AN EMPIRICAL ANALYSIS OF BUBBLES IN THE NIGERIAN STOCK EXCHANGE (1985-2018): A GENERALISED SUP AUGMENTED DICKEY-FULLER APPROACH

1Jamilu Iliyasu, 2Aliyu Rafindadi Sanusi, 3Dahiru Suleiman

Abstract

This study sets out to conduct an empirical analysis of the nature of bubbles in the Nigerian Stock
Exchange (NSE). The paper achieved this in three inter-related steps. In the first step, a battery of
tests, including the Generalised Supremum Augmented Dickey-Fuller(GSADF) test and Backwards
Supremum Augmented Dickey-Fuller(BSADF)were conducted using monthly data on (the nominal
and real values of the) All-Share Index of the NSE (NSE-ASI) to establish the existence of bubbles
during the period 1985 - 2018. In the second step, the paper employed the Component-GARCH-inmean
(CGARCH-M) model to estimate the bubbles risk premium as well as decompose returns
volatility into its transitory and permanent components. In the third step, the paper employs a Logistic
Regression Model to examine the influence of the permanent and transitory volatility components in
the likelihood of a bubble in the NSE. Results from the GSADF and BSADF tests show evidence of two
episodes of bubbles based on the Nominal All-Share Index (ASI) and three episodes based on Real All-
Share Index(RASI). Analysis of results obtained from the Component GARCH-M model shows
thatboth episodeswere associated witha positive risk premium. Estimates of the Logistic Regression
model suggest that periods dominated by permanent volatility were less likely to experience bubbles
episodes, while periods dominated by transitory volatility have ahigher risk of experiencing
bubbles.The paper, therefore, concludes that a prolonged period of a rising risk premium
characterised by transitory volatility that is driven by market sentiments are more prone tobubbles
than periods of higher volatility that is driven by fundamentals. Therefore, when a rise in the risk
premium and transitory volatility are observed, financial regulators such as Central Bank of Nigeria
(CBN) and Securities and Exchange Commission (SEC) should diagnose the Market for bubbles
using modern econometrics techniques such as the one used in this study.