UNPUBLISHED Work

Two unpublished studies estimate "normalized" US-PRC trade and propose comparative market methodologies for estimating the impact of extending MFN. Neither, however, actually calculates the MFN effect.

David Denny in an unpublished 1973 study proposes that with full trade normalization, the U.S. share of PRC exports to 24 developed countries should be the same as the U.S. share of other Asian developing countries' exports to the same 24 countries. That is, XOther Asia-US

XPRC-US

=

XPRC-Industrial world XOther Asia-Industrial World

The advantages of using other Asian countries as the reference suppliers are that the PRC and many of these countries are at approximately the same level of economic development, export similar products, and share similar transport costs to the United States. Denny suggests that the commodities he studied, disaggregated to the four-digit SITC level, be divided into categories according to whether the tariff differential is “negligible,” “moderate,” or “extreme." For all products with "negligible" tariff differentials, the normalized level of PRC exports to the United States (X'US-PRC) is calculated

Then for each ith product with a "non-negligible” tariff differential, the MFN effect on PRC export levels to the United States is

XUS-PRCUS-Other Asia Xind. World-Other Asia/

Nai-Ruenn Chen's approach 10 is essentially similar. He suggests that with full trade normalization, the PRC share of total U.S. imports should equal the PRC share of world imports, that is,

He is suggesting that PRC exports to the various industrialized countries are apportioned by country according to each (developed) market's proportion of total imports of all the industrialized countries.

Chen's proposed methodology for isolating the effect of lack of MFN tariff treatment is the same as Denny's.

Both Denny's and Chen's methodologies assume a fixed Chinese export supply Thus, a hypothetical increase in PRC exports to the United States due to MFN tariff extension implies a reallocation of exports among the PRC's other markets rather than any absolute increase in Chinese exports.

Denny, David L., "The Effect of Normalized Commercial Relations on PRC Exports to the U.S.," U.S. Department of Commerce, Bureau of East-West Trade, Washington, D.C., November 1, 1973.

10 Chen, Nai-Ruenn, "A Probabilistic Approach to the Study of U.S.-PRC Trade," U.S. Department of Commerce, Washington, D.C., October 1, 1973.

THE IMPACT OF U.S. MOST-FAVORED-NATION TARIFF TREATMENT ON PRC EXPORTS

BY HELEN RAFFEL, ROBERT E. TEAL, AND CHERYL MCQUEEN

A. 1976 top 50 PRC exports to The Twenty, with ad valorem equivalent
U.S. tariff rates.-

B. Estimating equations....

1. INTRODUCTION AND SUMMARY

848

849

This paper presents estimates of the amounts by which U.S. imports from the PRC would have exceeded their actual 1976 values had PRC products been assessed at U.S. most-favored-nation (MFN) tariff rates in that year. Our estimates are derived from a detailed econometric analysis of the tariff sensitivity of China's top 50 1976 hard-currency-earning exports, on a product-by-product basis. Because we do not expect that the product composition of China's overall hard currency export capabilities will alter rapidly in the near term, we believe that the 30-percent MFN-induced rise in total U.S. imports from the PRC predicted by this model may be taken as applicable to the near future, should the United States grant MFN to China during the next few years. The analysis of which products are likely to show the largest increases in exports to the United States should also be relevant for a number of years to come.

2

Our estimates of the impact of MFN were derived from a crosssectional multivariate regression model of the relative shares in specific product imports from the PRC of 19 hard currency areas (actually, 20 political units, Belgium and Luxembourg together constituting a customs union: the 20 importers will hereafter be called "The Twenty"). For each product, the share of each importing area

3

1 See the article by Kravalis, elsewhere in this volume, for in-depth discussion of China's export capabilities. 2 The products are the 1976 top 50 PRC hard currency exports at the 4- to 5-digit SITC level. SITC is the standard international trade classification nomenclature used by the United Nations. The highest level of disaggregation is the 4-digit level for some types of products and the 5-digit level for others. Successive digits added to a given numerical code signify successive subcategories of the class with the same initial digits. There are a total of 56 2-digit divisions, subdivided down to 1,312 items at the 4- to 5-digit level.

"The Twenty" are: Australia, Austria, Belgium, Canada, Denmark, the Federal Republic of Germany, France, Hong Kong, Indonesia, Italy, Japan, Luxembourg, Malaysia, the Netherlands, Norway, Singapore, Sweden, Switzerland, the United Kingdom and the United States.

was assumed to be explicitly dependent upon several variables, among them the tariff it raises against the particular product coming from the PRC. After the parameters of the regression relationship for a given product were estimated from actual data for The Twenty, the U.S. MFN rate, along with the U.S. values of the other independent variables, was inserted into the estimating equation and the revised U.S. share calculated."

