Dynamic Asset Allocation

Roadmap for future work in long-only portfolios

Information presented here is for educational purposes only and does not intend to make an offer or solicitation for the sale or purchase of any specific securities, investments, or investment strategies. Investments involve risk and there are no guarantees of any kind. Be sure to first consult with a qualified financial adviser and/or tax professional before implementing any strategy discussed here in. Past performance is not indicative of future performance.

Where we are right now

We have managed to incorporate in our retail portfolios, all features that one expects from a roboadvisor, and more.

 

 

 

 

We have also been able to:

      outperform the benchmarks

      demonstrate the capability of managing sizeable assets

      find our niche as a data driven and AI based asset manager

Risk Management Frequent Rebalancing
Tax Optimisation Algorithmic execution

How we got here

  • Aim was to follow best practices and bring institutional investing to retail clients
  • Started with hedge fund like strategies like risk parity, trend following, value etc
  • Incorporated features like risk management, HFT like execution
     
  • Roll-out of retail offering
  • Initially targeted data driven individuals for early adopters
  • Recommended portfolios as a custom mix from a huge pool of strategies
     
  • Incorporated deep learning based strategies
  • Moved to a fixed set of featured portfolios
     
  • Systematic CIO for allocating to strategies in featured portfolios
  • Long-short and long only portfolios for institutional clients

2015

Early 2016

Late 2016

 2017

We need our long only offerings to appeal to institutional investors too

There are a few aspects about our portfolios that institutional investors might not be comfortable with:
 

  • High correlation with market (or high beta)
  • Capacity constraining aspects
    • Our strategy and execution alphas should not be limited to smaller investments.
    • Our execution should be scalable, with little chances of a mistake.
  • Black box-like approach (AI models, no anxiety management, insufficient demystification)
  • No easy customisation (turnover, product set, and risk management)

This is how we are planning to tackle these problems

High Correlation:

  • Incorporate all strategies we have worked on so far in our long only portfolio.
    • Adapt the L/S deep learning strategies to long only portfolios
    • Payroll
    • US investor index strategy
       
  • Explore alternate data sources, to generate alternative risk premia
     
  • Make our portfolios resilient to market regime shifts  by improving the systematic CIO so that it can:
    • Detect regime and allocate accordingly
    • Diversify among existing alphas

​Customization and scalability

  • It shouldn't take long to run studies and subsequently move to production.
    • Parallelize and support caching in the backtesting framework.
    • Bring the prototyping and production frameworks in line with each other.
  • Tool to make a custom portfolios.
     

​Moving away from black-box approach

  • Make it easier for investors to understand what it would really feel like investing in this portfolio
    • Scenario analysis

This is how we are planning to tackle these problems

Long Only Portfolios

By qplum

Long Only Portfolios

  • 1,272