Estimating the output gap for Turkey: an unobserved components approach

Abstract

This study specifies a basic univariate and a bivariate unobserved components model to estimate potential output using information from observable aggregates and presents results for the Turkish economy. The first specification used in the study, i.e. univariate approach, decomposes actual output into potential output that follows a random walk with a time-varying potential growth rate and a stationary output gap. The univariate specifications commonly ignore some economic content, which might be relevant for the measurement of the output gap. In this respect, the univariate model is extended by utilizing the relationship between inflation and the output gap, namely the Phillips curve. Whereas both models give similar output gap estimates, signal extraction statistics suggest that incorporating the supply side to the system reduces the parameter uncertainty and the total standard error and improves the gap estimate.

Acknowledgements

The views expressed in this paper are those of the authors and do not necessarily correspond to the views of the Central Bank of the Republic of Turkey. We are particularly indebted to Kenneth N. Kuttner for providing us the GAUSS code and his excellent explanations and comments. Also we would like to thank Cevriye Aysoy, Zerrin Gürgenci, Ali Hakan Kara, Zelal Kotan and M. Gülenay Ongan for their helpful comments and suggestions.

Notes

¹ For example, as a special case, famous Hodrick-Prescott filter can be achieved by imposing the restrictions of no level shock (σ_η = 0), random walk trend growth rate with nonzero σ_ϵ (ρ = 1) and serially uncorrelated output gap (). Smoothing parameter of 1600 for quarterly data is equivalent to σ_ϵ = 0.025 σ_ξ.

² It is adjusted for the changes in the exchange rate.

³ Note that one of the important shortcomings of the filters is that the end-point estimates become imprecise due to the unknown direction of the future values, hence forecasted values are added for the last three quarters.

⁴ It means that, in the absence of shocks, output growth would converge to within 1% of the steady-state rate in just about 10 years.

⁵ Initially, following Watson (1986), an AR(2) process for the output gap is tested. However, estimation results indicate that the second term is insignificant.

⁶ The significance of the gap estimates truly depends on the statistical criteria used; explicitly other assumptions may designate different dates.

⁷ See Kuttner (1994) for further information.