The results are shown in table 1 of section 3. They indicate that U.S. imports from the PRC would have been about 30-percent higher than the actual 1976 value of $202 million had China enjoyed MFN status. Increases would have occurred primarily among the following product groups, in the order given: household linens, basketwork, silk fabrics, medicaments, porcelain and chinaware, underwear, cotton fabrics, and dried vegetables.

Although it is not known whether China would seek generalized system of preferences (GSP) status from the United States, we nevertheless inserted the still lower tariff rates associated with GSP into the relevant equations and calculated the effect of GSP on U.S. imports. The result was an estimated 34-percent rise in 1976 imports, not much larger than the MFN effect alone. The principal beneficiary of GSP would have been basketwork, followed, in absolute size of increase over the MFN effect, by medicaments and preserved fruit and nuts.

It should also be mentioned that the PRC (along with the Soviet Union) still suffers from an absolute U.S. ban on seven types of furskins. Because no other country invokes this ban, we were unable to statistically determine its quantitative impact. But since weasel and kolinsky (mink) are major PRC exports, we would expect that the removal of the ban would lead to significant U.S. fur-skin imports.

The estimated values of U.S. import responses to MFN and GSP presented in this paper may be biased downward. This is because the entire MFN-induced increase in imports is in each instance solely that attributed to a hard-currency-maximizing redirection of trade flows among the importing areas. Such an MFN-induced reaction would certainly benefit the PRC by augmenting its net hard currency earnings from a given total of exports, and its import capabilities would thereby be strengthened. If, in the longer run, the enhanced profitability of selling in the U.S. market should induce China to raise its total hard currency exports above what they otherwise would have been, and/or to change its export mix to favor those products which would benefit most from U.S. MFN, exports to the United States could be greater. (See section 4 for further discussion of this point.)

Finally, our results indicate that the geographical distribution of PRC exports is more tariff-sensitive than has sometimes been sug

♦ For a more detailed description of the model and the calculating procedure, see section 2. For a discussion of the derivation of the model used in this paper, a comparison with other models heretofore used to evaluate MFN impact, and an application of the model to Eastern European countries and the U.S.S.R., see Raffel, Rubin, and Teal, "The MFN Impact on U.S. Imports from Eastern Europe" in John P. Hardt, ed., East European Economies Post-Helsinki, Joint Economic Committee of Congress (Aug. 25, 1977) pp. 1396-1427. Under the GSP system of tariffs, which came into effect on January 1, 1976, zero tariff rates are assessed on a specified list of products imported by the United States from those countries which have been accorded "developing country" status by Presidential designation.

gested. There were, however, some notable exceptions. The greatest surprise was that none of the textile outergarment items showed statistically significant tariff sensitivity, although both of the underwear items and all of the fabrics did. On the other hand, distance from the PRC was a highly significant explanatory variable for the distribution of outergarments, and this suggests a tentative reason for the outergarment anomaly with regard to tariffs. If distance is a proxy for, among other things, cultural relatedness, then it may be that the style and quality of Chinese clothing (apart from underwear) is still too culture-bound to be widely acceptable far from the areas of traditional trade and large overseas Chinese populations."

Distance from the PRC was a highly significant variable in all but two of our equations, and this strong relationship was evidenced both with and without the inclusion of Hong Kong. In a previous study of the MFN impact on the East European countries and the U.S.S.R., distance was rarely a significant variable. As we have suggested, it may be that the generally strong inverse relationship between sales and distance of purchaser indicates that China is still relatively inhibited on the world economic scene by cultural isolation.

2. ESTIMATION OF MFN AND GSP IMPACT

A. The Model

The model used in this paper is identical with the model developed for an earlier paper in which the estimated effects of MFN on U.S. imports from various Eastern European countries were presented.' The relative shares of 19 importing areas ("The Twenty" are here estimated for a selected group of PRC export products. For each product, the share of each importing area is assumed to depend upon (1) its share in total world exports of that product to The Twenty, (2) the tariff rate raised by the importing area against the given product from the PRC, (3) the presence or absence of quantitative restrictions (QRs) on that product specifically directed against the PRC, and (4) the distance between the importing area and the PRC (both air and maritime distances were used, alternatively).

