Import Protection

Since the early 1960s, import protection in Latin America has resulted from the interdependence of tariffs and/or quantitative restrictions with some regional devices. Many times the latter would affect the ranking of incentives underlying the structure of national legislation.

The Tariff. The national tariff has been the principal instrument used in Latin America to control imports. The pattern of rates can be summarized as follows: for produced goods the rates are higher for consumer goods than for intermediate or capital goods, white for complementary imports the rates are lower. Another aspect of the structure of protection is related to tariff exemptions provided by generous fiscal incentives. Given their size, the importance of these “tax expenditure” aspects should be stressed; the proportion of tariff-exempt imports was in 1979/80 the highest for Brazil (74 percent), followed by Argentina (44 percent), Mexico (61 percent), and Colombia (42 percent) (Berlinski, Camelo and Pazmino, 1984). The findings presented in Table 2 also show the large difference between the average tariff paid, which ranged from 7 to 14 percent, compared with corresponding ad valorem rates for imports outside the region (excluding imports from the Latin American Free Trade Association—now the Latin American Integration Association—countries) of 18 to 28 percent.

Table 2.

Selected Latin American Countries: Tariff-Exempt Imports, 1979–80

Source: Berlinski, Camelo, and Pazmino (1984). Note: Imports were used as weights for estimating tariff averages.

Table 2.

Selected Latin American Countries: Tariff-Exempt Imports, 1979–80

	Argentina	Brazil	Colombia	Mexico
	In millions of U.S. dollars
C.i.f. value of imports	6,700	25,614	4,663	17,517
	In percent
Tariff paid	10	7	14	7
LAFTA imports	14	3	5	4
LAFTA tariff	5	7	7	7
Proportion of imports on which tariffs are paid	42	23	53	35
Tariff rates	22	28	26	18
Proportion of tariff-exempt imports	44	74	42	61

Source: Berlinski, Camelo, and Pazmino (1984). Note: Imports were used as weights for estimating tariff averages.

View Table

Table 2.

Selected Latin American Countries: Tariff-Exempt Imports, 1979–80

	Argentina	Brazil	Colombia	Mexico
	In millions of U.S. dollars
C.i.f. value of imports	6,700	25,614	4,663	17,517
	In percent
Tariff paid	10	7	14	7
LAFTA imports	14	3	5	4
LAFTA tariff	5	7	7	7
Proportion of imports on which tariffs are paid	42	23	53	35
Tariff rates	22	28	26	18
Proportion of tariff-exempt imports	44	74	42	61

Source: Berlinski, Camelo, and Pazmino (1984). Note: Imports were used as weights for estimating tariff averages.

A word should be added about the beneficiaries of these tariff exemptions, especially regarding the effect on domestic prices of close foreign substitutes or the reduction in import costs. The former corresponds to a case in which the size of imports is high, so that the imports may overflow to the open market and in so doing affect domestic prices downward; the latter corresponds to the more general case of dual markets, where cost savings on intermediate and capital goods increases the rate of return for the beneficiaries of tariff exemptions.

National tariffs were designed using as criteria the type of good involved and the domestic supply. Lower rates were applied to intermediate and capital goods, which basically are not produced in each country, while the highest rates were applied to consumer goods produced within domestic markets.

Nontariff Restrictions. National legislations introduced several instruments that affected the incentive content of the ad valorem tariff rates. This was done mainly by quantitative restrictions and import surcharges of different kinds, including the more common case of imposing consumption taxes, with lower rates on domestic supply than on competing imports. In addition, if for some products free trade within a subset of countries existed (owing to bilateral or multilateral agreements), those additional restrictions might have increased the level of protection provided to regional partners.

The existence of quantitative restrictions implies that the tariff may not be the relevant instrument of protection. In the absence of other instruments and if tariffs are not redundant, it is known that tariffs allow the domestic price of an imported good to increase above the border price level. But in several cases the implicit tariff might be different from the explicit one. The implicit tariff is reflected in the behavior of entrepreneurs; the explicit tariff reflects the nominal legal rates corrected (when required) for the existence of other price instruments (e.g., official prices of imports) or the ad valorem equivalent of other devices such as pre-import deposits. If quantitative restrictions become relevant, it will be found that the implicit tariff would be greater than the explicit one.

We have seen that the tariff was one among the instruments used for import protection. In many cases the effect of those nontariff restrictions set at the national level was of great importance. Therefore, they have to be seen together in order to have a correct understanding of the price signals concerning the profitability of investments. In addition, replacing quantitative restrictions with tariffs would provide a clear ceiling to domestic prices of import-competing goods; otherwise, implicit protection would fluctuate depending on shocks, most of them imposed by exogenous macrovariables.

Export Taxation (Protection)

Export taxation has been a traditional way of providing revenue to national governments, acting at the same time as an income tax substitute for groups generally hard to tax. Export taxes were also used to prevent windfall gains when export proceeds in domestic currency showed large fluctuations owing to increases in international prices or changes in the exchange rate. Also, setting nominal exchange rates in countries with a large dispersion in international competitiveness may require a complementary export tax. This is especially true when those products have some weight in the cost of living; that is, export taxes would act as a hidden subsidy on domestic consumption. In some countries, in addition to the taxation of traditional exports, quantitative restrictions have also been imposed. This seems to be related mainly to trying to keep foreign exchange proceeds under control.

One of the latest steps in the introduction of trade intervention measures was related to export promotion laws. These are generally intended to provide additional foreign exchange in trade regimes with high incentives for sales to the domestic market. While there are successful experiences, in several countries the laws were enacted too late and their coverage was too limited to have resulted in the removal of the discrimination against exports. We will come back to this in a later section.

