Asset Allocation and Rebalancing: The Decision That Does Most of the Work
Key takeaway
- The stock-bond mix decides how much your portfolio moves. Nothing you pick inside those buckets changes that much.
- Return rises gently as you add stocks. The worst year falls much faster — which is why the right mix is set by what you can hold, not by what you would like to earn.
- Portfolios drift towards risk on their own. A rebalancing band is the rule that resets them, and it forces you to sell what rose without needing a view.
Learning pathThe index investor's whole jobStep 11 of 12
Read before this:What a Bond Actually Is, and Why Its Price Moves the Wrong Way
Based on The Four Pillars of Investing — William J. Bernstein, 2002
The decision you make before you pick anything
How much sits in stocks versus bonds sets the shape of everything that follows. It is one number, and most people never write it down.
William Bernstein was a neurologist who taught himself finance and wrote The Four Pillars of Investing for people in exactly that position. His four pillars are theory, history, psychology and business — but the operational core of the book is short: decide the mix, hold it cheaply, and put it back when it moves.
Read the figure as a constraint rather than a menu. If a 45% fall would make you sell, then a mix that permits a 45% fall is not available to you at any expected return, because you will not still be holding it when the return arrives.
Why holding two things beats holding the better one
This is the one genuinely free thing in investing: combining assets that do not move together reduces the swings by more than it reduces the return. Not to zero, and not reliably in a crisis — see the limits below — but enough that a mixed portfolio beats the parts on the measure that decides behaviour.
Setting your own number, in one page
| Input | The question | Which way it pushes |
|---|---|---|
| Horizon | When will you need to spend this money? | Longer horizon supports more stocks |
| Capacity | If it fell 40%, would you still be able to leave it alone? | Depends on income stability, not on courage |
| Need | Does your plan actually require the extra return? | If not, taking the risk is unpaid work |
The third row is the one people skip and Bernstein is firm about: if you have already accumulated enough for the goal, adding equity risk raises the chance of not getting there. "When you have won the game, stop playing" is the same argument stated by another author, and it is the only allocation advice that gets more conservative as things go well.
How the mix usually gets set
- Whatever felt right after the last rally
- Copied from someone with a different horizon
- Never written down, so never reviewed
How to set it
- One number, in writing, with a date
- Chosen against a 40% fall, not a good year
- Reviewed when your life changes, not when prices do
Portfolios drift, and always in the same direction
The direction is not random. Whatever has risen most becomes the largest holding, so an untouched portfolio steadily concentrates into whatever has recently worked — which is exactly the position you least want to be in when that thing stops working.
Rebalancing: a rule that makes you do the uncomfortable thing
Rebalancing means selling some of what went up to buy what went down. It is trivial to describe and difficult to execute, which is why it has to be a rule.
Both forms work. A calendar rule (once a year, on a fixed date) is simpler and needs no monitoring. A band rule (act when a weight moves more than five percentage points from target) responds to what actually happened rather than to the calendar. Pick one and write it down; the failure mode is having neither and improvising.
- Rebalance with new money first. Directing contributions to the underweight side does the job with no sale and no tax event.
- Mind the tax account. In a taxable account, selling to rebalance realises gains. Bands wider than five points, or rebalancing only inside a tax shelter, are the usual answers.
- Do not rebalance into a broken thesis. The rule assumes both assets are things you still want to own. It is a weighting tool, not a reason to keep buying something whose case has changed.
Where the model is weaker than it looks
- Correlations rise in crises. The assets that diversify each other in a normal year often fall together in the week you most need them not to. Diversification reduces the ordinary variation more than the tail.
- The inputs are estimated from history. Expected returns and correlations come from the past, and the future is under no obligation. This is an argument for round numbers and wide bands, not for precision.
- Bonds are not automatically the safe half. They carry inflation risk, and a period of rising rates hurts them. "Safe" means "moves differently from stocks", which is a narrower claim than it sounds.
- None of it survives being abandoned. The best allocation is the one still in place after a bad year, which is why the honest input is your own behaviour rather than a risk questionnaire.
Try this week
- Write your target stock-bond mix as a single number, with today's date next to it.
- Work out what a 40% fall in the stock half would do to the total, and read that figure out loud.
- Look up your current actual weights. If they are more than five points from target, you have drifted.
- Write your rebalancing rule in one sentence: calendar or band, and which threshold.
Common questions
What is asset allocation?
It is how your money is split between broad asset classes — typically stocks and bonds. It is the decision that sets how much your portfolio moves, and it does more to determine your experience than which individual holdings you choose inside each bucket.
What is a good stock-bond split for a beginner?
There is no single answer, because the input is your horizon and your tolerance for a bad year rather than your age alone. The useful test is the 40% one: pick the mix whose worst plausible year you would still hold through, then write it down before you need it.
How often should I rebalance my portfolio?
Once a year on a fixed date, or whenever a weight drifts more than about five percentage points from its target. Both work. Rebalancing more often adds costs and taxes without adding much control.
Does rebalancing increase returns?
Not dependably. Its main job is keeping the portfolio's risk where you set it. Whether it also adds return depends on the assets and the period — in a long one-way run it can cost you by trimming the winner too early.
Should I rebalance in a taxable account?
Be more reluctant there, because selling realises gains. Direct new contributions to the underweight asset first, use wider bands, and do the bulk of any rebalancing inside a tax-sheltered account where you have one.