|No 54 Em Português||
Ministério do Desnvolvimanto, indústria e Comércio Exterior
Ministério da ciência e Tecnologia
e&e : the Organization and the Periodical
The e&e Organization
Ttext for Discussion:
Some Grounds that Explain the Robustness of our Commercial Balance in the Context of the Real Exchange Rate Valuation and Projections for 2006
MSc in Economy from the UnB
Brasília, August 18, 2005
Why our commercial balance continues to present considerable surplus and our exports have a strong growth pace in spite of the Real exchange rate valuation?
The solidity of our external accounts is one of the main fundamentals of our macroeconomic stability and it has contributed to stabilize market’s expectations even in the present political context. In the last two years the Brazilian exports presented large expansion, increasing on the average 17.4%, 10.00 percent points above the average world exports in the same period (7.6%). In current values, exports have grown with an average rate of 26.5% in the last two years, adding up to US$ 96.50 billion in 2004 (an average of US$ 8.00 billion per month). However, since June 2004 we have had a period of larger valuation of our effective Real exchange rate (Figure 1). The effect of a more valuated exchange rate on exports is recognized as negative and we have observed several manifestations of discomfort from the exporting sectors. Therefore, the question is weather there will be a reversal of this fundamental that could deteriorate the external perception of our solvency and the solidity of our external accounts.
The Secretariat of International Affairs – SAIN, of the Ministry of Economy is of the opinion that this will not happen within the next two years, 2005 and 2006. We do believe that there will be a reduction of our commercial surplus but there are several reasons that make us believe that our commercial balance will continue to have a considerable surplus and our exports will keep growing in spite of the Real exchange rate valuation.
Several econometric models that simulate the growth of our exports (IPEA, Pastore , Tendências, BACEN, SAIN) indicate that the (negative) effect of actual valuation of our currency on our exports is small. According to these models if the Real is valuated in actual terms, 1%, the volume or quantum of total exports drops 0.25%.
Figure 1 – Exported Quantum and Effective Real Exchange Rate
Movable Average 3 months
Movable Average 12 months
Source: Funcex and Brazilian Central Bank
Figure 2 –Price index of commodities and manufactures
Source: Funcex and IMF
The recent Real valuation (+22.3% between June 2004 and June 2005, measured by the effective actual exchange rate) has obviously a negative effect on exports (and, it should be noted, a positive one on our inflation) (see Part 1 of the present note). However, the negative effect, about which the exporting sector complains, is being compensated by the five following factors:
1. The considerable growth of the world demand. In 2003 and 2004 the world GNP grew 3.9% and 5.1% respectively and it is projected to be around 4% in 2005 and 2006, replacing more pessimistic projections of 3-4 months ago;
2. The price rise of commodities and manufactured products (see Figure 2). Between June 2004 and June 2005 they grew about 9%;
2. Profit reduction of exporters who prefer to remain in markets they have penetrated instead of reducing their exports. The FINCEX profitability index regarding exports dropped about 19.8% between June 2004 and June 2005 (Figure3);
3. The possibility of large exporters making long-term exchange rate contracts and using the exports financing to obtain additional financial profit. For example, by applying on CDIs using ACCs and derivatives. It is estimated that between June 2004 and June 2005 a financial gain of about 5.0% was obtained on the export contract value (Part II of the present note);
4. A probable structural change, not captured by the models, regarding the quality and penetration of our exports in new markets.
This structural change includes several components: (a) price stabilization and smaller exchange rate volatility changes the vision of entrepreneurs about long-term profitability of our exports and about the attractiveness of the external market; (b) the start of exports activity of a larger number of small- and medium-size enterprises; (c) the diversity of our exports agenda by product and by market.
The weight of each of these five factors can be measured using a simple calculation based on the observed variation, by decomposing the growth of our exports (in the following periods) as follows:
In a more rigorous way, using SAIN’s econometric models and assuming that the actual exchange rate valuation of 13.7% will remains in the present plateau, our exports, imports and economic balance projections for 2005 (similar to the average value of the market) indicate $ 38 billion for balance, $144 billion for exports and $76 billion for imports.
In summary, the present elements of the economic policy are producing a durable effect on our exports in spite of the relative actual exchange rate valuation. So it should be pointed out the importance of following an agenda that maintains the macroeconomic stability, increases productivity through microeconomic reforms to help sustaining the exporting sector (by decreasing taxes on capital goods and computers sectors) as well as looking for a larger opening regarding the large consuming markets (European Union, United States and Japan) including current commercial negotiations and implementation and strengthening the Doha Conference (World Trade Organization) results. Other initiatives  aiming at interfering in the floating exchange rate system would not have the same effect and could be counterproductive.
