Chapter

2 downloads 0 Views 296KB Size Report
that same period (Schmitz, Seale, and Buzzanell, 2002). There has been a phenomenal increase in the yield of ethanol from sugarcane. In 1999, roughly 5,500 ...
International Sugar Journal CVI,1271(2004):586-596.

Determinants of Brazil’s Ethanol Sugar Blend Ratios Andrew Schmitz, and James L. Seale, Jr, Troy G. Schmitz University of Florida, Gainesville, Florida University of Florida, Gainesville, Florida Arizona State University, Mesa, Arizona

INTRODUCTION Brazil is the world’s largest producer of sugarcane, the world’s largest exporter of sugar, and the world’s third largest consumer of sugar. Brazil produces sugarcane-refined sugar for human use as well as anhydrous and hydrous alcohol, which are used mainly as a blend in domestically consumed gasoline. Over 50 percent of Brazil’s sugarcane production is converted into fuel for automobile use. The Brazilian government affects Brazil’s sugarcane market through its alcohol fuel program. The government sets the blend ratio of blending alcohol with gasoline. Blend ratios change over time. This paper derives theoretically and empirically some of the determinants of the blend ratio changes. It provides insights into the factors the Brazilian government uses when setting the blend ratios.

THE BRAZILIAN SUGAR INDUSTRY Brazilian sugarcane is used to create refined sugar, anhydrous alcohol, and hydrous alcohol.

Anhydrous alcohol is used to blend with gasoline as mandated by the Brazilian

government, and hydrous alcohol is used as fuel for vehicles that are powered 100 percent by alcohol. Hydrous alcohol production has declined by more than 50 percent for the period



This research was partially supported by the International Agricultural Trade and Policy Center, Food and Resource Economics, University of Florida. Florida Agricultural Experiment Station Journal Series No. xxxxx.

1990/91-2002/03 while anhydrous-alcohol production has increased more than five-fold over that same period (Schmitz, Seale, and Buzzanell, 2002). There has been a phenomenal increase in the yield of ethanol from sugarcane. In 1999, roughly 5,500 liters were produced per hectare, while in 1975 per-hectare yield was only approximately 2,000 liters. This represents almost a three-fold increase in the efficiency of ethanol produced from sugarcane. Brazil exports between 0.5 and 1.0 billion liters of ethanol per year. The portion of sugarcane used for fuel alcohol in Brazil increased steadily from 1976 through 1985, peaking at 74 percent in 1989, but declined during the 1990s, reaching 55 percent in 2000 (Figure 1). The number of vehicles powered by hydrous alcohol has declined sharply over the years although these vehicles still exist due to the subsidies provided by the Brazilian government for rental cars, taxis, and some government vehicles powered by hydrous alcohol. Prior to 1998, the Brazilian sugar industry was highly regulated. The Institute of Sugar and Alcohol (IAA) was created in 1933 to deal with the sugar overproduction problem by setting the sugar production quotas and fixing prices to control the volume of exports. Brazil also had import tariffs and export taxes placed on sugar to ensure that alcohol-production targets were met (Schmitz, Seale and Buzzanell, 2002). The Brazilian National Alcohol Program (PROALCOOL) was created in 1975 in response to the 1973 oil crisis. Under this program, the IAA purchased anhydrous alcohol at an equivalency rate of 44 liters of alcohol per 60-kilogram bag of sugar while Petrobas, the state-owned oil company, controlled ethanol distribution.

Credit guarantees and low

fixed-interest-rate subsidies were also provided for the construction of distilleries and autonomous plants built adjacent to sugar mills.

2

70% 60% 50% 40% 30% 20%

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

0%

1972

10%

1970

Portion of Sugarcane Used for Alcohol

80%

YEAR

Figure 1. Portion of Sugarcane Used for Fuel Alcohol in Brazil (1970-2000) from 81/82 to 99/00 Informativo Datagro, data from IAA, MIR, MIC, MICT. DATAGRO Boletim Informativo Quinzenal sobre Cana, Açúcar e Álcool – Year 99, Number 14P

The monopoly enjoyed by Petrobas was removed as policy changes were enacted in 1997/98. First, anhydrous alcohol prices were liberalized in May of 1997, and hydrous alcohol prices were liberalized in February of 1999. Subsidies paid to hydrous-alcohol producers were reduced from 0.98 reals per liter to 0.45 reals per liter while subsidies paid to anhydrous-alcohol producers were eliminated (USDA, 2001). However, ethanol production is still regulated by government decree. Each year, a Presidential Decree sets an alcohol-to-gasoline blend-ratio range for the percentage of ethanol that must be used in Brazilian gasoline. Both the domestic sugar and ethanol markets are protected from competition from other low-cost exporters. A common external tariff of 20 percent on sugar imports and 30 percent on imports of ethanol was put in place in 2001. However, there is no tax on intra-zone trade of ethanol for Brazil’s Southern Cone Common Market (MERCOSUR) partners. Furthermore,

3

these tariff levels are relatively low when compared to trade barriers for sugar and ethanol in many other sugar-importing countries around the world.

