FAOSTAT is the source of forest products data and World Bank is the source of GDP, deflator and exchange rate data.
source("code/func.r")
print(load("enddata/EU27 paper products demand.rdata"))## [1] "paperProducts" "ppagg" "pptrade" "wb"
pp = paperProducts # Give a shorter name to the data frameIn the following file, we explore a data table containing paper and paperboard consumption data for 27 countries from 1961 to 2012.
In 2012, the overall EU consumption, production and trade in million Tons per item was:
## Item Consumption Production Net_Trade Import Export
## 205 Total Paper and Paperboard 80 92 12 49 61
## 206 Newsprint 7 8 0 5 5
## 207 Printing and Writing Paper 24 32 8 20 27
## 208 Other Paper and Paperboard 49 53 4 24 28
# Consumption, production, trade and net trade at the European Level
ggplot(data=subset(ppagg, !Element%in%c("Import_Value", "Export_Value","Price"))) +
geom_line(aes(x=Year, y=Value/1e6, colour=Item),size=1) +
facet_wrap(~Element) +
ylab("Million metric Ton") + theme_bw()
Net trade is slightly in favor of exports for most products except for newsprint. From bilateral trade statistics (not shown here), we know that most of the trade is happening inside the European Union.
Apparent consumption of Total paper and paperboard by country
ggplot(data=subset(pp, Item=="Total Paper and Paperboard"))+
geom_line(aes(x=Year, y=Consumption/1e+06, colour=Item), size=1) +
facet_wrap(~Country, scales = "free_y") +
xlab("Year") + ylab("Million metric Ton") +
theme_bw() + theme(legend.position="bottom")
The largest consumers of paper and paperboard products are France, Germany, Italy, Spain, United Kingdom.
# Create a new table with graphics papers following Lauri's comments
ppgraph = subset(pp, Item %in% c("Newsprint","Printing and Writing Paper"))
ppgraph = aggregate(ppgraph[c("Consumption", "Net_Trade")],
ppgraph[c("Year", "Country")], sum)
ppgraph$Item = "Graphics Paper"
ppgraph = rbind(ppgraph,
subset(pp, Item == "Other Paper and Paperboard",
c(Year, Country, Item, Consumption, Net_Trade)))
ggplot(data=ppgraph)+
geom_line(aes(x=Year, y=Consumption/1e+06, colour=Item), size=1) +
facet_wrap(~Country, scales = "free_y") +
xlab("Year") + ylab("Million metric Ton") +
# scale_colour_manual(values = c("Blue", "Green")) +
theme_bw() + theme(legend.position="bottom") 
Consumption of total paper and paperboard in
# add population data
pp = merge(pp, wb[c("Country","Population")])
ggplot(data=subset(pp, Item=="Total Paper and Paperboard"), aes(Year, Country)) +
geom_tile(aes(fill = 1000*Consumption/Population), colour = "white") +
scale_fill_gradient(low = "white", high = "steelblue")
ggplot(data=pp, aes(Year, Country)) +
geom_tile(aes(fill = 1000*Consumption/Population), colour = "white") +
scale_fill_gradient(low = "white", high = "steelblue") + facet_wrap(~Item)
Apparent consumption and Net Trade of Total paper and paperboard by country
ggplot(data=subset(pp, Item=="Total Paper and Paperboard"))+
geom_line(aes(x=Year, y=Consumption/1e+06, colour=Item, linetype = "Consumption"),
size=1) +
geom_line(aes(x=Year, y=Net_Trade/1e+06, colour=Item, linetype = "Net trade"),
size=1) +
facet_wrap(~Country, scales = "free_y") +
xlab("Year") + ylab("Million metric Ton") +
theme_bw() + theme(legend.position="bottom")
Large producing countries such as Finland and Sweden produce over 10 million tons of paper products per year and export most of their production. Germany is the largest consumer with over 20 million tons of paper products consumed annually, Germany's net trade is slightly positive. France is the second largest consumer and is slightly importing. The United Kingdom also consumes around 10 million tons annually and is importing more than half of its production.
# 1 plot per country
ggplot(data=wb, aes(x=Year, y=GDPconstantUSD/1e9)) +
geom_line(size=1) + xlab("Years") + ylab("GDP (constant) in 2010 Billion USD ") +
theme_bw() + theme(legend.position= "none") + facet_wrap(~Country, scales = "free_y")
Average trade price of paper products in the European Union
ggplot(data=subset(ppagg, Element=="Price")) +
geom_line(aes(x=Year, y=Value, color=Item)) +
ylab("Price in (2010) constant USD /Metric Ton")
Price has lowered since 2010, probably due to overcapacity in the market.
Prices are a ponderation of import and export prices as used in Chas-Amil and Buongiorno 2000. Prices are expressed in constant US dollars of 2010.
