Monday starts with exceptions, not a household rebuild

What Athena is — and what it refuses to be

Athena is household-coordinated, tax-aware ETF implementation across taxable, Traditional, and Roth accounts. It combines tax-loss harvesting, asset location, gain management, and banded risk control inside a long-only, 13-bucket ETF universe with designated peers for rotation. A harvest sells an eligible taxable lot and buys a dollar-matched peer in the same economic bucket so the household stays invested.

Athena rotates among pre-approved ETFs inside each asset class when a taxable loss is worth harvesting, then locates and rebalances the household without selling taxable lots just to make each account look balanced.

It is not stock-level direct indexing. It is not a daily mean-variance optimizer that chases every residual. It does not claim tax alpha, promise to maximize loss generation, or replace a client’s tax professional. Concentrated names outside the sleeve can count toward household equity; whether to sell them remains a client decision, not an automatic diversification promise.

Success is wash integrity, policy bands held (or a documented constraint), usable tax value after costs, and controlled turnover. Success is not every registration looking like 60/40 on its own statement.

Why copied models fail the tax bill

A 60/40 household is easy to describe and expensive to implement the lazy way. Taxable, Traditional, and Roth do not share one tax rule. Copy the same mix into every account and you ignore location. Rebalance every account back to the picture and you often sell the taxable lots that went up.

Household coordination flips that. Bonds and other tax-heavy holdings prefer Traditional. Growth prefers Roth. Tax-efficient equity can stay taxable, which is also where harvestable lots need to live. Existing appreciated taxable market value in a bucket is treated as a floor: Athena does not create a current capital-gain bill merely to make each account resemble the policy portfolio. Risk is measured at the household. Account cosmetics are not the mandate.

When at least two funded accounts span different tax treatments, location and wash controls matter most. A single-registration book can still run the sleeve; the household story is richest when the wrappers differ.

Where this sits vs peers (short and fair)

Tax-aware implementation spans a spectrum. A no-harvesting ETF book is simple and relies mainly on deferral. A simple ETF-pair harvest can rotate one or two exposures in taxable while IRAs stay on someone else’s model. Stock-level direct indexing usually creates the widest harvest surface because individual names disperse even when an index rises — and it brings sampling, tracking, and a thicker oversight load.

Athena sits in the middle: a 13-bucket ETF sleeve with eligible peers, household wash controls, location with taxable floors, inventory-aware gain management, and banded rebalancing. Peers already talk about tax. Orion Trading rebalances households and harvests; Advyzon Quantum and Black Diamond put tax columns and wash views in the suite; Personalized Indexing and Parametric harvest at the stock. Athena’s difference is the job, not a claim that those firms forgot taxes: one overnight household investment policy on the ETF sleeve, with exceptions in Mission Control instead of you operating a harvest queue. Read the full peer table on the tax-aware ETF implementation page.

The honest trade-off belongs in the first meeting. ETF rotations offer fewer independent loss opportunities than a broad direct-indexing book, especially when markets move together. In exchange you get simpler oversight, named peer substitutions, no stock-selection sampling problem, and coordination across registrations — including the wash risk when an IRA buy could impair a taxable loss.

Location, harvest, wash, and bands in advisor English

Location. After the household allocation is set, location asks which registration should hold the coupon, the growth, and the harvestable lots. Athena uses a reviewable ranking: tax-inefficient income tends to fill Traditional first; leftover high-tax-drag assets and selected growth can go to Roth; tax-efficient core equities are allowed to remain in taxable. Location runs when accounts span tax treatments. It does not sell appreciated taxable lots just to improve placement. Multi-account books can therefore sit structurally off a standalone 60/40 path inside a single wrapper while household equity stays inside the band. That is a design consequence, not a failed rebalance.

Harvest. Taxable lots are screened for depth and estimated benefit. When a lot qualifies, Athena sells it and buys an eligible peer in the same bucket for approximately the same dollars. Rotation is weight-neutral inside the bucket — a clean lot reset, not a stealth rebalance. Tax-advantaged accounts do not harvest, because losses there do not create a usable capital deduction. Some exposures have no designated substitute; those are ineligible for harvest rather than abandoned or washed. The method prefers fewer, deeper, usable losses over shallow tickets that consume wash locks and fail to pay for themselves.

Wash. Locks are direction-aware and scoped to the coordinated household for roughly thirty days. After a harvest sale, the household may not repurchase that ticker, use it as a substitute, or park cash in it during the lock. After a replacement is bought, that replacement can be topped up but is not immediately re-harvested. A purchase in Traditional or Roth can impair a taxable loss; Athena’s controls are built around that fact. Peer inclusion is an investment mapping, not legal certainty that two funds are substantially identical. Clients should consult tax counsel for spouses, trusts, held-away accounts, and entities the system cannot see.

Bands and overnight order. Athena trades when household equity or a subclass leaves its band, not because a residual is merely non-zero. Tax-advantaged overweights and available loss capacity come before taxable gain sales. Leftover cash deploys after harvest, location, and risk work so residual cash is real and a trade is not immediately undone. If a job has nothing to do, it holds. The run is one household investment policy, not four disconnected checklists. You supervise policy and exceptions; Athena executes the approved overnight work.

Production uses a hierarchy of legal and wash constraints, eligibility gates, policy bands, inventory capacity, gain budgets, and cooldowns. That is auditable gates, not a single solved optimizer that maximizes a fully specified utility function every night. Reporting should keep pre-tax portfolio results separate from estimated tax benefit. A harvesting path can lag before tax and still show estimated after-tax value if the benefit is usable — illustrative teaching cases are not a live composite or a forecast.

What changes Monday for the book

Reviews stop opening with “rebuild every account to 60/40.” You scan household drift, tax activity, wash constraints, cash, and exceptions with the household beside the decision. Quiet books near policy stay quiet; that is the process working, not inattention. Drawdowns may show a cluster of dollar-matched peer rotations and household locks — review wash integrity, substitution exposure, and ending policy risk together, not gross loss harvested as a vanity metric. Three-account households may show extreme account-level mixes while the aggregate stays appropriate. Concentrated non-sleeve equity that the client will not sell can leave a chronic overweight; document the election rather than masking it by repeatedly trading the sleeve.

Athena replaces a separate CRM for this book: Mission Control is where the household record, overnight activity, and exceptions live. You are not inventing a harvest queue overnight.

Honest limits

Athena does not provide tax or legal advice. Tax outcomes depend on rates, utilization, holding period, state law, filing status, future transactions, and disposition. A harvested loss can defer rather than eliminate tax; lower replacement basis can create a later gain. ETF peers can diverge on index construction, holdings, expenses, and liquidity during the substitution window. Policy bands permit tracking difference by design. Taxable floors and client concentration elections can keep a household from a cosmetic target. Location ranking today is a production rule with known omissions (state rates, RMDs, planned withdrawals, step-up, charitable intent, future rate paths) — present it as current practice, not a universal theorem.

Selected historical simulations in internal research are operational and illustrative. They are not live composite performance, not confidence intervals across thousands of household paths, and not permission to publish tax-alpha headlines. This page invents none of those numbers.