In what follows we will discuss the effects of the Actual Exchange Rate and Exports Financing on our commercial balance.
The Real Exchange Rate Factor
In spite of the actual exchange rate valuation, the new econometric projections carried out by SAIN indicate an increase of the commercial surplus, namely US$ 37.8 billion in 2005 as compared with US$ 33.7 billion in 2004. The results obtained are close to the average value of the average market projections (Table 1).
Exports growth is obviously negatively affected by the valuation of the Real but only moderately due to the low elasticity of the exports price (0.25). The SAIN’s econometric model estimates a low elasticity-price of the Brazilian exports (Table 2). These results are compatible with other models (IPEA, Cavalcanti and Frischtak ), and explain why the recent exchange rate valuation was not sufficient for reverting the growth pace of exports, together with the other five mentioned factors.
Table 1 – Projections of the Commercial Balance 2005
Elaboration: SAIN/MF. Source: SAIN, BCB, SPE and CSFB.
1.Tendências, LCA, Unibanco, CSFB, IPEA e UFRJ.
Table 2 – Elasticity of Brazilian Exports According to Different Models.
On the other hand, imports expansion will not be larger because of the effective protection tariff increase, which has doubled on the average (from 8.5% in 2001 to 17% in 2003) after changes in the PIS and Cofins, according to Kume and Souza ). This change has contributed, albeit temporarily, to a smaller growth of imports in the last two years. However, it had an effect on inflation, since the effective protection tariff increase raised the shadow price of imports, extending the margin that domestic producers have in order to elevate their prices without facing international competition.
The hypotheses made for projecting the estimated model are summarized in Table 3 below.
Table 3 – Hypotheses for Projecting SAIN’s Econometric Model
Based on the results obtained until June 2005, an important remark is that the projection results of imports can be underestimated and a US$ 74 billion value is possible, so that the commercial surplus could be US$ 40.6 billion.
Exports Financing Factor
Besides the poor effect of the actual exchange rate factor, we give now details of one important factor among the five compensating ones: profits from exports financing, from which the exporter can benefit.
Filling the export declaration form at the customs does not necessarily imply the incoming of currency since the exporter has a series of financial instruments that allow for advancing or postponing the incoming of part or the total value of the receipts obtained from exports. If all exports would be paid cash down then exchange rate registration would be equal to the exports’ physical registration (Table 4). The Central Bank classifies the exchange rate contracts for exports in three categories; anticipated payment, ACC/ACE and others (Table 4). In the latter are included the cash operations that represent a significant share of the total, in a situation of stability and exchange rate valuation, and other operations such as export note.
Table 4 – Exports and Contracted Exchange Rate for Exports
Source: BCB. Elaboration: SAIN/MF.
1. Share in the contracted exchange rate for export.
2. Includes cash market operations that represent almost the whole category.
Among the several forms of financing exports, it should be pointed out the Advances on Exchange Contracts (ACC) and Advances on Exports Contracts (ACE). The only difference between ACC and ACE is that the former is a loan given before the merchandise is shipped and the latter happens after shipping. Besides that, an ACC is converted in ACE through accounting operations after the merchandise is shipped, even if the ACC’s taker has not yet paid for it.
The exporter has three benefits by taking an ACC. The first one is the possibility of guaranteeing a larger Reais amount in a conjuncture of exchange rate valuation. For example, if the exchange rate is R$/US$ 2.80 whenever a US$ 100 million ACC is obtained and if it reaches R$/US$ 2.55 at the time the ACC is paid, the exporter will have an additional receipt of R$ 25 million (he has been protected against the exchange rate valuation).
Another reason for obtaining an ACC is the possibility of arbitrage gain between the internal and external interest rate. By obtaining an ACC the exporter receives immediately in Reais the equivalent of his export contract and he can apply it in the domestic market and receive Over-Selic interest rates. His opportunity cost is the interest rate of the loan that is established in dollars, plus the devaluation expectancy. For example regarding an advance of a US$ 100 million export in October 2004, the exporter would approximately have a financial gain of R$ 8.6 million (not considering CPMF and income tax). In case of domestic currency valuation in the period, the gains would be larger; in case of devaluation, the opposite would happen. Table 5 shows simulations with three scenarios: stability, exchange rate valuation and exchange rate devaluation.
The possibility of gain was more attractive at the time of administered exchange rate and for this reason the ACC share in the total contracted exchange rate has been decreasing since 1997 (Table 4). The predictability of gradual devaluation of the exchange rate between 1995 and 1998 generated the certainty of a determined financial gain since it was very unlikely that an exchange rate evaluation would generate a gain higher that that from the cash market. For example in scenario III in Table 5 the exporter gets a financial gain of R$ 7.3 million. He does not gain more from the opportunity cost point of view, because if he would sell his hard currency in the cash market at the end of the operation with an exchange rate of R$/US$ 3.0 he would get R$ 300 million. By choosing ACC he got 267.3 million but in advance.
