# ------------------------------------------------------------------------------
#  LMF_nfa.R  
#  Grabs Lane and Milesi-Ferretti database on net foreign assets, described in 
#  "The External Wealth of Nations," JIE, 2007 
#  See paper at:  http://www.imf.org/external/pubs/cat/longres.aspx?sk=18942.0
#  Data is large flat file, countries stacked up, 1970-2007 or subset 
#  Written by Espen Henriksen, September 2010 
#  Adapted by Dave Backus 
# ------------------------------------------------------------------------------
# 0. Preliminaries 
# clear workspace ("remove" everything in list generated by ls command) 
rm(list=ls()) 
# set working directory for output 
# NB: check your own directory, and make sure to use / not \ 
setwd("c:/Documents and Settings/dbackus/My Documents/Userdata/Papers/BCH/data/LMF")

# load libraries of apps (must be installed -- and the second needs perl, too)
library(gtools)
library(gdata)

# 1. Download data
# download data and save in "fname" (dated to avoid overwriting old data)
# url = internet source of data, destfile = filename for local version 
fname = paste("LMF99_",Sys.Date(),".zip",sep="")    # creates string with date
download.file(url="http://www.imf.org/external/pubs/ft/wp/2006/data/update/wp0669.zip",destfile=fname,method="auto")
# input data sheet (first one reads data, second one reads headers for variable names) 
lmf.data     <- read.xls(unzip(zipfile=fname,files="EWN II update for web (August 2009).xls",list=FALSE),header=TRUE,skip=1,sheet=4)
series.names <- read.xls(unzip(zipfile=fname,files="EWN II update for web (August 2009).xls",list=FALSE),header=FALSE,skip=0,sheet=4,nrows=1)

series.names.old <- series.names
print(t(series.names.old))

# why?  seems redundant 
series.names[1] <- "Country"
series.names[2] <- "IFS.code"
series.names[3] <- "Year"
series.names[4] <- "Portfolio equity assets"
series.names[5] <- "Portfolio equity liabilities"
series.names[6] <- "FDI assets"
series.names[7] <- "FDI liabilities"
series.names[8] <- "Debt assets (portfolio debt plus other investment)"
series.names[9] <- "Debt liabilities (portfolio debt plus other investment)"
series.names[10] <- "Financial derivatives assets"
series.names[11] <- "Financial derivatives liabilities"
series.names[12] <- "FX Reserves minus gold"
series.names[13] <- "Total assets"
series.names[14] <- "Total liabilities"
series.names[15] <- "NFA"
series.names[16] <- "Net IIP as officially reported"
series.names[17] <- "NFA (alternative FDI valuation)"
series.names[18] <- "GDP"
series.names[19] <- "Portfolio debt assets"
series.names[20] <- "Portfolio debt liabilities"
series.names[21] <- "Other investment assets"
series.names[22] <- "Other investment liabilities"
series.names[23] <- "FDI assets (other valuation)"
series.names[24] <- "FDI liabilities (other valuation)"
series.names[25] <- "Exchange rate vis-a-vis US dollar (period avg)"
series.names[26] <- "Exchange rate vis-a-vis US dollar (end of period)"

# assign var names to data, [[2]] means names apply to columns (see also colnames)
dimnames(lmf.data)[[2]] <- series.names  

# 2. Extract subset 
# extract a subset of columns (variables) 
# note use of curly brackets {} to run across lines 
lmf.subset <- {data.frame(Country = lmf.data$Country,
                          IFS     = lmf.data$IFS.code,
                          Year    = lmf.data$Year,
                          NFA     = as.numeric(gsub(",","",lmf.data$NFA)),
                          GDP     = as.numeric(gsub(",","",lmf.data$GDP)),
                          NFAGDP  = as.numeric(gsub(",","",lmf.data$NFA)))}
lmf.subset$NFAGDP <- lmf.subset$NFA/lmf.subset$GDP

# extract a subsect of countries (rows) 
# why not use IFS code??
countries <- c("Australia","Canada","China,P.R.: Mainland","Germany","France","Italy","Japan","Norway","United Kingdom","United States")

lmf.oecd<- NULL
for(i in 1:length(countries)){
  lmf.oecd <- cbind(lmf.oecd,lmf.subset[lmf.subset$Country == countries[i],]$NFAGDP)
}
dimnames(lmf.oecd)[[2]] <- countries
lmf.oecd <- ts(lmf.oecd,start=1970,frequency=1)

# Line plot of NFA/GDP for subset of countries
# color sequence 
{colseq <- c("black", "green", "blue", "orange", "yellow", "red", "aquamarine",
  "azure","brown", "burlywood")} 
{ts.plot(lmf.oecd,
  gpars=list(main=NULL,xlab=NULL,ylab="NFA/GDP", 
  mgp=c(2.5,1,0), lwd=c(1,2,3),
  lty=1, col=colseq[1:dim(lmf.oecd)[2]],
  axes=TRUE)
  )}
abline(h=0,lty=1,col="black")
dev.print(device=postscript, "linesnfagdp.eps") #, onefile=FALSE, horizontal=FALSE)
#dev.print(device=pdf, "linesnfagdp.pdf")

# Bar plot NFA/GDP for that subset of countries
tab <- rbind(lmf.oecd[11,],lmf.oecd[16,],lmf.oecd[21,],lmf.oecd[26,],lmf.oecd[31,],lmf.oecd[36,])
dimnames(tab)[[2]] <- c("AU","CA","CH","DE","FR","IT","JP","NO","UK","US")
barplot(tab, beside = TRUE,col=c("grey","blue","orange","darkblue","red","darkgreen"), ylab="NFA/GDP",ylim=c(-.6,.6))
legend("topleft", legend = c("1980","1985","1990","1995","2000","2005"), fill=c("grey","blue","orange","darkblue","red","darkgreen"),bty="n")
dev.print(device=postscript, "barsnfagdp.eps", onefile=FALSE, horizontal=FALSE)
dev.print(device=pdf, "barsnfagdp.pdf")