Shadow banking and systemic risk in South Africa

Date
2023-12
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH SUMMARY: Shadow banking, credit intermediation involving entities and activities outside the regular banking system, is growing globally and even at a faster rate in emerging economies. The development of shadow banking is driven by a demand for safe and liquid investments redeemable at par, investors’ pursuit of higher returns, regulatory arbitrage, financial sector development, technology, and economic growth. Shadow banks perform credit, liquidity, and maturity transformation without or with limited supervision and backstop support from the central bank, which among policymakers and regulators in South Africa and elsewhere led to concern about shadow banks’ potential threat to financial system stability or systemic risk. However, a lack of data impeded the effective regulation of shadow banking. This study is motivated by three critical gaps in the literature. Firstly, there are few studies on the contribution of shadow banking to systemic risk globally, but none focus on South Africa despite the unique structure and composition of shadow banking in the country. Secondly, there is scant information on the interconnectedness within the shadow banking system. The interconnectedness, through counterparty contracts and common assets and market exposure, creates a channel for contagion in financial systems. Lastly, efforts to prevent systemic risk prioritise the monitoring of large shadow banking entities; however, whether a relationship exists between the size of shadow banks and system risk remains an empirical issue. Nevertheless, large financial institutions are theoretically assumed to be more connected such that their failure could trigger systemic failure. To this effect, we investigate the contribution of shadow banking to systemic risk in South Africa using monthly returns data of fixed-income funds, funds-of-funds, money market funds, and multi-asset funds for the period of January 2015 to December 2021. We measure systemic risk using Conditional Value-at-Risk (CoVaR) and Delta Conditional Value-at-Risk (ΔCoVaR) and apply the quantile regressions with financial sector returns as an independent variable. We include a set of state variables (equity market return, market volatility, yield spread) as independent variables to produce the time varying CoVaR and ΔCoVaR. We determine the contribution of shadow banking to systemic risk, and also breakdown the analysis into pre-COVID-19 and during-COVID-19 periods. We analyse interconnectedness in subsamples of shadow banks with high marginal contributions to systemic risk and examine the relationship between the size of shadow banks and systemic risk. Insights from the study are presented in three empirical chapters to address the critical gaps identified. In the first chapter, we report the contribution of shadow banking to systemic risk in South Africa and highlight that all shadow banks contribute to systemic risk. Multi-asset funds provide the largest contribution to systemic risk, followed by funds-of-funds and fixed-income funds. Whilst money market funds do not contribute to systemic risk during whole study period, they are the largest contributors during COVID-19. Therefore, money market funds are found to be systemic only during turbulent periods. Altogether, the contribution of shadow banking to systemic risk follows patterns of macroeconomic and financial distress and is at an all-time high during the COVID-19 period. In the second empirical chapter, we focus more narrowly but also in more depth on the subsample with the highest ΔCoVaR and estimate interconnectedness using a Toda and Yamamoto Granger causality analysis. The level of interconnectedness in shadow banking is high, and linkages reach a peak during COVID-19 showing the impact of the pandemic on the financial sector. Multi-asset funds are more interconnected than all other funds. However, money market funds have more inbound connections. Generally, multi-asset funds are significant transmitters and receivers of systemic risk whereas money market funds are receivers. In the last empirical chapter, we measure the relationship between the size of shadow banks and systemic risk using ordinary least square regressions (OLS) and quantile regression. We find a positive linear relationship between the size of shadow banks and systemic risk, and the linkages are pronounced among multi-asset funds, especially retail multi-asset funds sponsored by asset managers. There is no relationship between the size of fixed-income funds, funds-of-funds, and money market funds and systemic risk. The size of shadow banking institutions does not increase the marginal contribution to systemic risk. Overall, we conclude that shadow banking contributes to systemic risk in South Africa. Among the shadow banking entities, multi-asset funds are the largest contributors to systemic risk, and money market funds are more systemic during turbulent periods. Notably, systemic risk and interconnectedness increase during periods of shocks, reaching an all-time high during the COVID-19 pandemic. Our results also confirm the relationship between the size of shadow banks and systemic risk, with the relationship being stronger for retail multi-asset funds sponsored by asset managers. The regulatory implications of our results point to the need for the regulator to strengthen macroprudential regulation of shadow banking to prevent and manage systemic risk.