Table 5 – Arbitrage Gain with ACC Loans
1. Refers to the 3-months period, between October 7, 2004 and January 10, 2005.
2. Refers to a 3-months operation contracted in October 2004.
The availability of resources in advance via ACC is important because they can be used as working capital. Small and medium size enterprises use ACC loans as working capital instead of applying them in the financial market. This is so because the ACC cost is lower (considering the exchange rate stability) relative to the working capital loan. The ACC average rate in dollars was 5.23% a. a. in February 2005 while the working capital was 32.18% a. a. (in Reais). In the advance payment item are included payments made in advance without using the credit and the financing granted by the importer to the exporter, for example supplier’s credit. Other credit facilities provided by the importer (buyer’s credit, forfainting and factoring) as well as financing provided by BNDES Exim may be distributed in the three items shown in Table 4.
Briefly, even in a valuation context, advance financing of export contracts (via ACC or other contract mechanisms with derivatives, given the high remuneration rate of financial assets in the short term in Brazil) can produce significant arbitrage gains for the exporter, compensating eventual losses from actual valuation of the currency.
Part I : Forecast of the Commercial Surplus in 2005 and 2006
The exchange rate valuation, started in 2004, its constancy until reaching 2.33 Reais per dollar and its maintenance in a valuated plateau started the debate about the exchange rate effect on the Brazilian commercial balance. In 2004 the commercial balance surplus set a record, US$ 33.7 billion, reaching US$ 11.7 billion (1.9% of the GNP) surplus in current transactions, the largest value in the historical series.
As a support for the debate, two regression estimates have been made: one for imports the other for exports. From these equations and considering a pre-determined path for the exogenous variables, exports and imports have been projected from 2005 to 2009. In the first note only the traditional equations have been considered:
where Qx and Qm are, respectively the exported and imported quanta, Y* is the product of the rest of the world , Y is the domestic product and er is the actual effective exchange rate. The estimated functions measure the sensibility of exports and imports relative to elasticity income and elasticity price. In this second estimate the export equations have also been estimated considering the profitability of exports as a substitution to the actual effective exchange rate, the effect of the price of commodities (without oil) and the impact of oil on exports. The first equation to be re-estimated was that of imports.
The data used are the quarterly one and the considered period is from the first quarter of 1991 to the second quarter of 2005, adding two observations to the first estimate. Imports were adjusted for working days and deflated relative to the July 2005 price. For this purpose the Funcex price index was used. So, the variables used were imports at June 2005 prices, the IBGE index for the GNP and the actual effective exchange rate calculated by the Brazilian Central Bank based on the domestic and international IPA.
The Box&Jenkins test was applied again on the series in order to detect the presence of possible structural breaks in the series. The results are shown in Table I.
Table I –Structural Breaks in the Imports Equation Series
The breaks detected in the exchange rate series are the same of the first note. The fist break detected indicates the change of the exchange rate system and the second one, a change in the process of generating data due to adverse shocks that have always occurred in the third quarter (immediate consequences of September 11, 2001, the 2002 elections and others) reverted in the first semester of the following year.
In the case of imports, it was detected a new break in the first quarter of 1997 corresponding to a value below the predicted one. The two other breaks have already been identified in the previous note, indicating the effects of changes in the exchange rate system that occurred in June 1994 and January 1999.
Besides the detected breaks two more ones have been considered: one for the exchange rate and another for imports. In the case of exchange rate it was considered a temporary change type break from the third to the fourth quarter of 1994, including a semi-flexible exchange rate period at the start of the Real Plan that was not detected in spite of the inclusion of the two 2005 quarters.
In the imports series it was added another break of the level shift type from de second quarter of 2003 on, in order to include the change in Cofins. According to Kume and Souza (2004), after PIS and Cofins have been imposed on imports, the tariff of effective protection of industrialized goods have doubled on the average (changing from 8.5% in 2001 to 17% in 2003). It was expected that the inclusion of two more observation would permit the software to detect this break that is statistically significant. The increase of the effective protection has increased the shadow price of imports so that this could be one of the components that explain the reduction of the effect of exchange rate valuation on prices. For that reason, among other factors, the effect of exchange rate valuation on consumer price was not immediate. Regarding the GNP, no structural breaks were detected.
The next procedure was to verify the presence of unitary root in the series using the ADF test. The results are in Table II.
