The Impact of the Divergence in Banks’ Loan Portfolios and Credit Risk Models on Asset Quality

[EN] After 2018, the ratios for non-performing loans (NPL) and closely monitored (Stage 2) loans of public banks have diverged positively compared to other banks in the sector, raising important discussions regarding the reasons for the divergence. This research note analyzes the drivers of this divergence using firm- and bank-based microdata sets. In brief, the study shows that banks' differing loan portfolios impact the divergence in their NPL ratios, while the use of different models by banks within the scope of IFRS-9 standards causes their Stage 2 loans ratios to diverge. Firstly, the NPL ratio analysis of commercial loans reveals that the NPL performances of the (joint) firms that take out loans from both public and other banks are similar, and the divergence in the NPL ratios is due to firms that are not in the joint portfolio of these two bank groups. The higher average net profit margins and financing expense coverage ratios of firms to which only public banks extend loans compared to firms to which only other banks extend loans may be a factor leading to lower corporate NPL ratios of public banks. Secondly, the NPL ratio analysis of retail loans suggests that the divergence may be driven by the fact that secured housing loans with positive repayment performance compared to other loans account for a larger share of public banks’ portfolios whereas unsecured general-purpose loans and credit cards have a larger share in other banks’ portfolios. Finally, the analysis of Stage 2 loans reveals that banks’ Stage 2 loan ratios have diverged following the introduction of the IFRS-9 standards in 2018. While public banks and other banks maintain similar ratios of loans classified as Stage 2 due to overdue payment, other banks are more likely to classify loans without overdue payment as Stage 2 loans than public banks, indicating that the divergence in Stage 2 loan ratios is due to the different IFRS-9 models used by banks. [TR] 2018 yili sonrasinda kamu bankalarinin tahsili gecikmis alacak (TGA) ve yakin izlemedeki kredi oranlari sektordeki diger bankalara kiyasla olumlu ayrismaya baslamis ve bu ayrismanin nedenleri onemli tartisma konularindan biri haline gelmistir. Bu calismada, firma ve banka bazli mikro veri setleri kullanilarak, bahse konu ayrismanin nedenleri analiz edilmektedir. calismada ozetle, bankalarin TGA oranlarinin ayrismasinda bankalarin farkli nitelikte kredi portfoylerine sahip olmasinin, yakin izlemedeki kredi oranlarinin ayrismasinda ise bankalarin TFRS-9 standartlari kapsaminda kullandiklari model farkliliklarinin etkili oldugu gosterilmektedir. Ilk olarak, ticari kredilere yonelik TGA analizlerinde, kamu ve diger olmak uzere her iki banka grubunca da kredi kullandirilan (ortak) firmalarin TGA performanslarinin benzer oldugu, TGA oranlarindaki ayrismanin ise ortak portfoyde olmayan firmalardan kaynaklandigi bulgulanmaktadir. Sadece kamu bankalarinin kredi kullandirdiklari firmalarin ortalamada net kâr marji ve finansman gideri karsilama oraninin yalnizca diger bankalarin kredi kullandirdiklari firmalara kiyasla daha yuksek olmasinin kamu bankalarinin firma TGA oranlarinin daha dusuk olmasinda etkili olabilecegi degerlendirilmektedir. Ikinci olarak, bireysel kredilere yonelik TGA analizlerinde, soz konusu ayrismada, kamu bankalarinin portfoylerinde teminatli yapida olan ve kredi geri odeme performansi diger kredilerden olumlu ayrisan konut kredilerinin, diger bankalarin portfoylerinde ise teminatsiz yapida olan ihtiyac kredisi ve kredi karti payinin daha yuksek olmasinin etkili olabilecegi gosterilmektedir. Son olarak yakin izlemedeki kredi oranlarina yonelik analizlerde, 2018 yilinda TFRS9 standardinin kullanilmaya baslamasi ile birlikte bankalarin yakin izleme oranlarinin ayristigi izlenmektedir. Kamu bankalari ile diger bankalarin gecikme nedeniyle yakin izlemeye alinan kredi oranlari benzerken, gecikmesi olmayan kredileri diger bankalarin kamu bankalarina kiyasla daha fazla yakin izlemede siniflandirdigi ve yakin izleme oranlarindaki ayrismanin bankalarin kullandigi model farkliligindan kaynaklandigi degerlendirilmektedir.