Rebalancing is the part of portfolio management most UK investors know they should be doing and put off because it feels fiddly. Most of the friction is not the act of selling and buying — it is figuring out whether to rebalance, which wrapper to do it in, and whether the resulting tax cost is worth the risk reduction.
This guide covers the practical mechanics: when to rebalance, how to do it tax-efficiently across ISA, SIPP and GIA, the difference between trackers that just show you drift and tools that actually suggest trades, and where robo-advisors fit. It is written for self-directed UK investors holding across multiple wrappers.
This is not financial advice. Past performance does not guarantee future returns. Consider speaking to an FCA-authorised financial adviser for personalised guidance.
TL;DR
- Rebalancing is risk control, not return enhancement. Done well, it stops your portfolio drifting into a risk profile you did not choose.
- Calendar rebalancing (annually or semi-annually) is simpler but can miss large moves. Threshold rebalancing (e.g. the 5/25 rule) responds to drift but creates more decisions.
- Tax-efficient sequence matters: rebalance inside the ISA and SIPP first, only touch the GIA when you must. Use the £3,000 CGT annual exemption deliberately.
- Most general portfolio trackers (Sharesight, Snowball) show you drift but do not suggest trades. Robo-advisors (Nutmeg, Wealthify, Moneyfarm) auto-rebalance for you. Self-directed simulators sit in between.
- The biggest mistakes UK investors make are over-trading inside the GIA, ignoring asset location across wrappers, and rebalancing into the same exposure they were trying to reduce.
What rebalancing actually does
Imagine you set a portfolio target of 60% equities and 40% bonds. After a strong equity year, the actual mix drifts to 70% equities and 30% bonds. Your portfolio is now riskier than the one you signed up for — not because you made a decision, but because you didn't.
Rebalancing brings the mix back to target. Sell some equities, buy some bonds (or direct new contributions to bonds until the ratio returns). The act itself is mechanical. The thinking that surrounds it is what most investors get wrong.
Two truths about rebalancing worth being clear on:
- It is risk control, not return enhancement. Academic research (Vanguard's "Best practices for portfolio rebalancing" and similar studies) consistently finds that rebalancing has minimal effect on long-term return — sometimes slightly positive, sometimes slightly negative. Its job is to keep your risk profile where you want it. Anyone selling rebalancing as a return booster is misreading the data.
- The rebalance bonus is real but small. When you systematically sell what has gone up and buy what has gone down, you naturally "buy low, sell high" at the asset-class level. Over decades this contributes a few basis points of additional return. That is not nothing, but it is not why you do it.
If your sole goal is maximum return, the cheapest answer is "don't rebalance and accept the drift". Most investors are not actually optimising for that — they are optimising for a portfolio that lets them sleep at night. Rebalancing serves that goal.
Calendar vs threshold rebalancing
Two main approaches, neither obviously better, with different operational profiles.
Calendar rebalancing. Pick a date or dates each year and rebalance regardless of drift. Common cadences: annually (typically end of tax year for UK investors), semi-annually, or quarterly. Simple, predictable, and minimises the number of decisions you need to make. Drawback: between rebalance dates, the portfolio can drift far from target, particularly in volatile years.
Threshold rebalancing. Rebalance whenever an asset class drifts beyond a defined band — e.g. when any holding is more than 5 percentage points away from its target weight, or when its allocation has moved by more than 25% of its target. This is the "5/25 rule" popularised by Larry Swedroe.
Quick example. Target weights: 60% global equity, 30% UK equity, 10% bonds. The 5/25 rule says rebalance when:
- Global equity goes above 65% or below 55% (5 percentage point drift), OR
- UK equity goes above 35% or below 25%, OR
- Bonds go above 15% or below 5%, OR
- Any holding's allocation has changed by more than 25% relative to its own target — for the 10% bond allocation, that's a move to above 12.5% or below 7.5%, which would trigger before the 5pp absolute rule.
The 5/25 rule mostly waits for genuine drift while triggering earlier on smaller positions. Most UK investors using threshold rebalancing settle on something close to it.
Hybrid. Many investors use a calendar floor with threshold triggers — "I will rebalance at least annually, and earlier if the 5/25 rule fires". This combines the simplicity of a fixed cadence with the responsiveness of drift-based rebalancing.
UK tax considerations: rebalance order across wrappers
This is where UK self-directed investors get the biggest practical edge if they think about it.
Inside an ISA or SIPP, rebalancing is tax-free. No CGT, no dividend tax, no reportable income. You can sell down a winner and buy a laggard with no friction beyond dealing fees and bid-ask spreads. The tool you use should ideally suggest these trades first.
Inside a GIA, every rebalance is potentially taxable. Selling a position with gains crystallises CGT. The 2026/27 annual exemption is £3,000 — anything above that is taxed at 18% (basic rate) or 24% (higher rate). Dividends realised on rebalance trades sit on top of your existing dividend income for the year.
