Long Memory in Turkish Unemployment Rates

ABSTRACT

In this article we have examined the unemployment rate series in Turkey by using long memory models and in particular employing fractionally integrated techniques. Our results suggest that unemployment in Turkey is highly persistent, with orders of integration equal to or higher than 1 in the majority of the cases. This implies lack of mean reversion and persistence of the shocks. We found evidence in favor of mean reversion in the case of female unemployment and this happens for all the groups of non-agricultural, rural, urban, and youth unemployment series. The possibility of nonlinearities are observed only in the case of female unemployment and the degree of persistence is higher in the cases of female and youth unemployment series. Important policy implications emerge from our empirical results. Thus, for example, positive shocks reducing unemployment will have permanent effects being good for the economy, but negative shocks increasing unemployment will also have permanent effects and strong measures should then be adopted to reduce it. Labor and macroeconomic policies will most likely have long-lasting effects on the unemployment rates.

KEY WORDS:

• fractional integration
• hysteresis
• NAIRU
• Turkey
• unemployment

JEL CLASSIFICATION:

• C22
• E24

Acknowledgments

Comments from the Editor and two anonymous reviewers are gratefully acknowledged.

Notes

1. Note that an I(d) process may also contain weak autocorrelation of the ARMA form (ARFIMA processes). Then, if the d-differenced process is, for example, AR(1) and the AR coefficient is close to 1, the process is highly persistent.

2. Note that fractional integration and structural breaks are issues which are very much related. See, for instance, Sibbertsen (2004), Gil-Alana (2008), Baillie and Morana (2009).

3. Some authors argue that the concept of mean reversion in nonstationary series is a misnomer (Phillips and Xiao 1999). However, the idea behind this is that under mean reversion the series should converge to its original long term projection.

4. There are several studies that deal with other aspects of the unemployment rates in Turkey. Two recent examples include Berument, Dogan, and Tansel (2009) and Sengul (2014). Berument, Dogan, and Tansel (2009) investigate how various macroeconomic policy shocks in Turkey affect unemployment in several sectors of economic activity. Sengul (2014) analyzes the rates of inflow to and outflow from unemployment in Turkey with the main finding that volatility of the inflow rates to unemployment is the main driving force of the change in the unemployment rate in Turkey.

5. The I(d) processes belong to a wider class of process long memory, named so because of the strong degree of association between observation which are far distant apart in time.

6. For more details, see Gil-Alana and Robinson (1997).

Additional information

Funding

Prof. Luis A. Gil-Alana gratefully acknowledges financial support from the Ministerio de Economia y Competitividad (ECO2014-55236).