SCENARIO FORECASTING OF TEXTILE SECTOR DEVELOPMENT

Abstract

This thesis develops a scenario-based recurrence model for forecasting textile-chain export composition and
chain depth over 2027–2031, distinguishing the contribution of incentive policy from that of exogenous market shocks.
Each chain link’s export growth rate is modelled as the sum of a policy-driven component, calibrated using marginal-response
coefficients, and a shock-driven component, estimated by regression against four exogenous disturbances specific
to each link. The model was validated retrospectively against 2017–2025 data and then applied to forward projections
under three scenarios distinguished by incentive allocation and institutional readiness. Retrospective validation yields a
mean absolute error of 6.25 percent and a Theil inequality coefficient of 0.052, indicating an adequate fit, with the largest
single-year error attributable to an administrative measure affecting exports that the model does not capture as a market
shock. Forward simulation shows that reallocating incentives toward the chain-compatible distribution nearly quadruples
the five-year gain in chain depth relative to maintaining the current allocation. Monte Carlo simulation further shows that this
gain is markedly more robust to exogenous shocks, with forecast dispersion under the reallocated scenario being less than
one-third of that under the current allocation framework.