VAR with 48 observations!!?? :(
Posted: Fri Nov 25, 2011 9:57 am
Hello there
I am sort of depressed... well currently i am trying to do an empirical study on the relation between fertility and economic growth for a developing country. Using a VAR 4x1 with proxies for fertility (births x month), money holdings (M2), unemployment and growth has provided interesting results and neat IR functions for some countries. I struggled to get at least 120 monthly observations for each of the 4 proxies... but births x month was so so so so troublesome to get... after asking a lot i could get 48 observations (2005-2008).... its awful, i mean, apparently the only way of getting more observations is by checking municipality records manually (around 400 municipalities and absolutely no resources for this
I was just wondering, this period is SO short (after all i would use a lag of 12 months, def something bigger than 9 months anyways, for obvious reasons) that
i) should i give up this project?
ii) continue with this thing and see what results i can get with my poor data
ideas?
Thank you in advance
I am sort of depressed... well currently i am trying to do an empirical study on the relation between fertility and economic growth for a developing country. Using a VAR 4x1 with proxies for fertility (births x month), money holdings (M2), unemployment and growth has provided interesting results and neat IR functions for some countries. I struggled to get at least 120 monthly observations for each of the 4 proxies... but births x month was so so so so troublesome to get... after asking a lot i could get 48 observations (2005-2008).... its awful, i mean, apparently the only way of getting more observations is by checking municipality records manually (around 400 municipalities and absolutely no resources for this
I was just wondering, this period is SO short (after all i would use a lag of 12 months, def something bigger than 9 months anyways, for obvious reasons) that
i) should i give up this project?
ii) continue with this thing and see what results i can get with my poor data
ideas?
Thank you in advance