Implement one-month T-Bill excess returns for bond portfolios

[Yuning598]

Methodological decision

Adopt the corporate bond return construction in Dickerson, Mueller, and Robotti (2023), Priced Risk in Corporate Bonds, Journal of Financial Economics 150(2), 103707, DOI: 10.1016/j.jfineco.2023.103707.

Daily clean price is the volume-weighted average of intraday bond transaction prices. Monthly total return is computed from clean price, accrued interest, and coupon payments:

R(i,t) = [P(i,t) + AI(i,t) + C(i,t)] / [P(i,t-1) + AI(i,t-1)] - 1

The approved monthly excess return is:

excess_return(i,t) = monthly_return(i,t) - one_month_us_tbill_return(t)

A duration-adjusted or duration-matched Treasury return is not the primary convention for this implementation. It may be considered later as a robustness specification, but it must not silently replace the adopted one-month T-Bill benchmark.

Current data dependency

The current dataset exposes monthly_return and monthly_return_winsorized, but does not yet expose the approved excess-return field. Portfolio code may be developed against monthly_return temporarily, provided that the return-field selection is explicit. Final reported bond long-short performance must use the refreshed excess_return field after #9 validates the updated dataset.

Acceptance criteria

• The risk-free series is identified, documented, monthly, and aligned to the bond-return month.
• The implementation reproduces the stated formula on hand-checked sample observations.
• Return units and missing-value rules are validated before subtraction.
• Portfolio code uses an explicit return-field configuration rather than silently falling back.
• Final published bond portfolio results use excess_return, not raw monthly_return.
• The methodology page cites the paper, formula, benchmark source, and sample timing.
• Duration-adjusted results, if added later, are labeled as a separate robustness analysis.

Dependency

Blocked by #9, which tracks delivery and validation of the refreshed bond-return fields through December 2025.

[Yuning598]

This methodological decision blocks the final bond long-short performance implementation tracked in #7.

[Yuning598]

Method update: adopt the Dickerson–Mueller–Robotti (2023) one-month U.S. T-Bill excess-return convention. The methodological choice is resolved; implementation remains blocked only by the refreshed data validation in #9.