Table II –ADF test for Imports Equation series
1. Test with constant and five lags of the difference
2. Test with five lags of the difference
3. Test with eleven lags of the difference.
The results of Table II show that all series are stationary at level (10). Therefore, as we estimate a regression among the variables this relationship cannot be a non sense. Table III shows the results of the regression (the graphic with the residues of regression and the adjustment grade are in Annex 1). The Annex 1 analysis indicates a difference between the observed value and that foreseen one in the second quarter of 2005, the former is below the latter. This might indicate that the imports projections are overestimated.
Concerning the initial results, it was observed a reduction of elasticity both for the income and the price. The elasticity income dropped from 2.65 (1st note) to 2.49 and the elasticity price changed from -0.81 (1st note) to -0.71. These results show the instability of the parameters.
The results continue to indicate that imports are elastic relative to income and inelastic relative to the exchange rate even with a value higher than o.5. So, even though the GNP has a larger effect on imports, the exchange rate effect is not negligible especially when one considers that the exchange rate volatility is larger (Annex II shows the interval within which the coefficients may vary). Based on this equation, the imports from 2005 to 2009 were projected.
Table III – Elasticity Price and Imports Income
The continuity of exports growth in spite of the actual effective exchange rate valuation, 18.6% between the second quarter of 2004 and the second quarter of 2005, has obliged the economists to review the conditions considered as ceteris paribus. Among the several reasons pointed out as errors regarding projections of the exports traditional equation, it is agreed that two variables have not been taken into account: price of commodities and the relevance that oil and its products have had in our exporting agenda in the last three years.
The argument used is that the rise of the commodities price from 2002 on would have compensated the exchange rate valuation. Pastore (08/03/2005) has made econometric simulations in order to show that expansion of exports between 2002 and 2003 was due to the exchange rate devaluation while the growth observed from 2003 on was due to the rise of the price of commodities.
The rise of the commodities’ relative price should be compensated by the actual effective exchange rate. However, by using a price index either for the consumer or the producer the commodities’ price is diluted in a series of other prices because it has a small weight in the price index. Therefore the actual effective exchange rate does not correctly compensate the price growth of the Brazilian exports. So, an actual effective exchange rate that would precisely incorporate the effect of price rise of commodities would be in a plateau higher than that presented by the calculated actual effective exchange rate using an index price. A hypothetical example to show the actual effective exchange rate bias would be the case where Brazil would export only soybean to Germany. The effective exchange rate is obtained using the relationship between the inflation for the producer (PPI) abroad and in Brazil; therefore, considering a series of goods and services other than soybean.
In order to mitigate this problem it has been included in the regression a regressive factor that considers only the price of commodities, excluding oil, calculated by the IMF. However, by including this regressive factor a multi-co-linearity problem is created since the effective exchange rate includes the price of commodities.
Another argument used is that the traditional exports equation does not properly consider the effect of oil and fuels exports, whose share in the total exports grew from 1.0% in 1999 to 8.5% in 2004. Sending oil abroad to be refined is due to the fact that there is not capacity to refine it in Brazil.
So, the option was to estimate oil exports in a separate regression, subtracting it from the total exports, which continued to be estimated considering actual exchange rate and income of the rest of the world. Oil exports were projected using an ARIMA (Auto-Regressive Integrated Moving Average) process.
Regarding the data, as well as those of imports, the observations are quarterly and the period considered is from the first quarter of 1991to the second quarter of 2005. The exports (without oil, fuels and products) were adjusted for working days and deflated according to July 2005 prices. For this purpose the Funcex price index was used. The actual effective exchange rate is the same used for imports, the price of commodities is given by the index calculated by the IMF that excludes oil and the profitability of exports is calculated by Funcex
The world’s GNP was calculated considering the Brazilian commercial partners and their share in our exports in 2004: European Union (25.0%), USA (21.0%), Argentina (7.6%), China (5.6%), Mexico (4.1%) and Japan (2.9%). These countries have received 66.2% of our exports. The weights were used so that they would add 100%. As it was not possible to obtain a series for the Chinese GNP or a long series of the Chinese industrial production, as well as of imports, an ad hoc adjustment was made. In the case of China, the solution was to consider an additional growth of 0.77p.p.using the world’s GNP without China, since the average Chinese growth was 9.1% and its weight, re-weighted in the Brazilian exports, was 8.5%. It was not possible as well to obtain a quarterly GNP series of the USA without seasonal adjustment so the USA imports were used as proxy of the GNP behavior.
Another variable used was the Funcex index regarding the exports profitability which is a measure of the actual exchange rate obtained from the production costs of the main exported goods and the price of exports.
As the profitability index is a measure of the actual exchange rate there is no reason for estimating regression where this variable and the effective actual exchange rate are exogenous variables, taking into account the economic theory. Concerning the econometric question, the two variables present a high correlation (0.85) and this would cause a problem of multi-co-linearity in the regression.
