A dynamic approach to analysing the effect of the global crisis on nonperforming loans: evidence from the Turkish banking sector

ABSTRACT

This article analyses the effect of the global crisis on the determinants of nonperforming loans (NPLs) in the Turkish banking sector by using dynamic panel estimation techniques. Empirical findings suggest that NPLs present persistence, which is more evident after the crisis, while other regressors have also persistent effects in the post-crisis period. Moreover, NPLs are mostly shaped by bank-specific variables before the crisis, whereas, after the crisis, NPLs are also driven by macroeconomic and policy-related variables. In particular, the post-crisis significance of GDP, policy rate and sovereign debt shows that robust economic activity, tight monetary policy and strong fiscal balances restrict NPLs, thereby enhancing financial stability. The significance of inflation in both sub-periods shows that commitment to price stability objective is indispensable for limiting NPLs and promoting financial stability. In the period ahead, the speed and the direction of normalization in global monetary policies may determine the course of financial conditions, which, therefore, have implications regarding NPL dynamics and financial stability.

KEYWORDS:

• Global crisis
• nonperforming loans (NPLs)
• Turkish banking sector
• dynamic panel estimation
• persistence
• financial stability
• price stability
• normalization

JEL CLASSIFICATION:

• C23
• E44
• E52
• G10
• G21

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ CBRT (2010), Selçuk (2010) and Afşar (2011) show that NPLs increased sharply, while Ganioğlu and Us (2014) and Us (2015a) report that the structure of the Turkish banking sector has changed after the crisis.

² Greenidge and Grosvenor (2010), Salas and Saurina (2002), Jiménez and Saurina (2006), Ranjan and Dhal (2003), Louzis, Vouildis, and Metaxas (2012), Rinaldi and Sanchis-Arellano (2006), Berge and Boye (2007), Nkusu (2011), Berger and DeYoung (1997), Podpiera and Weill (2008), Sinkey and Greenwalt (1991), Kwan and Eisenbeis (1995), Hassan (1993), Brewer, Jackson, and Mondschean (1996), Gallo, Apilado, and Kolari (1996), Angbazo (1997), Hassan, Karels, and Peterson (1994), Corsetti, Pesenti, and Roubini (1999) and Breuer (2006) are important works along this line.

³ Beck et al. (2013), Chaibi and Ftiti (2015), Klein (2013), Louzis, Vouildis, and Metaxas (2012), Abid, Ouertani, and Zouari-Ghorbel (2014), Salas and Saurina (2002), Jiménez and Saurina (2006) and Rinaldi and Sanchis-Arellano (2006) are previous studies that assume persistence in NPLs. This is because NPLs are not immediately written off from banks’ balance sheets, which implies that lagged NPL terms may have significant impact on current NPLs. This requires utilizing dynamic panel data estimation, which includes lagged dependent variable as regressor.

⁴ Us (2015b) provides an overview of the CBRT’s tight monetary policy conduct after the crisis, which also observes financial stability.

⁵ FX lending to resident individuals was prohibited in 2009, which caused FX-loans to decrease substantially.