- Paul Smith and Tatiana Lyubimova
Forecasting the Ethylene Market: Potential Savings using Particle.One AutoML tools
Updated: Apr 18, 2021

Chinese Ethylene Market
The Chinese ethylene market has experienced a decade of aggressive growth that is in lock-step with the country’s modern industrialization efforts. Ethylene is widely used in the chemical industry, with worldwide production exceeding other organic compounds. A significant part of the production goes toward polyethylene and plastics, agri industry, and textile manufacturing. According to expert forecasts, both the demand and production of Chinese-based ethylene will continue to grow at a rapid pace in the coming decade.
How We Add Value to Large Petrochemical Producers
Through our experience and expertise in quantitative modeling, Particle.One builds predictive models that accurately gauge both the benchmark and local prices of rapidly-expanding commodities such as ethylene.
Our modeling takes into account global supply, demand, and inventory levels of the commodity and local Chinese logistics data. We analyze methods of transport, accounting for local shipping within the country, as well as export data. We also factor in supply-chain disruptions, which include plant shutdowns.
As an integral part of our modeling process, we perform high-resolution estimates with frequencies ranging from hourly to weekly. Finally, we include any of the petrochemical producer’s proprietary information necessary to improve the quality of the price forecasting method.
Forecasting the Price of Ethylene
Our advanced modeling techniques have been refined in the quantitative hedge fund space, and we use these predictive analysis methods to improve the sale of refined ethylene products and raw materials. To complete the prediction process, we feed both regionally-based and aggregated Chinese supply and demand indicators into our model in order to determine the commodity’s price trend for the coming week.
Particle.One’s approach to forecasting the price of ethylene is predicated on advanced quantitative modeling practices combined with regionally-sourced data, and is designed to help you streamline your buying and selling practices in order to optimize your ROI.
Case study
Using our models for price prediction, we optimize the purchase or selling of a given amount of product over a given period of time.
Assuming you want to sell “1000 metric tons of ethylene over N weeks time”
Scenario 1:
You sell this amount in N equal amounts 1000 / N metric tons, one per week
The average price of this approach of buying 1000 metric tons of ethylene in 2012 to 2019 would be $1,191 USD/ton
Scenario 2:
You sell this amount in N unequal amounts based on the predicted price value for the next week
The average price of this approach of selling 1000 metric tons of naphtha would be $1,476 USD/ton
Following the second scenario, the gain would be 22%.
Ethylene price forecast construction
Predictors
Predictors consist of regional Chinese supply indicators and two aggregated Chinese supply and demand indicators:
Ethylene output (aggregated)
Ethylene output, Jiangsu province
Ethylene output, Shandong province
Ethylene output, Shanghai province
Ethylene output, Zhejiang province
Ethylene consumption (China, aggregated)

Target
CFR Southeast Asia Ethylene, spot price, high-end

Strategy
Predict 1-week forward returns
Trade once weekly

