EViews.com

Thanks, the above assisted but now I get another error if I just take data (yields) for one day....The error says "Insufficient Number of Observations" that this have to do with the data itself? Is it a requirement to have more data and if thats the case how will one get a daily yield curve using nelson siegel in a state space model? please help

Thank you startz, that did give me their final values, but it also gave me large standard errors. Do you have any advice?
rodneyleshaba, try to estimate the model without c(1) next to sv1 in your signal equation and then make state series.
To better understand the output I suggest you to read original Diebold, Rudebusch and Arouba (2006) paper and Eviews manual.

Do u mean the make state series output then gives me the zero coupon yields, meaning the data gives me the yield curve? Should I only move to trusting the output if the SV1, SV2 and SV3 are significant?

Hi everyone,

My current project is to estimate a Dynamic Factor Model (DFM) in order to nowcast quarterly GDP growth. So far I've gathered 16 indicators (PMI manufacturing, services, IP etc.) that are available on a monthly basis and that are correlated to economic activity. Having these indicators how would you guys implement a DFM in EViews respectively how to extract the common unknown factor(s)? How to determine the optimal number of common factors in a DFM? How to determine the lag structure of the unobserved factors and/or of the idiosyncratic error?

Thanks for your help in advance :D

hi,

I am trying to develop and state space model based on diebold li VAR model. Eviews is delivering the message below:

Unmatched Parenthesis.

Below there is my specification. There something wrong?

@signal fed3m=sv1+((1-exp(-3*c(1)))/(3*c(1)))*sv2+(((1-exp(-3*c(1)))/(3*c(1)))-exp(-3*c(1)))*sv3+[var=exp(c(2))]
@signal fed6m=sv1+((1-exp(-6*c(1)))/(6*c(1)))*sv2+(((1-exp(-6*c(1)))/(6*c(1)))-exp(-6*c(1)))*sv3+[var=exp(c(3))]
@signal fed9m=sv1+((1-exp(-9*c(1)))/(9*c(1)))*sv2+(((1-exp(-9*c(1)))/(9*c(1)))-exp(-9*c(1)))*sv3+[var=exp(c(4))]
@signal fed1y=sv1+((1-exp(-12*c(1)))/(12*c(1)))*sv2+(((1-exp(-12*c(1)))/(12*c(1)))-exp(-12*c(1)))*sv3+[var=exp(c(5))]
@signal fed15m=sv1+((1-exp(-15*c(1)))/(15*c(1)))*sv2+(((1-exp(-15*c(1)))/(15*c(1)))-exp(-15*c(1)))*sv3+[var=exp(c(6))]
@signal fed18m=sv1+((1-exp(-18*c(1)))/(18*c(1)))*sv2+(((1-exp(-18*c(1)))/(18*c(1)))-exp(-18*c(1)))*sv3+[var=exp(c(7))]
@signal fed21m=sv1+((1-exp(-21*c(1)))/(21*c(1)))*sv2+(((1-exp(-21*c(1)))/(21*c(1)))-exp(-21*c(1)))*sv3+[var=exp(c(8))]
@signal fed2y=sv1+((1-exp(-24*c(1)))/(24*c(1)))*sv2+(((1-exp(-24*c(1)))/(24*c(1)))-exp(-24*c(1)))*sv3+[var=exp(c(9))]
@signal fed30m=sv1+((1-exp(-30*c(1)))/(30*c(1)))*sv2+(((1-exp(-30*c(1)))/(30*c(1)))-exp(-30*c(1)))*sv3+[var=exp(c(10))]
@signal fed3y=sv1+((1-exp(-36*c(1)))/(36*c(1)))*sv2+(((1-exp(-36*c(1)))/(36*c(1)))-exp(-36*c(1)))*sv3+[var=exp(c(11))]
@signal fed4y=sv1+((1-exp(-48*c(1)))/(48*c(1)))*sv2+(((1-exp(-48*c(1)))/(48*c(1)))-exp(-48*c(1)))*sv3+[var=exp(c(12))]
@signal fed5y=sv1+((1-exp(-60*c(1)))/(60*c(1)))*sv2+(((1-exp(-60*c(1)))/(60*c(1)))-exp(-60*c(1)))*sv3+[var=exp(c(13))]
@signal fed6y=sv1+((1-exp(-72*c(1)))/(72*c(1)))*sv2+(((1-exp(-72*c(1)))/(72*c(1)))-exp(-72*c(1)))*sv3+[var=exp(c(14))]
@signal fed7y=sv1+((1-exp(-84*c(1)))/(84*c(1)))*sv2+(((1-exp(-84*c(1)))/(84*c(1)))-exp(-84*c(1)))*sv3+[var=exp(c(15))]
@signal fed8y=sv1+((1-exp(-96*c(1)))/(96*c(1)))*sv2+(((1-exp(-96*c(1)))/(96*c(1)))-exp(-96*c(1)))*sv3+[var=exp(c(16))]
@signal fed9y=sv1+((1-exp(-108*c(1)))/(108*c(1)))*sv2+(((1-exp(-108*c(1)))/(108*c(1)))-exp(-108*c(1)))*sv3+[var=exp(c(17))]
@signal fed10y=sv1+((1-exp(-120*c(1)))/(120*c(1)))*sv2+(((1-exp(-120*c(1)))/(120*c(1)))-exp(-120*c(1)))*sv3+[var=exp(c(18))]

@state sv1=c(31)+c(19)*sv1(-1) + c(20)*sv2(-1) + c(21)*sv3(-1) + [ename = e1, var = exp(c(22)]
@state sv2=c(32)+c(23)*sv1(-1) + c(24)*sv2(-1) + c(25)*sv3(-1) + [ename = e2, var = exp(c(26)]
@state sv3=c(33)+c(27)*sv1(-1) + c(28)*sv2(-1) + c(29)*sv3(-1) + [ename = e3, var = exp(c(30)]
@evar cov(e1,e2)=c(34)
@evar cov(e1,e3)=c(35)
@evar cov(e2,e3)=c(36)

Your state equations have mismatched parens at the end.

