Most investors think of their taxable account as one thing: "my portfolio." A more useful model is to think of it as three layers, each doing different work, with different cost structures, different risk profiles, and very different opportunities for tax savings.
This is how institutional managers describe these portfolios. It's also the architecture HarvestEngine builds against. Once you see it, you can't unsee it — and you'll find yourself making better allocation decisions almost immediately.
The three sleeves
| Sleeve | Holdings | Purpose | TLH role |
|---|---|---|---|
| Beta | Broad-market ETFs (VOO, VTI, IXUS, BND, AGG) | Cheap, efficient market exposure | Limited — single ETF = single harvest opportunity per drop |
| Long | Direct-index sleeve: 60-200 single stocks tracking an index | Same beta as Beta sleeve, but with per-stock granularity | The main TLH engine — every stock can be harvested independently |
| Short | Short positions on stocks not held long | Generate additional harvest surface; double the TLH potential | Additive — doubles harvest potential at cost of complexity |
Why three sleeves and not one
The classical question: "If the long sleeve already tracks the index, why hold any Beta ETFs at all?" Three reasons:
- Cost efficiency at scale. Holding $5M of VOO costs you the ETF's 0.03% expense ratio = $1,500/year. Holding $5M as a 100-stock direct-index sleeve has no expense ratio (you own the stocks directly), but each stock has a small bid-ask spread and a small commission cost when rebalancing. For broad exposure, ETFs are slightly cheaper. For tax-aware exposure, direct holdings dominate.
- Cash flexibility. The Beta sleeve can absorb inflows and fund outflows in single trades. Liquidating $200K of VOO is one transaction. Liquidating $200K from 100 single positions is 100 transactions, each with its own tax-lot selection.
- Tracking error budget. A direct-index sleeve tracks the index with some error (typically 0.5-1.5% annually). By holding part of your equity exposure as ETFs, you're saying "I want this slice to be exact." Most portfolios target something like 30-50% Beta, 50-70% Long. The blend gives you most of the TLH benefit while keeping overall tracking error tight.
The Short sleeve is its own argument. We covered it in the short overlay article in detail. Quick version: it doubles your harvest surface (every short can be harvested when it goes up, vs. the long sleeve which harvests when stocks go down) at the cost of margin requirements, infinite-loss potential, and §1233's ordinary-income tax treatment. Worth it for some portfolios; not for all.
How the three sleeves coordinate
The key insight is that the three sleeves are not independent. They share constraints and opportunities:
1. Constructive sale (§1259) connects long and short
You can't short a stock you're holding long. Doing so triggers constructive-sale treatment — the IRS treats it as if you sold the long position, immediately recognizing the gain. This means the Long sleeve and Short sleeve must hold disjoint stocks.
HarvestEngine enforces this at the planning layer: when you generate a short overlay, the candidate set excludes anything you already hold long.
2. Wash-sale (§1091) connects long and Beta
If you sell VOO at a loss in your Beta sleeve, you can't buy VOO in your Long sleeve within 30 days. (You also can't buy IVV or SPY — substantially identical.) This is the classic wash-sale trap when you have both ETFs and direct holdings; software has to track substitutes carefully.
The cleanest approach: Beta uses one fund family (e.g., Vanguard VOO/VTI), Long is direct stocks. There's no overlap, so wash sales are limited to within-sleeve events.
3. Drift correction connects all three
Over time, sleeve weights drift. Your Beta sleeve grows when ETFs appreciate. Your Long sleeve drifts as individual stocks move. Your Short sleeve unwinds as positions close.
The rebalance proposal looks at total household exposure, not just within-sleeve drift. If your overall tech exposure is 35% (target 30%) — even though it's split across both Beta and Long — the rebalance trims tech across both sleeves to bring you back to target.
What you actually allocate to each
The defaults that work for most accounts:
| Profile | Beta | Long | Short |
|---|---|---|---|
| Conservative ($100K-$500K) | 60% | 40% | 0% |
| Standard ($500K-$2M) | 35% | 65% | 0% |
| Aggressive TLH ($2M+) | 20% | 75% | 5% |
| Maximum tax-alpha ($5M+, advanced) | 15% | 75% | 10% |
The trade-offs are explicit: more Long sleeve = more TLH surface area, but more operational complexity (60-200 stocks to manage). More Short sleeve = additional alpha, but margin requirements + complexity. More Beta = simplicity, but smaller TLH surface.
The household consolidated view
For investors with multiple taxable accounts (yours, spouse's, joint), the three-sleeve framework applies at the household level:
- Your taxable Schwab: 30% Beta, 70% Long
- Spouse's taxable E*TRADE: 50% Beta, 50% Long
- Joint account at Fidelity: 100% Beta (cash management)
Wash-sale rules span the household. If you sell VOO at a loss in your account, your spouse's IRA can't buy VOO within 30 days either. The household consolidated view is what makes this operationally manageable. HarvestEngine's multi-account view is specifically built for this — it's not three separate views; it's one household.
The mental model
Stop thinking of your taxable account as "my portfolio" and start thinking of it as three sleeves. Beta does cheap market exposure. Long does the tax-aware heavy lifting. Short adds alpha if the operational complexity is worth it.
Once you have the model, every allocation question becomes clearer. "Should I add more international exposure?" — Beta sleeve, add IXUS. "I want more aggressive TLH" — shift more weight from Beta to Long. "I'm worried about a particular sector concentration" — the Long sleeve handles sector tilts naturally.
The three sleeves separate concerns. Each is doing its own job. Together, they compose into the most tax-aware structure available to a self-directed investor.