Time Series Forecast of Annual Financial Statement 2018-19

If you like this, Share!

The finale for the buzzing-B word which was hitting the headlines for the past fortnight has come to an end. The nail biting event anguished which the entire nation cataloging from the ubiquitous taxpayers to the investment giants.

Lets take a sneak peek into the quintessential strands in the ANNUAL FINANCIAL STATEMENT of the UNION BUDGET. The highlighted branches under the canopy of annual financial statement are

  • Capital Account Receipts
  • Capital Account Disbursement
  • Revenue Account Receipts
  • Revenue Account Disbursement
  • Public Account Receipts
  • Public Account Disbursement

Revenue Receipts covers the tax revenues (like income tax, excise duty) and non-tax revenue (interest receipts, profits)

Capital Receipts holds money from market borrowings, recovery of loans etc.

Public Receipts act as medium for the flow of transaction, where government is merely acting as a banker (for instance: provident fund and small savings)

For a quick analysis of the aforementioned statements, ARIMA modelling is used to forecast for the upcoming two years (ie; 2018 and 2019).

ARIMA MODEL TO FORECAST THE ANNUAL FINANCIAL STATEMENT

 Here is the synopsis of the CAPITAL ACCOUNT DISBURSEMENT using ARIMA:

The Capital Account  Revenue statement is taken as the analysis sample:

The time series data is subjected to Augmented Dickey Fuller test for check the stationarity of the data

>x1<-log(CAR$AMOUNT)

> adf.test(x1)

Augmented Dickey-Fuller Test

data:  x1

Dickey-Fuller = -1.6901, Lag order = 2, p-value = 0.6904

alternative hypothesis: stationary

Since the data is not stationary, multiple differentiation is applied

>x2<-diff(x1)
> adf.test(x2)        
Augmented Dickey-Fuller Test data: 
 x2Dickey-Fuller = -1.9368, 
Lag order = 2, p-value = 0.5965
alternative hypothesis: stationary 
>x3<-diff(x2)> adf.test(x3)         
Augmented Dickey-Fuller Test data:  
x3Dickey-Fuller = -3.3127, Lag order = 2, p-value = 0.0899
alternative hypothesis: stationary  
> x4<-diff(x3)> adf.test(x4)         
Augmented Dickey-Fuller Test data:  
x4Dickey-Fuller = -5.214, Lag order = 2, p-value = 0.01
alternative hypothesis: stationary

The third difference gave a stationary data series

Auto Correlation Function and Partial Auto Correlation Function is employed to find out the order of the data series.

>acf(x1)

> pacf(x1)

And the order of the data series is obtained as:
AR=1,
I=3,
MA=0

> model1
<-arima(x1,order=c(1,3,0))
> summary(model1)
Length Class  Mode     
coef       1     -none- numeric  
sigma2     1     -none- numeric  
var.coef   1     -none- numeric  
mask       1     -none- logical  
loglik     1     -none- numeric  
aic        1     -none- numeric  
arma       7     -none- numeric  
residuals 23     ts     numeric  
call       3     -none- call     
series     1     -none- character
code       1     -none- numeric  
n.cond     1     -none- numeric  
nobs       1     -none- numeric  
model     10     -none- list     
> model1$coef      
ar1 -0.514188

And ARIMA model is employed to data and prediction function is used for forecasting for the years 2018 and 2019.

>pred1
<-predict(model1,2)
> pred1
$pred
Time Series:
Start = 24 
End = 25 
Frequency = 1 
[1] 16.03107 16.27264
$se
Time Series:
Start = 24 
End = 25 
Frequency = 1 
[1] 0.3043931 0.8155953

And the forecasted values are plotted.

>x2<-c(x1,16.03107,16.27264)
> x3<-cbind(x1,x2)
>View(x3)
> x3[24:25]<-NA
>x4<-exp(x3)
> plot(x4[,1],lwd=10,type="l",xlab="YEAR",ylab="AMOUNT")
> lines(x4[,2],col="pink",lwd=3)

FORECASTED ANNUAL FINANCIAL STATEMENTS



$pred
Time Series:
Start = 24 
End = 25 
Frequency = 1 
[1] 16.03107 16.27264 
$se
Time Series:
Start = 24 
End = 25 
Frequency = 1 
[1] 0.3043931 0.8155953

PUBLIC RECEIPTS
 
 
REVENUE RECEIPTS
 
 

 

Particular FORECASTED AMOUNT(IN CR)
  2018 2019
Capital Account Receipts 9166536 11671260
Capital Account Disbursement 9413355 12218601
Revenue Account Receipts 10374796 12995712
Revenue Account Disbursement 11475796 13996711
Public Account Receipts 2419569 2708620
Public Account Disbursement 2437077 2734721

 

KEY INSIGHTS FROM THE FORECAST

 

GST surveillance will incorporate more taxpayers under the same canopy which is significantly reflected in the uptick of forecasted Revenue Account Receipts for the years 2018 and 2019.

There is a chance of more disinvestment of public sector units, trailed by hike in borrowings and recoveries of NPA which in turn outlines in the Capital Account Receipts.

Proclivity to improve interest rates for provident fund, to lure more investment from the public, there is a increment in the Public Account Receipts which justifies the forecast.

 

This article is written by Anjali UJ and Shabin Nahab.

Leave a Reply

1 Comment on "Time Series Forecast of Annual Financial Statement 2018-19"

  Subscribe  
newest oldest most voted
Notify of

Good job.