Scholarlink

<p>This study examines the long-term performance dynamics of the Malaysian property sector by integrating the Modified Internal Rate of Return (MIRR) with a two-state Markov chain model, analyzing stock market data from 62 Shariah-compliant property companies over the period 2010–2022. The empirical analysis reveals a transition probability matrix characterized by strong state persistence, with probabilities of 0.882 for remaining in bad states and 0.842 for maintaining good states. The stationary distribution demonstrates a contraction bias, with the system spending 57.2% of time in bad states versus 42.8% in good states a 14.4 percentage-point differential exceeding comparable emerging markets. The sector exhibits asymmetric transition dynamics, with decline from prosperity occurring 25% faster (1.75 periods) than recovery from downturns (2.34 periods). Mean sojourn times indicate extended durations in both states, with bad states persisting 34% longer (8.47 periods) than good states (6.33 periods). The entropy value of 0.570 bits (57.0% of maximum randomness) indicates moderate predictability, while the mixing time of 3.10 periods suggests rapid convergence to equilibrium. Sensitivity analysis reveals asymmetric policy leverage: recovery-enhancing interventions have approximately 34% greater impact on long-run good state probability than stability-enhancing interventions. Stress testing demonstrates that under severe financial crisis scenarios, the long-run probability of good states collapses from 42.8 to 15.4%, with required capital buffers increasing by 138%. The proposed two-state Markov model achieves superior predictive accuracy (RMSE = 0.038) compared to mixture transition distribution (0.051) and regime-switching neural network (0.041) alternatives, with 44-fold computational speed advantage over neural network approaches. These findings provide quantitative foundations for dynamic portfolio optimization, state-contingent capital adequacy planning, and evidence-based policy design for emerging market property sectors.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Modelling long-term investment dynamics using markov chain analysis: a two-state probabilistic approach

  • 期刊论文

  • 发表日期：

    2026年6月18日

查看全文
  • Hamza Abubakar,
  • Amani A. Idris Sayed,
  • Abubakar Balarabe Karaye,
  • Mukhtar Salisu Abubakar

摘要

This study examines the long-term performance dynamics of the Malaysian property sector by integrating the Modified Internal Rate of Return (MIRR) with a two-state Markov chain model, analyzing stock market data from 62 Shariah-compliant property companies over the period 2010–2022. The empirical analysis reveals a transition probability matrix characterized by strong state persistence, with probabilities of 0.882 for remaining in bad states and 0.842 for maintaining good states. The stationary distribution demonstrates a contraction bias, with the system spending 57.2% of time in bad states versus 42.8% in good states a 14.4 percentage-point differential exceeding comparable emerging markets. The sector exhibits asymmetric transition dynamics, with decline from prosperity occurring 25% faster (1.75 periods) than recovery from downturns (2.34 periods). Mean sojourn times indicate extended durations in both states, with bad states persisting 34% longer (8.47 periods) than good states (6.33 periods). The entropy value of 0.570 bits (57.0% of maximum randomness) indicates moderate predictability, while the mixing time of 3.10 periods suggests rapid convergence to equilibrium. Sensitivity analysis reveals asymmetric policy leverage: recovery-enhancing interventions have approximately 34% greater impact on long-run good state probability than stability-enhancing interventions. Stress testing demonstrates that under severe financial crisis scenarios, the long-run probability of good states collapses from 42.8 to 15.4%, with required capital buffers increasing by 138%. The proposed two-state Markov model achieves superior predictive accuracy (RMSE = 0.038) compared to mixture transition distribution (0.051) and regime-switching neural network (0.041) alternatives, with 44-fold computational speed advantage over neural network approaches. These findings provide quantitative foundations for dynamic portfolio optimization, state-contingent capital adequacy planning, and evidence-based policy design for emerging market property sectors.