There still remains a support

mechanism that compensates for sugarcane-cost differentials across regions that is well under the de minimis clause of the WTO agricultural agreement.

BLEND RATIOS AND BRAZILIAN SUGARCANE MARKETS The Brazilian government sets the portion of anhydrous alcohol that is used in gas-powered vehicles. This “blend ratio” is adjusted from time-to-time by government decree (Schmitz, Schmitz, and Seale, 2003). The blend ratio was 25 percent in 1970 before dropping to 11 percent by 1976. It increased to 22 percent in 1985, but then decreased to 13 percent by 1990. It reached as high as 25 percent in June 2002, but was then decreased to 20 percent in January 2003 but was increased again to 25 percent on July 1, 2003. The reason for the changes in the government mandated blend ratio is the scope of this paper.1

ETHANOL POLICY AND BRAZILIAN SUGAR Theory In a previous paper, we show how supply response is influenced by producers’ attitudes towards risk. As shown below, if producers are risk averse, there will be a positive supply response. Hence, the output of sugar is related to the blend ratio. In Figure 2, the supply of sugarcane is given by Ss. Total demand is given by DT (assuming a 20 percent blend ratio), which includes not only the domestic demand for sugarcane for fuel but also the domestic

1

An anonymous reviewer pointed out that a major factor influencing the Government’s choice of the blend ratio is the level of domestic sugar production. Blend ratios increase when sugar production is high. In part this is done to avoid dumping excess sugar on the world market. 4

demand for sugar (DS) and the export demand for sugar ED. In equilibrium, the price is p1, and q1 of output is produced. Suppose now that the Government decides to increase the blend ratio for fuel. This shifts the total demand for sugar to DT′. For sugarcane growers, the price increases to p2. Output increases to q2. However, what if producers are “risk averse” and respond to the policy under the perception that increased policy certainty is added to the market? Supply now shifts to Ss1 (Just, Hueth, and Schmitz, 1982). As a result, output increases further to q3. But note that, in this particular case, prices are not affected by the change in the blend ratio (the supply curve, in the short run, is much more inelastic, at least for prices below p1). As a result, exports are unaffected as is the domestic consumption of sugar. The use of sugarcane for fuel clearly increased. Here is a case where policy is production distorting but not trade distorting.

Ss Ss1 e

p2 d

c

p1

a

DT1

b DT ED q21

q31

DS q1

q2

Figure 2. Risk Averse Supply Response ad Procuersers

5

q3

In the above model, the blend ratio was not necessarily changed because of higher sugar output. A higher sugar output can result from a fixed blend ratio under risk-averse behaviour. Now consider the model presented in Figure 3. The supply of sugarcane is given by S. The aggregate derived demand for sugarcane consists of several components giving rise to total demand D. These include export demand for sugar Ed, domestic demand for sugar Dd, and demand for sugarcane for ethanol production De. We assume that De is based on a blend ratio of 20 percent. Note that total demand D is conditional on the magnitude of the blend ratio.

Figure 3. Sugar Production and Ethanol Demand.

6

Unlike in our previous work (Schmitz, Schmitz, and Seale, 2002), we assume that the supply of sugar is uncertain, and hence we deal with expected prices and quantities ex ante and compare these with ex post outcomes. In the model, the expected price of sugar at the farm level is pˆ and the expected output is qˆ . Now suppose that actual output (qa) exceeds the expected output. If the blend ratio of 20 percent is unaltered, sugar prices will fall to p1. Also exports will increase from x to x  . Suppose that in response to the higher than expected output, the government increases the blend ratio to 25 percent. This will shift the derived demand curve for sugar via an increase in ethanol demand to the right, shifting the total demand for sugar to D  , leaving sugar prices unaffected by the high supply. In other words, the ethanol program absorbs the output that exceeded the expected output. The ethanol program essentially props up sugar prices (Schmitz, Seale, and Schmitz, 2003). In the model, note that the sugarcane supply and blend ratio move together; as the supply increases, so does the blend ratio.2 In the above, the blend ratio was changed in response to sugar production. Now consider the effect of such factors as the price of oil. In Figure 6, assume that the expected output of sugar equals actual output, but the blend ratio is increased due to rising oil prices. In this case, aggregate demand for sugar increases to D  , causing the price of sugar to rise and output to increase (one could also consider the combined effects on the blend ratio of changes in oil prices and changes in sugar output).

