Paper: arXiv 2306.08105
Authors: Vadim Zlotnikov, Jiayu Liu, Igor Halperin, Fei He, Lisa Huang
Abstract
Crowding is widely regarded as one of the most important risk factors in designing portfolio strategies. In this paper, we analyze stock crowding using network analysis of fund holdings, which is used to compute crowding scores for stocks. These scores are used to construct costless long-short portfolios, computed in a distribution-free (model-free) way and without using any numerical optimization, with desirable properties of hedge portfolios. More specifically, these long-short portfolios provide protection for both small and large market price fluctuations, due to their negative correlation with the market and positive convexity as a function of market returns. By adding our long-short portfolio to a baseline portfolio such as a traditional 60/40 portfolio, our method provides an alternative way to hedge portfolio risk including tail risk, which does not require costly option-based strategies or complex numerical optimization. The total cost of such hedging amounts to the total cost of rebalancing the hedge portfolio.
Complexity vs Empirical Score
- Math Complexity: 4.0/10
- Empirical Rigor: 6.0/10 (capped from 8.0: transaction costs not modelled; see penalty flags)
- Quadrant: Street Traders — empirical and applied, lower math complexity (effective empirical rigor capped from 8.0)
Why this score: The paper focuses on practical network analysis and portfolio construction with empirical backtesting on real data, but the mathematics (primarily graph centrality measures and regression) is relatively accessible compared to advanced stochastic calculus or machine learning papers.
Research Flowchart
flowchart TD
A["Research Goal: Model-Free Hedge for Market Risk<br>Using Crowding Networks"] --> B{"Methodology"}
B --> C["Data: Fund Holdings<br>Stock Prices"]
C --> D["Network Construction"]
D --> E["Crowding Score<br>Computation"]
E --> F{"Portfolio Construction"}
F --> G["Long-Short Hedge<br>Portfolio"]
G --> H["Outcomes: Hedge added to 60/40<br>Model-free, No Options, Low Cost"]
subgraph Computational Process
direction LR
D --> E --> F --> G
end
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 graph eigenvector centrality to compute crowding scores for stocks and construct a model‑free, costless long‑short hedge portfolio without numerical optimization.
Source sentence
These scores are used to construct costless long-short portfolios, computed in a distribution-free (model-free) way and without using any numerical optimization, with desirable properties of hedge portfolios.
Assumes 5
- Fund holdings data are only available about 45 days after the quarter end. data
Source sentence
Considering that fund holdings data is only available about 45 days after the quarter end, we construct the portfolios 2 months after the quarter end.
- Portfolios are constructed 2 months after quarter end and evaluated over the next 1, 3, 6, and 12 months. execution
Source sentence
Considering that fund holdings data is only available about 45 days after the quarter end, we construct the portfolios 2 months after the quarter end. We observe the returns for the next 1, 3, 6, and 12 months.
- Factor exposures are limited to between -0.02 and 0.02 for beta, growth, momentum, volatility and size. model
Source sentence
we use a proprietary optimizer to limit factors exposure between -0.02 and 0.02 for the following Tive Barra factors: beta, growth, momentum, volatility and size.
- The long‑short portfolio construction relies only on linear algebra and does not use any optimization. model
Source sentence
construction of our long‑short portfolio does not rely on any optimization and uses only linear algebra.
Show 1 more
- The total cost of hedging equals the total cost of rebalancing the hedge portfolio. execution
Source sentence
The total cost of such hedging amounts to the total cost of rebalancing the hedge portfolio.
Data
- 13F fund holdings reports alternative public
Source sentence
…d hedge funds in 13F reports. The crowding signals were further used to construct a diversiTied portfolio that has desired characteristics during periods of market downturns.
- Russell 1000 market returns market-prices public
Source sentence
correlation with the market returns (Russell 1000)
Costs
Unclear - sources disagree
Source sentence (the authors' words)
The total cost of such hedging amounts to the total cost of rebalancing the hedge portfolio.
Finds 3
- Stocks with higher eigenvector centrality (more crowded) underperform less crowded stocks on average. performance
Source sentence
On average the most crowded quintile underperforms the less crowded quintiles with a s…
- The long‑short crowding portfolio provides protection for both small and large market price fluctuations, showing negative correlation with the market and positive convexity. performance
Source sentence
More specifically, these long-short portfolios provide protection for both small and large market price fluctuations, due to their negative correlation with the market and positive convexity as a function of market returns.
- Adding the crowding‑based long‑short hedge to a baseline 60/40 portfolio offers an alternative way to hedge tail risk without costly option‑based strategies. performance
Source sentence
By adding our long-short portfolio to a baseline portfolio such as a traditional 60/40 portfolio, our method provides an alternative way to hedge portfolio risk including tail risk, which does not require costly option-based strategies or complex numerical optimization.
Checked by others
Supported by
- Not All Factors Crowd Equally: Modeling, Measuring, and Trading on Alpha Decay
Compare the findings
This paper: Adding the crowding‑based long‑short hedge to a baseline 60/40 portfolio offers an alternative way to hedge tail risk without costly option‑based strategies.
That paper: Average returns are efficiently priced; crowding predicts tail risk rather than mean returns.
2 papers in the archive
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.