Creating high‐frequency national accounts with state‐space modelling: a Monte Carlo experiment

Abstract

This paper assesses a new technique for producing high-frequency data from lower frequency measurements subject to the full set of identities within the data all holding. The technique is assessed through a set of Monte Carlo experiments. The example used here is gross domestic product (GDP) which is observed at quarterly intervals in the United States and it is a flow economic variable rather than a stock. The problem of constructing an unobserved monthly GDP variable can be handled using state space modelling. The solution of the problem lies in finding a suitable state space representation. A Monte Carlo experiment is conducted to illustrate this concept and to identify which variant of the model gives the best monthly estimates. The results demonstrate that the more simple models do almost as well as more complex ones and hence there may be little gain in return for the extra work of using a complex model.

Should quarterly government finance statistics be used for fiscal surveillance in Europe?

“…In particular, we can summarize the main points highlighted in the paper as follows: (i) 4 Other approaches for modeling data at different sampling intervals are the methods based on regression techniques (Chow andLin, 1971, Guerrero, 2003), the MIDAS (MIxed DAta Sampling) approach (see Ghysels, Santa-Clara &Valkanov, 2004, Clements andGalvão, 2007), the state space approaches of Liu and Hall (2001) and Mariano and Murusawa (2003), or the ARMA model model with missing observations of Hyung and Granger (2008).…”

The University of Pennsylvania Models for High-Frequency Macroeconomic Modeling

“…Liu & Hall (2001) estimate monthly GDP for the US using Kalman filter methodology. See, Hamilton (1994aHamilton ( , 1994b, Harvey (1987Harvey ( , 1989, Kalman (1960Kalman ( , 1961, Kalman & Bucy (1961), Kim &Nelson (1999), Stock &Watson (1991 for the application of Kalman filters.…”

The Chow-Lin method extended to dynamic models with autocorrelated residuals

“…Di Fonzo (2003) notes for the earliest solutions that the algorithms needed to calculate the estimates and their standard errors seem rather complicated and not straightforward to be implemented in a computer program. And further demotivating the potential user, Liu and Hall (2001) show on US data that the gains of complexity may in fact be limited. Santos Silva and Cardoso (2001) overcome this difficulty and provide a straightforward solution to dynamic models under the assumption of white noise residuals.…”