According to the methodology used in imports, it was first examined the existence of structural breaks in the world GNP, in imports and in the profitability index of exports (the effective actual exchange rate is the same for imports). The Table IV shows the estimated breaks.
Table IV – Structural breaks in Exports Equation series
In the exports series, besides the two detected breaks, three more breaks were considered: an additive one (AO), in the first quarter of 2003, a change of level (LS) from the second quarter of 2003 on and another change of level from the second quarter of 2004 on.
The level change in 2003 was due to changes of Cofins and it eased the cost-cutting process, exempting the exporter not only from the PIS/Cofins payment during production as well as giving him better conditions to use the credit due to the cost-cutting of the previous phases.
The other level break adjusts the series regarding the effects of a structural change in investments favoring tradable goods and the conquest of new external markets. The 1999 change on the exchange rate system has caused a reallocation from the non-tradable goods (privileged during the privatization and exchange rate band period) to the tradable goods as a function of the relative prices. The development of new projects would have been delayed by the menace of electric energy shortage in 2001 and by the presidential election in 2002. For this reason exports changed to a new plateau. Table V shows the changes on the average that occurred in these two periods.
Table V –Additional Structural Changes in Exports
1. In US$ million at June 2005 prices, without oil and fuels.
In the commodities price series no break was detected. Once the breaks have been detected, one can perform the unitary root test. The results of the ADF (Augmented Dickey Fuller) test are shown in Table VI.
Table VI – ADF Test for the Exports Equation Series
1. Test with five difference lags.
2. Test with six difference lags.
3. Test with constant and without difference lags.
4. Test with constant and eight difference lags.
The results of Table VI show that the exports and world GNP series are stationary at (10) level while the price index of commodities is not. So, when estimating a regression among the variables it is necessary to use the growth rate of the commodities’ price index (1st difference).Table VII shows the regression results considering the real effective exchange rate (the graphic with the regression residues and the adjustment degree are presented in Annex III).
Table VII – Price and Income Elasticity of Exports (A)
Exports are inelastic relative to exchange rate, as well as imports, and the exchange rate effect on the former is much smaller and they have elasticity close to one relative to the world income (Annex IV shows the interval within which elasticity may vary). The t statistics of the commodities price growth rate is biased relative to the null hypothesis (statistical non-significance of the variable) as a function of multi-co-linearity. Based on this elasticity, exports were projected from 2005 to 2009.
Besides the actual effective exchange rate, a regression with the profitability index was estimated. Table VIII shows the results of the regression, and Annex V, the graphic with regression residues and its adjustment degree.
Table VIII – Price and Elasticity Income of Exports (B)
The elasticity price and income are close to those estimated using the effective actual exchange rate. Exports are inelastic relative to the exchange rate and have elasticity close to one relative to external income (according to the Wald test the probability of elasticity being equal to one is 63%). The t test both for profitability and commodities price is biased for accepting the null hypothesis. In this regression the profitability is calculated considering the exports price index.
From the theoretical point of view, regression with the actual effective exchange rate (regression A) presents a better adjustment than the equation with the profitability index (regression B) because the Akaike and Schwarz statistics have smaller values.
In order to project the commercial balance an exchange rate of R$/US$ 2.55 was considered for 2005 and a nominal devaluation of R$/US$ 0.11 in 2006 and R$/US$ 0.1 in the subsequent years. It was assumed a domestic inflation of 5.1%, in 2005 and of 5.0% in the subsequent years and a world inflation of 2.5% for the period. The Brazilian GNP grew 3.0% in 2005 and 4.0% in the other years. The expansion of the world GNP (concerning the six economies used to project the world GNP) considered was 4.3% in 2005, 4.0% between 2006 and 2008 and 3.8% in 2009. The results of the projections as well as those observed in previous years are shown in Table IX. It is important to point out that these are point projections at the center of the interval (Annex VI shows the quarterly projections).
Table IX – Projections of the Commercial Balance
NR a Posteriori: Values
Observed in 2005
1. At June 2005 prices.
The considerable growth of the Brazilian economy and the average 13.7% effective actual exchange rate valuation in 2005 caused a growth of imports, namely 22.3% in 2005 and 11.4% in 2006. On the other hand, the exchange rate valuation and the slow down of the world GNP growth decelerated the expansion of exports to 18.2%-18.8% in 2005 as compared to 32.0% in 2004. In spite of the fact that exports had a smaller growth rate than that of imports, the commercial surplus will be around US$ 38.0 billion in 2005 (at June 2005 prices). The result of the commercial balance, ceteris paribus, will decrease only in 2006 when it will be US$32.0 billion (a value close to that of 2004). Even with the commercial balance decrease, it will be larger than US$ 20.0 billion until 2009.