Four forms of the relationship were tested, two linear and two log linear. The theoretically preferred form of each pair imposed a coefficient of unity upon the share from the world, and in most cases

An International Trade Commission study, Implications for U.S. Trade of Granting Most-Favored Nation Tariff Treatment to the People's Republic of China (May 1977), concluded that the product mix of PRC exports to the United States was not highly correlated with U.S. tariff differentials between the non-MFN and MFN rates. This may be true, but it does not follow from this fact alone that MFN would have little impact. Whether a product will be shipped to a specific market depends, among other things, upon the actual tariff rate in that market and upon the size of that tariff rate relative to the rates assessed in other world markets in which the product might instead be sold. The economic profitability calculation will not depend upon a comparison of the actual tariff rate with a hypothetical, unavailable alternative.

The model presented in this paper attempts to estimate PRC exports to various world markets as a funetion of the actual tariff rates in those markets. Since, of the products tested, a large proportion in fact proved to be tariff-sensitive, and, of these, the value of PRC exports to the United States predicted by the regres sion parameters were generally very close to the actual values, we conclude that the PRC is not careless of tariff rates in making its export decisions. Nevertheless, we also conclude on the basis of our quantitative estimates that the increment in PRC exports to the United States consequent upon the receipt of MFN would not be large.

Dr. Nai-Ruenn Chen suggested to us that the overseas Chinese populations of The Twenty might be a significant explanatory variable for some of our products, but we were unable to obtain the data. See footnote 4 for reference.

See footnote 4.

10 These 19 importing areas-"The Twenty" political units—are itemized in footnote 3. Belgium and Lux. embourg together comprise a single importing area, hence the total number of observations for regression is 19. 11 See footnote 18 of Raffel, Rubin, and Teal, op. cit.

where the best fit equation did not meet this condition the coefficient was close to unity. Our estimating equations were:

(2.1) CSip=A+A1WS1p+A2D¿+A3(l+tip)+AQR1p (2.2) CS-WS1p=A+A2D1+A3(1+tip)+AQRP

(2.3) LnCS1, A2+A1LnWSip +A2LnD1+A3Ln(1+t1p)+A,LnQR {p (2.4) LnCS-LnWS=A+A2LnD+A3Ln(1+tip)+ALNQR CS=fractional share of importing area i in PRC exports of product p to The Twenty;

WS=fractional share of importing area i in world exports of product to The Twenty;

D-air distance or, alternatively, maritime distance between Peking and the capital of importing area i;

=

tip ad valorem equivalent tariff rate (expressed as a fraction) imposed by importing area i on PRC product p;

QR1pa dummy variable expressing the existence or absence of any type of quantitative restriction specifically directed against PRC product p by importing area i.12

These equations are assumed to provide values of the relative shares (CS). Thus, several of the manipulations do not preserve the sum of shares at unity (the log forms do not, as well as the subsequent insertion of MFN or GSP tariff values into the U.S. equation). Where the sum is not preserved, all values of CS were scaled proportionately to restore the sum to unity.

The world shares (WS) are taken to represent relative strengths of demand in the different importing areas. Distance is a proxy for freight charges, but it is also a proxy, albeit imperfect, for political and cultural relations, an index of length of acquaintance and degree of historical relatedness, as well as of ease of commercial contacts. Both air and maritime distances were therefore used, as alternatives. Maritime distance more frequently appeared in the equation of best fit (17 times for maritime distance, 6 times for air distance).

B. The Estimating Procedure

The set of products specifically examined was the top 50 PRC exports (by dollar value, at the 4- to 5-digit SITC level) to The Twenty in 1976.13 These top 50 items accounted for 64 percent of 1976 PRČ exports to The Twenty. From this sample of 50, we eliminated from consideration of MFN impact those items for which the difference between the discriminatory rate and the MFN rate did not exceed 5 percent. For each of the 27 remaining products, actual data for The Twenty for the year 1976 were entered into the estimating equations, Eqs. (2.1)-(2.4), plus variants of these four equations that omitted one or more of the independent variables. Ordinary least squares estimates

14

12 The use of a 2-valued dummy variable leaves a great deal to be desired. It does not distinguish between the various types of Q Rs nor does it indicate how stringently a QR is applied. The QR term in our regression equations turned out to be significant only for the cotton fabric categories (see Appendix B).

13 See Appendix A for the itemized list of products, their 1976 export values to The Twenty and to the United States, and the U.S. MFN and non-MFN tariff rates.

14 The ad valorem equivalent tariff rates used in this study are provided in Appendix A.