The practical sequence for a multi-wrapper UK portfolio:
- Use new contributions first. If you are still adding to your ISA or pension, direct new contributions to whatever asset class is below target. This rebalances by addition, with no selling required.
- Rebalance inside ISA and SIPP next. Sell down winners and buy laggards inside the sheltered wrappers. No tax consequences.
- Use the £3,000 GIA CGT allowance deliberately. If GIA positions still need rebalancing, sell only enough each tax year to stay within the £3,000 exemption. This works well if you can spread the rebalance over two or three tax years.
- Bed and ISA. Selling a position in the GIA and re-buying it inside the ISA on the same day uses your annual ISA allowance to convert taxable holdings into sheltered ones. You crystallise the gain in the GIA in the process — so this is still a CGT event, but with the upside that future growth is sheltered.
If you only remember one thing: inside ISA / SIPP first, GIA last. A tool that does not understand this distinction is making it harder for you to rebalance efficiently.
Tools that just track vs tools that suggest
Portfolio tools fall into roughly three buckets when it comes to rebalancing.
Drift trackers. Show you target weights vs actual weights. You see that you are 4.2% over on global equity and 3.1% under on bonds. The tool does not suggest specific trades — that is your job. Most general portfolio trackers (Sharesight, Snowball Analytics) sit in this bucket.
Trade simulators. Let you draft hypothetical trades and see the resulting allocation, gain/loss, and tax impact before executing. You ask "what happens if I sell £2,000 of VWRP and buy £2,000 of VAGS?" and see the answer. Stronger than a drift tracker because you can iterate — try a few combinations, pick the one that gets closest to target with the lowest tax cost. Invormed's planned simulator approach falls here.
Auto-rebalancers. Robo-advisors that hold your money and rebalance automatically on a schedule or when drift hits a threshold. You don't draft trades; you set a target portfolio and the platform does the rest. Nutmeg, Wealthify, Moneyfarm and similar UK platforms operate this way.
The bucket you want depends on how much control you actually want over the trades.
Robo-advisor auto-rebalancing
Auto-rebalancing platforms are the right answer for a meaningful slice of UK retail investors. They are not universally better, and they are not universally worse — they are a specific trade.
Nutmeg, Wealthify, Moneyfarm and similar:
- Hold your money in a managed portfolio matched to your risk score.
- Automatically rebalance when drift hits a threshold or on a fixed schedule.
- Charge a platform fee (typically 0.45–0.75%) on top of the underlying fund TERs.
- Wrap into ISA, SIPP, GIA and JISA structures.
When this is the right answer:
- You do not want to think about rebalancing at all.
- You hold your full portfolio with one provider.
- You are comfortable paying 0.45–0.75% platform fees on top of fund TERs.
- Your portfolio is small enough that the fee in absolute terms is bearable.
When this is the wrong answer:
- You hold across multiple providers and wrappers (the auto-rebalance only works inside the platform).
- You want to choose specific funds rather than accept the platform's portfolio.
- Your portfolio is large enough that the platform fee compounds materially over decades.
- You actually enjoy the analytical work of running your own portfolio.
For an investor with £15,000 across an HL SIPP, a Vanguard ISA and an ii GIA, a robo-advisor auto-rebalance does not solve the problem — it would only rebalance whichever sub-portfolio they moved to it. For an investor with £40,000 in a single Nutmeg account, it is genuinely the right tool.
Self-directed simulators
For investors keeping their accounts spread across HL, ii, AJ Bell, Vanguard or Trading 212 — i.e. most UK self-directed investors — the right tool is a consolidated tracker with rebalance simulation.
What "good" looks like:
- Pulls in CSVs from all your providers and shows the consolidated view.
- Maps each holding to its tax wrapper (ISA, SIPP, GIA, LISA, JISA).
- Shows current drift against your target allocation.
- Lets you draft hypothetical trades and see the resulting allocation, tax impact and cost.
- Suggests an order — rebalance inside ISA / SIPP first, GIA last.
- Looks through ETFs to underlying constituents so you don't accidentally rebalance into the same exposure.
Sharesight shows drift but does not produce specific rebalance suggestions. Good for the "what is my current allocation" question, less good for the "what should I trade" question.
Snowball Analytics is similar in shape — strong on dividend tracking, lighter on rebalance simulation.
Invormed's planned approach is the consolidated tracker plus rebalance simulator — wrapper-aware, with ETF look-through, designed for UK self-directed investors holding across multiple platforms. Currently in early access.
For most investors, a self-directed simulator is the better answer than either a drift tracker or a robo-advisor — assuming the investor has the discipline to actually execute the suggested trades on their broker platforms. That last clause is non-trivial.
Common rebalancing mistakes UK investors make
A few patterns that recur across self-directed UK portfolios.
