Hedge Fund

Methodology

• Defining the universe

Top 500 stocks from Emerging Market.

• Historical back-testing

Portfolios are checked for historical outperformance to ensure that only consistently outperforming strategies are selected.

• Research

The team does in-depth quantitative research using statistical and fundamental parameters used for constituent screening and constantly manages portfolio risk using max drawdown, skew kurtosis and other volatility parameter.

• Constituent Screening

The research team does individual stock picking after going through company reports and financials and decides on a host of qualitative and quantitative parameters to be considered while screening stocks for the respective strategy.

• Weighting

Following parameters are considered while assigning weights to stocks in this portfolio

• Market Capitalization.
• Management quality and vision.
• Theme exposure, revenues and earnings visibility.
• Rebalance

This portfolio has a real-time rebalance schedule. On a daily basis, the research team reviews this portfolio and realign the weights with the selected asset allocation strategy.

JMI Implementation

• JMI team collected the information of 100+ US-based biotech and pharma companies and analyzed each company on the basis of parameters like market cap, FDA approval, trial efficacy, and development stage.
• Screened 30 companies with a trial efficacy rate of more than 80% in clinical stage 2.
• JMI team also analyzed the past trend and other KPIs like revenue growth, profitability margin etc. of each company and screened 22 companies that meets the investment strategy.
• Back-tested the portfolio with different weights for portfolio optimization to maximize Sharpe ratio or Sortino ratio with less volatility.
• JMI team also tracked the portfolio for a month and identified that the portfolio is trading at a rolling average of 6-11% higher relative to Dow Jones US Pharma Index in the past month.

JMI Implementation

JMI was involved in conducting initial market research covering all these aspects. Key findings of the project are listed:-

• The market cap of all listed equities in the EM is US$ 3.2tn and is amongst the top-10 traded markets globally. Equity cash market trading volume is US$8.3bn/day and derivative trading volume is US$15bn/day.
• 20mn derivative trades and 15mn cash equities trades are executed on the exchange every day and more than 10000 instruments are available for trading.
• There are more than 3000 offshore FPI funds and 500 onshore funds trading into the EM using various structures, and a fund structure could be setup within 4-6 weeks with a set up cost of US$50,000.
• The EM has tax treaties with different jurisdictions and a favorable offshore structure could be set up. Long-term capital gain taxes are 10% and short-term capital gains taxes are 15%.
• Mean latency of 120.521 microseconds with 360mn order messages per day to the exchange. Co-location and Direct Market Access is available to perform high and mid-frequency trading.
• Average daily volatility of around 3% and annual volatility of 56% across instruments offered ample high and mid-frequency trading opportunities. 10-year index return of 100% with 30% annual volatility offered ample low-frequency opportunities.

JMI Implementation

• JMI team conducted deep research on Technology, Media and Telecom sector and identified the key factors including Government and Regulatory Changes, GDP growth, Inflation and Sectoral revenue growth etc.
• Tracked the market-driven factors like liquidity, momentum and fund flow for each stock and analyzed the co-variance in relation with mean return, industry growth etc.
• Conducted detailed fundamental research and analyzed the financial model of firms such as Intel, Apple, Sony, ASML, Live Nation etc. and figured out key metrics for each company.
• JMI team tracked & maintained the latest financials and key parameters for each company and updated the model accordingly
• Models were updated on a daily basis and any significant impact was highlighted to the client immediately.

JMI Implementation

• JMI team collected the macro-economic data like GDP growth, currency impact, crude oil price and sector trends etc. and analyzed its impact on portfolio companies.
• Conducted fundamental research on portfolio companies to get insights on KPIs and valuation ratios like PE ratios, PB ratios and, EV/EBITDA values.
• Analyzed the technical and pricing attributes like momentum and mean reversion of each stock.
• Analyzed the volatility of each stock based on historical trends.

JMI Implementation

• JMI team accessed a set of 2,000+ stocks and fixed-income instruments from the client.
• JMI developed a customized quantamental model that inherited the fundamental data of each stock, such as quarterly financials, institutional holding patterns, daily returns, etc.
• Leveraged ML model to combine core fundamental analysis with state-of-the-art quantitative techniques to predict the performance of a large number of stocks by comparing their relative performance.
• Investigated through three machine learning models using Feed-Forward Neural Networks (FNN), Random Forests (RF), and Adaptive Neural Fuzzy Inference Systems (ANFIS) for stock prediction based on fundamental analysis.
• The JMI model outperformed the standard ranking model used by the client earlier both in terms of accuracy and turnaround time as the JMI machine learning model was back tested against both long and short investment strategies for equities and fixed income assets over a long time.

JMI Implementation

• JMI team accessed the set of 5000+ stocks across three industry verticals.
• Extracted the data of each stock from DTCC which has a repository of data consisting of US Equities trade volumes in context of broader market activity on daily basis.
• Using JMI multilevel model, JMI measured volatilities, risk and returns trends by combining trends of multiple levels of trade concentration for each stock.
• Leveraged JMI model to compare the relative performance trends of industry and sector of the respective stocks.
• JMI deployed the model as a robotics process automation (RPA) to predict key metrics based on the daily feed of the data.
• JMI team filtered out the companies with ± 1 standard deviation in revenue growth relative to industry average.
• JMI team also identified that the 36 companies which are frequent in public news have high volatility and trading volume compared to industry average.

JMI Implementation

• Assessed client’s Time Series-based model and determined that biases and missing values were not handled optimally - disrupting prediction engines.
• Consolidated uncontrollable number of objective functions containing multiple lagging variables to much simpler objective functions - reducing model complexity and improving efficiency of portfolio returns using dynamic optimization techniques.
• Leveraged proprietary AI runtime to identify and eliminate survivorship, look-ahead, optimization and other biases and helped traders to transform the patterns.
• Using our trained and tested models, JMI helped clients to reduce overfitting because of fewer variables utilized - model performance stabilization.
• This resulted in reduction of false signals generated by earlier system due to missing values and data biases.

