The aim of this thesis is to model the volatility of some cryptocurrencies after a hacking incident. Starting from Grobys’ paper When the blockchain does not block: on hackings and uncertainty in the cryptocurrency market I replicated the results and went further trying to find a more suitable model that could fit better and minimize some information criteria. Moreover, I validated the results enlarging the dataset and adding new hacking incidents. In particular, the paper investigates the volatility response of Bitcoin and Ethereum when an exchange is attacked and portfolios containing Bitcoins are stolen. Because of the high correlation between the protocols, it studies the direct response on Bitcoin and the co-movement on Ethereum. Bitcoin shows an increase in volatility due to a contemporaneous effect at the day of the attack and another increment five days after. Ethereum, instead, has a variation in volatility only after a temporal lag of five days. Through an econometric analysis I replicated Grobys’ models but I found some critical differences; so I looked for some more effective models that could fit better. First of all, I highlighted that both cryptocurrencies show an asymmetric response to positive shocks: this result was not present in paper’s analysis, but expected in literature. In addition, I showed that Bitcoin has a reduction of the volatility two days after the stealing while Ethereum has a decrease one week after. Then, changing the model and the distribution of the log-returns, the error of the fitting and the information criteria have been reduced. Finally, I validated the results enlarging the dataset until September 2022 in order to demonstrate the better quality of the models.
Lo scopo della tesi è quello di modellizzare la volatilità di alcune criptovalute dopo un attacco hacker. Partendo dall’articolo di Grobys When the blockchain does not block: on hackings and uncertainty in the cryptocurrency market ho replicato i suoi risultati e cercato di trovare un modello più adatto che rappresentasse meglio i dati e riducesse alcuni criteri di scelta. Inoltre, ho validato i risultati allargando il dataset e aggiungendo gli attacchi hacker corrispondenti. In particolare, l’articolo, investiga la risposta della volatilità di Bitcoin ed Ethereum quando una piattaforma exchange viene attaccata e vengono rubati portafogli contenenti Bitcoin. Quindi l’autore studia la risposta diretta di Bitcoin e il co-movimento di Ethereum dovuto alla elevata correlazione tra le due valute. Bitcoin mostra un incremento nella volatilità dovuto ad un effetto immediato il giorno dell’attacco ed uno cinque giorni dopo. Ethereum, invece, ha una variazione solo dopo un arco temporale di cinque giorni. Attraverso un’analisi econometrica ho replicato i modelli presentati nell’articolo. Per rappresentare meglio i dati ed evidenziare alcune sostanziali differenze ho trovato altri modelli più efficaci. Innanzitutto, ho evidenziato come, en- trambe le monete presentassero una risposta asimmetrica agli shock. Questo risultato non è presente nell’articolo ma già dimostrato in letteratura. Inoltre, ho mostrato come Bitcoin subisse una diminuzione di volatilità due giorni dopo il furto mentre Ethereum una diminuzione una settimana dopo. Poi ho cambiato il modello e la distribuzione dei log-ritorni riducendo l’errore ed i criteri di scelta. Infine, ho validato quanto ottenuto ingrandendo il dataset fino a Settembre 2022 ottenendo risultati simili e dimostrando il miglioramento.
Modeling crypto's volatility after a hacking attack
FERRI, RICCARDO
2021/2022
Abstract
The aim of this thesis is to model the volatility of some cryptocurrencies after a hacking incident. Starting from Grobys’ paper When the blockchain does not block: on hackings and uncertainty in the cryptocurrency market I replicated the results and went further trying to find a more suitable model that could fit better and minimize some information criteria. Moreover, I validated the results enlarging the dataset and adding new hacking incidents. In particular, the paper investigates the volatility response of Bitcoin and Ethereum when an exchange is attacked and portfolios containing Bitcoins are stolen. Because of the high correlation between the protocols, it studies the direct response on Bitcoin and the co-movement on Ethereum. Bitcoin shows an increase in volatility due to a contemporaneous effect at the day of the attack and another increment five days after. Ethereum, instead, has a variation in volatility only after a temporal lag of five days. Through an econometric analysis I replicated Grobys’ models but I found some critical differences; so I looked for some more effective models that could fit better. First of all, I highlighted that both cryptocurrencies show an asymmetric response to positive shocks: this result was not present in paper’s analysis, but expected in literature. In addition, I showed that Bitcoin has a reduction of the volatility two days after the stealing while Ethereum has a decrease one week after. Then, changing the model and the distribution of the log-returns, the error of the fitting and the information criteria have been reduced. Finally, I validated the results enlarging the dataset until September 2022 in order to demonstrate the better quality of the models.| File | Dimensione | Formato | |
|---|---|---|---|
|
2022_12_Ferri_01.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
2.54 MB
Formato
Adobe PDF
|
2.54 MB | Adobe PDF | Visualizza/Apri |
|
2022_12_Ferri_02.pdf
non accessibile
Descrizione: Executive summary
Dimensione
576.46 kB
Formato
Adobe PDF
|
576.46 kB | 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/10589/196307