I am trying to see work Phillip curve using state space modelling .. the model seems to work fine when i do not include c(2)* inf (-1)... but i want to include it .... i do not se any reason why it should be included...I have attached a wok sheet for your reference.. please help me in this regard..

@signal inf = c(1) + c(3)*gdp_gap + [var = exp(c(4))] + c(2)* inf (-1)
@signal gdp = gdp_gap + gdp_pot
@state gdp_pot = gdp_pot (-1) + [var = exp(c(5))]
@state gdp_gap = gdp_gap(-1) + gdp_gap1 (-1) + [var = exp(c(6))]
@state gdp_gap1 = gdp_gap(-1)

Thanks a lot
Sunny

There might be a bug.

Just change the name of your series and use the new one instead:

The error spec in brackets has to go at the end of the equation.

Hi Glenn,

There are also spaces between the characters in the last term of first equation, which I think are due to copy-paste procedure. The problem with this model appears to be more than the use of incorrect syntax. Please run the code I have posted and see if using the variable name "inf" is causing the trouble.

Thanks, I just ran a quick test but with your latter example, which really doesn't help, does it? :)

I'll look into this more carefully when I get into the office, both with regards to INF and with regards to the brackets placement. I have a suspicion that INF may be causing problems for our math expression parser which is getting confused between the series and a numeric INF, but I'll double-check and let you know.

As Trubador pointed out, the parser has an issue discriminating between the series name INF and the numeric INF. As the original poster noted, this is not an issue for INF as the dependent variable (which doesn't go through the analytic function engine), and only shows up when the lags of INF appear on the right-hand side.

We're looking into seeing whether we can get the parser to understand the difference between the two concepts. For now, Trubador's suggestion of using a different dependent variable name will work.

I will also note that I was correct that the error specification must be at the end of the signal or state equation.

I'm doing this model, the Nelson Siegel model with ten maturity, when I run the progran appears a message, "Near to singular matrix"

thanks for help

Dear ALL

I am doing DieBold and Li and putting my sspace modelling, but showing NEAR SINGULAR MATRIX error, could anyone suggest me anything? am I making any mistakes on my equations?

PLS PLS help. Many tnx

@signal _3_month =sv1+( ( 1-exp(-c(26)*3) ) / (c(26)*3) )*sv2+( ( ( 1-exp(-c(26)*3) ) / (c(26)*3) ) - ( exp(-c(26)*3) ) )*sv3+[ename=e_3_month]
@signal _2_year =sv1+( ( 1-exp(-c(26)*24) ) / (c(26)*24) )*sv2+( ( ( 1-exp(-c(26)*24) ) / (c(26)*24) ) - ( exp(-c(26)*24) ) )*sv3+[ename=e_2_year]
@signal _5_year =sv1+( ( 1-exp(-c(26)*60) ) / (c(26)*60) )*sv2+( ( ( 1-exp(-c(26)*60) ) / (c(26)*60) ) - ( exp(-c(26)*60) ) )*sv3+[ename=e_5_year]
@signal _10_year=sv1+( ( 1-exp(-c(26)*120) ) / (c(26)*120) )*sv2+( ( ( 1-exp(-c(26)*120) ) / (c(26)*120) ) - ( exp(-c(26)*120) ) )*sv3+[ename=e_10_year]
@signal _20_year=sv1+( ( 1-exp(-c(26)*240) ) / (c(26)*240) )*sv2+( ( ( 1-exp(-c(26)*240) ) / (c(26)*240) ) - ( exp(-c(26)*240) ) )*sv3+[ename=e_20_year]
@signal _30_year=sv1+( ( 1-exp(-c(26)*360) ) / (c(26)*360) )*sv2+( ( ( 1-exp(-c(26)*360) ) / (c(26)*360) ) - ( exp(-c(26)*360) ) )*sv3+[ename=e_30_year]
@signal _40_year=sv1+( ( 1-exp(-c(26)*480) ) / (c(26)*480) )*sv2+( ( ( 1-exp(-c(26)*480) ) / (c(26)*480) ) - ( exp(-c(26)*480) ) )*sv3+[ename=e_40_year]

@state sv1 = c(10) + c(1)*( sv1(-1) - c(10) ) + c(2)*( sv2(-1) - c(11) ) + c(3)*( sv3(-1) - c(12) ) + [ename=ni_l]
@state sv2 = c(11) + c(4)*( sv1(-1) - c(10) ) + c(5)*( sv2(-1) - c(11) ) + c(6)*( sv3(-1) - c(12) ) + [ename=ni_s]
@state sv3 = c(12) + c(7)*( sv1(-1) - c(10) ) + c(8)*( sv2(-1) - c(11) ) + c(9)*( sv3(-1) - c(12) ) + [ename=ni_c]

@evar var(ni_l)=exp(c(13))
@evar var(ni_s)=exp(c(14))
@evar var(ni_c)=exp(c(15))
@evar cov(ni_l,ni_c)=c(16)
@evar cov(ni_s,ni_c)=c(17)
@evar cov(ni_l,ni_s)=c(18)

@evar var(e_3_month)=exp(c(19))
@evar var(e_2_year)=exp(c(20))
@evar var(e_5_year)=exp(c(21))
@evar var(e_10_year)=exp(c(22))
@evar var(e_20_year)=exp(c(23))
@evar var(e_30_year)=exp(c(24))
@evar var(e_40_year)=exp(c(25))

@param c(26) 0.0609