The comparison of national regulations regarding export promotion devices shows that they consist mainly of the following:

(a) Drawbacks, the general idea of which is to provide neutrality to exporting firms, regarding the system of protection through tariff suspension (or rebate), mainly on inputs and capital goods, (b) Rebates on the f.o.b. value or value added, where nontraditional exports are defined by providing either a negative list of traditional exports or a positive list of what should be considered nontraditional. The rebates are in the form of cash refunds, or endorsable certificates that may have some maturity period before they can be used to pay taxes, (c) The provision of pre-export and post-export financing, in many cases with subsidized rates and long financing periods, which then represent an important component of the promotion package.

Export promotion schemes are very similar; basically they provide tariff exemptions on inputs and capital goods or rebates on non-traditional exports (and/or their increments) and provide pre- and post-export financing. But this might reflect only the intentions of policymakers. In some countries the incentive is (in absolute and relative terms) insufficient to overcome the costs imposed by the trade regime; in others, foreign exchange regulations may in the short run call into question the usefulness of an instrument like export protection, since it may have replaced, for example, the share of export proceeds sold in a free foreign exchange market. This means overlooking the volatility of nominal exchange rates compared with the need of having schemes that are able to remove the tax content of traded inputs. We shall come back to this subject when the sources of anti-export bias are discussed.

Multilateral Trade Arrangements

The institution of several multilateral agreements since the 1960s was intended to overcome trade limitations imposed by national boundaries. Among the agreements, we will mention as examples only the Central American Common Market (CACM) and the Latin American Free Trade Association (LAFTA). By negotiating a common external tariff and free trade between the countries involved, a strong incentive to existing activities related to the enlargement of domestic markets was implied in the CACM agreement. The analysis of benefits and costs of economic integration in Central America made by Cline (1978), in the tradition of an earlier similar work by Balassa (1975) concerning the European Common Market, showed little importance for the trade diversion measured, which I think underestimates the real magnitude. Recent negotiations of a new common external tariff (excluding Honduras) has brought back some of the earlier discussions, but apparently the benefits associated with the enlargement of domestic markets is still a strong reason for pursuing that kind of policy in small countries.

Regarding LAFTA, the original idea was based on two basic principles: reciprocity and the most-favored-nation clause, the first to protect the least developed countries (discriminatory), the second to extend to all members any tariff advantage granted. The gradual elimination of trade restrictions among the members during the transition period of twelve years did not take place; also, some of the instruments as designed did not prove to be practical. This agreement was later replaced by the Latin American Integration Association, with two basic instruments: regional preferences and negotiated bilateral agreements that might be granted to other members. Table 2 provides figures of the tariff rate paid, as well as the limited amount of trade, measured from the import side. The average tariff rate on imports from the region varied between 5 and 7 percent, and the size of trade was 3 percent for Brazil and 14 percent for Argentina. Here, one of the main problems has been the erosion of the preference margins negotiated in each case because of existing tariff exemptions on imports from outside the region, mainly on government purchases and several promotion schemes on particular economic activities.

Exchange Rate Regulations

The best way of looking at the interaction of trade policies and exchange rate regulations is to see them as a system of multiple exchange rates. At one end, the lowest nominal exchange rate would correspond to traditional exports where export taxes are imposed, and at the other end, final goods (mainly consumer goods) would correspond to imports substituted under high tariff barriers; somewhere in between would be nominal exchange rate for intermediate products and capital goods and also the rate on promotion of non-traditional exports.

Trade regimes in many Latin American countries are associated with overvalued exchange rates, which are made possible by the high international competitiveness of traditional exports, as well as with high tariff levels and/or quantitative restrictions on the import side. In addition, the existence of tariff redundancy, as well as of quantitative restrictions, may affect real exchange rates, owing to the fact that implicit protection of import-competing activities becomes endogenous and therefore subject to shocks, many of them imposed by macroeconomic policies.

Sources of Bias

Local producers have to buy inputs (from local or imported sources) that are protected at prices higher than the border level. In the absence of export rebates, the result would be negative effective rates of protection for exports. At the same time, given import restrictions, sales in the domestic market are made at prices higher than the international level, following generally the made-to-measure principle starting from input protection. Taking into account these effects, one can consider two sources of bias against exports (leaving aside adjustments of overvalued exchange rates): the absolute bias related to export taxation on the input side, and the relative bias owing to higher nominal rates of protection applied to domestic sales as compared to exports.

The most frequent case to be found in Latin America regarding the tradable sector can be broken down into three activities: agriculture, agro-industries (Industry 1), and the remaining industries (Industry II). Absolute bias against exports measured by negative effective rates would be the highest in Industry II in the absence of export promotion, because of the tax imposed by input protection. The agro-industries internalize the international competitiveness of agriculture, receiving a high (many times the highest) positive effective protection for sales to the domestic markets; that is, in such a situation, even a uniform tariff on manufacturing will not produce uniformity in effective rates because Industry II is securing its tradable inputs at higher than border prices. In other words, while under the uniform nominal tariff hypothesis relative nominal bias against exports is the same for all industries, the effective bias is higher for Industry II.

In trying to provide estimates of absolute anti-export bias, Table 3 shows two examples of structures of protection that correspond mainly to the nonnatural resource-based industries of Argentina (1977) and Brazil (1981); they basically correspond to the so-called Industry II, since sectors excluded from the comparison are mainly primary production and processed foods. By computing the realized higher costs of inputs and an estimate of the corresponding protection to exports² the effective rates were obtained. The frequency distribution of the rates show that for Argentina only one third of the activities had positive rates, while 100 percent of Brazilian sectors had positive rates and about two thirds of them were in the range of 26 to 50 percent.

