PVC Operational Pricing Model

A baseline HTML model for estimating the PVC price across building blocks: cost floor, supply pressure, demand ceiling, and arbitrage. It includes three market regimes, scenario analysis, historical charts, a dedicated Netback / basis tab for Company, and an expanded description of the calculation methodology and formulas.

Parameter inputs
If the regime is unknown, leave it on Balanced. Then review the sensitivity when switching to Oversupply and Tight.
Cost Floor
Supply Pressure
Demand Ceiling
Arbitrage / policy
Calculation results
Model price
$650.6/t
benchmark: $760.0/t; regime: oversupply
Cost floor
$561.6
lower bound
Supply pressure
100.0
supply pressure index
Demand ceiling
86.3
demand ceiling index
Model below marketOversupply

Calculation formula

ModelPrice = w_cf·CF + w_sp·(CF + SupplyTerm) + w_dc·(CF + DemandTerm) + w_arb·(CF + ArbTerm) CF = 0.52·Eth + 0.12·Caustic + 0.18·Brent + 0.18·RegionalEnergy + 110 SupplyTerm = (SupplyPressure - 50) × 6 DemandTerm = (DemandCeiling - 100) × 5 ArbTerm = Arbitrage Weights(oversupply) = { cf: 0.55, sp: 0.28, dc: 0.07, arb: 0.1 }

Weights by regime

Regime Cost floor Supply Demand Arbitrage
Oversupply 0.55 0.28 0.07 0.10
Balanced 0.60 0.18 0.12 0.10
Tight 0.62 0.08 0.22 0.08

Diagnostics

Block Value Interpretation
Benchmark760.0External reference price for the selected region
Cost floor561.6Cost base below price: a normal configuration
Supply pressure100.0China / new capacity weighs on the market
Demand ceiling86.3Demand does not hold the price up on its own
Arbitrage97.6Arbitrage does not erode the price
Model gap-109.4Coefficient calibration is required
This is a transparent structural model. Its purpose is to decompose the price into its underlying mechanisms and test hypotheses. Higher accuracy requires historical data, lags, and subsequent calibration of the coefficients against actual time series.
Historical data
If no CSV is uploaded, the page uses a built-in demo series covering 18 months.

Expected CSV fields

date,pvc_price_asia,pvc_price_eu,pvc_price_us,ethylene_price,caustic_soda_price,electricity_eu,oil_brent,natgas_us,china_coal_price,china_pvc_export,china_capacity_util,global_capacity_growth,outages_capacity,china_construction_index,eu_construction_index,us_construction_index,india_pvc_import,scfi_index,freight_asia_eu,china_policy_dummy,antidumping_measures,regime
At a minimum, the charts only require the fields date, the regional PVC price, ethylene_price, china_pvc_export and construction indices. The rest improves diagnostics and the calculation of the model price over history.
Historical summary
Observations
18
Average error
$206
Average price
$810
Average cost floor
$585
Metric Value Comment
Average supply pressure54.5Supply pressure was not extreme
Average demand index91.4Demand was moderate
Average signed error-206.1On average, the model understates the price
Max absolute error272.1Shows the worst month for the fit
Actual price vs. model price
Cost floor vs. actual price
Supply pressure index
Demand ceiling index
Contract and basis
Routes are pre-populated based on your Company table. In custom mode, you can set the basis and all deductions manually.
Deduction matrix by basis
Item CIF → FCA CFR → FCA DAP → FCA FOB → FCA FCA → FCA
Marine freight − − 0 0 0
Insurance − 0 0 0 0
Destination port costs − − 0 / partial 0 0
Inland logistics to customer − if calculating back to the plant − if calculating back to the plant − 0 0
Export handling / terminal / docs − − − − 0
Plant to port / border haulage − − − − 0
Trader margin / other adjustments − − − − −
Route costs
Netback
$736.0/t
quoted: DAP; target: FCA; producer: Company
Total deductions
$44.0
the sum of all deductions
Margin vs Cost Floor
$149.5
netback − cost floor
Margin vs Model Price
$124.4
netback − model price

Calculation formula

Netback(FCA Company) = QuotedPrice(DAP) - TotalDeductions TotalDeductions = SeaFreight·0 + Insurance·0 + DestPort·0 + InlandDest·1 + ExportHandling·1 + PlantToPort·1 + ExportDuty + Other + TraderMargin InsuranceAmount = 780.0 × 0.00% × 0 = 0.0 ExportDutyAmount = 780.0 × 0.00% = 0.0

Deduction breakdown

Item Flag Amount, $/t Comment
Sea freightOFF0.0Not applicable for the selected basis pair
InsuranceOFF0.0Not included for this basis
Destination portOFF0.0Not deducted
Inland destinationON22.0Need to trace the price back from the market to the source
Export handlingON6.0Terminal / documentation / forwarding
Plant to port / borderON16.0Company → port/border leg
Export duty / taxesOFF0.0Not set
Other adjustmentsOFF0.0Other adjustments
Trader marginOFF0.0Commercial discount / intermediary margin
Detailed model description

1. Overall logic

The model treats the PVC price not as a direct function of a single feedstock factor but as the outcome of the interaction of four building blocks:

  • Cost floor — the lower bound set by ethylene, a chlorine proxy, and regional energy.
  • Supply pressure — supply pressure, primarily via China: exports, capacity utilization, and new capacity additions.
  • Demand ceiling — support from the demand side: construction and India's imports as an additional demand buffer.
  • Arbitrage — the inter-regional spread, accounting for freight and trade barriers.

