Application of three non-linear econometric approaches to identify business cycles in Peru

I use three non-linear econometric models to identify and analyze business cycles in the Peruvian economy for the period 1980:1-2008:4. The models are the Smooth Transition Autoregressive (STAR) model suggested by Teräsvirta (1994), the extended version of the Markov-Switching model proposed by Hamilton (1989), and the plucking model of Friedman (1964, 1993). The results indicate strong rejection of the null hypothesis of linearity. The majority of models identify quarters concentrated around 1988-1989 and 1990-1991 as recession times. Other important events which happened in the Peruvian economy (natural disaster in 1983, e¤ects of the Asian and Russian crises in 1990s, terrorist activities in 1980s) are not selected except as atypical observations. Most of models also identify the period 1995:1-2008:4 as a very long and stable period of moderate-high growth rates. From the perspective of the Peruvian economic history and from a statistical point of view, the MSIAH(3) model is the preferred model.