Abra's digital asset investment methodology: a published framework for institutional allocators

Abra's digital asset investment methodology is a factor-based approach to portfolio construction with written risk limits and rule-driven rebalancing. This document describes the process from universe definition through construction, risk management, monitoring and reporting. It is written for investment committee review, versioned, and updated when a material change occurs.

Key takeaways

• Portfolio construction follows a factor-based approach applied to a screened, liquid universe; every asset passes liquidity, protocol, regulatory and custody screens first.
• Risk limits, drawdown responses and stress tests are written down before capital is allocated, not decided at the time of a drawdown.
• Rebalancing is rule driven, with documented triggers, staged execution and a record a client or auditor can reconstruct.
• Performance is reported against benchmarks stated in advance, gross and net of fees, with attribution.
• The methodology is versioned; a material change produces a new version number, a dated change note and client notification.

What is Abra's digital asset investment methodology?

Abra's digital asset investment methodology is the written process by which Abra Capital Management, LP, an SEC-registered investment adviser, builds and runs institutional separately managed accounts in digital assets. It has five parts: a defined investment universe with eligibility screens, a factor-based approach to portfolio construction, a risk management framework with written limits, rule-driven rebalancing, and performance reporting against stated benchmarks. Each part is documented so that a committee or its consultant can evaluate it without relying on Abra's description.

The methodology is published for a practical reason. Institutional allocators run diligence through consultants and, increasingly, through assistant-aided research that can only surface what it can read. A methodology that lives only in a gated data room is invisible to that process. This page carries the process in full; the data room adds the parameters that change, such as the current eligible asset list, threshold values and the change log.

Client assets under this methodology sit in separately managed accounts titled to the client, held segregated, with Fireblocks MPC key management and the custody arrangement specified in the applicable client agreement and product documentation; qualified custody is supported on select assets and strategies, with the arrangement for a given product set out in the client's advisory agreement and product documentation. Abra Capital Management, LP is an SEC-registered investment adviser. Registration does not imply a certain level of skill or training.

How is the investment universe defined and screened?

The investment universe is defined by written eligibility criteria before any asset is considered for a portfolio. An asset must have a public, verifiable protocol; sufficient trading depth on venues Abra can access; custody support through the firm's custody infrastructure; and a regulatory classification that the compliance team has documented. Assets that fail any criterion are excluded regardless of return expectations. The universe is reviewed on a scheduled basis and between reviews when a material event occurs.

Screening proceeds in stages. A liquidity screen removes assets whose traded volume or venue coverage falls below the thresholds set in the methodology note. A protocol screen assesses code maturity, governance and concentration of supply. A regulatory screen documents the classification and any restrictions on holding the asset for particular client types. A custody screen confirms that the asset can be held in a segregated account titled to the client with no rehypothecation.

Bitcoin and other large, established digital assets form the core of most institutional portfolios under this methodology. Smaller assets are eligible only where every screen is satisfied and where the client mandate permits them. The current eligible list is provided in the data room rather than on this page because it changes.

How are portfolios constructed?

Portfolio construction uses a factor-based approach in which each eligible asset is characterized by a set of exposures rather than treated as a single position. Factors considered include market beta to bitcoin, liquidity, network growth, and yield source where applicable. Target weights are set at the mandate level in consultation with the client, and the factor model is used to keep realized exposures close to what the mandate intends.

The allocation framework starts from the client's target weight to digital assets and the constraints in the investment policy statement. Within that weight, assets are sized by a rules-based process that combines factor exposures with liquidity limits and concentration limits. Optimization, where used, is constrained so that the output cannot drift far from a transparent baseline such as market-capitalization weighting, and the constraints are documented so that a committee can reproduce the result.

The tradeoff is stated plainly. A constrained, reproducible process gives up some of the return an unconstrained optimizer might claim in exchange for a result a committee can audit. Abra takes that trade because an allocation a committee cannot explain to its board is one it will not hold through a drawdown.

How do yield sleeves fit into the methodology?

Yield strategies, including BTCaf and USDaf, are treated as distinct sleeves with their own eligibility rules rather than as inputs to the factor model. Each yield source must pass the same liquidity, protocol, regulatory and custody screens as any other asset, and the compliance team documents its classification before the investment team may size a position. Where a new source is added to a sleeve, the sleeve's stated target range is reviewed and the change is recorded in the change log.

The target ranges are planning ranges, not forecasts. BTCaf targets roughly 2 to 5% APY and USDaf targets 4 to 10% APY. Both targets are variable and not guaranteed, and both strategies can lose principal. Target ranges are not floors or guarantees, and realized results may fall outside them. Realized yield is reported next to the target so that a committee can see the gap in either direction. Yield strategy tokens are issued by AbraFi and accessed through Abra Capital Management.

Yield sleeves carry risks that a spot allocation does not, including smart-contract, protocol and counterparty risk, and the methodology treats them as a separate risk budget for that reason. A committee that wants spot exposure only can exclude yield sleeves at the mandate level without changing the rest of the process.

Related: How Abra generates yield: methodology and risk management

What risk management framework applies?

Risk management is defined by written limits that apply before, during and after portfolio construction. Pre-trade limits cover maximum position size, maximum exposure to any single venue or counterparty, and minimum liquidity for any asset added. Ongoing limits cover portfolio concentration, realized volatility relative to mandate, and drawdown from peak. Post-trade controls cover reconciliation between internal records and custodian records at a set cadence.