The values projected for 2005 are close to the forecast for the average market. The Brazilian Central Bank estimates exports of US$ 108 billion, imports of US$ 78.0 billion and commercial balance of US$ 30.0 billion for 2005. Table X shows the comparison of the results for 2005
Table X – Projections of the Commercial Balance for 2005
Elaboration: SAIN/MF. Source: SAIN, BCB, SPE e CSFB
1. Tendências, LCA, Unibanco, CSFB, IPEA e UFRJ.
Projections indicate a descending path of the commercial balance accumulated in 12 months, since until July the surplus was US$ 39.9 billion. The exports accumulated in 12 months until July totaled US$ 108.9 billion and imports, US$ 69.0 billion. Therefore, according to the projections exports should grow on the average US$ 1.14 billion monthly and imports, US$ 1.56 billion monthly, on the average.
The US$ 1.56 billion monthly growth of imports may be overestimated since in the last six months the value accumulated in 12 months has grown US$ 1,0 billion monthly, imports totaled approximately US$ 74.0 billion in 2005 and this would produce a commercial surplus between US$ 40.0 billion and US$ 40.6 billion.
References Part I
CAVALCANTI, Marco Antônio F. H. e FRISCHTAK, Cláudio R. (2001) “Crescimento
econômico, balança comercial e a relação câmbio-investimento”. IPEA -
Texto para Discussão nº 821.
KUME, Honório e SOUZA, Evaldo “Mudanças na Cofins e no PIS-PASEP e a estrutura de
incentivos à produção doméstica”. Texto interno do MF/SEAE.
PASTORE, Afonso C. (2005) “O Mistério das Exportações”, Informe Especial da A.C.
Pastore e Associados, August 3.
Annexes of Part I
Annex I – Residues and Adjustment Grade of Regression of Imports
Annex II – Interval of Variation of Elasticity Income and Elasticity Price of Imports (with a significance level of 95%)
C(2) is the elasticity income and C(3) the elasticity price.
Annex III – Residues and Grade of Adjustment of Regression of Exports - A
Annex IV – Interval of Variation of Elasticity Income and Elasticity Price of Exports (with a significance level of 95%)
C(2) is the elasticity income and C(3) the elasticity price.
Annex V – Residues and Grade of Adjustment of Regression of Exports - B
Annex VI - Projections of Exports and Imports
(US$ million at June 2005 prices)
Elaboration: SAIN/MF. NR: a posteriori values for 2005
Exports: III 33.042, IV 31.589; Imports III 20.052, IV 19.471
Part II: Exports Financing
In the last two years the Brazilian exports presented a considerable expansion with an average increase of 17.4%, 10.0 p.p. above the average world expansion in the same period (Table III). In current values, exports have grown at an average rate of 26.5% in the first two initial years of Lula’s Administration, summing US$ 96.5 billion in 2004 (average of US$ 8.0 billion monthly).
Table I –Brazilian and World Exports
Source: MDIC and IMF. Elaboration: SAIN/MF.
The registration of exports at the customs does not necessarily imply takings of foreign exchange since the exporter has a series of financial instruments that permit advance or postponement of bringing in part or the total value of the money obtained by the exports. So an exchange registration may be divided in three parts: cash terms, exports advance or financing. If all exports would be paid as cash terms then the exchange registration would be equal to the physical exports registration. Table IV shows the comparison between physical exports (Secex) and the exports exchange contracts.
Table II – Exports and Exports Contracted Exchange
Source: BCB. Elaboration: SAIN/MF.
1. Share in the contracted exchange for export.
2. Includes cash terms market operations.
Foms of Exports Payment.
When a producer exports a product he can be paid in four different ways: (i) in advance; (ii) by letter of credit; (iii) by documental collection; and (iv) by remittance without draft.
In the first form, advance payment, the importer finances the exporter and this is more used in the exports of machines and equipment under purchase order. In this modality, first a business contract is signed between the Brazilian exporter and the importer. Then the importer makes a payment in a foreign bank that remits the currency to the exporter’s bank (bank authorized by the Brazilian Central Bank to operate with foreign currency). Afterwards the exporter sends the commodity abroad. In this case the exporter’s risk is zero. Table IV shows that the advance payment represents 22% of the exports exchange contracts. In this statistics are included financing granted by the importer to the Brazilian exporter through the Supplier’s Credit, Buyer’s Credit.