Table 3.

Argentina and Brazil: Frequency Distribution of Absolute Anti-Export Bias

Source: Based on estimates of realized effective rates of protection on exports (Berlinski, 1977; and Tyler, 1983. Note: Effective protection estimates are conventional partial equilibrium figures of realized protection using the Corden criteria for the treatment of nontraded inputs. No adjustment for overvaluation was introduced, since the figures involved for those years were very close (about 20 percent). To provide a similar basis for comparison of both sets of estimates, some Brazilian sectors were not included here; they correspond mainly to natural resource-based activities—primary production and processed foods. Estimates for Argentina do not include the subsidy content of pre- and post-export financing, owing partly to data shortcomings but mainly to frequent changes in procedures and rates.

Table 3.

Argentina and Brazil: Frequency Distribution of Absolute Anti-Export Bias

	Argentina (1977)			Brazil (1981)
Rates	Total	Intermediate Goods	Capital Goods	Total	Intermediate Goods	Capital Goods
–49 to –25	7.5	9.1	8.3	0.0	0.0	0.0
–24 to 0	60.0	68.2	41.7	0.0	0.0	0.0
1 to 25	32.5	22.7	50.0	22.9	28.0	0.0
26 to 50	0.0	0.0	0.0	62.9	56.0	100.0
> = 51	0.0	0.0	0.0	14.3	16.0	0.0
Total (in percent)	100.0	100.0	100.0	100.0	100.0	100.0
Sectors	40	22	12	35	25	7
Mean	–6.2	–9.7	–2.2	37.0	37.3	35.9
Standard deviation	12.2	12.1	13.0	25.7	29.8	6.0
Dispersion	195.2	125.1	580.2	69.4	80.0	16.7

Source: Based on estimates of realized effective rates of protection on exports (Berlinski, 1977; and Tyler, 1983. Note: Effective protection estimates are conventional partial equilibrium figures of realized protection using the Corden criteria for the treatment of nontraded inputs. No adjustment for overvaluation was introduced, since the figures involved for those years were very close (about 20 percent). To provide a similar basis for comparison of both sets of estimates, some Brazilian sectors were not included here; they correspond mainly to natural resource-based activities—primary production and processed foods. Estimates for Argentina do not include the subsidy content of pre- and post-export financing, owing partly to data shortcomings but mainly to frequent changes in procedures and rates.

View Table

Table 3.

Argentina and Brazil: Frequency Distribution of Absolute Anti-Export Bias

	Argentina (1977)			Brazil (1981)
Rates	Total	Intermediate Goods	Capital Goods	Total	Intermediate Goods	Capital Goods
–49 to –25	7.5	9.1	8.3	0.0	0.0	0.0
–24 to 0	60.0	68.2	41.7	0.0	0.0	0.0
1 to 25	32.5	22.7	50.0	22.9	28.0	0.0
26 to 50	0.0	0.0	0.0	62.9	56.0	100.0
> = 51	0.0	0.0	0.0	14.3	16.0	0.0
Total (in percent)	100.0	100.0	100.0	100.0	100.0	100.0
Sectors	40	22	12	35	25	7
Mean	–6.2	–9.7	–2.2	37.0	37.3	35.9
Standard deviation	12.2	12.1	13.0	25.7	29.8	6.0
Dispersion	195.2	125.1	580.2	69.4	80.0	16.7

Source: Based on estimates of realized effective rates of protection on exports (Berlinski, 1977; and Tyler, 1983. Note: Effective protection estimates are conventional partial equilibrium figures of realized protection using the Corden criteria for the treatment of nontraded inputs. No adjustment for overvaluation was introduced, since the figures involved for those years were very close (about 20 percent). To provide a similar basis for comparison of both sets of estimates, some Brazilian sectors were not included here; they correspond mainly to natural resource-based activities—primary production and processed foods. Estimates for Argentina do not include the subsidy content of pre- and post-export financing, owing partly to data shortcomings but mainly to frequent changes in procedures and rates.

Within this general picture, Argentina’s capital goods industry (with fewer observations) shows a better position but with a higher dispersion relative to intermediate goods. In Brazil, average rates are about the same for both types of goods, but with a substantially lower dispersion in capital goods (also with fewer sectors). The subsidy content of export financing for manufacturing was 8 percent in Brazil (see Tyler, 1983), which was not computed in the figures for Argentina, not because of lack of data but mainly because of the frequent changes in procedures and rates. This does not seem to affect the basic message of stronger export promotion in Brazil but might reduce the bias measure in favor of Argentine capital goods.

As mentioned earlier, another source of anti-export bias is related to the lack of symmetry between nominal rewards for sales to the domestic market and to the export market. The relevant calculations for Argentina and Brazil are presented in Tables 4 and 5, where the bias measure was the ratio between protection coefficients (1 + rate), the numerator being the ratio of protection to domestic sales. The first set of columns corresponds to the nominal bias coefficient, which for Argentina (Table 4) shows the highest concentration of activities in the brackets of up to 1.5 (between 80 and 90 percent); this percentage is strongly reduced when we look at the relevant effective coefficients, which incorporate both bias measurements. The average bias for capital goods is lower, with a higher dispersion relative to the coefficients for intermediate goods.

Table 4.