2

In an earlier work (Schmitz, Schmitz, and Seale, 2002), an anonymous reviewer commented that a major factor influencing the Government’s choice of the blend ratio is the level of domestic sugar production. The blend ratio increases when sugar production is high. Thus there is a degree of simultaneity between sugar production and the blend ratio. The model above, which uses an expectation framework, can account for this observation. Our empirical results strongly suggest this is the case. Interestingly, the Brazilian Minister in charge stated that the reason for increasing the blend ratio, due to the huge sugar crop of 2002/03, was to prevent the sugarcane crop from flooding the international sugar market. 7

Regression Analysis In the section above, when supply is uncertain, it is possible for actual supply of sugarcane to be greater than expected supply. If the blend ratio is unchanged, sugarcane prices will fall due to the higher than expected quantity of sugarcane production. Exports of refined sugar would also increase. The analysis suggests that the government, if it were interested in supporting sugarcane and sugar prices, would raise the blend ratio. Thus, there should be a negative relationship between sugar prices and the blend ratio. There should also be a positive relationship between exports and refined sugar production and the blend ratio. Also, if the government changes the blend ratio in response to oil prices, there should be a positive and significant relationship between the blend ratio and oil prices. To explore these relationships, we regress the average annual blend ratio on either current or lagged real sugar prices, real oil prices, export quantities, and sugar-quantity production starting in 1975, when PROALCOOL was created, until 2003. There should also be a positive relationship between anhydrous production and sugarcane production. However, we do not have a full set of data for these latter two variables. As shown in Table 2, however, the correlation among sugar-quantity variables is positive and highly significant. As such, inclusion of any of these quantity variables for which we have a full data set should essentially give us the same statistical results as if we included the quantity of sugarcane produced. For example, the correlation between sugarcane production and sugar production (exports) is .78 (.71). The same is true for the correlation of lagged sugarcane production and lagged sugar production (exports) with a correlation of .79 (.69). Tobit regression is utilized because the blend ratio is truncated below zero and is also truncated above by technology constraints at somewhere around 30 percent. We regress the

8

average annual blend ratio on current or lagged (or some combination of these) real sugar prices, real oil prices, or either sugar production or sugar exports. We do so is to see whether or not the parameter estimates are sensitive statistically to inclusion or exclusion of the explanatory variables. We do not include sugar production or sugar exports together in the regression analysis because of the high collinearity between these two variables. We report the parameter results in Table 1 with standard errors reported in parentheses. Table 1. Correlation matrix of Brazilian-sugar-industry quantities, 1975-2000. Correlation Matrix

SUGCAN ANHYDRO SUGAR EXP DCOMP

SUGCAN 1.00000 0.76883 0.78183 0.71276 0.81095

ANHYDRO

SUGAR

EXP

DCOMP

1.00000 0.91120 0.90290 0.78133

1.00000 0.08027 0.87902

1.00000 0.81213

1.00000

Correlation Matrix SUGCAN(-1) ANHYDRO(-1) SUGAR(-1) EXP(-1) DCOMP(-1) SUGCAN 1.00000 ANHYDRO 0.79478 1.00000 SUGAR 0.78991 0.89418 1.00000 EXP 0.69118 0.87167 0.97612 1.00000 DCOMP 0.85119 0.76842 0.83580 0.74466 1.00000 SUGCAN = sugarcane production; ANHYDRO = anhydrous alcohol production; SUGAR = refined sugar production; EXP = sugar exports; DCOMP = domestic consumption of sugar. Data sources: Brazilian centrifugal sugar production, sugar exports, and domestic consumption are from USDA. From 1975 to 1976 : 1981 Commodity Year Book; from 1977 to 1981-2 : 1986 Commodity Year Book; from 1982-3 to 1885-6 : 1989 CRB Commodity Year Book; from 1986-7 to 1990-1 : 1993 CRB Commodity Year Book; from 1991-2 to 1998-9: 2001 CRB Commodity Year Book; 1999-00 to 2003-04: Haley and Suarez, 2004 http://www.fas.usda.gov/htp/sugar/2003/November%202003%20PSD.pdf . Sugarcane and anhydrous alcohol production data are from 70/71 and 80/81, Nastari (1983); 81/82 to 99/00; Informative Datagro, IAA, MIR, MIC, MICT.

