad

The Leontief test is a subject of intense scrutiny and debate

Not surprisingly, the first reflex of the economist community at the time was to reject (“push back”) Wassily Leontief’s test. The epithet of “paradox” associated with it reveals the test status as an “anomaly.” Thus, no element of the test escaped criticism. By seeking to make the results of his observation compatible with the HO theorem, W. Leontief ultimately helped to open up new avenues of thought, some of which went far beyond the factorial approach. The ‘paradox’ associated with the Leontief test has triggered a rethinking of the reigning paradigm and hinted at the potential for a ‘scientific revolution’ in T. Kuhnʼs phrase, ushering in a new era of understanding in international trade theory.

Introduction

Here, we shall present a comprehensive overview of the criticisms leveled at the Leontief test, followed by an exploration of the many conceptual overruns to which the “Leontief paradox” has led. Neither the test hypotheses nor the statistics used, nor W. Leontief’s method has escaped criticism.

Introduction

6. An erroneous grasp of the historical context

3. An incorrect understanding of production factors

4. Incorrect understanding of exports and imports

5. A too-aggregated view of trading partners

2. A questionable methodology

1. Contradictory tests of empirical evidence

Contents

01

Contradictory tests of empirical evidence

Source

Several economists have applied W. Leontief’s method to analyze international trade in other countries. Table 20 summarizes some of the results obtained. The table above shows that the results align with the HO theorem where a<1 in the case of the UK, Canada, Japan, and West Germany.

1. Contradictory tests of empirical evidence

The UK, Canada, Japan, and West Germany were relatively more abundant in capital than labor in the reference years, and the ratio of capital intensities of imports to exports was logically less than 1.By contrast, the results for East Germany, Japan (abundant in labor in the 50s), and France confirm the paradox. As the table is synthetic, not all studies are included. However, several results invalidate the factorial theory.Examples include Ranganath Bharadwaj’s test, which confirms the paradox in the case of trade between India and the United States, or the work of Robert M. Stern and Keith E. Maskus, which confirms the paradox in the case of U.S. foreign trade for the year 1958, but lʼinfirm it for the year 1972. Lastly, David S. Clion and William B. Marxsenʼs study shows that the foreign trade of certain countries is in line with the HO theorem (Australia, Ireland, Japan, Korea, New Zealand, United States). In contrast, others are in a “paradox” situation for the study reference years (Israel, Kenya, United Kingdom).

1. Contradictory tests of empirical evidence

There is no point in multiplying empirical references insofar as they contradict each other. The logical conclusion is that it is impossible to confirm or disprove the HO theorem simply by observing the facts. From then on, criticism focused on other elements of the “Leontief paradox.”

1. Contradictory tests of empirical evidence

02

A questionable methodology

Using a table of industrial exchanges (TIE) to assess a countryʼs foreign trade is highly questionable. Stefan Valavanis-Vail, for example, insists that TIEs are based on the persistence of technical coefficients of capital and labor for each production unit. Factor theory explains the transition from a situation before the opening of international trade to a situation after the opening of international trade. Each country then undergoes a movement of specialization (according to its factor endowments), which modifies its production structure and technical coefficients.While the HO approach is dynamic, the W. Leontief test, on the other hand, fits into a static framework that is too rigid. For S. Valavanis-Vail, fixed technical coefficients do not allow the factorial model to be tested. However, Jean-Louis Mucchielli points out that TIE can be considered only if carried out over several years.

2. An incorrect understanding of production factors

03

An incorrect understanding of production factors

W. Leontief assumes that the United States is relatively better endowed with capital than the labor factor. This assumption amounts to saying that the author does not empirically verify the “relative scarcity of factors” referred to by E. Heckscher. Instead, he focuses his test on measuring the factorial “intensity” of products traded by the United States. As a result, the paradox may not be one since the HO model has not been tested here. To find out, we would have to measure the United States’ factor endowments before trade opening, but then a major problem arises: how can we measure the relative scarcity of a country’s factors of production in a situation of autarky when the latter is already open to international trade at the time of the test? We then come back to remarks of the same type as those developed by Jagdish N. Bhagwati concerning the limits of empirical tests of Ricardian theory.

3. An incorrect understanding of production factors

W. Leontief has also been criticized for excluding natural resources from his reasoning. For Boris C. Sweerling, Norman S. Buchanan, or Muhammed A. Diab, this "oversight" has the unfortunate consequence of introducing systematic errors in the determination of the factorial intensities of traded products. Jaroslaw Vanek, who makes the lʼhypothesis that natural resources are complementary to capital but that labor is substitutable for them, considers that American imports are so capital-intensive because they are intensive in “natural products.”W. Leontief acknowledges that natural resources are “invisible in all the tables, but very present as a third factor.”

3. An incorrect understanding of production factors

Furthermore, N. S. Buchanan challenges the very apprehension of the capital factor. W. Leontief calculates the value of capital directly or indirectly required to obtain one million dollars of production increase or the investment requirements. He obtains the indicators he requires, therefore, as coefficients of investment needs per unit of output, which are only equal to the coefficients of investment under the assumption of a uniform life of capital in all industries. This assumption of capital homogeneity (dictated by the concern for data aggregation) is precisely rejected by N. S. Buchanan, who points out that the rate of capital depreciation is far from identical for the different branches of the same country and also for the productive units of the same branches. Seev Hirsch, for his part, believes that the paradox would never have existed if capital intensity had been measured by finer indicators like the one proposed, for example, by Hal B. Lary (ratio, for each industry, of value-added - net of wage costs - to the number of workers employed).

