pith. machine review for the scientific record. sign in

arxiv: 2604.16773 · v1 · submitted 2026-04-18 · 💱 q-fin.PM

Recognition: unknown

Topological Risk Parity

Dnyanesh Kulkarni, El Mehdi Ainasse, Revant Nayar

Authors on Pith no claims yet

Pith reviewed 2026-05-10 07:27 UTC · model grok-4.3

classification 💱 q-fin.PM
keywords topological risk parityportfolio constructionlong-short strategiescorrelation distancemarket neutraltree topologyhierarchical risk parityrisk parity
0
0 comments X

The pith

Topological Risk Parity allocates long-short portfolio weights by propagating signed signals across a sparse rooted tree built from asset correlation distances.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces Topological Risk Parity as a method that extracts a tree structure from historical return correlations and then maps cross-sectional signals into weights using a propagation rule that respects the direction of each signal. Unlike binary hierarchical risk parity, the approach leaves partial weight at each parent node because correlations between parent and child are imperfect. This design targets market-neutral and hedged strategies where explicit control over market, sector, or factor exposures matters. The authors present both a fully data-driven minimum-spanning-tree version and a semi-supervised version that anchors the root at SPY and the next layer at sector ETFs. The resulting portfolios are claimed to remain more stable when macroeconomic shocks alter within-cluster correlations.

Core claim

TRP extracts a sparse rooted topology from a correlation-distance graph and applies a propagation law that converts signed signals into portfolio weights while retaining a fraction of exposure at each parent node to reflect imperfect parent-child correlations; two variants are given, one purely unsupervised via minimum spanning tree and one that fixes SPY as root and sector ETFs as the second layer, allowing economically motivated hierarchies to be imposed directly.

What carries the argument

A sparse rooted topology extracted from a correlation-distance graph, together with a propagation law that maps signed signals into portfolio weights without full transfer to children.

If this is right

  • TRP preserves the sign of each signal instead of forcing all positions positive.
  • Exposure at parent nodes is retained rather than fully delegated to leaves.
  • Sector or factor hierarchies can be imposed explicitly to enforce neutrality at those levels.
  • The method is intended for market-neutral equity stat-arb and long-short trend strategies.
  • Robustness to crises is improved because the hierarchy does not assume perfect within-cluster correlation.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Dynamic re-estimation of the tree at regime-change points could further reduce exposure drift.
  • Combining the topology with explicit factor loadings might tighten neutrality constraints beyond what the tree alone achieves.
  • The same propagation rule could be tested on non-equity assets such as futures or credit instruments where signed signals also arise.
  • Portfolio turnover may be lower than fully reoptimized methods because the tree changes only when the correlation graph changes materially.

Load-bearing premise

A tree structure learned from past correlations will continue to describe how risk should be allocated when markets experience sharp changes in correlations.

What would settle it

During a period of macroeconomic stress in which within-cluster correlations rise sharply, a TRP portfolio exhibits larger drawdowns or larger unintended market exposure than a comparable HRP or equal-weighted long-short portfolio.

read the original abstract

We develop \emph{Topological Risk Parity} (TRP), a tree-based portfolio construction approach intended for long/short, market neutral, factor-aware portfolios. The method is motivated by the dominance of passive/factor flows that naturally create a tree-like structure in markets. We introduce two implementation variants: (i) a rooted minimum-spanning-tree allocator, and (ii) a market/sector-anchored variant referred to here as \emph{Semi-Supervised TRP}, which imposes SPY as the root node and the 11 sector ETFs as the second layer. In both cases, the key object is a sparse rooted topology extracted from a correlation-distance graph, together with a propagation law that maps signed signals into portfolio weights. Relative to classical Hierarchical Risk Parity (HRP), TRP is non-binary and designed for signed cross-sectional signals and hedged long-short portfolios: it preserves signal direction while using return-dependence geometry to shape exposures. It accounts for the fact that there is imperfect correlation between parent and child nodes, and thus does not propagate weights entirely to the children. We can also impose economically motivated hierarchy that involves industries, sub-industries or factors, etc. This makes it much more robust to macroeconomic shocks and crises, where within-cluster correlations might spike. These features make TRP well suited for market-neutral, equity stat-arb or L/S trend-type strategies, where enforcing neutrality or limiting exposures at the market, sector or factor level is extremely important.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper proposes Topological Risk Parity (TRP), a tree-based portfolio construction method for long/short, market-neutral, factor-aware portfolios. It extracts a sparse rooted topology from a correlation-distance graph, either via minimum spanning tree or a semi-supervised variant anchored at SPY and the 11 sector ETFs, and applies a propagation law that maps signed signals to weights while respecting imperfect parent-child correlations. The central claim is that this non-binary structure improves on binary Hierarchical Risk Parity (HRP) by preserving signal direction, limiting higher-level exposures, and providing greater robustness to macroeconomic shocks where within-cluster correlations may spike.

