A Closer Look at the Oil Price - Gasoline Price Relationship in Ontario

In a recent post, Myles Harrison investigated the relationship between Ontario average gasoline prices and US oil prices. He concluded by showing a strong positive relationship between the two. The recent drop in oil prices has led to a drop in gasoline prices, but my question is whether gas prices are as low as they should be given that oil is currently a few cents below $60 per barrel.

In this post, I will conduct further analysis by extending his work in several directions. First, Myles loaded the gasoline price data from the Ontario government website using a call to wget. This is certainly an easy way to pull down data from websites. The only drawback is that two separate programs are being used: one to load the data, and another to analyze the data. I will load the data directly into R. In addition I will load additional data on oil prices, exchange rates, and the Canadian CPI. My approach results in code that is more lengthy than Myles, but has the advantage of being contained in one program. The data set is monthly starting in January 1990 and continuing until March of 2015. Second, I will investigate the gasoline price - oil price relationship using 4 different specifications of the data. From these relationships I will make a forecast of current gasoline prices based off of current oil prices. Third, I will provide some tests for residual stationarity to ensure that the the variables are cointegrated.

The gasoline price data comes from an Ontario government website. Additional data on oil prices, exchange rates, and Canadian consumer price index (CPI) comes from the Federal Reserve of St. Louis database (FRED data).

First. here are some plots of Ontario gasoline prices and US oil prices. As expected, the two series move together closely.



 I estimate four specifications.

S1: gasoline prices ~ oil prices (nominal values not adjusted for exchange rates)
S2: log(gasoline prices) ~ log(oil prices) (nominal values not adjusted for exchange rates)
S3:  gasoline prices ~ oil prices (adjusted for exchange rates and inflation)
S4: log( gasoline prices) ~ log(oil prices) (adjusted for exchange rates and inflation)

Specification S1 is the most basic and does not adjust for inflation or different currencies. Notice the strong positive correlation between gasoline prices and oil prices. The simple correlation between the two is 0.973.

Using this specification to forecast gas prices (for oil prices of $50, $55, $60, $65, labelled as points 1 through 4 respectively) results in the following table. For example, at oil prices of $60 (point 3) the forecast of average Ontario gasoline prices is 90.9 cents/liter. The most recent actual value is $1.16 which is beyond the upper 95% confidence level. Based on this analysis, gas prices are too high.

Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
1           82.5  74.3  90.6  70.0  95.0
2           86.7  78.5  94.9  74.2  99.2
3           90.9  82.8  99.1  78.4 103.4
4           95.1  87.0 103.3  82.6 107.7

Also notice the increase in variability towards the right of the plot. This can be controlled for by estimating a log - log specification (S2).

Specification 4 is the one that is most consistent because each variable is expressed in real Canadian dollars. The log - log specification accounts for changing variability in the data.




Here is a table comparing the gasoline price forecasts from each of the four specifications.

                      S1     S2     S3     S4
Oil coefficient    0.845  0.477  0.702  0.439
t stat            72.871 81.483 55.576 51.631
R-squared          0.946  0.957  0.911  0.899
Forecast (oil=50) 82.472 86.766 86.606 88.295
Forecast (oil=55) 86.696 90.801 90.116 92.068
Forecast (oil=60) 90.921 94.648 93.626 95.653
Forecast (oil=65) 95.146 98.330 97.136 99.074

The estimated coefficient on the oil price variable is positive and statistically significant in each specification. The dependent variable in each specification is different so it is not possible to compare across models, but clearly for each specification, oil prices have a statistically significant positive impact on Ontario gasoline prices. The gasoline price forecast for oil prices at $60 range between 90.9 cents/liter to 95.7 cents/liter. The interesting point is that the current price of gasoline at $1.16 per liter is higher than the upper 95% confidence bound in all cases.

Lastly, the table below shows the p-values from an ADF test on the residuals from each specification. The residuals from each specification are stationary at 5% which is consistent with a cointegrating relationship between gasoline prices and oil prices.

