Wind Speed Modelling for Derivatives Pricing

Butali, Enrico (2026) Wind Speed Modelling for Derivatives Pricing. [Laurea magistrale], Università di Bologna, Corso di Studio in Greening energy market and finance [LM-DM270], Documento full-text non disponibile
Il full-text non è disponibile per scelta dell'autore. (Contatta l'autore)

Abstract

This thesis develops a wind-speed modelling framework for the pricing and risk management of wind-linked derivatives, motivated by the volumetric exposure of German onshore wind producers. Daily wind speed is Box-Cox transformed and deseasonalised, then modelled through an AR(4) conditional mean, a GARCH(1,1) volatility layer and a continuous-time CAR(4) embedding with Normal-Inverse-Gaussian innovations; a Heston-type stochastic-volatility extension is derived theoretically. The pipeline is built on a low-cost proxy dataset (Bologna, 10 m, 2015-2018) and applied unchanged to ERA5 reanalysis data aligned with the hedged quantity (Nordfriesland, 100 m hub height, 2015-2024), linked to SMARD onshore generation and day-ahead prices. Under a zero market-price-of-risk benchmark, the fair strike of a synthetic wind variance swap admits a closed form: the product of the GARCH conditional variance and the seasonal variance level. All fitted models beat persistence in Diebold-Mariano tests but are equivalent at a one-day horizon, so they are compared on pricing and hedging criteria. The Value of Information of the data upgrade is large: the proxy understates the fair strike by 93.8%, and still by 88.0% after a wind-profile height correction, indicating that geographic mismatch rather than measurement height drives the error. More fundamentally, Nordfriesland hub-height wind explains 58.7% of the daily variation in German onshore generation, whereas the proxy explains essentially none. A Cumulative Accumulated Wind Speed (CAWS) instrument, evaluated on non-overlapping annual observations, pays out in the calm years to which the variance swap is indifferent, acting as a left-tail complement. The central conclusion is that data alignment is a precondition for, not a refinement of, wind-derivative pricing: a wind series should be shown to explain the cash flows it is meant to hedge before any instrument is priced on it.

Abstract
Tipologia del documento
Tesi di laurea (Laurea magistrale)
Autore della tesi
Butali, Enrico
Relatore della tesi
Scuola
Corso di studio
Indirizzo
CLIMATE&BUSINESS SCIENCE
Ordinamento Cds
DM270
Parole chiave
wind derivatives; wind speed modelling; variance swap; renewables production; CAR(4) process; GARCH; stochastic volatilty; value of information; energy risk management; CAWS; German electricity market
Data di discussione della Tesi
28 Ottobre 2026
URI

Altri metadati

Gestione del documento: Visualizza il documento

^