3. An incorrect understanding of production factors

Finally, for J. Michael Finger, there is no independence between the factor intensities of exporting and importing industries. Indeed, by classifying industries according to their relative capital and labor intensity and their “trade balance,” the author finds no significant difference in factor intensity between importing and exporting industries. Therefore, the paradox’s invalidation or confirmation would only result from chance.

3. An incorrect understanding of production factors

04

Incorrect understanding of exports and imports

The previous point ties in with another criticism of the homogeneity assumption of branch production. For Thomas Balogh and J. M. Finger, for example, the aggregation of branches carried out in W. Leontief’s test is too large, which introduces a bias in the reasoning. Indeed, at such a level of aggregation, a single branch necessarily covers a group of very different products, some of which are exportable while others are not. Boris C. Sweerling adds that the choice of branches selected for the test is questionable. W. Leontief selects the seven most essential branches (in absolute value) in total U.S. imports and exports.

4. Incorrect understanding of exports and imports

But in doing so, it calculates the absolute importance of branches of activity in foreign trade and not the relative importance of foreign trade in relation to the production of each national branch, which introduces distortions (certain branches are strongly represented in total American exports or imports while, at the same time, the share of their exported or imported products is marginal in relation to their total production). To overcome this bias, B. C. Sweerling suggests choosing branches according to the structure of domestic consumption, an argument Erik Homayer made.

4. Incorrect understanding of exports and imports

Another criticism relates to the apprehension of imports by national productions supposedly “substitutable” for the former. W. Leontief uses such domestic substitutes for imports, as he does not possess the TEIs of the various foreign trading partner countries of the United States. Behind this choice lies a powerful assumption: foreign production structures are identical to American structures. But it is precisely this assumption that is questionable.

4. Incorrect understanding of exports and imports

The HO model suggests that the United States imports products for which it has a comparative disadvantage, i.e., labor-intensive products (a scarce and expensive factor in the United States). Usually, the United States would have had to abandon its domestic production completely corresponding to the imported products. W. Leontief's role in the critique is significant. He finds enough to calculate his million-dollar “substitutes” for imports, which are necessarily part of competitive domestic products that have been able to withstand foreign competition. It is not unreasonable to think that these “surviving” domestic products are precisely capital-intensive (a factor that is abundant and inexpensive in the United States). By choosing these first, W. Leontief introduces a significant bias into his test since he does not consider a substantial portion of real labor-intensive imports that have no equivalent in the United States. This bias is not just a theoretical concern. It is confirmed empirically by Edward E. Leamer, who compares the factor intensity of American production and consumption and shows that the former is relatively more capital-intensive than the latter.

4. Incorrect understanding of exports and imports

For his part, P. T. Ellsworth formulates a comparable critique, referring directly to one of D. Ricardo’s intuitions: differences in techniques as determinants of countries’ comparative advantages. For P. T. Ellsworth, it is because countries use different production methods that the assumption of homogeneity of production structures used by W. Leontief is false. This argument prefigures the neo-technological overtaking of the HOS model.

4. Incorrect understanding of exports and imports

05

A too-aggregated view of trading partners

W. Leontief sees the United States’ trading partners as an undifferentiated whole.As we have just seen, not all countries are alike. This underscores the need for a more nuanced view of trading partners, which would provide a clearer understanding of the dynamics of international trade.To verify the HO model, it would be wiser to disaggregate the block of trading partners trading partners.

5. A too-aggregated view of trading partners

It is what Masahiro Tatemoto and Shinichi Ichimura did, who had found, before disaggregation, a paradoxical result for Japan relatively better endowed with labor in the 1950s. Once the disaggregation is carried out (division into two major zones of Japan’s trading partners: developed and developing countries), W. Leontief’s paradox disappears insofar as their test shows that :

• Japan exports capital-intensive products to developing countries (¾ of total Japanese exports in 1951).
• And labor-intensive products to developed countries (¼ of total Japanese exports).

5. A too-aggregated view of trading partners

The authors thus highlight that “Japan’s place in the economy is intermediate between advanced and underdeveloped countries.”A series of bilateral tests confirmed their intuition: United States/Canada (Wahl), India/United States (Bharadwaj), and United Kingdom/United States (Hodd). In the case of France, the intermediate position is confirmed by the Bernard Lassudrie-Duchêne & Jean-Louis Mucchielli test. At the same time, it remains paradoxical for 11 bilateral relations in the Prouteau test.This hierarchy of comparative advantages according to a given countryʼs trading partners, initiated by M. Tatemoto and S. Ichimura, will be widely taken up in neo-factorial models.

5. A too-aggregated view of trading partners

06

An erroneous grasp of the historical context

W. Leontief's choice of the year 1947 for his test has been criticized. This year, being too close to the end of the Second World War, is atypical. It does not represent the 'normal' state of international trade, let alone American foreign trade. This underscores the importance of considering historical context when analyzing trade patterns. Moreover, during the immediate post-war period, the industries most heavily protected (tariff and non-tariff protection) by the American authorities were labor-intensive industries, limiting the imports of labor-intensive products.

6. An erroneous grasp of the historical context

It seems that in the 1950s, American domestic demand focused massively on capital-intensive durable goods. This would partly explain the significant need to import this type of goods to offset the shortfall in domestic production. This highlights the role of domestic demand in shaping trade patterns. But this last point is mainly open to debate. Hendrik S. Houtakker (1924-2008) has shown that American demand was very close to that of its trading partners during this period. J. L. Mucchielli, moreover, points out that the higher a country’s income, the greater the national demand for services, most of which are labor-intensive.

6. An erroneous grasp of the historical context

Read back

Answer the questions below

Want to test your knowledge?