In johngil/mscidata: MSCI USA and Min Vol Data Analysis

Data Analysis

Sector Weights

Sector weights were calculated over time for both EUSA and USMV. Plots were made and displayed to compare the relative sector weights of EUSA and USMV. This was done to get a sense of what industries may be inherently more "low risk," as well as making sure the data is resilient, and each sector weighting is within 5% as specified by the Barra Optimizer.

Sector Weight Summary Statistics:

data(usa_percent)
data(minvol_percent)

## Summary statistics of EUSA sector weights
head(usa_percent)
tail(usa_percent)
summary(usa_percent)

## Summary statistics of USMV sector weights
head(minvol_percent)
tail(minvol_percent)
summary(minvol_percent)

Sector Weights for EUSA and USMV:

library(ggplot2)
library(mscidata)

## Energy
Eng1 <- usa_percent[which(usa_percent$sector_name=="Energy"), ]
Eng2 <- minvol_percent[which(minvol_percent$sector_name=="Energy"), ]
ggplot(Eng1, aes(date, percent, colour = "USA")) + geom_line() +  
ggtitle("EUSA and USMV Energy Sector Weights") + xlab("Time") + ylab("Sector Weight") +
geom_line(data = Eng2, aes(x=date, y=percent, colour="Min Vol"),show.legend = TRUE)

## Finacials
Fin1 <- usa_percent[which(usa_percent$sector_name=="Financials"), ]
Fin2 <- minvol_percent[which(minvol_percent$sector_name=="Financials"), ]
ggplot(Fin1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Financial Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + 
geom_line(data = Fin2, aes(x=date, y=percent, colour="Min Vol"),show.legend = TRUE)

## Consumer Staples
ConStap1 <- usa_percent[which(usa_percent$sector_name=="Consumer Staples"), ]
ConStap2 <- minvol_percent[which(minvol_percent$sector_name=="Consumer Staples"), ]
ggplot(ConStap1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Consumer Staples Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + geom_line(data = ConStap2, aes(x=date, y=percent, 
colour="Min Vol"),show.legend = TRUE)

## Consumer Discretionary
ConDis1 <- usa_percent[which(usa_percent$sector_name=="Consumer Discretionary"), ]
ConDis2 <- minvol_percent[which(minvol_percent$sector_name=="Consumer Discretionary"), ]
ggplot(ConDis1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Consumer Discretionary Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + geom_line(data = ConDis2, aes(x=date, y=percent, 
colour="Min Vol"),show.legend = TRUE)

## Health Care
Health1 <- usa_percent[which(usa_percent$sector_name=="Health Care"), ]
Health2 <- minvol_percent[which(minvol_percent$sector_name=="Health Care"), ]
ggplot(Health1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Health Care Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + geom_line(data = Health2, aes(x=date, y=percent, 
colour="Min Vol"),show.legend = TRUE)

## Industrials
Ind1 <- usa_percent[which(usa_percent$sector_name=="Industrials"), ]
Ind2 <- minvol_percent[which(minvol_percent$sector_name=="Industrials"), ]
ggplot(Ind1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Industrials Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + geom_line(data = Ind2, aes(x=date, y=percent, 
colour="Min Vol"),show.legend = TRUE)

## Information Technology
IT1 <- usa_percent[which(usa_percent$sector_name=="Information Technology"), ]
IT2 <- minvol_percent[which(minvol_percent$sector_name=="Information Technology"), ]
ggplot(IT1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Information Technology Sector Weights") + 
xlab("Time") + ylab("Sector Weight") + geom_line(data = IT2, aes(x=date, y=percent,
colour="Min Vol"),show.legend = TRUE)

## Materials
Mat1 <- usa_percent[which(usa_percent$sector_name=="Materials"), ]
Mat2 <- minvol_percent[which(minvol_percent$sector_name=="Materials"), ]
ggplot(Mat1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Materials Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + geom_line(data = Mat2, aes(x=date, y=percent, 
colour="Min Vol"),show.legend = TRUE)

## Utilites
Util1 <- usa_percent[which(usa_percent$sector_name=="Utilities"), ]
Util2 <- minvol_percent[which(minvol_percent$sector_name=="Utilities"), ]
ggplot(Util1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Utilities Sector Weights") + xlab("Time") + 
ylab("Sector Weight") + geom_line(data = Util2, aes(x=date, y=percent, 
colour="Min Vol"),show.legend = TRUE)

## Telecommunication Services
Telecom1 <- usa_percent[which(usa_percent$sector_name=="Telecommunications"), ]
Telecom2 <- minvol_percent[which(minvol_percent$sector_name=="Telecommunications"), ]
ggplot(Telecom1, aes(date, percent, colour = "USA")) + geom_line() + 
ggtitle("EUSA and USMV Telecommunications Sector Weights") + 
xlab("Time") + ylab("Sector Weight") + geom_line(data = Telecom2, aes(x=date, 
y=percent, colour="Min Vol"),show.legend = TRUE)

EUSA Constituent Trailing Volatilities

Data was collected from the past 10 years of the EUSA index. The data was collected from 12/31/2006 to 12/30/2016, from Wharton Research Data Services (WRDS) for the 908 historical constituents of the USA Equal Weight (EUSA) index, of which USMV is derived. Each tickers' 252-day (annual) trailing volatility was calculated and a month end spagetti plot was produced.

library(ggplot2)

head(monthly_rolling_vol)
tail(monthly_rolling_vol)
summary(monthly_rolling_vol)

# Spaggeti plot for monthly trailing vol
ggplot(monthly_rolling_vol, aes(Date, Volatility, group = Ticker)) + geom_line()

EUSA Constituent Trailing Betas

Data was collected from the past 10 years of the EUSA index. The data was collected from 12/31/2006 to 12/30/2016, was collected from WRDS for the 908 historical constituents of the USA Equal Weight (EUSA) index, of which USMV is derived. Each tickers' 252-day (annual) trailing beta was calculated and a month end spagetti plot was produced.

library(ggplot2)

head(monthly_beta_values)
tail(monthly_beta_values)
summary(monthly_beta_values)

# Spaggeti plot for monthly Beta
ggplot(monthly_beta_values, aes(Date, Beta, group = Ticker)) + geom_line()

EUSA Constituent Price to Book Ratios

Data was collected from the past 10 years of the EUSA index. The data was collected from 12/31/2006 to 12/30/2016, from WRDS for the 908 historical constituents of the USA Equal Weight (EUSA) index, of which USMV is derived. Each tickers' Price to Book ratio was calculated in two ways, to ensure accuracy.

library(ggplot2)

head(book_value_data)
tail(book_value_data)
summary(book_value_data)

Data Distribution

The distribution of the predictor variable, if the stock is in the index now, with respect to each response variable is shown below.

Index now vs. Trailing Volatility

ggplot(monthly_final, aes(x=volatility, y=index_now)) + geom_point()

Index now vs. Trailing Beta

ggplot(monthly_final, aes(x=beta, y=index_now)) + geom_point() 

Index now vs. Price to Book Ratio

ggplot(monthly_final, aes(x=price_to_book, y=index_now)) + geom_point() 

Index now vs. Index 6 months ago

ggplot(monthly_final, aes(x=index_before, y=index_now)) + geom_point()

johngil/mscidata documentation built on May 19, 2019, 5:14 p.m.