The Journal of Trading

Volatility Forecasting

Haim A. Mozes — 2018-08-16

This paper provides a perspective on volatility forecasting. The basic idea is that a number of factors are leading to volatility having a lower baseline expected value than in prior years. These factors include lower earnings uncertainty, greater market efficiency, better market-marking, and the fact that volatility trading itself tends to reduce volatility.

Dark Pools, Fragmented Markets, and the Quality of Price Discovery: Commentary

Robert Schwartz — 2018-08-20

This commentary is on a paper published in 2010. Few would wish to roll the markets back to where they were eight years ago, but have the issues that were debated then been adequately resolved? Are today's markets acceptably efficient? Can we relax about market quality? My answer to each of these is "no." What I wrote in 2010, I stand by now. Along with revisiting my previous discussions on dark pools, fragmentation, price discovery and liquidity, this commentary presents my newer thoughts concerning the definition of the term "liquidity" and the existence of an illiquidity premium.

If Best Execution is a Process, What Does that Process Look Like? Commentary by Wayne Wagner, Mark Edwards, and Steven Glass, 2018

Wayne Wagner — 2018-08-20

Active investment management is in a fight for competitive survival. Excellent idea generation will succeed only if the process is implemented effectively. The markets are where "the rubber meets the road," and effective trading forms the foundation for securing the benefits of excellent research and strategy.

Reflections on Cluster Analysis for Evaluating Trading Strategies

Jeffrey Bacidore — 2018-08-22

Our paper on Cluster Analysis was inspired by our need to group client data by trading strategy, when the data we were provided did not contain any information on trading strategy whatsoever. We ended up relying on a well-known statistical technique, k-means, which surprisingly had not been used widely in trading applications. At the time, non-quant traders were still reluctant to use quantitative techniques, especially black box applications like k-means. Fortunately, a lot has changed since that time, as quants are now using much more sophisticated techniques, like deep learning. And even more important, non-quant traders and business leaders have become much more accepting of such techniques, making it easier for such advanced techniques to be incorporated into trading applications.

A Retrospective Look: Phantom Liquidity And High Frequency Quoting

Ben Van Vliet — 2018-09-04

In this paper we take a retrospective look at our paper "Phantom Liquidity and High Frequency Quoting" and discuss the context of the research in light of our broader inquiry into the nature of the high frequency trading industry. The data presented in this paper appears to show that limit order cancellations of high frequency traders are associated with price discovery and liquidity provision, rather than some manner of systematic taking-advantage-of other market participants. These firms are acting as rational, profit-seeking business, and we believe time has shown this view to be correct. In the years since publication, HFT has matured, and consolidated into fewer, lower-cost providers of efficiency and liquidity services, mush like we would expect in any other industry.

Retrospective: "Toward Greater Transparency and Efficiency in Trading Fixed Income ETF Portfolios

Ananth Madhavan — 2018-09-05

In our original JOT paper, we described a logical approach to developing and implementing the intraday intrinsic value estimate. The approach is "bottoms up" or bond-by-bond, based on adjustments to previous quotes or trade prices for subsequent movements in the individual bond's yield curve plus an adjustment for changes in the credit spread. Adding in accrued interest and the fund's cash, we can then derive a portfolio level estimate of the fund's value. In this retrospective piece, we provide () some new evidence about the applications of our approach; and (2) We further examine the possibility that the industry coalesce around improving iNAV to reach an industry standard calculation for ETF Intrinsic Value that adjusts for staleness, as proposed in our Journal of Trading article.

Trends in Volume Forecasting: Developments & Applications

Venkatesh Satish — 2018-09-05

Authors examine their 2014 publication, "Predicting Intraday Trading Volume and Volume Percentages" and discuss subsequent changes in trading that validated the models outlined in the paper and prompted updates. The original models accommodate the general shift to passive investing and the trend toward ETF investing. Analyzing imbalance information has become more important to institutional traders as relative participation in closing auctions has increased.

Authors discuss the evolution of analytical software platforms since the paper and outline expected trends in both volume forecasting and trading analytics. A major application of enhanced volume forecasts relates to the trend of buy-side clients performing scientific experiments to select algorithms and inform parameter selection. Specifically, volume profile error, a metric examined in the paper, provides context to compare broker algorithm performance and real-time volume forecasts can be used in algorithm routing decisions.

Machine Learning for Algorithmic Trading and Trade Schedule Optimization

Robert Kissell — 2018-09-20

In this paper we present a machine learning technique that can be used in conjunction with multi-period trade schedule optimization used in program trading. The technique is based on an artifiical neural network (ANN) model that determines a better starting solution for the non-linear optimization routine. This technique provides optimization time improvements from 30% faster for small baskets (n= 10 stocks), 50% faster calculation times for baskets of (n=100 stocks) and up to 70% faster calculation times for large baskets (n>300 stocks.) Unlike many of the industry approaches that use heuristics and numerical approximation, our machine learning approach solves for the exact problem and provices a dramatic improvement in calculation time.

Trader Alpha Frontier: A Look Back and Forward

Vlad Rashkovich — 2018-09-24

Since the introduction of Trader Alpha Frontier, this framework has been adopted by asset managers of all sizes, to monitor their trading performance. The next logical step is for Chief Investment Officers to incorporate Trader Alpha Frontier into their main view of porfolio returns. The author visualizes how CIOs can get a full insight in all alpha sources throughout the investment value chain including Analysts, Portfolio Managers, Traders and Brokers.

Beyond th Black Box Revisited: Algorithmic Trading and TCA Analysis using Excel

Robert Kissell — 2018-09-25

In this paper we revisit techniques from "Creating Dynamic Pre-Trade Models: Beyond the Black Box" (Kissell, 2011) which was awarded the Journal of Trading's Best Paper of the Year Award in 2011. We provide investors a pre-trade of pre-trade modeling technique that can be used to decipher broker and vendor models, and can be used to calibrate a customized investor specific market impact model. We also provide a suite of Excel TCA Add-In functions that can incorporate investor specific market impact parameters and allow investors to perform TCA analysis on their own desktops within Excel, and with the added level of security and comfort that their investment decision process will not be reverse engineered because they do not need to upload or transmit any of their proprietary information and valuable trade information to a third-party website or API for analysis. Techniques in this paper enable investors to create their own customized TCA analyses within Excel to assist with both trading decisions and portfolio analysis and optimization.

Space Unicorns and the Intermarket Trading System: Revisiting Myths

Jamie Selway — 2018-10-02

The author reviews the original article, "Five Myths about Listed Trading," published in 2012, and provides tree thoughts for consideration to today's readers.

A Market Structure that Fits the Needs of Portfolio Managers

Evan Schulman — 2018-10-02

Trading "these" securities for "those" (portfolio trades) can be expensive if done through our current continuous markets. This article compares a broker-implemented blind bid solution to this problem in a continuous market setting versus a combined value computerized call market that maximizes available liquidity to create balanced trades between such lists. The technology is known: combined value markets are in use today servicing markets in logistics contracts, emissions permits, spectrum licenses, and aerospace procurement. Should not financial concerns, such as custodial banks, be currently offering such services to their clients?