AFRIKAANSE OPSOMMING: Skadubankdienste, oftewel kredietbemiddeling deur middel van instellings en aktiwiteite anders as in die gebruiklike bankdienstestelsel, neem wereldwyd toe en in ontluikende ekonomiee des te vinniger. Skadubankdienste se ontwikkeling word aangedryf deur ʼn vraag na veilige en likiede instrumente wat teen sigwaarde afgelos kan word, beleggers wat hoer opbrengste najaag, regulatoriese arbitrage, ontwikkeling van die finansiele sektor, tegnologiese ontwikkeling en ekonomiese groei. Skadubankinstellings pas krediet-, likiditeit- en uitkeeromvorming toe sonder enige of slegs met beperkte oorsig en eindborgsteun van die sentrale bank, wat beleidmakers en reguleerders in Suid-Afrika en elders besorg maak oor sistemiese risiko, oftewel die potensiele bedreiging wat skadubankdienste vir die bestendigheid van die finansiele stelsel inhou. Effektiewe regulering van skadubankdienste word egter deur ʼn gebrek aan data gekniehalter. Hierdie studie word deur vier kritieke leemtes in die literatuur gemotiveer. Eerstens is daar min studies oor die bydrae van skadubankdienste tot sistemiese risiko wereldwyd en niks wat op Suid-Afrika fokus nie, ondanks die unieke struktuur en samestelling van skadubankdienste in die land. Tweedens is min inligting beskikbaar oor die ineengeskakeldheid van die skadubankstelsel. Die ineengeskakeldheid, deur teenparty-ooreenkomste en gemeenskaplike bates en markblootstelling, skep ʼn kanaal waardeur besmetting in ʼn finansiele stelsel kan verprei. Laastens geniet die monitering van groot skadubankinstellings voorkeur in pogings om sistemiese risiko te verhoed, maar of daar ʼn verband is tussen die grootte van skadubankinstellings en sistemiese risiko bly ʼn empiriese vraagstuk. Nietemin, groot finansiele instellings word teoreties as meer ineengeskakel beskou, sodat as hulle misluk dit ʼn sistemiese ineenstorting kan veroorsaak. Gevolglik ondersoek ons die bydrae van skadubankdienste tot sistemiese risiko in Suid-Afrika. Ons gebruik data oor die maandelikse opbrengste van vasteinkomstefondse, fondse-van-fondse, geldmarkfondse en multibatefondse vir die tydperk Januarie 2015 tot Desember 2021. Ons meet sistemiese risiko deur voorwaardelike waarde-op-risiko (Conditional Value-at-Risk [CoVaR]) en Delta voorwaardelike waarde-op-risiko (Delta Conditional Value-at-Risk [ΔCoVaR]) te gebruik en pas die kwantielregressies met finansielesektoropbrengste as ʼn onafhanklike veranderlike toe. CoVaR is die bydrae tot sistemiese risiko en ΔCoVaR die marginale bydrae tot sistemiese risiko. Ons sluit ʼn stel standveranderlikes in (aandelemarkopbrengs, markongestadigheid, opbrengsspreiding) as onafhanklike veranderlikes om die tydwisselende CoVaR en ΔCoVaR te produseer. Ons bepaal die bydrae van skadubankdienste tot sistemiese risiko vir die hele studietydperk en verdeel die ontleding verder in die tydperk voor COVID-19 en tydens COVID-19. Ons ontleed die kwessie van ineengeskakeldheid in substeekproewe van skadubankinstellings wat hoe marginale bydraes tot sistemiese risiko lewer en ondersoek die verband tussen die grootte van skadubankinstellings en sistemiese risiko. Insigte uit die studie word in drie empiriese hoofstukke aangebied om die kritieke leemtes wat geidentifiseer is, te hanteer. In die eerste hoofstuk doen ons verslag oor die bydrae van skadubankdienste tot sistemiese risiko in Suid-Afrika en beklemtoon ons dat alle skadubankinstellings tot sistemiese risiko bydra. Multibatefondse lewer die grootste bydrae tot sistemiese risiko, gevolg deur fondse-van-fondse en vasteinkomstefondse. Hoewel geldmarkfondse nie in die hele studietydperk tot sistemiese risiko bydra nie, is hulle tydens COVID-19 die grootste bydraers. Daar word dus bevind dat geldmarkfondse slegs in onstuimige tye sistemies is. Oor die algemeen volg skadubankdienste se bydrae tot sistemiese risiko die patroon van makro-ekonomiese en finansiele nood en is op ʼn allerhoogtepunt tydens die Covid-19-tydperk. In die tweede empiriese hoofstuk is ons enger maar ook meer in diepte toegespits op die substeekproef met die hoogste ΔCoVaR en word ineengeskakeldheid geraam deur ʼn Toda en Yamamoto Granger-kousaliteitsontleding te doen. Die vlak van ineengeskakeldheid in skadubankdienste is hoog en die skakeling bereik ʼn hoogtepunt tydens COVID-19, wat aanduidend is van die impak van die pandemie op die finansiele sektor. Multibatefondse is meer ineengeskakel as al die ander fondse.. Geldmarkfondse het egter meer inwaartse skakeling. Oor die algemeen is multibatefondse beduidende verspreiders en ontvangers van sistemiese risiko, terwyl geldmarkfondse ontvangers is. In die laaste empiriese hoofstuk meet ons die verband tussen die grootte van skadubankinstellings en sistemiese risiko deur gewone kleinstekwadraatregressie (OLS) en kwantielregressie toe te pas. Ons vind ʼn positiewe lineere verhouding tussen die grootte van skadubankinstellings en sistemiese risiko en die verband is veral merkbaar onder multibatefondse, veral kleinhandelmultibatefondse wat deur batebestuurders geborg word. Daar is geen verband tussen die grootte van vasteinkomstefondse, fondse-van-fondse of geldmarkfondse en sistemiese risiko nie. Die grootte van skadubankinstellings verhoog nie die marginale bydrae tot sistemiese risiko nie. In die geheel is ons slotsom dat skadubankdienste tot sistemiese risiko in Suid-Afrika bydra. Onder die skadubankentiteite is multibatefondse die grootste bydraers tot sistemiese risiko en geldmarkfondse is meer sistemies in onstuimige tye. Opmerklik is dat sistemiese risiko en ineengeskakeldheid in tye van makro-ekonomiese skokke toeneem en tydens die COVID-19-pandemie ʼn allerhoogtepunt bereik. Ons resultate bevestig ook die verband tussen die grootte van skadubankinstellings en sistemiese risiko, met die verband sterker vir kleinhandelmultibatefondse wat deur batebestuurders geborg word. Die regulatoriese implikasies van ons resultate dui op die noodsaaklikheid dat die reguleerder makrovoorsorgmaatreels vir skadubankdienste moet versterk om sistemiese risiko te voorkom en te bestuur.
Description
Thesis (PhD)--Stellenbosch University, 2023.
Keywords
Citation