Research

Stress Testing Megacap Portfolios with Serenity for investors to assess the resilience of their portfolios

This paper undertook an empirical analysis of delta hedging strategies for Bitcoin options using Serenity’s comprehensive derivatives analytics and volatility surfaces, considering the unique characteristics of the digital asset market. By comparing two prominent delta hedging strategies – the classic Black-Scholes delta and Serenity’s smile-adjusted delta – the study shed light on the intricate dynamics of risk management in the digital assets space

In this paper, we provide an early review of how risk factor models within Serenity enables systematic trading of digital assets while effectively managing risk exposure and turnover.

This paper presents a comprehensive methodology for generating profit and loss (P&L) scenarios within Serenity risks tool, with a focus on options in the digital asset markets.

A review of factor models and Principal Component Analysis (PCA) for building risk models for digital assets. We examine the benefits of PCA and its feasibility as an alternative to factor models.

This articles reviews the methodology for constructing volatility surfaces for option markets using the Stochastic Volatility Inspired (SVI) parameterization and looks at how a well-behaved volatility surface is an essential component for risk analytics.

We review various approaches to risk neutral discounting for options on crypto currencies. In particular, we examine sources of crypto lending rates as well as benchmark indices constructed to represent a risk free rate of return for crypto markets. We compare these rates with the discount rates implied from future and option prices on Deribit. We conclude with a discussion of the approach to discounting used in Serenity.

A quantitative and statistical review to compare two taxonomy systems, Digital Asset Taxonomy System (DATS) and Digital Asset Classification Standard (DACS).

A review of risks faced by digital assets investors in 2022.

FTX’s collapse has been a black swan event that came with severe consequences for many investors but also much to learn from. Cloudwall did a review via a sectoral lens of using statistical tools to generate insights t the FTX crash

Luna’s crash has been a black swan event that came with severe consequences for many investors but also much to learn from. If there is anything we can take away from this, is to ask and answer the question: can we use statistical insights to gain early warnings about such events? Here, we use our risk factor model to decompose risk into factor specific and idiosyncratic risk, as well as study risk contribution amongst individual factors. This gives us insights into how different factors influence the volatility of LUNA, especially right before the crash. We observe a notable divergence in the volatility trend for LUNA at the start of the week of the crash, which is very different from the rest of the assets. More notably, we also observe a regime change for LUNA’s factor risk decomposition, wherein the Momentum factor driving LUNA’s volatility gets replaced by liquidity and volatility factors, serving as an early warning of the liquidity shock.

In this paper we’ll show how portfolio managers are usually evaluated through performance attribution. As an example, we’ll introduce the Brinson–Fachler model. Instead of evaluating a portfolio manager, we’ll use it to explore DACS, the token taxonomy introduced by CoinDesk. Our objective is to see if using sectoral diversification there’s value to be uncovered. The same methodology can be used to compare e.g., crypto VCs investing in liquid tokens and see if their out- or under-performance is due to asset selection, sector allocation, or the interaction of the two.

While factors have become the foundation of investing, it is increasingly difficult to observe these factors. With a nascent market like crypto, it becomes imperative that we identify and successfully extract targeted factors driving our asset returns. In this article, we come up with a Three Risk Factor Framework, which involves identification: based on economic justification, observable and resistant return patterns, longevity, robustness, and cost-effective implementation. This is followed by a Rationale framework by which asset returns are driven by systematic risk that reflect compensations for providing insurance, exploit market inefficiencies due to persistent investor behavior biases and accommodate structural market imbalances. Lastly, we have a factor extraction framework which includes idiosyncratic factors, macro factors and statistical factors.

Usually, risk factor literature assumes that these factors driving returns are observable (at least theoretically) and are usually obtained by a regression model. An alternative approach is to assume that factors are infact unobservable and generate these out of pure statistical methodology. One such technique is Principal Component Analysis (PCA), which can be used to generate orthogonal (or independent) factors that drive variability of returns. In this paper, we apply PCA to extract factors out of a top 1000 (by market cap) crypto asset universe, after data cleaning and processing, leaving a net 103 assets studied. We find that the first 10 factors explain about 60% of variability in returns, which is much lower than the usual explain ability with mature markets like equities, possibly suggesting additional factors in the crypto universe or hinting at a yet amateur market. We also see the effect of survival bias at play by observing a notable increase in total variability when we reduce the data coverage constraints. Lastly, we construct eigen portfolios and compare performance of these with an equally weighted benchmark.

Our last paper briefly studied the statistical properties of Bitcoin’s price process. In this paper, we take a step forward and make a thorough study of stylized facts across the crypto market. The paper begins with a statistical analysis of a top 10 token - market cap weighted index. Next, we use the DACS taxonomy to construct sectoral indices and compare statistical behaviour across the 5 sectoral indices. A range of techniques and metrics are utilized such as distribution fitting, Q-Q plots for distribution comparisons, skewness, kurtosis, GARCH fitting and autocorrelation plots to confirm or deny the common stylized facts.