Lending booms and bank fragility : the South African experience

Maphosa, Michael (2020-03)

Thesis (PhD)--Stellenbosch University, 2020.

Thesis

ENGLISH SUMMARY : This thesis empirically examined the link between credit booms and bank fragility in South Africa. Fundamentally, the thesis looked at how the current developments in the domestic credit market affect the banking system, and in particular financial system stability in South Africa. The past two or three decades have seen an unprecedented increase in the level of domestic credit to the private sector. We have used mostly South African Reserve Bank and World Bank time series data for the three empirical studies. The thesis applied the robust autoregressive distributed lags (ARDL) approach by Pesaran, Shin and Smith (2001) and the nonlinear autoregressive distributed lags (NARDL) methodology of Shin, Yu and Greenwood-Nimmo (2014). The thesis contains three empirical studies. The first empirical study investigated the aggregate drivers of credit booms in South Africa using the causality tests based on the ARDL and Error Correction Model (ECM). Credit growth was analysed in relation to economic growth, types of loans, composition of credit by economic sector, debt-to-income ratio and the business cycle phases. Statistical evidence showed that South Africa has had a strong persistent growth in domestic credit over the past three decades with evidence of procyclical credit provision. The ARDL and ECM results showed that foreign capital inflows, mortgage loans, real interest rates and GDP per capita were important drivers of credit booms in South Africa. The second empirical study investigated whether excessive credit growth signalled future vulnerabilities in the South African banking sector. The main objective was to examine the growth-risk nexus in bank lending, given the credit booms currently experienced in South Africa. The business cycle was included in the model to reinforce the growth-risk nexus by allowing the study to develop a tri-variate model. The study found that credit risk management was still backward-looking and procyclical even though there were strong moves towards countercyclical models as suggested by the Basel Committee on Banking Supervision (Basel III accord). The ARDL model revealed the presence of a long-run relationship between credit risk, credit booms and the business cycle while the NARDL model established the presence of an asymmetric cointegration between the three variables. Negative shocks on the business cycle have a higher and more pronounced effect on credit risk than positive shocks while positive shocks to credit have a negative effect on credit risk in South Africa. The third empirical paper explored the relationship between credit booms, banking sector finance sources and its implications for financial stability in South Africa. It was noted that it was important for the study to identify the sensitivity of the banking sector to funding sources in South Africa. It was established that, like all other banking systems around the world, South African banks also tapped into wholesale funds to satisfy growing local demand for credit. The empirical results revealed a strong presence of an asymmetric relationship between credit booms and banking sector funding sources. Specifically, the study revealed that in the long run, positive developments in the wholesale funds market had a positive effect on the ability of the banking sector to satisfy credit demand; however, statistical evidence revealed that wholesale funds were highly volatile and susceptible to negative public signals. On the other hand, the study established that in the long run, positive developments in the domestic deposit market had positive effects on credit booms, while in the short run positive developments also had a positive effect on credit booms. Finally, negative shocks in domestic deposits in previous years had negative effects on credit booms. Based on the above, the study believes that credit booms are too risky to be left alone, and that appropriate monetary policy is a major instrument that is capable of curbing credit booms and limiting over-indebtedness in South Africa. The increase in the level of indebtedness beyond sustainable levels is a potential trigger of financial fragility in the economy. Strong fiscal policy capable of stimulating the finance and the real sector is important if and when a credit bust occurs. It is also important to note that fiscal discipline is required during the upswing since credit booms do not only flatter the balance sheets of banks and consumers that they extend credit to, but they also flatter government financial accounts.

