The core difficulty the researchers faced is philosophical as much as technical. Substitution, by its nature, can never be directly observed, because it only exists relative to a counterfactual: a version of history in which the substitute product was never produced, or produced at a different rate. Econometric studies can measure how markets responded to past price shocks, but they cannot answer the fundamental question of what would have happened had a product simply not existed. Life cycle assessment, the standard tool for comparing the environmental footprints of products, sidesteps the problem entirely by implicitly assuming perfect substitution—that one unit of product A displaces exactly one unit of product B—without ever measuring whether markets actually behave that way. Prior attempts to estimate displacement rates across various sectors have produced averages of roughly 40 to 60 percent, but with enormous variation and deep uncertainty rooted in the price elasticities that drive the calculations.

To break through this impasse, the team adopted system dynamics modeling, a simulation technique that represents a system’s behavior through feedback loops rather than static statistical correlations. Their model, named SubText, covers the three fiber categories that together account for more than 90 percent of the global fiber market: synthetic fibers, represented by polyester, which makes up 85 percent of synthetics; regenerated cellulosic fibers, represented by viscose, which accounts for 90 percent of that category; and cotton. The model is built on conventional microeconomic supply and demand equations, with demand responding to prices, income, and population, and supply responding to prices, feedstock availability, and feedstock costs. The crucial innovation lies in the inclusion of empirically estimated cross-price elasticities, which capture how strongly the consumption of one fiber responds to price changes in competing fibers—the higher the cross-price elasticity, the closer the substitutes.

The model’s architecture reflects the peculiar dynamics of the textile market. Balancing feedback loops link price to demand, price to supply, and price to profit to production capacity, generating the oscillating behavior typical of commodity markets. A single reinforcing loop transmits substitution effects between fiber pairs through their mutual prices. The researchers also introduced parameters rarely seen in market models, including an elasticity for the share of fast fashion, which serves as a proxy for a production strategy emphasizing quantity over quality, and a supply constraint for cotton that reflects the fact that the land area devoted to cotton cultivation has remained nearly unchanged for decades. Supply in the model is pushed by feedstock availability: crude oil extraction for synthetics, cotton seed harvest for cotton, and pulpwood harvest for cellulosic fibers. Time delays of one to two quarters for supply adjustments and a 22-year average mill lifetime for capacity divestments complete the picture of a market that responds sluggishly and asymmetrically to shocks.

The calibration process is where the approach earns its credibility. The researchers first assigned plausible parameter values from the literature and expert judgment, then ran a Powell optimization procedure with up to 5,000 iterations per variable to minimize the mismatch between modeled and statistical production and price data for the period 2004 to 2021. The model successfully reproduced long-term trends for all three fibers, smoothing out only the price peaks caused by extreme shocks. This calibration to historical data is what allows the method’s most novel step: recreating alternative histories. In scenarios P1 through P3, the researchers simulated the past development of the market with one of the three fibers effectively removed, comparing the resulting trajectories against the actual historical baseline. The difference between the two reveals, for the first time in a calibrated framework, how much the remaining fibers would have expanded had one competitor never grown.

The results are striking. The strongest substitution effect ran between regenerated cellulosic fibers and cotton: had viscose production not increased since 2004, 32 percent of that unrealized increase would have been compensated by higher cotton production, accompanied by a higher cotton price. In the forward-looking scenarios, a one-unit increase in cellulosic fiber production led to a 31 percent reduction in cotton production, closely matching the backward-looking estimate. Every other substitution pair produced far weaker effects, with a maximum of 8 percent displacement in supply-shock scenarios. Cotton proved the most responsive fiber overall, likely because it is the most commoditized of the three, while synthetics and cellulosics—still in the growth phase of their product life cycles—barely reacted to market shocks. Most tellingly, synthetics appeared almost completely isolated from the other fibers: changes in the production and price of cotton or viscose had essentially no effect on polyester output.

The environmental implications follow directly from these modest displacement rates. When the researchers recalculated displacement factors using their empirical rates instead of the conventional one-to-one assumption, the picture darkened considerably. Because a unit increase in one fiber displaces at most a third of a unit of another fiber, the majority of any supply increase simply adds to total consumption rather than replacing anything. When the researchers combined these empirical displacement rates with cradle-to-gate life cycle inventory data from ecoinvent, the global warming potential flipped from a net avoided emission to a net caused emission in all applicable substitution cases. The only exception was cellulosic fibers substituting for cotton, where the displacement rate was high enough to preserve avoided impacts in marine and terrestrial eutrophication, water use, and freshwater ecotoxicity. In every other case and most impact categories, expanding the supply of the supposedly greener fiber increased the sector’s overall environmental load.

The sensitivity analysis reinforced both the strengths and the limits of the method. The backward-looking scenarios proved remarkably stable, because any alternative parameterization would have degraded the fit between modeled and observed historical data, providing a built-in check against arbitrary assumptions. The forward-looking scenarios, by contrast, remained volatile, sensitive to the growth trajectory of fast fashion—which the researchers identified as the single largest source of variance—and prone to nonlinear responses when supply constraints or profitability thresholds were hit. Notably, without incorporating the fast fashion effect, the model’s baseline production would have fallen 47 percent short for synthetics, 45 percent for cellulosics, and 13 percent for cotton by 2021, suggesting that the explosive growth of cheap, disposable fashion, along with population and income growth, drives fiber production far more than any substitution dynamics do.

The authors are careful to frame their estimates as orders of magnitude rather than precise numbers, and to note that displacement rates are specific to the period, region, and market structure studied. Still, the broader message is hard to escape. For bioeconomy strategies that lean heavily on substitution as a pathway away from fossil resources, the textile case delivers a sobering verdict: the fossil-based synthetic fiber market is almost completely insulated from the production of alternative fibers, and cellulose-based fibers mostly substitute for each other rather than for polyester. Policies that encourage supply growth, such as subsidy schemes for novel fibers, are therefore unlikely to reduce the sector’s environmental footprint on their own. The researchers argue that alternative pathways—emission pricing, feedstock supply quotas, and measures targeting the overconsumption fueled by fast fashion—deserve far more attention, and that the alternative-history simulation method they pioneered should now be replicated across other sectors to test how widely these conclusions hold.

Subject of Research: Quantifying displacement rates and substitution dynamics among cotton, synthetic, and regenerated cellulosic textile fibers using system dynamics simulation

Article Title: Quantifying the unquantifiable: simulation of displacement rates among textile fibers

Article References: Hurmekoski, E., & Boiger, T. (2026). Quantifying the unquantifiable: simulation of displacement rates among textile fibers. Journal of Industrial Ecology. https://doi.org/10.1007/s44498-026-00161-6

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00161-6

Keywords: substitution, displacement rate, textile fibers, system dynamics, life cycle assessment, fast fashion, cotton, viscose, polyester, counterfactual scenarios, industrial ecology, bioeconomy