Argentina: Frequency Distribution of Relative Anti-Export Bias, 1977

Source: See Table 3. Note: The bias against exports is the ratio between protection coefficients (1 + rate) to domestic sales and exports, while nominal and effective refer to the corresponding protective rates.

Table 4.

Argentina: Frequency Distribution of Relative Anti-Export Bias, 1977

Bias Coefficient	Nominal			Effective
	Total	Intermediate Goods	Capital Goods	Total	Intermediate Goods	Capital Goods
<1.0	17.5	9.1	16.7	17.5	9.1	16.7
1.0 to 1.5	70.0	81.8	66.7	35.0	36.4	50.0
1.6 to 2.0	12.5	9.1	16.7	27.5	36.4	16.7
2.1 to 2.5	0.0	0.0	0.0	7.5	4.5	8.3
2.6 to 3.0	0.0	0.0	0.0	7.5	9.1	8.3
3.1 to 3.5	0.0	0.0	0.0	2.5	4.5	0.0
> =3.6	0.0	0.0	0.0	2.5	0.0	0.0
Total (in percent)	100.0	100.0	100.0	100.0	100.0	100.0
Sectors	40	22	12	40	22	12
Mean	1.3	1.3	1.2	1.7	1.7	1.4
Standard deviation	0.3	0.2	0.3	0.9	0.6	0.6
Dispersion	21.8	17.8	24.0	54.2	38.4	42.0

Source: See Table 3. Note: The bias against exports is the ratio between protection coefficients (1 + rate) to domestic sales and exports, while nominal and effective refer to the corresponding protective rates.

View Table

Table 4.

Argentina: Frequency Distribution of Relative Anti-Export Bias, 1977

Bias Coefficient	Nominal			Effective
	Total	Intermediate Goods	Capital Goods	Total	Intermediate Goods	Capital Goods
<1.0	17.5	9.1	16.7	17.5	9.1	16.7
1.0 to 1.5	70.0	81.8	66.7	35.0	36.4	50.0
1.6 to 2.0	12.5	9.1	16.7	27.5	36.4	16.7
2.1 to 2.5	0.0	0.0	0.0	7.5	4.5	8.3
2.6 to 3.0	0.0	0.0	0.0	7.5	9.1	8.3
3.1 to 3.5	0.0	0.0	0.0	2.5	4.5	0.0
> =3.6	0.0	0.0	0.0	2.5	0.0	0.0
Total (in percent)	100.0	100.0	100.0	100.0	100.0	100.0
Sectors	40	22	12	40	22	12
Mean	1.3	1.3	1.2	1.7	1.7	1.4
Standard deviation	0.3	0.2	0.3	0.9	0.6	0.6
Dispersion	21.8	17.8	24.0	54.2	38.4	42.0

Source: See Table 3. Note: The bias against exports is the ratio between protection coefficients (1 + rate) to domestic sales and exports, while nominal and effective refer to the corresponding protective rates.

For Brazil (Table 5), the concentration of activities in the same two brackets (up to 1.5) is also high (between 80 and 90 percent) for nominal as well as effective bias. But the composition within them is different: about 50 percent of the sectors are in the <1.0 bracket, which means that nominal as well as effective incentives to exports are higher than those realized for sales to the domestic market. The average bias shows fewer differences between types of goods but also a higher dispersion for capital goods.

Table 5.

Brazil: Frequency Distribution of Relative Anti-Export Bias, 1981

Source: See Table 3. Note: The bias against exports is the ratio between protection coefficients (1 + rate) to domestic sales and exports, while nominal and effective refer to the corresponding protective rates.

Table 5.

Brazil: Frequency Distribution of Relative Anti-Export Bias, 1981

Bias Coefficient	Nominal			Effective
	Total	Intermediate Goods	Capital Goods	Total	Intermediate Goods	Capital Goods
< 1.0	45.7	48.0	57.1	48.6	52.0	57.1
1.0 to 1.5	42.9	44.0	28.6	34.3	32.0	28.6
1.6 to 2.0	11.4	8.0	14.3	14.3	16.0	0.0
2.1 to 2.5	0.0	0.0	0.0	2.9	0.0	14.3
2.6 to 3.0	0.0	0.0	0.0	0.0	0.0	0.0
3.1 to 3.5	0.0	0.0	0.0	0.0	0.0	0.0
> =3.6	0.0	0.0	0.0	0.0	0.0	0.0
Total (in percent)	100.0	100.0	100.0	100.0	100.0	100.0
Sectors	35	25	7	35	25	7
Mean	1.0	1.0	1.1	1.1	1.0	1.2
Standard deviation	0.3	0.3	0.4	0.5	0.4	0.6
Dispersion	31.5	29.9	37.3	43.9	41.9	52.9

Source: See Table 3. Note: The bias against exports is the ratio between protection coefficients (1 + rate) to domestic sales and exports, while nominal and effective refer to the corresponding protective rates.

View Table

Table 5.

Brazil: Frequency Distribution of Relative Anti-Export Bias, 1981

Bias Coefficient	Nominal			Effective
	Total	Intermediate Goods	Capital Goods	Total	Intermediate Goods	Capital Goods
< 1.0	45.7	48.0	57.1	48.6	52.0	57.1
1.0 to 1.5	42.9	44.0	28.6	34.3	32.0	28.6
1.6 to 2.0	11.4	8.0	14.3	14.3	16.0	0.0
2.1 to 2.5	0.0	0.0	0.0	2.9	0.0	14.3
2.6 to 3.0	0.0	0.0	0.0	0.0	0.0	0.0
3.1 to 3.5	0.0	0.0	0.0	0.0	0.0	0.0
> =3.6	0.0	0.0	0.0	0.0	0.0	0.0
Total (in percent)	100.0	100.0	100.0	100.0	100.0	100.0
Sectors	35	25	7	35	25	7
Mean	1.0	1.0	1.1	1.1	1.0	1.2
Standard deviation	0.3	0.3	0.4	0.5	0.4	0.6
Dispersion	31.5	29.9	37.3	43.9	41.9	52.9

Source: See Table 3. Note: The bias against exports is the ratio between protection coefficients (1 + rate) to domestic sales and exports, while nominal and effective refer to the corresponding protective rates.