9

From the evidence, it is clear that both current and real sugar prices have a negative and significant effect on the blend ratio no matter what other explanatory variables are included or excluded in the regression. For example, the parameter estimate on current (lagged) real sugar prices is approximately -.1 (-.8) for any combination of the other variables. It is also clear that sugar production and sugar exports, whether current or lagged, have a positive and significant relationship to the blend ratio. It should also be noted that these parameter estimates are stable across equations.

10

H:\1RESEARCH\ARTICLES\1PUBLISHED\brazil blend ratio\paper\1-Brazil's Ethanol Blend Ratios 043041.doc @2pm Table 2. Tobit regression results, blend ratio on real prices and quantity variables, 1975-2002. Independent Variables BLEND RATIOS

a

Sugar price

Sugar price(-1)c Oil priceb Oil price(-1)d Sugare Sugar(1)f

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

-.102 (041)i

-.105 (.042)

-.111 (.040)

-.118 (.040)

-.086 (.041)

-0.89 (.038)

-.099 (.040)

-.108



— .082 (.064)

— .067 (.064)

Constant

.045 (.063)

-.079 (.032)

-.086 (.034)





.084 (.065)

.068 (.066)

.074 (.063)

.046 (.065)



.405 (.109)





.466 (.137)

16.156 (2.221)

16.626 (2.276)

.128 (.065)

.116 (.060)

.101 (.063)





.358 (.109)







.529 (.1380 — 16.964 (1.722)



.428 (.1580



18.194 (1.632)

19.140 (1.641)

14.307 (2.358))

.433 (.126)

14.389 (2.498)

(13)

(14)



-.074 (.035)

.342 (.107)





-.072 (.034)

.133 (.063)

.349 (.121)

(12)







(11)





Exportsg Exports(-1)h

.073 (.061)

(10)



(15)

(16)



-.068 (.031)

-.070 (.032)

-.078 (.030)

-.086 (.032)

.154 (.062)

.149 (.064)

.138 (.060)

.126 (.064)



.423 (.105) —





.483 (.140)



.555 (.135)

.516 (.164)





.449 (.161)





.552 (.160)

17.634 (1.822)

15.720 (2.239)

17.871 (1.653)

18.804 (1.656)

13.643 (2.204)

16.350 (1.631)

16.927 (1.724)

.366 (.123)

16.187 (2.301)

.455 (.122)

13.667 (2.344)



a

Sugar price is real sugar price. Oil price is real oil price. Real sugar price lagged one time period. d Real oil price lagged one time period. e Sugar production in Brazil. f Lagged Brazilian sugar production. g Exports of raw Brazilian sugar. h Logged exports of raw Brazilian sugar. i Standard errors are in parentheses. b c

Data Sources:From 1970 to 1982 prices are for Spot Raw Sugar International Sugar Agreement World Price. From 1983 to 2003 prices are for World Raw Sugar Price. The primary source of both prices is the Economic Research Service, USDA. Crude Oil Nominal Prices: Saudia Light 34 API prices are usually F.O.B. at the foreign port of landing. Prices are as of the Friday that is closest to January 1, except in 1987, when prices are as of the first Friday of February. Sources: 1970-1978—Petroleum and Energy Intelligence Weekly, Inc., Petroleum Intelligence Weekly; 1979 forward—Energy Information Administration, Weekly Petroleum Status Report. http://www.eia.doe.gov/emeu/aer/txt/ptb1107.html Nominal prices are in U.S. dollars and are converted into real prices using the Consumer Price Index: Bureau of Labor Statistics—All Urban Consumers — (CPI-U). ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt Brazilian centrifugal sugar production, sugar exports, and domestic consumption are from USDA. From 1975 to 1976 : 1981 Commodity Year Book; from 1977 to 1981-2 : 1986 Commodity Year Book; from 1982-3 to 1885-6 : 1989 CRB Commodity Year Book; from 1986-7 to 1990-1 : 1993 CRB Commodity Year Book; from 1991-2 to 1998-9: 2001 CRB Commodity Year Book; 1999-00 to 2003-04: Haley and Suarez, 2004 http://www.fas.usda.gov/htp/sugar/2003/November%202003%20PSD.pdf .