2. Core formula

ModelPrice = w_cf·CF + w_sp·(CF + SupplyTerm) + w_dc·(CF + DemandTerm) + w_arb·(CF + ArbTerm)

Weights here w_cf, w_sp, w_dc, w_arb depend on the market regime. In oversupply, the model responds more strongly to supply pressure. In a tight regime, demand carries more weight.

3. Price floor formula

CF = 0.52·Ethylene + 0.12·CausticProxy + 0.18·Brent + 0.18·RegionalEnergy + 110

This is not a physically precise production costing but a structural proxy. Its role is to approximate the level of the price floor and check whether the market is systematically drifting below a reasonable cost boundary.

4. Supply pressure index

SupplyPressure = clamp(50 + 18·ExportScore + 15·UtilScore + 8·GrowthScore - 10·OutageRelief + 10·Policy, 0, 100)

Where:

  • ExportScore — normalized PVC exports from China.
  • UtilScore — a penalty for low capacity utilization. The lower the utilization relative to the baseline level, the higher the risk of oversupply.
  • GrowthScore — the effect of new capacity additions.
  • OutageRelief — easing of pressure due to capacity shutdowns.
  • Policy — a binary amplifier if China's policy supports export pressure.

5. Demand ceiling index

DemandCeiling = clamp(0.45·ChinaConstruction + 0.25·EUConstruction + 0.20·USConstruction + IndiaBoost, 70, 130)

The logic here is conservative: the model does not always make demand the primary driver but gives it more weight only in a tight regime.

6. Arbitrage

Arbitrage(region) = RegionalSpread - Freight + BarrierAdjustment

If Asia is materially cheaper on a freight-adjusted basis, the model flags the risk of import pressure on the local market. Where anti-dumping measures are in place, this effect is partly offset.

7. Market regimes used

Regime When to use Meaning
Oversupply Weak demand, high Chinese exports, low capacity utilization The price is driven more by oversupply than by cost base or demand
Balanced No clear skew Default regime for an initial estimate
Tight High utilization, supply disruptions, strong construction Demand and tight supply keep the price above the cost floor

8. What the charts mean

  • Actual price vs. model — shows how well the model's structure tracks the market at all.
  • Cost floor vs. actual price — tests the hypothesis that the price should not persistently sit below the lower bound.
  • Supply pressure — visualizes when the market entered a supply-pressure regime.
  • Demand ceiling — shows whether there were periods when demand could genuinely support the price.

9. Netback methodology by basis

Tab Netback / basis does not forecast the market price but answers a different question: what price, on a basis of FCA Company an external quote on a different basis implies. This is the layer of commercial deal viability, not the layer of market formation.

Netback(FCA\ Company) = QuotedPrice - \sum(Deductions\ by\ basis) - TraderMargin

The calculation approach depends on the quote's starting basis:

  • CIF → FCA: subtract sea freight, insurance, the destination port, inland logistics in the destination market, as well as export logistics and terminal costs from the price if they were included in the original quote.
  • CFR → FCA: the same, but without insurance, because it is not included in a CFR quote.
  • DAP → FCA: subtract the leg from the DAP point to the plant/origin border and the related upstream costs.
  • FOB → FCA: only export costs up to the port/terminal and documentation are subtracted.
  • FCA → FCA: there are no additional logistics deductions; only commercial adjustments remain.
TotalDeductions = SeaFreight·I_{sea} + Insurance·I_{ins} + DestPort·I_{port} + InlandDest·I_{inland} + ExportHandling·I_{exp} + PlantToPort·I_{precarriage} + ExportDuty + OtherAdjustments

Where I — binary indicators of whether an element applies, depending on the pair quoted basis → target basis. This is exactly why the tab includes a basis matrix.

10. Linking netback to the model price

The tool now includes two independent but connected layers:

  • Model Price — a structural estimate of the PVC market price.
  • Netback (FCA Company) — back-calculation of an external quote to the price at Company's plant.

Comparison of netback with Cost Floor and Model Price lets you quickly distinguish three different situations: the market looks acceptable, the market formally exists but does not deliver a plant-level margin, or the quote simply does not clear Company's economics.

11. Model limitations

  • The coefficients are currently set as a structural hypothesis rather than a statistically calibrated estimate.
  • The model does not account for lags explicitly. In the real market, some effects arrive with a 1–3 month delay.
  • Chinese data and industry estimates can be noisy.
  • Netback is sensitive to the quality of the route rates. If the wrong basis or logistics leg is selected, the error can exceed that of the market model itself.
  • For a production version, it would be advisable to add CSV upload, rolling error, lagged variables, calibration against real historical data, and a dedicated library of Company routes by country/port/rail lane.