The comparison of anti-export biases between activities of Industry II type (nonnatural resource-based) for Argentina and Brazil shows that in Argentina there is stronger input discrimination (absolute bias) as well as an important differential reward to domestic sales (relative bias).

Changes in Anti-Export Bias

When the effect on growth of the import substitution industry slowed down, governments introduced several measures to promote nontraditional exports. As we have seen in an earlier section, the measures are generally designed to compensate the by-product effect of trade regimes on export profitability. Thus, drawbacks allow firms to secure inputs at international prices; rebates on f.o.b. values are also intended to compensate the high cost of inputs and in general they did not go beyond “tax” removal. Using our classification of bias, these devices are oriented toward the reduction of absolute bias.

Foreign exchange regulations and the setting of nominal exchange rates have also emphasized absolute bias, disregarding that in the medium run the relative bias measure is important in changing the profit equation of the firm. Otherwise, exports will be driven mainly by domestic recession, and the moment the recession is over the trade balance will be worse because of increased imports associated with the increase in domestic activity and decreased exports for the same reason. This would not be true if there was more symmetry between prices of products sold in the domestic market and those for the export market, and capacity may even be stretched to satisfy the demands of both markets.

A recent example of testing this positive correlation between incentives and the growth of exports can be seen in Balassa and others (1986) concerning Greece and the Republic of Korea. On the other hand, Teitel and Thoumi (1986), in trying to correlate the high rates of growth in manufacturing exports of Argentina and Brazil during the 1970s, did not find this conventional association. But the data on Brazilian incentives that they used correspond to legal nominal rates for 1977 (Tyler, 1983), compared with realized protection rates computed by the same author for 1980–81 and on which Tables 3 and 5 are partly based. This might change Teitel and Thoumi’s conclusions, given the generalized high redundancy of tariffs found (Tyler, 1985). The result for Argentina is puzzling because the additional evidence provided in the earlier portion of this section is that the anti-export bias was still there. We were also aware that not including the interest subsidy content in those calculations might have influenced the results, especially for capital goods where the financing period is longer. The hypothesis developed by the authors is basically related to learning by doing, first selling to domestic markets and at a later stage becoming exporters. It is my impression that in testing the association of incentives and exports a high explanatory power would be found in the actual practice of marginal cost pricing for exports, in activities where competition among the few allowed them to charge fixed costs to domestic buyers.

Going back to the main discussion, absolute bias would have been reduced by export rebates, the provision of inputs at international prices, or by changes in the real exchange rate. While relative bias might have been reduced by the first measure if its magnitude went beyond “tax” removal, it would have remained about the same for any real devaluation given its protective effect on import-competing activities. Another way of reducing relative bias is through changes in import protection, which is discussed in the next section.

Initial Conditions

The initial conditions describe the points of reference for the evaluation of the reform proposals. This is important when the adjustments required imply a strong divergence from present conditions. Also, recent studies made of the Southern Cone liberalization programs show that credibility is crucial to success (Corbo and de Melo, 1985). The initial conditions should include not only the structure of protection categorized by anti-export bias (see preceding section) and high dispersion of protective rates but also the institutional framework and the functioning of markets for products and factors.

In Latin America it has become commonplace to discuss in limited circles some of the underlying problems of the present trade regimes: (a) the conflict between agriculture and industry, (b) the functioning of the bankruptcy laws, (c) the generous industrial promotion mechanisms, and (d) imperfections of markets. This seems to be what Keynes (quoted by Gardner, 1969, p. xi) has called inside opinion: “For there are, in the present time, two opinions; not, as in former ages, the true and the false, but the outside and the inside; the opinion of the public voiced by the politicians and the newspapers, and the opinions of the politicians, the journalists, and the civil servants, upstairs and backstairs and behind-stairs, expressed in limited circles.”

It is an open question as to what extent this apparent dichotomy (outside-inside opinions) is real or simply represents the way pressure groups behave in order to defend their interests. In many cases I have found a rhetorical mix of vested-interest string-pulling and the existence of an old script followed by the bureaucracy, which is supposed to make decisions involving private rates of return.

Going back to the problems mentioned above, the conflict between agriculture and industry is best expressed by the process of setting the nominal exchange rates. This becomes difficult once relative domestic prices for some goods differ substantially from border prices. The distance between nominal exchange rates for agriculture (adjusted for export taxes) and those for industry-cum-protection is such that setting the exchange rate raises permanent tension between technical and political decisions.

Next, the bankruptcy laws seem to be functioning only for medium-sized and small firms—a strong restriction for any intended rationalization of the incentive system. The problem is more difficult when, in addition, projects are approved under the general belief that for several reasons protection is to be maintained for long periods of time. As usual, it is difficult to generalize, but my impression is that this protection network soon becomes a source of quasi-rents.

Another common feature is the implementation of a generous system of “tax expenditures,” which supported private capital formation for long periods, extending private rates of return beyond conventional pay-off periods. In most cases, prices used in these evaluations were not even corrected to allow for the discrepancy between private and social opportunity costs, which describes the worst of the worlds.

