Montée industrielle du Maroc et valorisation de la Bourse de Casablanca : une modélisation prospective par réseaux de neurones

Morocco's promotion to the top of the African Development Bank (AfDB) Africa Industrialisation Index in May 2026, ahead of South Africa, marks a structural turning point for the Moroccan economy. This paper offers a prospective analysis of the spillover effects of this industrial rise on the Casablanca Stock Exchange. Using a comparative panel of eight African and emerging stock exchanges over the period 2010–2025, extended to a semi-annual frequency to enlarge the training sample, we deploy artificial neural network modelling (multilayer perceptron and recurrent LSTM network) to project the trajectory of market capitalisation, the MASI index, and sectoral performance towards 2030. Explanatory variables incorporate the AfDB industrialisation index, manufacturing value added, industrial investment flows, and a set of international macro-financial indicators (interest rates, market liquidity, global financial conditions). Hyperparameter optimisation (grid search and time-series cross-validation), an explicit train/validation/test split, and overfitting controls are detailed to ensure the robustness of the approach. Our results suggest that a sustained rise in the industrialisation index would translate into a projected growth in market capitalisation of around 45 to 60 % by 2030, reported with confidence intervals and driven mainly by the industrial, automotive, and aerospace sectors. The predictive robustness of the LSTM model, confirmed against benchmark econometric models (OLS, ARIMA) and alternative machine-learning algorithms (Random Forest, XGBoost), validates the relevance of the connectionist approach for capturing the nonlinear dynamics between industrial transformation and stock-market valuation