H:\1RESEARCH\ARTICLES\1PUBLISHED\brazil blend ratio\paper\1-Brazil's Ethanol Blend Ratios 043041.doc @2pm There is little evidence that the Brazilian government changed the blend ratio in response to changes in real oil prices over the period of 1975 through 2003. In all cases, the parameter estimates of current real oil prices are positive but are statistically the same as zero ( =.05). Only in the cases where lagged real oil prices are included with lagged real sugar prices do the parameter estimates on real oil prices become both positive and statistically significant. It is not surprising that there is only a weak relationship between blend ratios and oil prices. Brazil is close to being self-sufficient in oil. Also, huge reserves of natural gas have been found. This does not imply, however, that oil prices were not important in the development of the ethanol industry in Brazil. According to Knight (2003), the reason why the Brazilian sugar industry expanded so much is due to the quadrupling of the oil prices in the 1970s. Our analysis did not include this period. Likewise, Haley, Suarez, and Bolling (2003) point out that there has been rapid growth in the use of sugarcane as alcohol in 2003/04, taking up 53 percent of the sugar forecast crop. They argue this is due to the high international petroleum prices where alcohol is the principal alternative fuel. This period is not included in our analysis.

CONCLUSIONS The Brazilian government changes from time to time the alcohol blend ratio. For example, the government set the alcohol content in gasoline at 25 percent as of July 1, 2003, up from the 20 percent share set on February 1, 2003 (Haley, Suarez, and Bolling, 2003). Some gasoline companies and stations have even used a higher proportion of anhydrous alcohol in the mixture than the government-mandated 25 percent. In terms of the future use of alcohol as fuel, automobile manufacturers have launched “flex-fuel” models, which are fitted with engines that can run equally well on either gasohol or

hydrous alcohol, or any mixture of the two. Some suggest that 70 to 80 percent of all new cars will be fitted with “flex-fuel” engines. This added flexibility may further add to the demand for ethanol produced from sugarcane. In addition, Brazil is seeking avenues to increase ethanol exports. Brazil is now the leading exporter of both sugar and fuel alcohol, with India becoming a major importer.

13

REFERENCES FNP. (2002) Agrianual 2001 Anuario da Agricultura Brasileira. Sao Paulo, Brazil. JOB Economica. (2002) Weekly Sugar and Alcohol Report (4 April). Sao Paulo, Brazil. Peter Buzzanell & Associates, Inc. (2000) The Brazilian Sugar and Alcohol Industry: 2000 and Beyond. Reston, VA (June). Haley, Stephen, and Nydia R. Suarez. (2004) Sugar and Sweeteners Outlook, USDA/ERS SSS-239, United States Department of Agriculture, Economic Research Service, January 30. http://www.fas.usda.gov/htp/sugar/2003/November%202003%20PSD.pdf Haley, Stephen, Nydia R. Suarez, and Christine Bolling. (2003) Sugar and Sweeteners Outlook, USDA/ERS SSS-238, United States Department of Agriculture, Economic Research Service, September 30. Available online at http://www.ers.usda.gov. Just R., D. Hueth, and A. Schmitz. Applied Welfare for Economics and Public Policy: Prentice Hall, 1982. Knight, Patrick. (2003) Brazil—What a Difference a Year Makes. International Sugar and Sweetener Report, Volume 135, No. 34, 607-610, November 18. Schmitz, Andrew, James L. Seale, Jr., and Troy G. Schmitz. (2003) Sweetener-Ethanol Complex in Brazil, the United States, and Mexico: Do Prices Matter? International Sugar Journal 105 (1259): 505-513. Schmitz, Troy G., Andrew Schmitz, and James L. Seale, Jr. (2003). Brazil’s Ethanol Program: The Case of Hidden Sugar Subsidies. International Sugar Journal 105 (1254): 254-265. Schmitz, Troy G., James L. Seale, Jr., and Peter J. Buzzanell. Brazil’s Domination of the World Sugar Market. In Sugar and Related Sweetener Markets in the 21st Century:

14

International Implications, edited by Andrew Schmitz, Thomas Spreen, Charles B. Moss, and William Messina. Wallington, England: CABI Publishers, November 2002. USDA (United States Department of Agriculture). (Various years) Sugar and Sweetener Situation Outlook Yearbook. Market and Trade Economics Division, Economic Research Service, Washington, DC. _____. (2001) Brazil Sugar Semi-Annual Report. Sao Paulo, Brazil: U.S. Agricultural Trade Office, October.

15