Concerning the effects of market imperfections on products, firms were allowed a liberalized period of adjustment to the new situation. The asymmetry of rewards to domestic and export markets resulted in firms reducing exports based on reduced protection to domestic markets where the markets behaved as price leaders (Berlinski, 1982). Regarding the markets for factors of production, the effects of “tax expenditures,” subsidized credit lines, and social security taxes in favoring more capital-intensive techniques are well known.

Corbo, de Melo, and Tybout (1985) provided a very comprehensive evaluation of the reasons for the failure of recent reforms in the Southern Cone. They mention (a) the policy inconsistencies associated with strong incentives to borrow from outside, the conflict between foreign exchange and trade policies affecting first the profitability of exports, and the lack of fiscal restraint, and at the micro level the failure to eliminate important distortions; (b) the adjustment lags related mainly to the thought that the rate of domestic inflation would rapidly converge to the rate of international inflation and the rate of devaluation; and (c) evasions of the intended reforms, related to the lack of long-term adjustment of the firms and their concentration in financial activities. Among the lessons drawn in the conclusions, we would have liked to see more about the adjustment costs associated not only with technical inconsistencies of the programs but also with other restrictions (bankruptcies, tax expenditures, imperfect markets) that may undermine the stability of any future policy package.

We have seen that recent changes in import protection were approached with inadequate attention to initial conditions when dismantling exercises started from highly distorted situations. For example, if the bankruptcy law could not be enforced for large firms, then credibility about sustaining those policies in the medium run would be the basic point of reference for entrepreneurs. Some of the entrepreneurs, while applauding the policy in public (outside opinion) but pointing out contradictions in limited circles, were postponing the decision to close down (if this would have been the case) by increasing their borrowing. The hope was that the snowball effect that follows this type of generalized attitude would lead to a policy reversal.

There is a need to look closely at outside and inside opinions when designing policy reforms, which will necessarily lead us to a compromise between the “best” policy generally represented by the economist’s point of view and the one that might provide stable rules for the decision maker.

New Reform Proposals

Import substitution industrialization started in Latin America under high tariffs. The general pattern was to assign higher rates for consumer goods than for intermediate and capital goods, with a built-in escalation owing to the varying degrees of fabrication plus setting rates for complementary imports at a low fiscal floor.

It is well known that a tariff increases the price of imports, and in so doing may provide incentives to import-competing activities by increasing their factor rewards. But, if inputs are provided at prices higher than international levels, value added under protection increases less than the margin allowed by output protection. The measure that incorporates both effects is precisely the effective protection rate. The patterns of realized effective rates of protection for domestic sales used in Tables 4 and 5 will be presented here as examples: average rates in Argentina are higher for intermediate products than for capital goods and the opposite holds in Brazil. Their frequency distributions show a high concentration of activities (85 percent) on both ends (effective rates up to 25 percent and higher than 76 percent) for Argentina, while 69 percent of Brazilian sectors were located in the same bracket; that is, what makes the larger relative difference in anti-export bias between those countries is basically the export promotion system.

The rationalization of the structure of protection is related mainly to private rates of return of import-competing activities, fiscal revenues, and the effect on consumers. That is why the removal of special regimes for tariff exemptions is also involved, but, primarily, the rationalization should be concerned with the anti-export bias induced by tariff and non tariff measures. More specifically, this rationalization implies necessarily reducing the present dispersion of effective rates, which does not seem to be justified on normative grounds. This dispersion is related basically to differentials in protection of outputs and inputs, so that the adjustment should be oriented toward the reduction of high rates on consumer goods, increasing the so-called fiscal floor for imported inputs. This set of rules might also require changes in the exchange rate to maintain trade balances within desired limits.

How easy is the transition between an old and a new trade policy? That is a difficult question, not only on technical grounds but also looking at the political economy of the adjustment. The need to increase exports is a must in searching for sources of growth beyond cyclical fluctuations. This is leading us to look at the rationalization of incentives as a sequential process, where only a strong increase in exports will generate the credibility needed to introduce the required changes in import protection. In the meantime, the higher costs of traded (domestic and imported) and nontraded inputs (absolute bias) should be removed, and selective rebates to overcome the relative anti-export bias should be provided. Trade policies should be oriented toward the minimization of the domestic resource cost of earning or saving foreign exchange, giving priority to consolidating those activities with a comparative advantage based on actual exports. But, as pointed out by Cline (1984), as soon as export-led growth becomes increasingly popular, one should also be concerned with the possible aggregate implications regarding the protectionist responses.

Looking for similar points of view regarding the proposal of a phased increase in exports before putting more rationality into incentives to domestic sales, we have found Diaz-Alejandro (1975) deeply concerned about the transition between two trade regimes: how much of the old system has to be dismantled, as well as what is to be expected from the new trade policies? Another complementary view concerned with international regulations is that of Snape (1984) when analyzing the rules of the General Agreement on Tariffs and Trade regarding the code of subsidies. He suggests a change in Article 14 in order that a developing country may grant export subsidies aimed at compensating the distorting effects of import barriers as a step in a program of import liberalization.

Sterie T. Beza

Julio Berlinski’s paper addresses a subject that has been traditional in the debate on economic policy in Latin America; one that has recently been the subject of renewed interest by specialists—that is, trade policy in the context of a growth strategy. But just as in previous decades the focal point was the role of an import substitution policy, in recent years emphasis has been placed on the need to expand exports.