Rebalancing in the GIA when the ISA had room. Selling £4,000 of an over-weighted holding in the GIA, triggering CGT — when the same holding existed in the ISA and could have been sold there with no tax impact at all. Pure friction loss, fixable by checking all wrappers before placing trades.
Rebalancing into the same exposure. Selling VWRP and buying VHVG to "diversify". Both are global equity ETFs with 60%+ US exposure and ~70% overlap in their underlying holdings. Without ETF look-through, the trade looks like rebalancing; with look-through, it is barely a rebalance at all.
Over-trading inside the GIA without using the CGT allowance. Realising £8,000 of gains in a single tax year, paying CGT on £5,000 — when spreading the same trades over three tax years would have used three £3,000 allowances and minimised the tax cost.
Ignoring asset location. Holding the bond portion in the ISA and the equity portion in the GIA, then realising the bonds were tax-efficient inside the wrapper while the equities (with their growth and dividend yield) were exposed to CGT and dividend tax.
Rebalancing too often. Quarterly rebalancing for a long-term passive portfolio adds dealing fees, spread costs and (in the GIA) potential tax events with little risk-control benefit. Annually, or 5/25 threshold-driven, is plenty for most investors.
Using calendar dates that aren't tax-efficient. Rebalancing on 1 January means any GIA gains crystallised land in the current tax year. Rebalancing in late March (just before tax year end) gives you the option to spread realisations across two tax years and use both £3,000 CGT allowances.
A practical workflow
A reasonable rebalancing workflow for a UK self-directed investor:
- Set targets once a year. Start of tax year is a natural moment. Decide your target allocation across asset classes (equity / bonds / cash / alternatives) and within each (UK / global / EM / etc).
- Check drift quarterly. Pull updated CSVs, look at consolidated drift, decide whether anything triggers your threshold rule.
- If trigger fires, rebalance in sequence. ISA / SIPP first. GIA only if absolutely necessary, and only enough to stay within the £3,000 CGT annual exemption.
- Use new contributions to rebalance by addition. When possible, don't sell — just direct fresh contributions to the under-weighted asset class.
- Look through ETFs before trading. Make sure the trade actually changes underlying exposure rather than swapping one US large-cap proxy for another.
- Record the trades for cost basis purposes. Particularly important in the GIA, where realised gains feed into Self Assessment.
Most UK investors who follow something like this rebalance two or three times in an average year, often through a combination of new contributions and small ISA-side trades, and rarely touch the GIA except at year-end.
FAQ
How often should I rebalance my UK portfolio?
For most long-term passive investors, once a year is sufficient. Threshold-based rebalancing (e.g. the 5/25 rule) can trigger more often in volatile markets and is fine if you have the discipline to act on it. Quarterly rebalancing is usually overkill and adds friction without clear benefit. The exception is if your portfolio is small and you are still in accumulation — directing new contributions to under-weighted asset classes is a form of rebalancing that happens naturally each month.
Should I rebalance in ISA, SIPP, or GIA first?
Inside the ISA or SIPP first, almost always. These are sheltered from UK tax, so rebalance trades have no CGT or dividend tax consequence. Only touch the GIA when you must, and try to use no more than the £3,000 CGT annual exemption per tax year if you can spread the rebalance.
What's the 5/25 rule?
A threshold rebalancing rule: rebalance when any asset class is more than 5 percentage points away from its target weight, or when its allocation has changed by more than 25% relative to its own target — whichever fires first. The 25% relative trigger catches drift in smaller positions before the 5pp absolute trigger would.
Does Sharesight tell me when to rebalance?
Sharesight shows you target vs actual allocation if you set targets, so you can see drift. It does not produce specific rebalance suggestions or model the tax impact of hypothetical trades. For drafting actual rebalance trades you need a simulator-style tool or a spreadsheet.
Will rebalancing trigger CGT in my GIA?
Yes — any GIA disposal is a potential CGT event. The 2026/27 CGT annual exemption is £3,000. Gains above that are taxed at 18% (basic rate) or 24% (higher rate). Spreading the rebalance over multiple tax years can keep you within the exemption each year.
Are robo-advisors better than self-rebalancing?
For an investor who wants to outsource the entire decision and is happy paying a 0.45–0.75% platform fee on top of fund TERs, robo-advisors are the right tool. For a self-directed investor who already holds across multiple providers and wrappers, robo-rebalancing only works inside the platform — it does not solve the multi-wrapper rebalancing problem. Different tools for different shapes of portfolio.
Can I simulate a rebalance before doing it?
Yes — using a portfolio simulator that lets you draft hypothetical trades and see the resulting allocation, tax impact and total cost. Sharesight does not do this in a tax-aware way; Invormed's rebalance simulator (in early access) does. A spreadsheet can also work if you build it carefully.
Want this kind of analysis on your real portfolio? Invormed is in early access — join the early-access waitlist.