Jump risk premia in the presence of clustered jumps
Francis Liu Natalie Packham Artur Sepp · 2025
Read Jump risk premia in the presence of clustered jumps by Francis Liu Natalie Packham Artur Sepp (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-21297.txt.
Introduction
Francis Liu, Natalie Packham, and Artur Sepp build an option pricing model where prices can jump in clusters, with upward jumps and downward jumps able to excite each other over time. Clustered jumps create lopsided, changing return patterns that simpler pricing models tend to miss.
The model uses a two-sided Hawkes process for the clustering and adds two parameters for how the market prices the risk of upward and downward jumps. Those risk prices can be read off option prices, letting the model match observed patterns in how option prices tilt. Fitted to Bitcoin options, the inferred risk prices help predict both the cost of carrying Bitcoin futures and the returns of hedged option trades.
Catalog record
| CREATOR | Francis Liu Natalie Packham Artur Sepp |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 113 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-21297
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Jump risk premia in the presence of clustered jumps Francis Liu Natalie Packham Artur Sepp Abstract This paper presents an option pricing model that incorporates clustered jumps using a bivariate Hawkes process. The process captures both self- and cross-excitation of positive and negative jumps, enabling the model to generate return dynamics with asymmetric, time-varying skewness and to produce positive or negative implied volatility skews. This feature is especially relevant for assets such as cryptocurrencies, so-called “meme” stocks, G-7 currencies, and certain commodities, where implied volatility skews may change sign depending on prevailing sentiment. We introduce two additional parameters, namely the positive and negative jump premia, to model the market risk preferences for positive and negative jumps, inferred from options data. This enables the model to flexibly match observed skew dynamics. Using Bitcoin (BTC) options, we empirically demonstrate how inferred jump risk premia exhibit predictive power for both the cost of carry in BTC futures and the performance of delta-hedged option strategies. JEL classification: D81, G13 Keywords: Volatility risk premium, clustered jumps, Hawkes process, cryptocurrencies 1 Introduction Jumps are acknowledged as an important risk factor in asset pricing, derivatives pricing, and risk management. Although there is a large body of literature investigating jumps in financial markets for traditional assets, the emergence of option markets for cryptocurrencies and retail-driven trading activity in options on “meme” stocks provide a new environment for this field of study. In this paper, we focus on the dynamics and options market for cryptocurrencies and, specifically, for Bitcoin (BTC). Empirically, \mathbb{P} -dynamics of BTC exhibit both large positive and negative jumps, while option prices under \mathbb{Q} -dynamics exhibit changes in implied volatility skew from negative (when demand for protective put options increases) to positive (when demand for upside leverage rises). Figure 1 illustrates the dynamics of Bitcoin (BTC) daily returns, at-the-money (ATM) implied volatilities and call and put skews from April 2019 to May 2024. First, we observe extreme returns ranging from -30\% to 20\% , with significant clustering. The implied at-the-money volatility for options with maturity of one month and Black-Scholes delta of 0.5 ranges from 40\% to 140\% . The call skew is computed as the difference between implied volatility of a call option with 0.25 Black-Scholes delta and implied volatility of a call with 0.5 delta. The put skew is computed as the difference between implied volatility of a put option with -0.25 Black-Scholes delta and the implied volatility of a put with -0.5 delta. Positive call and put skews indicate the higher demand for calls and puts, respectively. We observe that on average both call and put skews are positive. In addition, call skews exceed put skews in most periods, indicating a higher demand for calls than for puts. This behaviour contradicts traditional patterns in options on stocks and stock indices where the call skew is negative, while the put skew is consistently positive. Figure 1 : Subplot (A) shows Bitcoin daily returns from April 2019 to May 2024 with extreme volatility and pronounced clustering during market stress periods. Subplot shows ATM Volatility daily with average of 68.3\% . Subplot (C) shows BTC 25-Delta Call/Put Skews with corresponding averages of 3.0\% and 1.7\% and with blue/pink regions corresponding to periods with relatively higher call/put skews respectively. To account for empirical features of crypto options markets, we develop a derivatives pricing model that incorporates two Hawkes processes, one for positive and one for negative jumps. These processes capture the clustering of jumps, driven by self- and cross-excitation, effectively capturing the structure of jump arrival times. The model can be calibrated to the joint dynamics of BTC prices and of BTC implied volatility surfaces. In particular, we explore the resulting positive and negative jump risk premia, defined as the difference in expected jump magnitudes between the statistical measure and the risk-neutral measures. Importantly, our model accommodates changing market preferences for skewness risk, when investors shift their demands between put options to hedge the downside risk and call options to capture positive upside. The preference for call options with upside potential is consistent with evidence from the equity market literature indicating some investors’ preferences for assets offering lottery-like payoffs, see, e.g. Bali et al. 2011 , Thaler and Ziemba 1988 , Kumar 2009 . In cryptocurrency option markets, such preferences for capturing the upside manifest in inflated prices of out-of-the-money call options during bullish periods, which produces positive risk premia of positive jumps. Furthermore, the positive implied volatility skew is also present in the market data of so-called “meme” stocks. Options activity in these stocks is characterised by strong demand and high volumes for call options and by a positive spot-volatility correlation. Gupta and Pascale 2025 introduce the “Equity Euphoria Indicator”, which measures the percentage of stocks with a positive implied skew ranging from 0\% to 100\% . The most recent values, 10% as of January 2025 and 11.9% as of September 2025, are close to the levels observed during the Dot.com bubble in late 1990 and the post-Covid rally in 2021. We notice that implied skews on “meme” stocks change sign to negative during severe bear market periods, so that structurally we cannot model the dynamics of these stocks using a stochastic volatility model with a constant correlation parameter. Adding jumps arriving at a constant intensity rate, as for the Bates model ( Bates 1996 ), would not improve the structural consistency of the model dynamics, unless we specify stochastic intensity rates for positive and negative jumps. Thus, our model represents a simple construction to model the dynamics of stocks with varying signs of implied skew. It is straightforward to incorporate stochastic volatility in our framework; however, we skip this feature to reduce the number of model parameters. Our work contributes to the existing literature in several areas. First, the model of the price process allows one to identify and understand the factors that drive variations in jump risk premia, in particular, how jump intensities adapt in response to the occurrence of jumps. Extending the work of Hainaut 2016 ; Hainaut 2017 on univariate Hawkes process, we assume that the dynamics of these jump intensities are governed by a bivariate Hawkes process. Jump intensities increase with jump arrivals and decrease with time. As such, the model accounts for changes in asset prices and option prices stemming not only from movements in the underlying price, but also from variations in jump intensities. Second, alongside specifying the pricing model, we derive the measure change between the statistical and risk-neutral measures. This allows us to derive the jump risk premia associated with positive and negative jumps as the difference between expected jump magnitudes under the risk-neutral and statistical measures. A jump risk premium therefore quantifies the difference between market expectations (from options prices) and market realisations (from the price history) and can be interpreted as a measure of market sentiment. As option prices (under the risk-neutral measure) reflect expectations about future jumps and their intensities, in our setting, investors pay a risk premium for call options due to positive jumps if they are optimistic (positive skewness of option prices). Likewise, investors pay a risk premium for put options due to negative jumps if they are pessimistic (negative skewness of