The second method of payment is trough a letter of credit that is conditioned money order, that is, the exporter will receive the payment if he fulfils all the requirements of the contract. The sequence of actions is shown in Annex 1. In this form of payment, the exporter’s risk is also small since it involves an operation guaranteed by one or more banks. It is important to note that when the exporter’s bank accepts the letter of credit it has not the commitment of receiving directly from the importer but from the bank that emitted the letter. So, if the importer does not pay for the commodity, the problem is between him and the bank that granted him the letter of credit. The letter of credit can be in cash terms or long term.
Regarding the documental collecting, the Brazilian exporter, after signing the trade contract, ships the commodity and then issues a bill of exchange for the importer. This letter is delivered to the exporter’s bank that delivers it to the importer’s bank. The latter presents the documents to the importer and charges him. Once the debt is acknowledged by the importer, the bank draws the money from his account, but sends it to the exporter’s bank only after the commodity is cleared by the customs of the country that imported it (Annex II). At the time the exporter delivers the bill of change to his bank he can have an advance on the exchange contract. (ACC – explained further in the text). The documental collecting can be in cash terms or long term.
At last the remittance without draft form, where the exporter ships the commodity, sends the documents and the importer clears the commodity at the customs of his country before paying. In this form no letter of exchange is issued after shipping and it is more used when the exporter and the importer are interconnected enterprises. Due to a larger risk, the Brazilian legislation imposes restrictions on this form of payment in our exports.
Even though the first two forms of payment have almost no risk and are speedy, it is possible to make an ACC in all cases, which is also a way of financing exports.
Forms of Financing Exports.
The different forms of financing exports can be divided in three categories: (i) advance of resources operations; (ii) operations of discounting letters of exchange already acknowledged by the importer; and (iii) exporter’s issue document.
Among the advance of resources operation we point out: Advance of Exchange Contract (ACC)/ Advance of Export Contract (ACE), BNDES Exim and Proex. The two main financing forms are ACC and BNDES Exim, even though the ACC volume (average of US$ 2.7 billion monthly) is much larger than that of BNDES Exim (average of US$ 325 million monthly).
The operations of discounting letters of exchange already acknowledged by the importer are: Supplier’s Credit, Buyer’s Credit, Forfainting and Factoring. However, data regarding their volume are not available because they are offered by foreign banks.
Exporter’s issue documents for financing exports are: Export Notes, Exchange Debentures, Warrants Discount and Exports Security Provision. The most important of them is Export Notes but its average volume (US$ 3.0 million of new concessions each month) is much lower than the credit through ACC.
The difference between ACC and ACE is that ACC is a loan received before shipping the commodity and ACE is received after shipping it. Besides that, an ACC is converted to ACE by accounting operation after the shipping of the commodity, even if the ACC borrower has not paid it yet.
The banks authorized by the Brazilian Central Bank to carry out exchange operation grant to the exporter advance money on the exchange contract (partial or total advance) delivering an amount in Reais equivalent to the amount in foreign currency bought forward by the bank. In order to give this loan the bank has a credit facility abroad (Figure I).
Figure I – Advance of Exchange Contract
The ACC operations cannot be transferred, that is, the bank cannot negotiate this credit in the secondary market and it is obliged to keep it until it expires. From August 1999 on the BCB has doubled the term permitted for the ACC operations: 360 day before shipping and 210 days after shipping. In spite of the possibility of one year term, the average term observed in the last twelve months was around 100 days.
The first benefit of making an ACC is the possibility of guaranteeing larger receipts in Reais in an exchange rate valuation conjuncture. For example, if the exchange rate is R$US$2.90 when a US$ 100 million ACC is made and it reaches R$US$2.60 at the time the ACC is paid, the exporter would have gained R$ 30 million (protecting him against exchange rate valuation).
Another reason for making an ACC is the possibility of arbitrage gain between the internal and external exchange rate. By making an ACC the exporter receives Reais immediately and he can invest them in the domestic market and receive Over-Selic interest rates. His cost of opportunity is the loan’s interest rate, established in dollars, plus the devaluation expectation. For example, for an advance on exports of US$ 100 million, the exporter would have a financial gain of approximately R$ 8.7 million (without considering CPMF and income taxes). In case of valuation of the domestic currency in the period, the gains will be larger and in case of devaluation, the opposite is true.
For small and medium enterprises, the ACC is used as working capital instead of investment in the financial market. This is so because of the lower ACC cost (considering the stability of the exchange rate) relative to loans for working capital. The average ACC rate in dollars was 5.23% a.a. in February 2005 while that of working capital was 32.18% a.a. (in Reais). Annex IV shows the average interest rate in dollars of the ACC loans and the movable three-month average of the new concessions (in US$) of ACC. Table V shows a comparison between new ACC concessions and exports in the month as well as the average cost of an ACC operation and other credit facilities for legal entities that have banking resources for free investment.