AFRIKAANSE OPSOMMING : Hierdie tesis doen ’n empiriese ondersoek na die verband tussen kredietbloeitye en bankbreekbaarheid in Suid-Afrika. Fundamenteel het die tesis gekyk na hoe die huidige ontwikkelings in die binnelandse kredietmark die bankstelsel raak, en in die besonder finansiële stelsel-stabiliteit in Suid-Afrika. Die afgelope twee of drie dekades het ’n ongehoorde toename in die vlak van binnelandse krediet tot die privaatsektor gesien. Ons het meestal Suid-Afrikaanse Reserwebank- en World Bank-tydreeksdata vir die drie empiriese studies gebruik. Die tesis benut die kragtige outoregressief verspreide vertragings (ARDL)-benadering deur Pesaran, Shin en Smith (2001) en die nie-lineêre outoregressief verspreide vertragings (NARDL)-metodologie van Shin, Yu en Greenwood Nimmo (2014). Die tesis bevat drie empiriese studies. Die eerste empiriese studie het die algehele drywers van kredietbloeitye in Suid-Afrika ondersoek met gebruik van die kousaliteitstoetse gebaseer op die ARDL en Foutkorreksiemodel (ECM). Kredietgroei is ontleed in verhouding tot ekonomiese groei, tipes lenings, samestelling van krediet deur die ekonomiese sektor, skuld-tot-inkomste-verhouding en die sakesiklusfases. Statistiese bewys het getoon dat Suid-Afrika ’n sterk aanhoudende groei in binnelandse krediet oor die afgelope drie dekades gehad het met bewys van prosikliese kredietvoorsiening. Die ARDL- en ECM-resultate het getoon dat buitelandse kapitaalinvloei, verbandlenings, vaste-eiendomkoerse en BBP per capita belangrike drywers van kredietopbloei in Suid-Afrika was. Die tweede empiriese studie het ondersoek of oormatige kredietgroei op toekomstige kwesbaarhede in die Suid-Afrikaanse banksektor dui. Die hoofdoel was om die groeirisiko-neksus by banklenings te ondersoek, in die lig van die kredietopbloei wat tans in Suid-Afrika ondervind word. Die sakesiklus is by die model ingesluit om die groeirisiko-neksus te versterk deur die studie toe te laat om ’n trivariaatmodel te ontwikkel. Die studie het bevind dat kredietrisikobestuur nog terugwaarts gerig en prosiklies was, selfs al was daar sterk bewegings in die rigting van teensikliese modelle, soos gesuggereer deur die Basel Committee on Banking Supervision (Basel III-akkoord). Die ARDL-model het die teenwoordigheid van ’n langdurige verhouding tussen kredietrisiko, kredietbloeitye en die sakesiklus bekend gemaak, terwyl die NARDL-model die teenwoordigheid van ’n asimmetriese mede-integrasie tussen die drie veranderlikes gevestig het. Negatiewe skokke op die sakesiklus het ’n hoër en meer uitgesproke effek op kredietrisiko as positiewe skokke, terwyl positiewe skokke tot krediet ’n negatiewe uitwerking op kredietrisiko in Suid-Afrika het. Die derde empiriese skripsie het die verhouding tussen kredietbloeitye, banksektor-finansieringsbronne en die implikasies daarvan vir finansiële stabiliteit in Suid-Afrika ondersoek. Daar is kennis geneem dat dit vir die studie belangrik was om die sensitiwiteit van die banksektor vir befondsingsbronne in Suid-Afrika te identifiseer. Daar is vasgestel dat Suid-Afrikaanse banke, soos alle ander bankstelsels oor die wêreld heen, ook op aanvraaggrondslag groothandelfondse aangebied het om die groeiende plaaslike aanvraag vir krediet te bevredig. Die empiriese resultate het ’n sterk teenwoordigheid van ’n asimmetriese verhouding tussen kredietbloeitye en banksektor-befondsingsbronne getoon. Die studie het spesifiek getoon dat positiewe ontwikkelings op die lange duur in die groothandel-befondsingsmark ’n positiewe uitwerking op die vermoë van die banksektor gehad het om kredietaanvraag te bevredig; statistiese bewys het egter aan die lig gebring dat groothandelfondse hoogs ongestadig en vatbaar vir negatiewe openbare tekens was. Aan die ander kant het die studie vasgestel dat positiewe ontwikkelings in die binnelandse depositomark op die lange duur positiewe gevolge vir kredietopbloei gehad het, terwyl positiewe ontwikkeling oor die korter termyn ook ’n positiewe effek op kredietopbloei gehad het. Oplaas het negatiewe skokke in binnelandse deposito’s in voorafgaande jare negatiewe gevolge op kredietbloeitye gehad. Gebaseer op bovermelde, glo die studie dat kredietbloeitye te riskant is om alleen te los, en dat toepaslike monetêre beleid ’n belangrike instrument is wat daartoe in staat is om kredietbloeitye aan bande te lê en oormatige skuldlas in Suid-Afrika te beperk. Die toename in die vlak van skuldlas bo volhoubare vlakke is ’n potensiële sneller vir finansiële breekbaarheid in die ekonomie. Streng fiskale beleid wat in staat is tot stimulering van die finansies en die reële sector, is belangrik indien en wanneer ’n krediet-ineenstorting voorkom. Dit is ook belangrik om daarop te let dat fiskale dissipline vereis word tydens die opswaai omdat kredietbloeitye nie net die balansstate van banke en verbruikers aan wie hulle krediet verleen vlei nie, maar hulle vlei ook die finansiële rekeninge van die regering.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/108051
This item appears in the following collections: