Page 1 of 1

generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 1:52 am
by honeygupta
hello

i am working over expenditure of countries . i am facing problem with data. under various variables values are not given in starting years and end years. how can i generate this please tell

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 8:12 am
by EViews Gareth
Your question isn't clear.

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 11:20 am
by honeygupta
Sir ,

I have different variables like health expenditure and education of brics countries and data series is from 1990 to 2014. But values are missing from 1990 to 93 and from 2013-14.
I wanted to know how can I generate these missing value so that I can apply my tools.

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 11:22 am
by EViews Gareth
There's two parts to that question:
  • What econometric/statistical method should I use to make up values for my missing data.
  • How do I apply that method in EViews
Once you've figured out the first part, we can help with the second.

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 11:31 am
by honeygupta
Sir,
I really need answer to first question if you can help. I will be very grateful to you.

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 11:39 am
by startz
What tools can't you apply if some of the data is missing?

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 11:47 am
by honeygupta
I am going to see the impact of expenditures on economic growth so I want to apply var and ols for which we need minimum 20 to 25 years data. So, if through any methodology like back casting or something to make up data.

Or is there any other tool available for 15 year data period. So data make up problem can be solved. Please suggest. :cry: :roll:

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 12:16 pm
by startz
I am going to see the impact of expenditures on economic growth so I want to apply var and ols for which we need minimum 20 to 25 years data. So, if through any methodology like back casting or something to make up data.

Or is there any other tool available for 15 year data period. So data make up problem can be solved. Please suggest. :cry: :roll:
There's no rule that you need 20 to 25 years data for a var or ols, although more data is always better.

If you can't get results without making up more data, then won't your results depend on the made up data? This doesn't seem like a very good idea.

Re: generating missing values in starting and end of series

Posted: Wed Aug 05, 2015 10:49 pm
by honeygupta
sir ,
actually data of mostly variables are given after 1995 only annually and for some variables data is from 1988. so i decided to take data from 1990 to 2014 annualy to find impact of expenditure on gdp , national income and human development index(index was started from 1990 ,another reason to take data from 1990 ),

so please suggest how can i generate data for starting years like for 1990 to 1995.

suggestions are most welcome. :cry: :D :?: :?:
attaching one file

Re: generating missing values in starting and end of series

Posted: Thu Aug 06, 2015 10:26 pm
by honeygupta
please reply, if you have some suggestion. this is really important for me