CODEWITHSUNDEEP
r
Alejandro Jiménez Rico
Alejandro Jiménez Rico

Reputation: 457

R: Arguments imply differing number of rows using financial data

I am intending to build a code to check the cointegration of two time series of financial price data in order to make better forecasting of one of them.

For this purpose, I chose this two time series: BBVA historical price and IBEX35 and build this little code:

ibex <- new.env()
bbva <- new.env()

library(quantmod)

getSymbols("^IBEX", env = ibex, src = "yahoo", from = as.Date("2010-01-04"), to = as.Date("2015-11-30"))
getSymbols("bbva", env = bbva, src = "yahoo", from = as.Date("2010-01-04"), to = as.Date("2015-11-30"))
ibex <- ibex$IBEX
ibex <- ibex$IBEX.Adjusted
bbva <- bbva$BBVA
bbva <- bbva$BBVA.Adjusted

ldbbva <- diff(log(bbva))
ldibex <- diff(log(ibex))

mean <- mean(ldbbva, na.rm = TRUE) 
ldbbva[is.na(ldbbva)] <- mean

mean <- mean(ldibex, na.rm = TRUE) 
ldbbva[is.na(ldibex)] <- mean

library(urca)

jotest=ca.jo(data.frame(ldbbva,ldibex), type="trace", K=2, ecdet="none", spec="longrun")

At that point when I try to make a data frame from my time series, I face this error arguments imply differing number of rows: 1488, 1514.

What can I do?

Upvotes: 0

Views: 670

Answers (1)

Roland
Roland

Reputation: 132706

You need to join the time series to have them properly aligned.

dat <- merge(ibex, bbva)

dat <- diff(log(dat))

#mean imputation
dat <- na.aggregate(dat)

library(urca)
jotest=ca.jo(dat, type="trace", K=2, ecdet="none", spec="longrun")
###################################################### 
## Johansen-Procedure Unit Root / Cointegration Test # 
###################################################### 
#
#The value of the test statistic is: 619.1603 1473.644

Upvotes: 3

Related Questions