                            S1       S2       S3     S4
Residuals adf p value 0.000867 0.000264 0.000044 0.0464

The main take-away is that Ontario gasoline prices and oil prices are cointegrated and this is robust to several different regression specifications. Based on this analysis, gasoline prices are higher than the upper 95% confidence band.



Here is the R code.

#########################################################
#  Economic forecasting and analysis
#  Perry Sadorsky
#  June 2015
#  Forecasting Ontario gasoline prices
##########################################################

## basic reference is from this post
## http://www.r-bloggers.com/what-a-gas-the-falling-price-of-oil-and-ontario-gasoline-prices/


# load libraries
library(fpp)
library(quantmod)
library(reshape2)
library(urca)


##########################################################
## collect data from Ontario government
##########################################################

furl = c("http://www.energy.gov.on.ca/fuelupload/ONTREG1991.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1992.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1993.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1994.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1995.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1996.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1997.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1998.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG1999.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2000.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2001.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2002.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2003.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2004.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2005.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2006.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2007.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2008.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2009.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2010.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2011.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2012.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2013.csv",
         "http://www.energy.gov.on.ca/fuelupload/ONTREG2014.csv"        
         )

gasp = c(1,2,3,4,5,6,7,8,9,10,11,12)
xx = read.csv("http://www.energy.gov.on.ca/fuelupload/ONTREG1990.csv",skip=1)
xx = head(xx,-3)
xx = tail(xx,12)
xx = xx[,"ON.Avg"]
gasp = cbind(gasp,xx)


for (i in 1:24) {
  xx = read.csv(furl[i],skip=1)
  xx = head(xx,-3)
  xx = tail(xx,12)
  xx = xx[,"ON.Avg"]
  gasp =cbind(gasp,xx)
}

## output is a nice matrix
View(gasp)


## stack the data into a column
s_gasp = melt(gasp)
s_gasp = s_gasp[-c(1:12),] 
View(s_gasp)


on_avg = s_gasp[,3]
tail(on_avg)
View(on_avg)


## append 2015 data
# "http://www.energy.gov.on.ca/fuelupload/ONTREG2015.csv"

yy = read.csv("http://www.energy.gov.on.ca/fuelupload/ONTREG2015.csv",skip=1)
yy = head(yy,-3)
yy = tail(yy,6)
yy= yy[,"ON.Avg"]
yy

on_avg_a = append(on_avg,yy)
tail(on_avg_a)
View(on_avg_a)

gas = ts(on_avg_a, start=1990, end=c(2015,3), frequency=12)
plot(gas,main="Ontario gas prices (cents per liter)", xlab="",ylab="")


##########################################################
## now load data on oil prices and FX
##########################################################

# load data from FRED into new environment
symbol.vec = c("MCOILWTICO" , "EXCAUS", "CANCPIALLMINMEI")
data <- new.env()
getSymbols(symbol.vec, src="FRED", env = data)


(data$MCOILWTICO["1990-01-01::"])-> oil
(data$EXCAUS["1990-01-01::"])-> fx
(data$CANCPIALLMINMEI["1990-01-01::"])-> cpi

# rebase cpi
tail(cpi,14)
cpi2014 = mean(  cpi[(300-11):300]  )
cpi_2014 = cpi/cpi2014 * 100
mean(  cpi_2014[(300-11):300]  )


par(mfrow = c(3, 1))
plot(oil,main="Oil prices")
plot(fx,main="$C/$US")
plot(cpi_2014,main="CPI")
par(mfrow = c(1, 1))


df_fred = cbind(oil,fx,cpi_2014)
head(df_fred)
tail(df_fred)
## latest cpi value is for March

df_fred_1 =  ts(df_fred, start=1990, end=c(2015,3), frequency=12)
nrow(df_fred_1)
length(gas)

oil_n = df_fred_1[,1]

par(mfrow = c(2, 1))
plot(gas,main="Ontario gas prices (cents per liter)", xlab="",ylab="")
plot(df_fred_1[,1],main="Oil prices ($/bbl)", xlab="",ylab="")
par(mfrow = c(1, 1))

cor(gas,df_fred_1[,1])


