Paper: arXiv 2308.07763

Authors: Purushottam Parthasarathy, Avinash Bhardwaj, Manjesh K. Hanawal

Abstract

We revisit the online portfolio allocation problem and propose universal portfolios that use factor weighing to produce portfolios that out-perform uniform dirichlet allocation schemes. We show a few analytical results on the lower bounds of portfolio growth when the returns are known to follow a factor model. We also show analytically that factor weighted dirichlet sampled portfolios dominate the wealth generated by uniformly sampled dirichlet portfolios. We corroborate our analytical results with empirical studies on equity markets that are known to be driven by factors.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 4.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is dense with mathematical derivations, including factor models, growth rate bounds, and inequalities like AM-GM, which yields a high math complexity score. However, the empirical evidence is described generically (’empirical studies on equity markets’) without providing specific backtest results, datasets, or implementation details, leading to a lower empirical rigor score.

Research Flowchart

  flowchart TD
  A["Research Goal: Revisit Online Portfolio Allocation<br>Does Factor-Weighted Dirichlet Dominate Uniform Dirichlet?"]
  B["Methodology: Analytical Derivation<br>Factor Weighting & Growth Bounds"]
  C["Data: Historical Equity Market Returns<br>Known to be Factor Driven"]
  D["Computational Process: Backtesting<br>Simulate Universal Portfolios via Dirichlet Sampling"]
  E["Outcome 1: Analytical Proof<br>Factor-Weighted Strictly Dominates Uniform"]
  F["Outcome 2: Empirical Validation<br>Higher Portfolio Growth & Outperformance"]

  A --> B
  A --> C
  B --> D
  C --> D
  D --> E
  D --> F

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes using factor‑weighted Dirichlet sampling to create universal portfolios that outperform those using uniform Dirichlet sampling.

Source sentence
We propose factor weighing of the dirichlet sampling to construct universal portfolios that out-perform those using uniform dirichlet sampling.

Assumes 2

  • Asset returns follow a factor model. model
    Source sentence
    When the returns follow a factor model, we establish a lower bound on the average portfolio growth rate.
  • A constant re‑balanced portfolio is fully invested with weights summing to one over m assets and n trading periods. market
    Source sentence
    A portfolio b = (b1, b2, b3...bm), bi ≥0, Σbi = 1 is the allocation of wealth to assets. The symbol t stands for transpose. Xi is the ith relative price vector.

Data

  • Nasdaq 100 constituents (71 assets) · EOD HistoricalData market-prices commercial
    August 2007 to June 2023 · Subset of Nasdaq 100 index constituents
    Source sentence
    In this example we used a selection of 71 assets (listed cash equity) from the US stock market that are a subset of the Nasdaq 100 index constituents
  • FTSE 100 constituents (87 assets) · London Stock Exchange web page market-prices public
    June 2023 (data back‑sampled to 4,000 tr · Subset of FTSE 100 index constituents
    Source sentence
    Now we use a selection of 87 cash equity assets from the stocks that are listed on the London Stock Exchange (LSE) and are part of the FTSE 100 as of June 22nd 2023.

Costs

Not modeled A trading or execution result is reported without costs.

Finds 3

  • Factor‑weighted Dirichlet sampled portfolios generate higher wealth than uniformly sampled Dirichlet portfolios. theoretical
    Source sentence
    We analytically establish that the wealth generated by the factor weighted dirichlet sampled portfolios dominate the wealth generated by the uniformly sampled Dirichlet portfolios.
  • In back‑tests, momentum and size factor Dirichlet portfolios outperform the naive Dirichlet portfolio in growth rate. performance
    Source sentence
    The back-test (Chart Figure 1A) shows out-performance both for the momentum and size dirichlet portfolios in terms of the growth rate of the portfolio.
  • All factor‑based Dirichlet portfolios achieve higher Sharpe ratios than Cover’s benchmark, except the momentum factor. performance
    Source sentence
    The cumulative moving average sharpe ratio chart (CMAS - Figure 1B) shows that all factor return streams dominate Cover’s portfolio on a sharpe ratio basis, except the momentum factor.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 23 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive (1 by independent authors)

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.