Significance. If the robustness claims are empirically validated, TRP would represent a useful methodological extension of HRP to signed cross-sectional signals and hedged portfolios, allowing economically motivated hierarchies (industries, factors) while avoiding full weight propagation. The approach correctly identifies that imperfect correlations between parent and child nodes should be modeled explicitly rather than assumed binary, which addresses a practical limitation in existing tree-based allocators for stat-arb and L/S strategies.

major comments (2)
  1. [Abstract] Abstract: the claim that TRP 'makes it much more robust to macroeconomic shocks and crises, where within-cluster correlations might spike' is load-bearing for the paper's contribution, yet the manuscript supplies no back-test results, out-of-sample performance metrics, error analysis, or comparison against HRP (or other baselines) during stress periods. Without such evidence it is impossible to assess whether the imposed hierarchy actually stabilizes risk budgets when intra-cluster correlations rise.
  2. [Abstract] Abstract: the rooted tree is extracted from the same correlation-distance structure that would be used to evaluate portfolio performance, creating a circularity risk. The propagation law (which incorporates a scaling factor for imperfect correlation) therefore risks becoming a data-dependent fit rather than an out-of-sample prediction; no independent validation period, rolling-window test, or analytic bound on tree stability is provided.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments on our manuscript. The points raised regarding empirical validation and potential circularity are important for strengthening the paper's claims. We respond to each major comment below and outline the specific revisions we will make.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that TRP 'makes it much more robust to macroeconomic shocks and crises, where within-cluster correlations might spike' is load-bearing for the paper's contribution, yet the manuscript supplies no back-test results, out-of-sample performance metrics, error analysis, or comparison against HRP (or other baselines) during stress periods. Without such evidence it is impossible to assess whether the imposed hierarchy actually stabilizes risk budgets when intra-cluster correlations rise.

    Authors: We agree that the robustness claim is central to the paper's contribution and requires direct empirical support to be fully convincing. The current manuscript emphasizes the methodological innovation—the non-binary rooted tree and the propagation law that scales weights by imperfect parent-child correlations to avoid full propagation and cap higher-level exposures. However, we acknowledge that the abstract's assertion about stability during macroeconomic shocks lacks accompanying back-tests, out-of-sample metrics, or stress-period comparisons in the presented sections. In the revised version, we will add a new empirical subsection with rolling out-of-sample back-tests focused on known crisis intervals (such as 2008 and early 2020), including performance metrics, risk-budget stability analysis, and direct comparisons to HRP and other baselines. This will allow readers to evaluate whether the imposed hierarchy indeed limits exposure spikes when intra-cluster correlations rise. revision: yes

  2. Referee: [Abstract] Abstract: the rooted tree is extracted from the same correlation-distance structure that would be used to evaluate portfolio performance, creating a circularity risk. The propagation law (which incorporates a scaling factor for imperfect correlation) therefore risks becoming a data-dependent fit rather than an out-of-sample prediction; no independent validation period, rolling-window test, or analytic bound on tree stability is provided.

    Authors: The referee correctly identifies a potential circularity issue, since the topology is derived from the same correlation-distance matrix used in performance evaluation. The propagation law's scaling factor is explicitly designed to reflect observed imperfect correlations rather than assume binary dependence, which is a deliberate modeling choice to improve robustness. Nevertheless, we accept that this does not automatically guarantee out-of-sample validity. We will revise the manuscript to incorporate rolling-window validation experiments, in which the tree is constructed on an earlier training window and the resulting allocations are evaluated on a subsequent, non-overlapping test window. We will also add analysis of tree-structure stability across successive windows and, where possible, simple analytic bounds on how correlation perturbations propagate to the rooted topology. These additions will demonstrate that the method functions as a predictive allocator rather than an in-sample fit. revision: yes

Circularity Check

0 steps flagged

No circularity in the derivation chain

full rationale

The provided abstract and text define TRP as an algorithmic construction: extract a sparse rooted topology from a correlation-distance graph, then apply a propagation law that maps signed signals to weights while accounting for imperfect parent-child correlations. No equation, step, or claim is shown to reduce a 'prediction' or result to the inputs by construction. The robustness statement regarding macro shocks is a qualitative property asserted for the non-binary tree structure, not a quantitative output forced by fitting the same correlations used for evaluation. The derivation is self-contained as a proposed method rather than a tautology.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The approach rests on standard correlation metrics and graph algorithms but introduces a custom propagation law whose functional form is not specified; the tree structure itself is treated as given by market data.

free parameters (1)
  • propagation scaling factor for imperfect correlation
    The law that maps signals to weights must incorporate a factor less than one to reflect imperfect parent-child correlation, but no explicit value or fitting procedure is stated.
axioms (1)
  • domain assumption Passive and factor flows create a natural tree-like structure in asset return correlations.
    This premise directly motivates the choice of minimum-spanning-tree or sector-anchored topologies.

pith-pipeline@v0.9.0 · 5567 in / 1368 out tokens · 62600 ms · 2026-05-10T07:27:53.650272+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

5 extracted references

  1. [1]

    Markowitz

    H. Markowitz. Portfolio Selection.The Journal of Finance, 7(1):77–91, 1952

    1952

  2. [2]

    R. N. Mantegna. Hierarchical Structure in Financial Markets.The European Physical Journal B, 11:193–197, 1999

    1999

  3. [3]

    J. B. Kruskal. On the Shortest Spanning Subtree of a Graph and the Traveling Salesman Problem.Proceedings of the American Mathematical Society, 7(1):48–50, 1956

    1956

  4. [4]

    R. C. Prim. Shortest Connection Networks and Some Generalizations.Bell System Technical Journal, 36(6):1389–1401, 1957

    1957

  5. [5]

    Lopez de Prado

    M. Lopez de Prado. Building Diversified Portfolios that Outperform Out of Sample.The Journal of Portfolio Management, 42(4):59–69, 2016. Disclaimer This document is provided for informational and educational purposes only and does not constitute investment advice, a recommendation, or an offer to buy or sell any securities, financial instruments, or inves...

    2016