##########################################################
## linear regressions
##########################################################

table = matrix( NA, nrow=7, ncol=4)
table2 = matrix( NA, nrow=1, ncol=4)

lm1 = lm(gas ~ oil_n)

lm1s = summary(lm1)
table[1,1] = lm1s$coefficients[2,1]
table[2,1] = lm1s$coefficients[2,3]
table[3,1] = lm1s$r.squared


plot(gas ~ oil_n,
     ylab="Ontario gas prices (cents per liter)", xlab="WTI oil prices ($/bbl)")
abline(lm1)
adf1 =  ur.df(lm1$residuals, type="drift",selectlags="AIC", lags=12)
adf1@testreg$coefficients
table2[1,1] = adf1@testreg$coefficients[2,4]

## some forecasting
fcast <- forecast(lm1, newdata=data.frame(oil_n=c(50, 55, 60, 65)))
fcast
plot(fcast)


table[4,1] = fcast$mean[1]
table[5,1] = fcast$mean[2]
table[6,1] = fcast$mean[3]
table[7,1] = fcast$mean[4]
table



## log- log specification
lm2 = lm(log(gas) ~ log(oil_n))
summary(lm2)

lm2s = summary(lm2)
table[1,2] = lm2s$coefficients[2,1]
table[2,2] = lm2s$coefficients[2,3]
table[3,2] = lm2s$r.squared

plot(log(gas) ~ log(oil_n),
     ylab="Log of Ontario gas prices (cents per liter)", xlab="Log of WTI oil prices ($/bbl)")
abline(lm2)
adf2 =  ur.df(lm2$residuals, type="drift",selectlags="AIC", lags=12)
adf2@testreg$coefficients
table2[1,2] = adf2@testreg$coefficients[2,4]

## some forecasting
fcast <- forecast(lm2, newdata=data.frame(oil_n=c(50, 55, 60, 65)))
fcast
plot(fcast)

exp(fcast$mean)

table[4,2] = exp(fcast$mean[1])
table[5,2] = exp(fcast$mean[2])
table[6,2] = exp(fcast$mean[3])
table[7,2] = exp(fcast$mean[4])



## now do this in real Canadian dollars
gas_r = gas/df_fred_1[,3]*100
oil_r = (df_fred_1[,1]*df_fred_1[,2])/df_fred_1[,3]*100


lm3 = lm(gas_r ~ oil_r)
summary(lm3)

lm3s = summary(lm3)
table[1,3] = lm3s$coefficients[2,1]
table[2,3] = lm3s$coefficients[2,3]
table[3,3] = lm3s$r.squared

plot(gas_r ~ oil_r,
     ylab="Ontario real gas prices (cents per liter)", xlab="Real oil prices ($/bbl)")
abline(lm3)
adf3 =  ur.df(lm3$residuals, type="drift",selectlags="AIC", lags=12)
adf3@testreg$coefficients
table2[1,3] = adf3@testreg$coefficients[2,4]

## some forecasting
fcast <- forecast(lm3, newdata=data.frame(oil_r=c(50, 55, 60, 65)))
fcast
plot(fcast)

table[4,3] = fcast$mean[1]
table[5,3] = fcast$mean[2]
table[6,3] = fcast$mean[3]
table[7,3] = fcast$mean[4]



## log- log specification in real Canadian dollars
lm4 = lm(log(gas_r) ~ log(oil_r))
summary(lm4)

lm4s = summary(lm4)
table[1,4] = lm4s$coefficients[2,1]
table[2,4] = lm4s$coefficients[2,3]
table[3,4] = lm4s$r.squared


plot(log(gas_r) ~ log(oil_r),
     ylab="Log of real Ontario gas prices (cents per liter)", xlab="Log of real oil prices ($/bbl)")
abline(lm4)
adf4 =  ur.df(lm4$residuals, type="drift",selectlags="AIC", lags=12)
adf4@testreg$coefficients
table2[1,4] = adf4@testreg$coefficients[2,4]