Berlinski makes a contribution in this field from the perspective of economic policy execution in a world dominated by pressure groups, in which the problems of credibility are of primordial importance.

The starting point of his analysis is a statistical relationship, noted over twenty years ago by Simon Kuznets, between the relative size of a country’s foreign trade on the one hand and the national product and per capita income on the other. Accepting this relationship. Berlinski appears to take it as given that, to promote the process of economic growth, it is desirable to increase the openness of the economy. This seems a reasonable proposition, although it is curious that in reproducing Kuznets’ analysis for a recent period, Berlinski finds that the relationship is not valid for two of the four subgroups of countries analyzed. It would be interesting to know what the explanation of this phenomenon might be. He will possibly address it in a future paper.

Turning to the paper’s central theme, the author emphasizes the fact that the high degree of tariff protection in many Latin American countries produces an anti-export bias that acts as an obstacle to any increase in foreign trade and, in the end, to any sustained growth in the economy. Doubtless the prices at which local producers buy their inputs are higher than international prices to the extent that imports are subject to taxes or other barriers, and in the absence of any offsetting measures, exporters are faced with negative rates of effective protection. This is what the author calls absolute anti-export bias. He also points to the existence of a relative bias when prices in the domestic market are higher than international prices as a consequence of tariff protection on exportable goods.

Once the problem has been identified, the question arises as to how to deal with it. Berlinski’s proposal in this regard is, to my mind, the key point in the paper—but at the same time one of the most debatable. In principle, it would seem logical to think that the best solution would be to change the import system so as to make it more liberal. However, the author thinks that such a measure can be politically viable only if it is preceded by an expansion of the export sector. Such an expansion would be brought about by means of promotion mechanisms (such as tax refunds or subsidies) which would neutralize the anti-export bias of import restrictions. As a final objective the author favors the rationalization of incentive systems, in particular a tariff reform. However, he believes that this goal can only be reached in a sequential process, in which an increase in exports would enable the necessary credibility to be generated to successfully carry out such a tariff reform (and to eliminate other import barriers).

The author’s postulates call for a number of comments. In the first place, one wonders what he has in mind when he says that the import liberalization effort will be credible only if it follows an expansion of the export sector. Is he thinking of the need for an expanding export sector to offset any possible short-term negative effect on the overall level of economic activity caused by the liberalization of imports? Or is he perhaps thinking of the creation of a new pressure group of exporters which could offset the protectionist pressures of the sectors that have developed under the protection of import restrictions? At any rate, it is important to note that the problems of credibility frequently arise from incompatibilities between different policies. From this he derives the importance of the different policy measures forming a harmonious whole.

There is a risk that the authorities may act as though the export promotion measures are without cost. One problem is the pressure that these measures bring to bear on public finances. This seems to be one of the factors that leads the author to recommend the eventual dismantling of the import-protection-cum-export-promotion scheme. But beyond this problem, unless there are sufficient unused resources that can be swiftly mobilized, the need arises to reduce consumption in order to leave room for the expansion of exports and to transfer the resources intended for the production of nontradable goods or of import substitution goods to the production of exportable goods. If export subsidies are high enough to make the export of import substitution goods profitable, there need be no change in the structure of production; that is, there would be no transfer of resources from inefficient industries to other more efficient ones. If, on the contrary, export subsidies lead to a transfer of resources to new export sectors, there could be positive results in terms of standards of living. However, there would still be adversely affected groups, who logically would tend to oppose the change. In both cases, to facilitate the expansion of exports, it would be necessary to reduce domestic consumption (at least temporarily), through fiscal policy measures, for example.

Another important consideration is that it would be extremely difficult in practice to establish a system of export refunds that could adequately neutralize the anti-export bias of protectionist measures, insofar as the effects of these measures vary from product to product. Obviously, a uniform system of refunds would not eliminate the distortions caused by those protection measures that produce differential effects. Furthermore a neutral tariff protection system offset by export subsidies would be very difficult to administer, even assuming that such a one could be designed. There would doubtless be strong incentives to evade the costs or to take advantage of the benefits of the system, with the result that in reality it would probably be difficult to avoid the proliferation of fraudulent practices and a negative fiscal effect.

It should also be emphasized that a sequential process of the kind favored by Berlinski would over time prolong the import restriction measures, which not only have an anti-export bias but also give rise to other distortions. There is no doubt that the great number of nominal and actual protection rates applied in many Latin American countries has a harmful effect on the allocation of resources, which to a large extent overflows into exports. In this area, it should be pointed out that the argument in favor of tariff protection based on the optimum tariff theory does not in general seem applicable to Latin American countries, owing to the nonexistence of monopsonies in the region. The justification for protectionist practices based on the existence of external economies may be valid, but it is generally accepted that the objectives sought may be achieved more efficiently by other means, for example, by means of direct subsidies to the production it is wished to promote.

Berlinski’s paper includes a comparison of the protection schemes of Argentina and Brazil, and an estimate of the degree to which the anti-export bias has been offset by direct export promotion measures. This analysis seems to lead to the conclusion that there are no major differences between the two countries in the area of tariff protection policy, but that Brazil seems to have offset its anti-export bias to a much larger extent than Argentina. In this regard, it should be pointed out that, apart from the tariff protection on which Berlinski’s analysis is based, both countries have used other protectionist measures, such as quotas or import prohibitions. Although the author recognizes the existence of such practices, for obvious reasons they are not reflected in his quantitative analysis. This, as well as the frequent use of tariff exemptions, may seriously affect the validity of the results of the analysis.