Drawdown controls specify what happens when a portfolio falls by a stated amount from its peak. Responses range from a formal review to a reduction in exposure, and the response is decided in advance, not at the time of the drawdown. Correlation monitoring tracks how assets within the portfolio move together and how the digital asset sleeve moves against the client's other holdings, since diversification assumptions are a common point of failure.

Stress testing applies historical scenarios drawn from prior digital asset drawdowns and hypothetical scenarios such as a venue failure, a stablecoin dislocation, or a sharp change in dollar liquidity. Results are reported to the investment committee and shared with clients on request. Operational risk is addressed separately through custody of SMA assets through the custody arrangement specified in the applicable client agreement and product documentation with Fireblocks MPC infrastructure, separately managed accounts titled to the client, and no rehypothecation for Abra's own account, with strategy-specific deployment or use governed by the applicable client authorization and product disclosures, with qualified custody supported on select assets and strategies.

Related: Abra strategy and security: a governed, multi-layer control framework

How does rebalancing work?

Rebalancing follows written triggers rather than discretion. A portfolio is reviewed for rebalancing when any asset weight moves outside a tolerance band around its target, when the digital asset sleeve moves outside the band set in the client's investment policy statement, or on a fixed calendar date, whichever occurs first. Cash flows into or out of the account are used to move toward target before any additional trades are placed.

Execution is staged to limit market impact and to avoid concentrating trades at a single price. Orders are routed across the venues on Abra's approved list, and trades in less liquid assets are spread over a longer window. Every rebalance is documented with the trigger that caused it, the trades executed, and the resulting weights, so that a client or auditor can reconstruct the decision.

Rule-driven rebalancing is also where a cycle framework earns its place. A committee that has documented phase definitions can use them to pace initial funding or to trigger a pre-approved rebalance, which is the use institutional allocators have made of such frameworks in practice. The rebalance still follows the written trigger; the framework supplies the context.

Related: Bitcoin market cycle analysis: a framework for institutional allocators

How is performance reported and benchmarked?

Performance is reported against benchmarks stated in advance for each mandate. For a bitcoin-only mandate the benchmark is a named, administered bitcoin reference rate identified in the mandate, with the fee structure disclosed alongside it; there is no single authoritative published “spot price of bitcoin”, because spot prices differ across venues at any instant. The industry-standard form is an administered index such as the CME CF Bitcoin Reference Rate, a once-a-day benchmark that aggregates trade data from multiple Bitcoin-USD markets and is published at 4:00 p.m. London time (CF Benchmarks, “CME CF Bitcoin Reference Rate,” cfbenchmarks.com (retrieved 24 Sept 2026)). For a multi-asset mandate the benchmark is a published index or a composite of published indices chosen to match the mandate's eligible universe. For yield sleeves the reference is the stated target range, with realized yield reported next to it.

Attribution separates the effect of allocation decisions, asset selection and yield from the effect of market movement. Returns are shown gross and net of fees, and the fee schedule used is stated on each report. Where a composite is presented, the report explains which accounts are included and why. All performance presentations carry the disclosures and presentation requirements of the SEC marketing rule, including net-of-fee performance alongside any gross performance and results for the prescribed time periods (SEC Rule 206(4)-1 under the Investment Advisers Act of 1940 (the marketing rule), adopted Dec. 22, 2020 (Release No. IA-5653)).

Reporting cadence is set in the client agreement, with a standard periodic statement and a fuller quarterly review. Reports are available through the client portal and by direct delivery, and institutional clients can request data files for their own systems. The reporting explainer describes what each document contains and how figures reconcile to custodian records.

Related: How institutional crypto reporting works: statements, attribution, tax and audit

How is the methodology documented, versioned and updated?

The methodology is version controlled, and the version in force is stated at the head of this page. A material change, such as a change to eligibility criteria, factor definitions, risk limits or rebalancing triggers, produces a new version number, a dated change note describing what changed and why, and notification to institutional clients and to allocators who have requested the data room.

Non-material changes, such as clarifying language or a data vendor substitution that does not alter a series, are recorded in the change log without a version increment. Prior versions are retained and available on request so that any historical decision can be read against the rules in force when it was made. The full change log is provided in the data room.

The methodology note in the data room documents the screens, factor definitions, constraints and baseline weighting in enough detail for a committee or its consultant to reproduce a portfolio from the eligible list and the mandate parameters. Where a proprietary input is used, the note identifies it and describes its role so that its effect can be bounded. The DDQ library's investment questionnaire cross-references this page and that note.

Related: Crypto DDQ library: due diligence questionnaires for evaluating digital asset managers

Disclaimer. Abra Capital Management, LP is an SEC-registered investment adviser. Registration does not imply a certain level of skill or training. This document describes an investment process and is not investment advice or an offer of advisory services. Target yield ranges are variable and not guaranteed, and past performance does not predict future results. Yield strategy tokens (BTCaf, ETHaf, SOLaf, USDaf, XRP yield) are issued by AbraFi and accessed through Abra Capital Management. Digital assets involve a high degree of risk, including loss of principal, and are not FDIC or SIPC insured.