## some forecasting
fcast <- forecast(lm4, newdata=data.frame(oil_r=c(50, 55, 60, 65)))
fcast
(fcast$mean)
(fcast$lower)
(fcast$upper)
# plot(fcast)
exp(fcast$mean)
exp(fcast$lower)
exp(fcast$upper)

fcast4 = cbind(exp(fcast$mean),exp(fcast$lower),  exp(fcast$upper))
colnames(fcast4) = c("Forecast", "Lo 80", "Lo 95", "Hi 80", "Hi 95")
fcast4

table[4,4] = exp(fcast$mean[1])
table[5,4] = exp(fcast$mean[2])
table[6,4] = exp(fcast$mean[3])
table[7,4] = exp(fcast$mean[4])


colnames(table) = c("S1", "S2", "S3", "S4")
rownames(table) = c("Oil coefficient", "t stat", "R-squared", "Forecast (oil=50)", "Forecast (oil=55)", "Forecast (oil=60)", "Forecast (oil=65)")

options(digits=3, scipen=10)
table


colnames(table2) = c("S1", "S2", "S3", "S4")
rownames(table2) = c("Residuals adf p value")
table2

Seach Interest in the Tar Sands Peaked in 2006

Here are some charts showing how Google searches of terms tar sands, oil sands, and fracking compare. On a regional basis, searches for terms like tar sands or oil sands are mostly from Canada. This is a bit of a surprise, since the assumption here in Canada is that the world is very interested in the tar sands. It appears that there is much more interest in fracking.

Renewable Energy Moving Forward

The Pew Charitable Trusts recently published a research report on the state of renewable energy. They are  predicting that global revenue from the installation of renewable energy technologies would grow at a compound annual rate of eight per cent from $200bn in 2012 to $327bn by 2018.  This would create  cumulative revenue of $1.9tr. The compound annual rate of 8% is consistent with other projections (eg. IEA). In general, most analysis shows that renewable energy usage is the fastest growing component of the energy mix.

The Pew report points out that while there are tremendous opportunities for countries to profit from this trend in renewable energy investment, countries without a well formulated energy policy are likely to lose out. The report highlights the case in the US  but this equally applies to Canada. There are 118 countries with renewable energy targets. Unfortunately, neither Canada or the US is among them.

Here are a few interesting figures from the Pew report showing that G20 countries are leading the way in clean energy investment and how the cost of solar energy modules has fallen dramatically.

In order to get an idea of how renewable energy depends upon income and CO2 emissions, I gathered some data on world energy consumption, CO2 emissions, GDP, and renewable energy production from the World Bank on line database.Renewable energy includes biomass, wood waste, geothermal, solar, wind, tide, wave, etc. but excludes hydroelectric power.

The units for my variables are:

energy is measured in millions of kt of oil equivalent
renewable energy is measured as electricity production from renewable sources, excluding hydroelectric (billions of kWh)
CO2 is measured in millions of kt of carbon dioxide emissions
GDP is measured in trillions of 2005 international dollars

Looking at year over year % changes indicates that energy use, GDP, and CO2 emissions track each other very closely. Notice that renewable energy tends to have much greater fluctuations.

Here is how renewable energy correlates with CO2 emissions. Both variables are measured in natural logarithms.

Along the best fit line, a 1% increase in CO2 emissions is associated with a 4.59% increase in renewable energy.


Here is how renewable energy correlates with GDP. Both variables are measured in natural logarithms.

Along the best fit line, a 1% increase in GDP is associated with a 2.99% increase in renewable energy. The high income elasticity is consistent with some of my previous research on renewable energy consumption in developed and emerging economies (here, here).

By comparison, a 1% increase in GDP is associated with a 0.58% increase in total energy consumption.

As the world economy rebounds from the Great Recession, economic activity will increase and GDP will increase. Increases in GDP have a bigger impact on renewable energy consumption than total energy consumption, so expect to see further increases in renewable energy in the future. Unfortunately, without a reasonable energy policy, Canada and the US will be left on the sidelines as other countries capture competitive advantage in the renewable energy sector.