Apart from that, these results do not appear necessarily to support Berlinski’s position in favor of a sequential process. On the one hand, although Brazil has doubtless had a more dynamic export sector than Argentina and greater economic growth in general over the last few years, there are no indications that it has made any progress in dismantling its import restrictions. As the author very rightly warns, the perpetuation of the system may give rise to problems both on the fiscal side and as a consequence of possible trade reprisals by other countries.

Another factor worth bearing in mind in evaluating programs aiming to liberalize imports, and which may have had implications in both Argentina and Brazil, is the existence of a broad sector of protected public enterprises. Unless the authorities are ready to act energetically to submit these enterprises to the same discipline as private enterprises, the import liberalization process may be obstructed. It is obvious that in those countries where the public enterprise sector carries a lot of weight, to exempt it from adjustment is costly in terms of both the direct impact of the exemption and the negative effect on the credibility of government policy.

Apart from the cases of Argentina and Brazil, Berlinski’s paper does not give country experiences except in a very summary form; this despite the fact that some countries seem to have made some progress in rationalizing their import regimes and to have been able to develop somewhat dynamic nontraditional export sectors. These successes were certainly not achieved without difficulty, and the methods used included a combination of measures, including direct export promotion. However, it seems only fair to mention that the problem of import restrictions has been attacked more directly than the author suggests, and with a certain degree of success.

Many countries’ experience also points to the importance of exchange policy. As a matter of fact, the experience of Argentina, Chile, and Uruguay over the past decade suggests that the difficulties they may have experienced in handling their economies, and specifically in their efforts to liberalize imports, were intimately linked with the overvaluation of their currencies. On the other hand, an active exchange policy less subject to ups and downs has probably been a key factor in the results Brazil has achieved. This is a point that the author does not ignore, but perhaps emphasizes less than he should. It must be stressed that, apart from encouraging the transfer of resources from the sector that competes with imports to the export sector, it is also necessary to ensure that there is a transfer of resources from the sector that produces nontradable goods to the sector that produces tradable goods.

By way of conclusion, I will take the liberty of turning back to some points that seem important to me.

First, although it cannot be denied that a sequential process could help reach the import liberalization objective, there are abundant reasons to doubt that this is the best way to do it. Berlinski’s paper does not seem to adduce evidence that this method has been used with success thus far.

Second, any liberalization policy must undoubtedly lay emphasis on being credible. In doing this, it is essential that the various policies, including the exchange rate policy and the financial (above all, fiscal) policy, are mutually compatible. When the authorities announce their liberalization timetable in advance, which is clearly desirable in many cases, their credibility also will require them to strictly adhere to the established timetable.

Last, too much emphasis should not be placed on the relationship between trade and fiscal policies. It is important that the approach adopted of rationalizing trade policy is compatible with the objective of not negatively affecting public finance, and fiscal policy should try to facilitate the transfer of resources to the export sector. Furthermore, one cannot ignore the danger of the objective of rationalizing trade policy being thwarted by lack of action consistent with this objective in regard to public enterprises.

Germán Botero A.

Berlinski’s paper summarizes existing trade policies in Latin America and concludes by suggesting that the way to return to free trade is first to promote exports and then to remove import tariffs. This commentary summarizes the distortions mentioned by the author using a model of the two sectors (one consumer good and one capital good) and two factors (capital and labor). The distortions are modeled as changes in the technique (capital/labor ratio) of production vis-à-vis the technique that would exist under free trade, and it is shown how the growth rate of consumption is maximized if distortions are eliminated. According to the author, the existence of distortions is explained by political pressures from groups of producers of the protected goods.

Let us assume an open economy that produces capital goods (importable good) and consumer goods (exportable good): its factors of production are capital that produces or imports and manpower. With free trade, the relative prices of the two products are equal to those on the world market, which means that there is a Pareto-optimal frontier on the economy’s capacity for consumption. Preferences between present and future consumption determine the point on the consumption frontier at which the economy is located: ceteris paribus, an increase in the discount rate over time increases present consumption, reduces the accumulation of capital goods, reduces the capacity for producing goods in the following period, and consequently reduces future consumption.

An increase in domestic production of capital goods in the short term can be achieved by restricting trade through taxes on their importation. The production of consumer goods declines, and the whole frontier of consumption possibilities turns back to a lower Pareto level than the original frontier. Assuming normal goods, the reduction of income reduces demand: the country consumes less, accumulates less capital, and reduces the net present value of its consumption. Berlinski suggests that this loss of present and future consumption is caused by the (industrial) pressure groups that succeed in making governments keep up their protectionist policies.

The author recognizes that restrictions on trade must disappear if growth is to be promoted, but he emphasizes that the political viability of this aim depends on what compensation is given to the beneficiaries of the current customs tariff structure. He suggests that this might be achieved through an export promotion scheme, so that after exports increase the pressure groups can agree to a reduction of protection of the industrial sector.

Berlinski does not specifically spell out how to implement his proposal. It is obvious that the main restraint is where to find the resources to be used to increase the production of exportable goods. Ideally, as the long-term objective is to reduce distortions, they should not be obtained by means of indirect taxation. A choice must be made between direct taxes on the producers or taxes on the consumer’s income. The former choice is surely not feasible, because of the political pressures mentioned by the author; the latter also runs up against political difficulties because taxes are usually already very high.

To sum up, the author describes existing distortions, points out the political difficulties of eliminating them, and formulates a strategy for reducing them slowly. However, his strategy calls for additional taxes, which may also encounter major political difficulties. The other alternative is to confront the pressure groups and gradually reduce trade restrictions.