Domestic Public Borrowing and Private Credit Dynamics in Türkiye: Evidence from an ARDL Approach

Özet

Purpose: This study examines whether domestic public borrowing is associated with the displacement of private credit in Türkiye and distinguishes a conventional quantity-based crowding-out mechanism from public-borrowing flows, bank sovereign exposure, and broader macro-financial conditions. Design/Methodology/Approach: Monthly data from January 2010 to November 2023 are analysed using an ARDL/UECM reference specification, short-run OLS models with HAC standard errors, alternative credit outcomes and monetary controls, structural-break diagnostics, and Toda-Yamamoto directionality tests. The commercial-credit series is constructed consistently for all 167 months as total loans minus consumer loans and credit-card balances.

Findings: The bounds test supports a long-run level relationship, but adjustment is very slow. The normalized long-run coefficient on real domestic debt is negative and statistically insignificant, while borrowing cost and exchange-rate depreciation are negatively associated with credit and real activity is positively associated with credit. Official net domestic borrowing does not display a negative short-run association with commercial-credit growth. By contrast, increases in banks' government-securities exposure are associated with weaker commercial-credit growth, and this result remains significant under alternative financing-condition controls. Directionality tests do not establish a one-way causal relationship, and break analysis indicates regime sensitivity. Discussion: The evidence does not support a stable, mechanical full-sample crowding-out relationship based on the quantity of domestic public debt. It instead points to a narrower portfolio-allocation association within the sovereign-bank nexus and to a debt-credit relationship that depends on financing conditions and macro-financial regimes. The results are therefore interpreted as associations rather than structural causal effects.