How to Conduct an Annual Portfolio Review

Kabir opens the statements and starts at the wrong end

Kabir does an annual review, which puts him ahead of most people. What he actually does is open the year's performance figures, find the holding that did worst, and think hard about whether to replace it.

He is a careful investor and this is still the wrong first move, for a reason that has nothing to do with the quality of his judgement. Starting at performance means every question he asks afterwards is framed by a single year of returns, which is the least informative thing in front of him and the most emotionally loud. By the time he reaches the questions that matter — whether his goals have changed, whether he is contributing enough, whether his cover is still adequate — he has already spent his attention.

The review should run from the plan outwards to the products, not from the products inwards. The order is most of the method.

Start with the life, not the portfolio

The first section of a review has no investment content at all, and it is the section most likely to produce a change worth making.

The first check is whether the goals are still the goals. Amounts move, dates move, and some goals quietly cease to be real without anyone noting it. What each one now costs gets updated, using actual current prices rather than last year's estimate — which is the annual correction that estimating the future cost of a goal argues is worth more than the precision of the original figure.

Then the household facts. Has income changed, and has the contribution changed with it? Is the emergency reserve still the right size given what a month now costs? Is the insurance still matched to the people who depend on him, and has anything happened — a birth, a loan, a parent needing support — that changes what a loss would mean? Is there new debt?

This part takes twenty minutes and it is where the year's genuinely important findings usually are. A contribution that has not risen in three years matters more than any fund selection, and it is invisible from a performance report.

Reconcile before you judge

The second step is bookkeeping, and skipping it is how people reach confident conclusions from incomplete pictures.

Kabir needs everything valued on the same date, in one place: every account, every fund, every deposit, the balances and what went in and out over the year. Only once that exists can he see his actual allocation, and the actual allocation is frequently a surprise — money accumulating in a savings account is an allocation decision nobody made, and a holding that has run well has become a larger share than intended, which is the drift described in how rebalancing controls risk.

This is also where a funding map gets checked: each holding assigned to exactly one goal. Assets counted towards two goals are a shortfall that has not surfaced yet.

Then judge performance, carefully

Now the returns, and the discipline here is to compare each holding against the job it was given rather than against whatever did best.

A fund chosen to track an index should be assessed on how closely it tracked, not on whether the index went up. A stable holding should be assessed on whether it stayed stable. Comparing either against the year's strongest asset class produces a conclusion, and the conclusion is wrong.

The useful distinction is between a poor result and a poor process. If a holding lagged because its whole asset class lagged, that is the market and nothing has been learned. If it lagged because of cost, or because it drifted from what it claimed to do, or because it was more concentrated than Kabir realised, that is a defect in the holding and it is actionable. The first is far more common and far more often acted on.

One year of ranking is close to no evidence about a long-term strategy — for the reasons set out in why past performance is not enough to choose a fund — and Kabir already knows this in the abstract; the difficulty is that a table of last year's returns makes it very hard to hold on to. The question is not whether every component outperformed. It is whether the goals are on track, and a portfolio can lag a popular index while being perfectly on course, or beat it while the contributions are too small for the goal.

Limit what changes

A review that produces many changes has usually gone wrong somewhere, and it is worth having a bias against action built into the process.

Where the allocation has drifted, correct it with cash flows first, per rebalancing with new investments first. Where there is genuine duplication — two funds doing the same job — consolidating is worthwhile, but only after checking what exiting costs in tax, in exit loads and in lock-ins, because the tidiness is worth less than those can be.

Update assumptions because something has been learned about the world, not because the last year was good or bad. An inflation assumption revised upward after a year of high prices, and downward after a quiet one, is not an assumption; it is an extrapolation of the most recent twelve months.

And apply one rule that prevents portfolios growing indefinitely: nothing gets added without a statement of which job it does, and if that job is already covered, something else has to go or be redefined. Portfolios accumulate because every addition was individually defensible and none was ever assessed against what already existed.

Finish with a written record

The review is not complete when Kabir understands what to do. It is complete when it is written down.

A short list is enough: what he decided, why, who does it and by when. The "why" is the part that earns its keep, because next year's review is enormously easier when the previous year's reasoning is visible — it turns a repeated act of judgement into a record he can check himself against, and it exposes assumptions that have quietly stopped being true.

Then the next date gets set, and the looking stops. Continuous monitoring is not a more diligent version of an annual review; it is a different activity, one that generates decisions rather than findings, and it reliably produces more trading and worse outcomes.

What to take away

Review the plan before the products. Goals, cash flows, reserves, cover and debt first; then value everything on one date and see the allocation you actually have; then assess each holding against the job it was given rather than against the year's winner.

Expect most years to end in a short list and a decision to continue, and treat that as the review working rather than as an anticlimax. Correct drift with new money, require every addition to displace something, write down what was decided and why, and then set the next date and leave it alone.