Gourav Roy
Bangladesh Institute of Capital Market (BICM), Dhaka, Bangladesh

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Connectedness, Stress Amplification, and Predictive Spillovers across Dhaka Stock Exchange Index Segments Gourav Roy
Journal of Enterprise and Development (JED) Vol. 8 No. 3 (2026): September - December
Publisher : Faculty of Islamic Economics and Business of Universitas Islam Negeri Mataram

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20414/jed.v8i3.16151

Abstract

Purpose: This study examines whether DSEX, DS30, and DSES form a stable long-run system, whether cross-index forecast uncertainty intensifies under stress, and whether cross-index information improves next-day segment-rotation forecasts.Method: The analysis uses 2,383 synchronised daily log returns from 3 January 2016 to 19 July 2026. It combines unit-root and cointegration tests, a stable return VAR, generalised forecast-error variance decomposition with bootstrap inference, rolling connectedness with HAC and moving-block bootstrap tests, externally dated stress windows, and a chronological train-validation-test forecasting design.Result: The three index levels have cointegration rank zero despite return correlations above 0.91. The 10-day total connectedness index is 63.45% (95% bootstrap interval: 62.87%-63.97%). DSEX is a small but statistically positive net transmitter, DSES is a net receiver, and DS30 is not statistically distinguishable from zero. Connectedness rises by 2.00 percentage points under volatility-defined stress and 1.54 points under externally dated stress. Full cross-index features do not improve confirmatory out-of-sample forecasts after dependence-robust inference and multiplicity adjustment.Practical Implications for Economic Growth and Development: The findings caution against treating formal index labels as independent risk exposures. Rolling connectedness can support surveillance, margin design, and portfolio stress testing, helping reduce diversification errors and strengthen capital-market intermediation.Originality/Value: The study jointly separates long-run integration, forecast-error connectedness, stress amplification, directional uncertainty, and incremental predictive value across all three official DSE index segments while applying dependence-robust and model-selection-aware inference.