Private Wealth Management interview preparation

Questions

100

Traced to a firm

22

Firms

13

Updated

September 2026

Asked at

All firmsAllianceBernstein4Goldman Sachs4Northern Trust3J.P. Morgan2MSMorgan Stanley2Scotiabank2AMAres Management1BMBNY Mellon1Carlyle Group1Invesco1Neuberger Berman1SCSchroders1UBS1

Topic

All topicsClient discovery5Risk profiling4Asset allocation and rebalancing7Investment policy statement3Tax and asset location6Concentrated positions3Estate, succession and philanthropy6Fiduciary and trusts3Alternatives and liquidity4Products and platforms7Fees and conflicts4Bank economics and risk2Behavioural finance3Family governance3Onboarding and compliance3Business development6Fit and career15Markets and economy9Case and estimation7

Level

AnyCoreIntermediateHard

Type

AnyTechnicalCaseMarket viewFitBrainteaser

1. 012Walk me through mean-variance optimisation, and tell me why you would not hand the output to a client.Asset allocation and rebalancingHardsuperdayFamily officesWealth management

   Say this

   You feed in expected returns, volatilities and correlations, and it gives you the mix with the highest expected return for a given volatility. The problem is that it is an error-maximiser: tiny changes in the expected return inputs produce wildly different and usually absurd portfolios.

   Then walk it

   1. The mechanics: for each level of risk, the optimiser finds the weights that maximise expected return, and the set of those points is the efficient frontier. You then pick the point that matches the client's risk budget.
   2. The first failure is input sensitivity. Expected returns are estimated with huge error, and the optimiser loads up on whichever asset you happened to be most optimistic about. Michaud called it error maximisation and the name is fair.
   3. The second failure is corner solutions. Unconstrained, it will hand you 40 percent in emerging market small caps and zero in domestic large caps, which no client will hold and no committee will approve.
   4. The third is that correlations are unstable and rise in crises, which is exactly when diversification is supposed to pay. The matrix you optimised on is a fair-weather matrix.
   5. What I would actually do: use it as a diagnostic, not a decision. Constrain the weights to sensible ranges, use reverse optimisation or a Black-Litterman approach so the starting point is the market portfolio rather than my own return forecasts, and resample to see how stable the answer is.
   6. And the fourth failure, the one that matters most for a private client: variance is not the risk the client cares about. It ignores taxes, illiquidity, drawdown path and the fact that he may sell at the bottom. A portfolio that is 30 basis points off the frontier but that he will hold beats the optimal one he abandons.

   Where candidates lose it

   Describing the frontier competently and stopping. The question has 'why would you not hand it to a client' in it. Name error maximisation and the fact that variance is not the client's risk measure, or you have answered half the question.

   Expect next

   • What is Black-Litterman doing differently?
   • How do you handle illiquid assets in an optimiser?
   • What constraints would you impose and why?
2. 015After a two-year rally the portfolio is 75 percent equity against a 60 percent target, and selling means a large capital gains bill. What do you do?Asset allocation and rebalancingHardcase studyIndian wealth managementFamily offices

   Say this

   Rebalance, but pay as little tax as possible getting there. The 15-point drift is a real risk problem: the portfolio now loses meaningfully more than the client signed up for. I would close most of the gap using flows and tax-efficient lots, and accept some tax rather than letting the tail wag the dog.

   Then walk it

   1. Quantify the risk first so the decision is not abstract. At 75 percent equity, a 40 percent equity fall costs about 30 percent of the portfolio against 24 at target. On 20 crore that is over a crore of extra loss. Compare that with the tax bill before deciding.
   2. Then work through the low-tax levers in order. Stop reinvesting into equity and direct all new money and income to debt. Take the client's withdrawals from the equity sleeve. Use the annual Rs 1.25 lakh long-term equity gains exemption in each family member's hands.
   3. Harvest losses where they exist, including in the debt and global sleeves, to net against the gains. Sell the highest-cost lots first where the fund house or platform supports lot selection.
   4. Consider whether you can change exposure without selling: satisfy a charitable pledge by donating appreciated shares rather than cash, or gift appreciated units to an adult child who is in a lower bracket, remembering the cost and holding period carry over.
   5. Then hedge the residual instead of selling it if the tax cost is genuinely prohibitive: a modest index put spread or a short futures position buys time to bleed the exposure down over two financial years.
   6. And a straight answer on the arithmetic, because interviewers want it: long-term equity gains at 12.5 percent cost you 12.5 paise per rupee of gain, once. Being 15 points over-risked into a bear market can cost several rupees. Tax is a cost to minimise, not a reason to abandon the policy.

   Where candidates lose it

   Either extreme. 'Never let tax drive the decision' sounds principled and ignores a real, quantifiable cost. 'Do not sell because of the gains' is how clients end up 80 percent equity at the top. Show the comparison in rupees, then act.

   Expect next

   • How do you spread the sale across two financial years?
   • Would you use derivatives to manage the exposure instead?
   • What if the concentration is in one stock rather than the market?

Firm tags come from public, anonymous candidate reports on Wall Street Oasis: strong signal, not sworn testimony. Firms are named as the places a question was reported, not as partners of Fin Maverick. Answers are written for this page to show how to think out loud; they are not scripts to recite.