The RATS Software Forum

Hi,

I would like to estimate GDP Plus in Rats. See Aruoba, Diebold, Nalewaik, Schorfheide and Song (ADNSS, 2013), "Improving GDP Measurement: A Measurement-Error Perspective", Philly Fed working paper 13-16 (http://www.phil.frb.org/research-and-da ... p13-16.pdf). In the ADNSS model, true GDP is not observed and follows an AR(1) process:

GDP* = mu(1 - rho) + rho GDP*{1} + e

The observed expenditure (GDP) and income (GDI) measures are noisy measures of GDP*:

GDP = GDP* + u
GDI = GDP* + v

where corr(e,u) and corr(e,v) are negative etc. Thus, the random errors in the state and measurement equations are correlated.

Thanks,

Anthony

Just move e into the state vector, which will add a zero row and column to A and an extra weight of 1 to C. The SV will then be empty. Read the discussion starting after equation 2.2 in the attached (excerpted from the State Space e-course).

Thanks.

Anthony