newsfilter.io
Interview

David Harding: The Future of Quantitative Investing

  • Pivoting from Trend-Following Orthodoxy
    • Harding rejects the "pioneer" label for trend-following, noting the strategy predates his work by 15 years, with roots in 1960s punch-card computing and 1972 launches by John Henry and Richard Campbell.
    • A primary career driver was battling "trend deniers" who adhered to Eugene Fama's Efficient Market Theory, which asserted zero market autocorrelation.
    • Harding notes a shift in public sentiment: after decades of denying trend existence, the narrative has swung to assuming the strategy is "obvious" and easily replicable.
  • Deepening Historical Data Analysis
    • The firm's research methodology is driven by a compulsion to extend back-testing horizons to better estimate statistical properties and validate strategy robustness against drawdowns.
    • Historical tracking has expanded from early 1980s 10-year simulations to examining data sets extending as far back as 1919, mirroring industry benchmarks like the Barclay Equity and Gilt Study.
    • This approach reinforces competence by demonstrating that current market volatility is not unprecedented compared to historical episodes (e.g., events eight years prior being more severe).
  • Stance on AI and Machine Learning
    • Harding defines current AI initiatives strictly as deep learning and multi-level neural networks.
    • He remains skeptical due to the failure of the 1990s neural network wave, including major bank initiatives like a $25 million Citibank investment.
    • While acknowledging new breakthroughs since 2014-2015 led by figures like Geoffrey Hinton, he estimates the current public debate is "nine parts hype to one part substance."
    • The firm views its existing capability of translating inputs into intelligent investment outputs as a technology already in practice for 30 years.
  • Organizational and Market Risk Management
    • Harding identifies four conflicting constituencies that must be balanced: customers, staff, regulators, and shareholders, noting that their interests rarely align perfectly.
    • The firm maintains very low leverage levels, adhering to Warren Buffett's principle of prudence when other market participants act recklessly.
    • Investment philosophy prioritizes preparation for "unexpected" catastrophic events over probabilistic normal distribution models.
    • Specific risk targets include "bulletproofing" the firm against overnight market drops of 35%, similar to the crash events of 1929 and 1987, while acknowledging the impossibility of insuring against existential threats like meteorites.
  • Personal Resilience and Firm Trajectory
    • Harding acknowledges the difficulty of founding Winton after exiting AHL, specifically citing a difficult initial month and the October 2002 drawdown.
    • He reframes these negative experiences as essential for learning, stating that subsequent drawdowns were less severe, thereby validating the firm's long-term strategy.
    • The firm operates on the conviction that mathematical and computational modeling has a distinct, non-controversial application to investment management, countering the "balkanization" between quantitative methods and efficient market theory.