Price = (Import_Value + Export_Value)/ (Import_Quantity + Export_Quantity) / DeflUS *1000)
Countries with a particularly high price were removed from the set.
library(scales)
ggplot(data=subset(pp, !Country%in%c("Slovakia", "Malta"))) +
geom_line(aes(x=Year, y=Price, color=Item)) + facet_wrap( ~ Country) +
ylab("Price in (2010) constant USD /m3") + scale_y_continuous(labels = dollar)
ggplot(data=subset(pptrade, !Country%in%c("Slovakia", "Malta"))) +
geom_line(aes(x=Year, y=Price_Trade, color=Item, linetype = Trade))+
facet_wrap( ~ Country) +
ylab("Price in (2010) constant USD /m3")
# Sweden Germany France Finland
ggplot(data=subset(pptrade, Country%in%c("Sweden", "Germany", "France", "Finland"))) +
geom_line(aes(x=Year, y=Price_Trade, color=Item, linetype = Trade)) +
facet_wrap( ~ Country) + ylab("Price in (2010) constant USD /m3")
Number of countries for a given demand volume in million m3
# 1980, 1990, 2000, 2010
pp4hist = subset(pp, Year%in%c(1980, 1990, 2000, 2010)&Item=="Total Paper and Paperboard")
ggplot(data=pp4hist) + geom_histogram(aes(x=Consumption/1e6), binwidth=1) + facet_wrap(~Year)
Number of countries for a given demand volume in log of the volume in m3
# Consumption in log
ggplot(data=pp4hist) + geom_histogram(aes(x=log(Consumption)), binwidth=1) + facet_wrap(~Year)
ggplot(data=pp4hist) + geom_histogram(aes(x=log(Price)), binwidth=0.1) + facet_wrap(~Year)
ggplot(data=pp4hist) + geom_histogram(aes(x=log(GDPconstantUSD)), binwidth=1) + facet_wrap(~Year)
We will estimate the model $$ log(Consumption) = \beta_0 + \beta_1 log(GDP) + \beta_2 log(Price) + \beta_3 log(Consumption_{t-1}) $$
Lets look at the relationship between log(Consumption) and log(GDP) first.
Data for 2000 and beyond in red
plot(log(Consumption) ~ log(GDPconstantUSD), data=subset(pp, Item=="Total Paper and Paperboard"))## Warning: NaNs produced
points(log(Consumption) ~ log(GDPconstantUSD),
data=subset(pp, Item=="Total Paper and Paperboard"&Year>1999), col="red")
Sort countries by Net_Trade, then display a color for each country
d=data.frame(s=c(45,43))
ppc = subset(pp, Item=="Total Paper and Paperboard"&Year==2012)
ppc = ppc[order(ppc$Net_Trade),]
pp$Country = factor(pp$Country, levels= ppc$Country, ordered=TRUE)
p = ggplot(pp, aes(x=log(GDPconstantUSD), y=log(Consumption))) + facet_wrap(~Item)
p + geom_point(aes(alpha=Year, color=Country)) 
Check the ggplot shape parameter and the scale_shape_identity to improve this graph with a "+" sign for positive net trade and a "o" sign for negative net trade. Plot only Printing and Writing Paper for this graph.
p = ggplot(data=subset(pp, Item=="Printing and Writing Paper"),
aes(x=log(GDPconstantUSD), y=log(Consumption))) + facet_wrap(~Item)
p + scale_shape_identity() +
geom_point(aes(color=Country, shape=1 + 2*(Net_Trade>0) ))
plot(log(Consumption) ~ log(Price), data=subset(pp, Item=="Total Paper and Paperboard"))## Warning: NaNs produced
points(log(Consumption) ~ log(Price),
data=subset(pp, Item=="Total Paper and Paperboard"&Year==2012), col="red")
p = ggplot(pp,aes(x=log(Price), y=log(Consumption), alpha=Year)) + facet_wrap(~Item)
p + geom_point(aes(color=Country, alpha=Year)) 
Countries that trade less seem to have higher prices.
Change point shapes with a "+" sign for positive net trade and a "o" sign for negative net trade. Show only Printing and writing paper for this graph.
p = ggplot(data=subset(pp, Item=="Printing and Writing Paper"),
aes(x=log(Price), y=log(Consumption))) + facet_wrap(~Item)
p + scale_shape_identity() +
geom_point(aes(color=Country, alpha=Year, shape=1 + 2*(Net_Trade>0) )) 
Over the 1960 - 1992 estimation period.
EU15 = unique(wb$Country[wb$EU15==1])
# Create a column consumption at year t-1
pp = ddply(pp, .(Country,Item), mutate,
Consum_t_1 = c(NA,Consumption[-length(Consumption)]))## Warning: duplicated levels in factors are deprecated
# Select same data as in the chasamil paper
chasamil = subset(pp, Year>=1969&Year<=1992&Country%in%EU15 &
Country!="Ireland" & #Remove Ireland because it doesn't have GDPConstantUSD Data
Item=="Printing and Writing Paper", select=-Item)
# Set data as panel data
pchasamil <- plm.data(chasamil, index=c("Country","Year"))
pooling <- plm(log(Consumption) ~ log(GDPconstantUSD) + log(Price) + log(Consum_t_1),
data=pchasamil, model= "pooling")## Warning: NaNs produced
duplicate couples (time-id)
## Error:
printSummaryTable(pooling)## Error: object 'pooling' not found
summary(pooling)## Error: object 'pooling' not found
fixed <- plm(log(Consumption) ~ log(GDPconstantUSD) + log(Price) + log(Consum_t_1),
data=pchasamil, model= "within")## Warning: NaNs produced
duplicate couples (time-id)
## Error:
printSummaryTable(fixed)## Error: object 'fixed' not found
random <- plm(log(Consumption) ~ log(GDPconstantUSD) + log(Price) + log(Consum_t_1),
data=pchasamil, model= "random")## Warning: NaNs produced
duplicate couples (time-id)
## Error:
printSummaryTable(random)## Error: object 'random' not found
The estimation of the random effect model tells me that there is an Error in swar(object, data, effect) : "the estimated variance of the individual effect is negative".