A Sovereign Wealth Fund for Canada

The Canadian International Council has recently released a new report entitled “Nine Habits of Highly Effective Resource Economies.” (see here for the report).  According to the report, Canada is flush with valuable natural resources but lacks the capabilities and foresight to exploit these natural resources in a sustained value enhancing manner that will create long term growth and prosperity for Canadians. Canada's exploitation of natural resources can be nicely categorized as "rip and ship".

The details of the nine habits are in the report, but here is a listing of the nine habits.

Foreign ownership of Canada's natural resources is one area that is particularly worrisome. In general, Canadian natural resource companies grow to a reasonably  large size and, rather than striving to become global leaders, choose instead to sell out to a foreign competitor. A recent example of the sell to foreigners approach is China’s state-owned CNOOC Ltd.$15-billion bid for Calgary-based oil producer Nexen Inc.

For a country that has a lot of natural resources, Canadian resource companies are not prominently listed among the world's biggest resource companies. Canada is the world’s top producer of potash and titanium and ranks among the top 10 producers of forest products, uranium, aluminum, natural gas, sulphur, tungsten, diamonds, asbestos, nickel, platinum, crude oil, molybdenum, zinc, and gold. With such an impressive production record, one would expect Canadian natural resource companies to rank among the biggest in the world. Potash is the world's largest potash company, but most Canadian resource companies do not rank among the global giants.

Foreign ownership is particularly evident in the Alberta tar sands where by some estimates more than two-thirds of all tar sands production in Canada is owned by foreign entities (see here). This sends a majority of the profits from oil produced from the tar sands outside of  Canada. So, Canada needs a different approach if it is going to benefit from its natural resource wealth. 

A sovereign wealth fund (SWF) is one approach that Canada can expand upon. After all, Alberta has the Heritage Fund, so why not create a natural resource based sovereign wealth fund for Canada as a whole.

Here is a ranking of the world's top sovereign wealth funds. The Alberta Heritage Fund ranks 29th. Notice that Alberta's fund was started in 1976, the same year that Alaska started theirs. The Alaska fund is, however, 3 times larger than Alberta's. Australia started their fund in 2006, and look at how much money it already has. When compared against other SWFs, the Alberta Heritage Fund doesn't seem to be doing so well. A recent Toronto Star article on Norway points out that at least from Norway's perspective, Canada is a nice place with lots of natural resources, but badly managed.

Data sourced from Sovereign Wealth Fund Institute

The current situation in Canada can be described as too much foreign ownership and too low of a tax base for Canada to effectively generate wealth from its natural resources. This makes Canadian natural resources vulnerable to international rent seekers.

So what is Norway doing so well?

Norway's SWF was setup in 1990 and currently has slightly over $650 billion dollars. The fund is on track to amass $1 trillion by the end of this decade. The fund is an excellent case study on portfolio investing (see here). Norway faced foreign ownership problems in the oil business as well but they responded with a 90% marginal tax and focused on training their own citizens to be the primary source of employment in the oil and gas sector. In comparison, Alberta has a miniscule 10% royalty tax and companies working in Alberta outsource as much capital and labour as they can with the predictable result that Alberta is earning a fraction of what it should be from its valuable natural resource base. Norway is also not afraid of starring down carbon nay-Sayers. Norway recently announced that it would increase its current carbon tax on offshore oil companies by £21 to £45 per tonne of carbon. Norway also has a carbon tax imposed on the fishing industry.

Perhaps Canada needs to re-think the concept of a state owned oil company. 

Ontario Electricity Generation on a Hot Day

Today is a hot day in Toronto. The temperature is 34 degrees and it is only 2 pm. Toronto's Medical Officer has upgraded the heat alert to an extreme heat alert. Cities across the province are finding it very hot as well.
On hot days like today, demand for air conditioning swells. Air conditioning requires electricity, so what fuel sources are being used to generate electricity?

