With our strong equity expertise, from event-driven to long-term phenomenons, and our big data and IT skills, we design and implement a market-neutral, adaptive factor-based, and worldwide equity support: ERAAM Long Short Equity ©

Philosophy

Our views on factor investing

A simple idea

  • Some factors explain abnormal return
  • Many well-documented factors in the academic world, invested on for decades

In real life, many challenges on the way

  • Overfitting, execution costs, correlations
  • Differences in design, assembling and implementation lead to very different results

Our values

  • Scientific rigour and transparency
  • Cutting-edge and systematic implementation

True scientific approach and state-of-the-art engineering are key to achieve outperformance

Factor investing

US equities ranked by metrics, 15% extreme quantiles portfolios

US equities ranked by metrics, 15% extreme quantiles portfolios

Factor Long/Short portfolios show consistent returns over time

Factors universe

Factors universe

Broad, diversified and evolving factor universe

Investment process

Process overview

Process overview

It all starts with data

Heterogeneous data

  • Referential: North America + Europe + Asia Pacific universe, 6K+ alive stocks, 70K+ since 1990
  • Market daily: price + volume, market capitalization, corporate actions, mostly since 1990
  • Fundamentals: company accounting reports (yearly + quarterly, 100+ fields per stock per report), analysts estimates + recommendations, mostly since 1990
  • Intraday: Trade and Quote Level 1 + 1min bars, since 2008

15+ Providers

  • Varying quality and cost (High-End: Capital IQ, S&P Compustat, Zacks, etc. , Low-End: IB, QM, SH, AT, etc.)
  • Updated pool: providers technical upgrades follow-up, continuous selection

Challenges

  • Potential biases: survivorship, timestamping, quality
  • Quality control: systematic quality tests + protocols for human reviews (monitoring, alerting) + 5 years of live data recording. (We won awards for helping providers clean their dataset, eg. Compustat Quality Program Award)
  • Quantity control: 100GB of daily datas + 2TB of intraday + TaQ datas => proprietary architecture for optimized fetch + access + storage

Comprehensive and reliable database

Factors design

Our scientific factor selection process:

  1. Economic rationale, statistical significance and long-term robustness analysis
  2. Selection of most robust factors while limiting overfitting
  3. Embedded costs and correlation measurements in factor performance analysis

Factors design

Factors design is key to alpha generation

Factors selection

  • Single factor strategies show significant although moderate Sharpe Ratios
  • Returns are cyclical : factors can show long periods of underperformance

Factors selection

Need to assemble diversified factors to achieve long-term Sharpe ratio > 1 ...

But be careful with the factor selection bias !

Factors assembling

  • Traditional approaches: equal-weighted, risk-parity, Markovitz, etc.
  • Correlation clustering: how to avoid stacking up risks ?

Factors assembling

Strong correlations inside factor families and between some families as well

Portfolio construction, step 1: factors weighting

Step 1: Defining the weight of each factor, MLRC proprietary model

  • Measure of each factor predictive power
  • Factors relative weighting, dealing with correlations: multilinear regression
  • Rolling methodology: adaptive to factor trends / cycles

Portfolio construction, factors weighting

Portfolio construction, step 2: stocks selection

Step 2: Bottom-up (blended signals) portfolio construction

  • Returns forecasting with our MLRC model
  • One final forecast per stock
  • Execution costs embedded in the forecast

Portfolio construction, stocks selection

One signal per stock, embedding all factors predictive powers and costs

Portfolio construction, step 2: top-down vs bottom-up

Portfolio construction, top-down vs bottom-up

Blended signals allocation increases Inf. Ratio by 30% vs. mono portfolios assembling

Portfolio construction, step 3: risk management

  • Universe ranked by forecasted 3m return
  • Long / Short portfolio per region : Long the first decile, Short last the one
  • Beta hedge : hedge portfolio to offset market exposure

Portfolio construction, risk management

  • Each region (North America, Europe, Asia Pacific) is divided into 3 subportfolios
  • Each subportfolio holding period is 3 months
  • Trading executed over 3-5 days depending on market liquidity

Portfolio construction, risk management rebalancing

Reduces market impact, operating risk and path dependency

Execution

  • Low yearly turnover: 400%
  • VWAP algorithmic execution, continuously controlling beta and market cap exposure
  • Real-time monitoring
  • Market impact post-trade analysis

Execution

Rigorous execution and monitoring

Our competitive edge

Our competitive edge

Portfolio

Statistical properties of returns

  • Daily returns with good statistical properties: no black swans, no extreme events
  • Stable volatility over time

Statistical properties of returns

Stable returns, limited drawdowns

Stocks universe

  • North America + Europe + Asia Pacific (ex-China) (developed markets)
  • Filter: Market Cap > 100 M€, Liquidity > 10 M€ daily
  • Blacklist: based on Ottawa and Oslo conventions

Stocks universe

Our investable universe covers 90% of the world’s market cap (ex-China)

Portfolio concentration

Portfolio concentration

Highly diversified portfolio, idiosyncratic risk reduced to its minimum

Country and sector exposures

  • Global gross exposure < 150%
  • Country net exposure < 5%, sector net exposure < 15%

Country and sector exposures

Ex-post portfolio factors exposure as of 07/2018

Ex-post portfolio factors exposure

Pure factor investing

Performance, costs and opportunity in the L/S systematic world

Performance, costs and opportunity

100% systematic process + state-of-the-art technology to maximize cost effectiveness and opportunity