JMI Implementation

Assessed the client’s quant strategy using in-house advanced quant platform and recorded following parameters:

• Sharpe Ratio: 1.75
• Max Drawdown: 8.3%
• Success Ratio: 35%
• Annualized return: 17% over a 20-year high frequency data

Ran the Strategy on JMI’s proprietary quant platform, across multiple sections of in-sampled and out-sampled data and recorded strategy result dataset which showed varied performance across time frames.

On further statistical analysis across time periods, JMI identified that the back tested result dataset was influenced by high kurtosis (+9.5) and negative skew (-1.7), and this led to high deviation in actual and back tested performance of the long short strategy.

Optimized the long short strategy parameters by using multi-variate regression and statistical modelling.

Added new volatility-based indicator to the strategy which led to a significant improvement in overall strategy and actual results in line with the back tested results.

JMI Implementation

• Following the client’s primary ideas on the real-time tracking and optimize entry/exit of stocks, our team conducted a detailed research and study.
• Collected 20-years of daily, weekly and monthly exchange data for all the listed samples and systematically applied approx. 30,000 stock-specific corporate action adjustments to streamline the data.
• Performed regression analysis to identify key quant driven price and volume factors for real-time tracking of all 1000+ listed stocks.
• Using JMI’s automated AI/ML platform, JMI back tested different values of the factors and developed a detailed model to track and generate multiple real-time quant signals across the stock universe.
• Developed a connector and linked the model to real-time exchange data feed which would keep models updated in real-time, and generate sound/visual triggers to monitor the markets without any human intervention.

JMI Implementation

• The JMI team understood the total long-short exposure requirement from the client and mapped it to the liquidity of various index and stock futures and options. Based on ADTV (Average Daily Trading Value) data, the team proposed to develop an index Long-Short strategy.
• Collected 20 years of tick-by-tick index trading data from the exchanges and sampled the data into 1 hour, 2 hours, daily, and weekly time series.
• Identified key price, volume, and volatility-based quantitative factors using statistical methods, developed a hypothesis, and back-tested it on in-sampled data.
• Using JMI’s automated AI/ML platform, the JMI team ran millions of scenarios based on the values of the factors.
• JMI developed an optimized strategy as per the client’s needs, which showed consistent results across timeframes.
• Re-tested the strategy on out-sampled data and risk and return parameters matched with in-sampled data and validated that.

The Challenges

• A financial company approached JMI with a research platform challenge. The application brought together market data with their own research and opinions into one view to create and evaluate their strategies.
• Their legacy implementation ran on several servers as a monolithic application, with a hand managed and complex infrastructure.
• The overall performance was poor, which is unacceptable in a trading research environment where timeliness, speed and accuracy are paramount. In addition, the ongoing cost of this setup was hard to justify.

JMI Implementation

• Deploying resources: JMI offered a comprehensive coverage expansion support for the client through a dedicated team of equity research analysts to chase new investment opportunities.
• Rigorous research: JMI performed fundamental research for in-depth analysis to build a detailed financial model and write investment memos.
• Continuous support: JMI actively supported the portfolio manager to evaluate various new investment opportunities across multiple sectors.
• Attending meetings: JMI team participated in calls with company management, sell-side analysts and industry experts to gain deep insights into the company and industry
• New ideas: JMI team tracked non-core coverage to identify new investment opportunities.
• Deliverables: JMI team developed primers, tearsheets and models along with key assumptions and worked on preparing notes updates for earnings, corporate actions, and news releases.
• Portfolio tracking: JMI actively monitored companies under portfolio coverage and updated investment thesis with the impact event on portfolio entities and update existing models a significant event.

JMI Implementation

• Public Data: The initiation coverage reports included scanning the company websites, SEC filings, investor presentations, press releases and conference call transcripts, with an aim to gain deeper insights about background of companies, products and services, business strengths, etc.
• Financial Model: JMI prepared plug & play financial models for a period of ten years (including projections for five years) which enabled us to capture the effects of various industry cycles and seasonality of the company’s businesses.
• Strategy: JMI mapped the companies competitive positioning, growth strategies and forecast capital/operating expenditure.
• Other Aspects: JMI also prepared other aspects of the reports like industry analysis, impact on sustainability, risk assessment, financial overview and company outlook.
• Analysis: JMI prepared driver-based financial forecasts, industry-specific valuation, and scenario and sensitivity analysis.
• Investment Recommendation: JMI worked on the thesis for the investment opportunity.

JMI Implementation

• Market sizing: By combining proprietary research with client insights and using our subscribed databases, we estimated a TAM for the company.
• Competitor analysis and business model changes: JMI analyzed key players in the industry from the perspective of both a competitor and a potential buyer in a divestiture scenario and also identified a recent pivot to the business model which, while not explicitly a part of the client’s original thesis, could potentially generate similar long term returns if executed optimally.
• Ownership breakdown: JMI started by understanding the ownership structure and how that would impact any potential divestiture decisions.
• Segment wise analysis and valuation: We individually broke down each business segment, benchmarked them to competitors depending on the type of product they produced and estimated their spin-off potential. For each individual segment, we forecasted future growth, came up with comparable companies, and assigned estimated future multiples.
• Mile markers and sensitivity analysis: JMI worked on identifying mile markers like market highs, continued underperformance in certain segments and Industry M&A and also performed analysis which showed asymmetric upside if management executes either opportunity (divestitures or new business model) optimally.