Source	MW	%
Nuclear	10857	45.31
Hydro	4658	19.44
Gas	5410	22.58
Coal	1722	7.19
Wind	441	1.84
Other	873	3.64
Total	23961	100
source:http://www.ieso.ca/
Fuels used to meet demand Jun. 20 - 12:00-13:00

Nuclear is the most important fuel. Nuclear, hydro and natural gas provide 87.3% of the fuel. Other refers to wood waste, biogas, solar, etc. Electricity generation in Ontario is very dependent on nuclear. For the year to date, nuclear has accounted for 56.9% of the fuel used to generate electricity in Ontario. Renewables make up a relatively small proportion of total fuel source. Notice that coal, which used to be a big fuel source, now accounts for just 2.7% of total fuel usage.

Today's electricity demand is projected to peak is 24,267 MW at 5 pm.
The summer record was 27,005 MW on August 1, 2006.

Gasoline Prices Around the World

Bloomberg has put together a nice slideshow of gasoline prices around the world. I have created a table from the data in the slideshow. The data on gasoline prices were collected between April 2 and April 11 of this year.  Premium gasoline was used in order to account for differences in octane. The most expensive gasoline is in Norway ($9.69) while the cheapest gasoline is in Venezuela ($0.09). Canada ranks at #38 ($5.75) below Japan ($7.58) and above the US ($4.19). Compared to other developed economies, gasoline prices in Canada and the US are actually fairly cheap. Bloomberg also calculates something which they call the pain at the pump (the percentage of daily income needed to buy a gallon of gasoline). In Canada, a gallon of gasoline costs just 4% of average daily income. Gasoline consumers in UAE, Kuwait, Saudi Arabia and Venezuela have cheap gasoline and very low pain at the pump. For other countries, the pain at the pump values vary a lot between developed and developing countries.

Rank Country Price per gallon of premium % of average daily income Pain at the pump rank 1 Norway $9.69 3.6 48 2 Denmark $9.37 5.3 42 3 Italy $9.35 9.1 29 3 Netherlands $9.35 6.5 37 5 Greece $9.23 12 23 6 Sweden $8.97 4.8 44 7 Hong Kong $8.89 8.8 31 8 Portugal $8.85 na 21 9 United Kingdom $8.84 7.8 34 10 Belgium $8.82 6.5 37 11 France $8.72 7 35 12 Finland $8.59 6 40 13 Germany $8.56 6.9 36 14 Ireland $8.34 6.1 39 15 Switzerland $7.95 3.7 47 16 Slovakia $7.93 16 19 17 Hungary $7.69 19 11 18 Czech Republic $7.59 13 22 19 Japan $7.58 5.8 41 20 South Korea $7.57 11 26 21 Spain $7.55 8.1 32 22 Slovenia $7.54 10 28 23 Austria $7.45 5.2 43 24 Malta $7.32 12 25 25 Latvia $7.26 21 9 26 Luxembourg $7.24 2.1 51 27 Lithuania $7.24 19 12 28 Estonia $7.05 14 20 29 Poland $7.01 18 15 30 Cyprus $6.97 8 33 31 Bulgaria $6.94 33 5 32 Australia $6.75 3.8 49 33 Singapore $6.70 4.6 45 34 Romania $6.59 25 7 35 Chile $6.54 17 17 36 Brazil $6.41 18 13 37 India $6.06 135 1 38 Canada $5.75 4 46 39 South Africa $5.72 24 8 40 Seychelles $5.53 18 14 41 Argentina $5.44 18 16 42 China $5.31 34 4 43 Thailand $4.96 31 6 44 United States $4.19 3.1 50 45 Indonesia $4.11 40 3 46 Russia $3.71 9.1 29 47 Malaysia $3.49 12 24 48 Mexico $3.20 10 27 49 Iran $2.78 16 18 50 Nigeria $2.33 53 2 51 UAE $1.89 1 53 52 Egypt $1.73 20 10 53 Kuwait $0.88 0.7 54 54 Saudi Arabia $0.61 1.1 52 55 Venezuela $0.09 0.3 55