Efstathios Panayi

Trends in Crypto-Currencies and Blockchain Technologies: A Monetary Theory and Regulation Perspective

The internet era has generated a requirement for low-cost, anonymous and rapidly verifiable transactions to be used for online barter, and fast settling money has emerged as a consequence. For the most part, electronic money (e-money) has fulfilled this role, but the last few years have seen two new types of money emerge — centralized virtual currencies, usually for the purpose of transacting in social and gaming economies, and cryptocurrencies, which aim to eliminate the need for financial intermediaries by offering direct peer-to-peer (P2P) online payments. We describe the historical context that led to the development of these currencies and some modern and recent trends in their uptake, in terms of both usage in the real economy and as investment products. As these currencies are purely digital constructs, with no government or local authority backing, we discuss them in the context of monetary theory, in order to determine how they may be valued under each. Finally, we provide an overview of the state of regulatory readiness in terms of dealing with transactions in these currencies in various regions of the world.

Stochastic Simulation Framework for the Limit Order Book Using Liquidity Motivated Agents

In this paper we develop a new form of agent-based model for limit order books based on heterogeneous trading agents, whose motivations are liquidity driven. These agents are abstractions of real market participants, expressed in a stochastic model framework. We develop an efficient way to perform statistical calibration of the model parameters on Level 2 limit order book data from Chi-X, based on a combination of indirect inference and multi-objective optimisation. We then demonstrate how such an agent-based modelling framework can be of use in testing exchange regulations, as well as informing brokerage decisions and other trading based scenarios.

Liquidity commonality does not imply liquidity resilience commonality: a functional characterisation for ultra-high frequency cross-sectional LOB data

We present a large-scale study of commonality in liquidity and resilience across assets in an ultra high-frequency (millisecond-timestamped) Limit Order Book (LOB) data-set from a pan-European electronic equity trading facility. We first show that extant work in quantifying liquidity commonality through the degree of explanatory power of the dominant modes of variation of liquidity (extracted through Principal Component Analysis) fails to account for heavy-tailed features in the data, thus producing potentially misleading results. We employ Independent Component Analysis, which both decorrelates the liquidity measures in the asset cross section, but also reduces higher order statistical dependencies. To measure commonality in liquidity resilience, we utilise a novel characterisation proposed by Panayi et al. [Market resilience, 2014] for the time required for return to a threshold liquidity level. This reflects a dimension of liquidity that is not captured by the majority of liquidity measures and has important ramifications for understanding supply and demand pressures for market makers in electronic exchanges, as well as regulators and HFTs. When the metric is mapped out across a range of thresholds, it produces the daily Liquidity Resilience Profile for a given asset. This daily summary of liquidity resilience behaviour from the vast LOB data-set is then amenable to a functional data representation. This enables the comparison of liquidity resilience in the asset cross section via functional linear sub-space decompositions and functional regression. The functional regression results presented here suggest that market factors for liquidity resilience (as extracted through functional principal components analysis) can explain between 10 and 40% of the variation in liquidity resilience at low liquidity thresholds, but are less explanatory at more extreme levels, where individual asset factors take effect.

Agent-Based Modelling of Stock Markets Using Existing Order Book Data

We propose a new method for creating alternative scenarios for the evolution of a financial time series over short time periods. Using real order book data from the Chi-X exchange, along with a number of agents to interact with that data, we create a semi-synthetic time series of stock prices. We investigate the impact of using both simple, limited intelligence traders, along with a more realistic set of traders. We also test two different hypotheses about how real participants in the market would modify their orders in the alternative scenario created by the model. We run our experiments on 3 different stocks, evaluating a number of financial metrics for intra- and inter-day variability. Our results using realistic traders and relative pricing of real orders were found to outperform other approaches.

Survival models for the duration of bid-ask spread deviations

Many commonly used liquidity measures are based on snapshots of the state of the limit order book (LOB) and can thus only provide information about instantaneous liquidity, and not regarding the local liquidity regime. However, trading in the LOB is characterised by many intra-day liquidity shocks, where the LOB generally recovers after a short period of time. In this paper, we capture this dynamic aspect of liquidity using a survival regression framework, where the variable of interest is the duration of the deviations of the spread from a pre-specified level. We explore a large number of model structures using a branch-and-bound subset selection algorithm and illustrate the explanatory performance of our model.

Statistical Modelling for Precision Agriculture: A Case Study in Optimal Environmental Schedules for Agaricus Bisporus Production via Variable Domain Functional Regression

Quantifying the effects of environmental factors over the duration of the growing process on Agaricus Bisporus (button mushroom) yields has been difficult, as common functional data analysis approaches require fixed length functional data. The data available from commercial growers, however, is of variable duration, due to commercial considerations. We employ a recently proposed regression technique termed Variable-Domain Functional Regression in order to be able to accommodate these irregular-length datasets. In this way, we are able to quantify the contribution of covariates such as temperature, humidity and water spraying volumes across the growing process, and for different lengths of growing processes. Our results indicate that optimal oxygen and temperature levels vary across the growing cycle and we propose environmental schedules for these covariates to optimise overall yields.

Liquidity Commonality Does Not Imply Liquidity Resilience Commonality: A Functional Characterisation for Ultra-High Frequency Cross-Sectional LOB Data

We present a large-scale study of commonality in liquidity and resilience across assets in an ultra high-frequency (millisecond-time stamped) Limit Order Book (LOB) dataset from a pan-European electronic equity trading facility. We �?rst show that extant work in quantifying liquidity commonality through the degree of explanatory power of the dominant modes of variation of liquidity (extracted through Principal Component Analysis) fails to account for heavy tailed features in the data, thus producing potentially misleading results. We employ Independent Component Analysis, which both decorrelates the liquidity measures in the asset cross-section, but also reduces higher-order statistical dependencies. To measure commonality in liquidity resilience, we utilise a novel characterisation proposed by [PPDZ14] for the time required for return to a threshold liquidity level. This reflects a dimension of liquidity that is not captured by the majority of liquidity measures and has important rami�?cations for understanding supply and demand pressures for market makers in electronic exchanges, as well as regulators and HFTs. When the metric is mapped out across a range of thresholds, it produces the daily Liquidity Resilience Pro�?le (LRP) for a given asset. This daily summary of liquidity resilience behaviour from the vast LOB dataset is then amenable to a functional data representation.This enables the comparison of liquidity resilience in the asset crosssection via functional linear sub-space decompositions and functional regression. The functional regression results presented here suggest that market factors for liquidity resilience (as extracted through functional principal components analysis) can explain between 10 and 40% of the variation in liquidity resilience at low liquidity thresholds, but are less explanatory at more extreme levels, where individual asset factors take effect.