Emerging financial markets: spatial risks, elicitability of risk models, and shape shift contagion
Date
2020
Authors
Junior, Peterson Owusu
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The need to compare and contrast emerging market economies (EMEs) has never been greater,
especially following shocks to the global financial system. The quest to characterise EMEs does
not only emanate from risk mitigation purposes but also to maximise investment returns and to
articulate appropriate policy reforms even amidst financial turmoil. Moreover, increasing
integration of otherwise segmented economies has been largely amplified by the impact of these
market disturbances as seen in episodes of contagion. While the extant literature is ripe with studies
aimed at unearthing the differences and similarities in EMEs, most of them have centred on all too
ubiquitous dimensions such as returns and volatility behaviour, macroeconomic fundamentals, and
even cultural ethos, among others. Although these attributes are helpful, they fail to provide deeper
insights into EMEs for the rich understanding needed to make the best of investment and policy
decisions. This thesis provides empirical examination of additional dimensions by which to
characterise EMEs in order to provide a broader insight for prudent investment and policy
strategies. The thesis comprises three empirical studies on spatial risks, elicitability of risk models,
and “shape shift-contagion” of emerging markets equities, as the subtle extents by which to
compare and contrast EMEs, using advanced econometric techniques.
The first empirical study contrasts time-varying and spatial risks in order to assess systemic
vulnerabilities that affect emerging markets equity returns as well as to aid portfolio
diversification. The study spanned between Eurozone Crisis-Global Financial Crisis (EZC-GFC)
and post-GFC to represent turbulent and tranquil periods. With the joint (VaR, ES) models and
Global Liquidity Indicators (GLIs) for 12 EMEs which indicate time-varying and time-invariant
(spatial) risk measures, respectively, the study provides evidence that, the ranking of joint (VaR, ES) forecasts and spatial autocorrelations differ significantly. The results reveal that the overall
spatial autocorrelation between the 12 EMEs is smaller and negative for post-GFC as opposed to
positive and bigger for EZC-GFC periods. This suggests that EMEs may have employed prudent
liquidity policies to enhance their resilience to systemic susceptibilities as bitter lessons learnt from
crises experiences. The implications are that (VaR, ES) forecasts rankings are irrelevant, since
time-invariant systematic debacles have no respect for time-varying tail risks. For investors,
international portfolio diversification tends to yield its expected risk-minimising outcomes during
the post-GFC period. The study posits “Financial distance” as an extension of Cultural,
Administrative/Political, Geographic, and Economic (CAGE) distance dimensions to characterise
markets. This study is a subtle departure from the use of returns and volatilities used in describing
economies.
The second study examines the dynamics of emerging markets tail risk modelling and selection
behaviour under comparative back testing requirement in the Basel III paradigm. The study does
not only contribute to the growing need to correctly forecast and select the best tail risk model for
internal risk management purposes, but it also fits well into the aim of reducing regulatory
arbitrage. Across three market periods signifying tranquil and turbulent times, the study finds
evidence of time-, percentile-, equity-, and market period-dependent Superior Set Models (SSMs)
for 24 EMEs. These imply homogeneous vis-à-vis heterogeneous risk models which provide
portfolio diversification impetus for specific markets. Further, while some of the equities show
similar SSMs, there is no definite factor (such as either size of the market, geographical proximity,
and financial market maturity, among others) that can be attributed to this pattern. This study throws a further challenge to the mechanism of “bucketing” different markets into one class – a
typical practice of indexing institutions.
The last study investigates the role of higher moments in establishing the levels of connectedness
and contagion in EMEs under time-varying conditions. The findings surmise that EMEs respond
differently to both asymmetric and extreme returns across the spectrum of tranquil and turbulent
market periods. The novel rolling-window based generalised lambda distribution (GLD),
combined with the wavelet multiple correlation (WMC) and wavelet multiple cross correlation
(WMCC), and Baruník and Křehlík (2018) (BK18) spillover techniques show frequency
dependent connectedness and time-dependent fleeting higher moment contagious episodes which
are removed from the EZC and GFC periods. The results also expose the dominance of some
EMEs in the transmission of shocks instead of the United States, for instance. Nonetheless, the United States emerges a net transmitter of shocks rather than a net recipient. These dynamics sound caution to policy makers and investors alike to be more wary of shocks emanating from
EMEs as compared to those from the United States and by extension, other large developed
markets. Finally, the study establishes “shape shift-contagion” in emerging markets equities in
the short-term post both EZC and GFC episodes. This is consistent with shift-contagion and
delayed shift-contagion hypotheses. Moreover, it corroborates the notion that shock transmissions
tend to amplify even for an appreciable lapse in time after crisis episodes have died-off.
Nevertheless, this phenomenon varies from one EME to another. Hence this points to usefulness
of employing higher moments shocks to augment the mechanisms of classifying market
economies.
The results from all three empirical studies have one thing in common: EMEs are alike albeit
dissimilar in terms of spatial liquidity susceptibilities, the behaviour of equities risk models
pertaining to the reduction of regulatory arbitrage, and how they respond to financial market
shocks. These are some important fronts to distinguish EMEs from each other that are hitherto
missing from the literature. A number of investment and policy recommendations arising from the
findings in this thesis are offered for stakeholders to take advantage of the unique characteristics
of EMEs.
Description
A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy to the Faculty of Commerce, Law and Management in the Graduate School of Business Administration, Wits Business School, University of the Witwatersrand, Johannesburg, 2020