This dissertation shows how to apply a particular mathematical tool called wavelet in different financial problems where the final purpose is to analyze the accuracy, the efficiency and the robustness of the obtained results. Firstly, wavelets are introduced by describing the features needed to reach the final aim. Then, three applications are investigated and, in all the three cases, wavelets are used as a new basis for approximating a particular function. The first application regards a method for pricing European options through the discounted expected payoff pricing formula and wavelets are used to approximate the density function of the underlying process with a finite combination of Haar or B-splines wavelets basis functions whose coefficients are recovered from its characteristic function. The second application is related to the computation of four risk-measures under Vasicek one-factor model, in particular the considered risk-measures are Value at Risk (VaR), Expected Shortfall (ES), risk contributions to VaR (VaRC) and risk contributions to ES (ESC). In this case, Haar wavelets are used to approximate the cumulative distribution function (CDF) of the loss whose coefficients are computing by inverting its Laplace transform. Finally, the last application covers the valuation of the tranches of a synthetic collateralized debt obligation (CDO) when Vasicek one-factor model is used as framework. Then, wavelets are used to approximate the CDF of the loss with a finite combination of Haar or B-splines wavelets basis functions whose coefficients are recovered from its characteristic function. In all the applications, the accuracy is obtained by computing the relative error between the wavelet method and other pricing methods or Monte Carlo simulations; the efficiency is measured by comparing the computational costs and, finally, the robustness is tested by stretching different parameters in the first application and by considering a wide sized variety of portfolios or pools for the other two applications.
Questa tesi si propone di applicare in diversi ambiti finanziari un particolare strumento chiamato wavelet, valutando poi l’accuratezza, l’efficienza e la robustezza dei risultati ottenuti. Lo svolgimento descriverà inizialmente le caratteristiche delle wavelets utili al raggiungimento dello scopo finale, per poi analizzare tre diverse applicazioni che le sfruttano come nuova base per approssimare funzioni. Inizialmente le wavelets verranno impiegate in un metodo finalizzato a prezzare opzioni Europee attraverso la formula del valore atteso del payoff scontato. In questo primo utilizzo, la funzione densità del sottostante sarà approssimata da una combinazione finita di Haar o B-splines wavelets i cui coefficienti saranno ricavati dalla funzione caratteristica del processo che guida il sottostante. Successivamente, utilizzando il modello di Vasicek a un fattore, misure di rischio, tra cui Value at Risk (VaR) e Expected Shortfall (ES), saranno calcolate approssimando la funzione di ripartizione (CDF) delle perdite con le Haar wavelets, i cui coefficienti verranno determinati tramite inversione della trasformata di Laplace della CDF stessa. Infine, sfruttando nuovamente il modello di Vasicek a un fattore, le tranches di un CDO (collateralized debt obbligation) verranno valutate utilizzando Haar o B-splines wavelets per approssimare la funzione di ripartizione (CDF) delle perdite, ma, diversamente dalla precedente applicazione, i coefficienti si ricaveranno dalla trasformata di Fuorier della CDF. In tutte queste applicazioni, l’accuratezza sarà ottenuta calcolando l’errore relativo tra il metodo implementato con le wavelets e altri metodi di pricing o, in alternativa, simulazioni Monte Carlo; l’efficienza si misurerà comparando i costi computazionali dei metodi presi in esame e la robustezza sarà testata stressando diversi parametri nella prima applicazione e considerando una grande varietà di portafogli in termini di grandezza e tipologia nelle altre due.
Wavelets applications in finance
CARRARA, JESSICA
2016/2017
Abstract
This dissertation shows how to apply a particular mathematical tool called wavelet in different financial problems where the final purpose is to analyze the accuracy, the efficiency and the robustness of the obtained results. Firstly, wavelets are introduced by describing the features needed to reach the final aim. Then, three applications are investigated and, in all the three cases, wavelets are used as a new basis for approximating a particular function. The first application regards a method for pricing European options through the discounted expected payoff pricing formula and wavelets are used to approximate the density function of the underlying process with a finite combination of Haar or B-splines wavelets basis functions whose coefficients are recovered from its characteristic function. The second application is related to the computation of four risk-measures under Vasicek one-factor model, in particular the considered risk-measures are Value at Risk (VaR), Expected Shortfall (ES), risk contributions to VaR (VaRC) and risk contributions to ES (ESC). In this case, Haar wavelets are used to approximate the cumulative distribution function (CDF) of the loss whose coefficients are computing by inverting its Laplace transform. Finally, the last application covers the valuation of the tranches of a synthetic collateralized debt obligation (CDO) when Vasicek one-factor model is used as framework. Then, wavelets are used to approximate the CDF of the loss with a finite combination of Haar or B-splines wavelets basis functions whose coefficients are recovered from its characteristic function. In all the applications, the accuracy is obtained by computing the relative error between the wavelet method and other pricing methods or Monte Carlo simulations; the efficiency is measured by comparing the computational costs and, finally, the robustness is tested by stretching different parameters in the first application and by considering a wide sized variety of portfolios or pools for the other two applications.File | Dimensione | Formato | |
---|---|---|---|
Wavelets applications in finance.pdf
non accessibile
Descrizione: Thesis text
Dimensione
1.88 MB
Formato
Adobe PDF
|
1.88 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/134939