Table III – Ratio between the ACC volume and exports and loan taxes
Source: BCB, Secex, Fed. Elaboration: SAIN/MF.
1. The annual interest rates are average ones.
Annexes of Part II
Annex I – Letter of Credit
Annex II – Documental Collecting
Annex III –Interest Rate and ACC New Concessions
ACC Interest Rate (% a.a.) and USA Government Bond of 3 months (% a.a.)
New ACC concessions in US$ million–with seasonal and working days adjustment
Weighted average of several bilateral actual exchange rate, as a function of the commerce flow by country. Reduction of the effective actual exchange rate in Figure 1 means actual “valuation’.
 In order to eliminate the actual exchange rate peaks (for example, such as those observed in the main Asian economies before the 97 and 98 crises), it is first necessary to diagnose if the valuation is temporary or permanent and if it is being produced by long (commercial surplus) or short (speculative capital flows. After this diagnosis, the classical measures are: (a) carry on the flexible exchange rate system; (b) buy reserves with or without sterilization; (c) make additional fiscal adjustment; (d) increase the compulsory deposits; and (e) control the short-term capital (as Chile has done). It is controversial issue the efficiency of this last measure. The exchange rate valuation can also have positive effects on other areas of the economy, for example, the anticipated payment of the public and private sectors foreign debt.
 They are certificates issued by the exporter for financing the export’s production without a pre-defined term for shipping the exports. The average volume of new concessions in 2004 was US$ 3.0 million per month.
 ACC is an operation for advancing future exports receipts, in which the banks that are authorized by the Brazilian Central Bank to perform exchange operations grant as an advance on the exchange contract (partial or total anticipation) an amount in Reais equivalent to foreign currency bought by the bank. In order to make this loan, the bank has a credit facility abroad.
 Considering 3 months as the loan period (average period observed) , the accumulated internal interest rate in the period 4.23 %, average ACC loan rate 5.11% a.a. (in dollars) and average exchange rate observed in the period (R$/US$ 2,85 in October 2004 and 2.69 in January 2005)
 Credit granted by BNDES Exim in 2004 totaled US$ 3.8 billion.
 The elasticity income, ceteris paribus, measures how much imports are affected by an increase of 1% in the GNP PIB. So, if the estimated elasticity income is 2.6, it means that an increase of 1% in the GNP produces an increase of 2.6% in the total imports. The same reasoning is valid for elasticity income of exports but changing the DGP for the world one. The elasticity price, ceteris paribus, measures how much imports are affected by an exchange rate valuation (devaluation) 1%. For example, if the estimated elasticity price is 0.8, it means that an exchange valuation of 1% produces an increase of 0.8% in the total imports.
 The structural changes are of four types: additive (AO), temporary or transient (TC), level change (LS) and innovation (IO). The break test presents one restriction: detecting the break at the end of the series.
 The raise of the commodities price has also contributed to a less sharp inflation decrease, in spite of the exchange rate valuation.
 The main consequence of multi-co-linearity is increase of variance of the regression, biasing the t test to accept the null hypothesis, but the minimum square estimator continues BLUE (best linear unbiased estimator).
 The value obtained by adding the petroleum and coal exports and the fuels and the products.
 According to FUNCEX: “The profitability of total exports is calculated from the average nominal exchange rate of the month (R$/US$) corrected by total exports price index and the total export cost index ratios. The total cost index includes 31 sectors and is elaborated from the cost structure of the productive sectors obtained from the 1995 Inter-industrial Matrix of IBGE (MRI-95). The cost index of each sector is calculated based on the share of input of domestic origin, imported inputs and salaries and taxes, according to the matrix. The price indexes associated with these three components are: (i) wholesale sectorial indexes published by Fundação Getúlio Vargas; (ii) the FUNCEX import price index multiplied by the nominal exchange R$/US$; and (iii) an average hourly salary based on the total paid salaries and hours from FIESP”.
 The imperfect multi-co-linearity causes increase of variance of variables that present the problem.
 Previously de-taxing was made on a presumed credit calculated on the collected IPI.
 The series is I(1). The estimated value of the ADF test of the first difference of the series was
–4.64 and the critical value at 1% ,-3.55.
Exchange Contract: expensive bilateral act (regulated by the Law 4.595/64 and formalized by circular 2.231/92 of the BCB) by which the exporter sells to the bank the foreign exchange that he will receive both in cash terms and in long term.
 Considering 3 months as the loan term (average term observed, internal exchange rate 19,5 % a.a., ACC loan rate 5% a.a. (in dollars) and constant exchange rate in the period (R$/US$ = 2,60)
Graphic Edition/Edição Gráfica:
Tuesday, 11 November 2008.