Portfolio Management interview preparation

Questions

100

Traced to a firm

40

Firms

24

Updated

September 2026

Asked at

All firmsBLBlackRock4Vanguard4WMWellington Management4Amundi3ACAQR Capital Management3Neuberger Berman3SCSchroders3Man Group2MSCI2Northern Trust2AllianceBernstein1Apollo Global Management1Blackstone1BMBNY Mellon1Carlyle Group1Fidelity Investments1Goldman Sachs1Invesco1Millennium Management1MSMorgan Stanley1NUNuveen1PIMCO1SSState Street1TPTPG1

Topic

All topicsPortfolio theory5Factor models8Asset allocation11Rebalancing3Portfolio construction7Benchmarks and tracking error5Performance measurement8Risk management6Fixed income and LDI5Currency and global3Implementation and costs5Active versus passive6India markets7Brainteasers5Career and fit16

Level

AnyCoreIntermediateHard

Type

AnyTechnicalCaseMarket viewFitBrainteaser

1. 001What does modern portfolio theory actually say, and what does it get wrong?Portfolio theoryIntermediatetechnicalNorthern TrustAsset Management · Chicago · 2025

   Say this

   It says risk and return are properties of the portfolio, not of the asset, so the only thing that matters about a holding is what it adds to the whole. What it gets wrong is the inputs. It assumes you know expected returns, variances and correlations, and you do not.

   Then walk it

   1. The core insight: combine assets whose returns do not move together and total risk is lower than the weighted average of the parts. Diversification reduces risk without giving up expected return.
   2. That produces the efficient frontier, the set of portfolios with the highest expected return at each level of volatility, and the claim is that a rational investor holds one of them.
   3. So the decision rule changes. You stop asking whether an asset is good and start asking what it adds to what you already own. A 25 percent volatility asset with low correlation can lower portfolio risk.
   4. Where it breaks: expected returns are estimated with huge error, correlations are unstable and rise exactly when you need them low, and returns are fat-tailed and skewed rather than normal.
   5. It is also single-period and uses one risk measure. Real investors care about drawdown, liquidity and the path, because they have to fund something along the way.
   6. So my honest position is that the framework for thinking is right and still how the industry is organised, but the mechanical optimiser built on it is not trustworthy. That is why shrinkage, Black-Litterman and risk-based methods exist.

   Where candidates lose it

   Reciting the assumption list, normal returns and rational investors and no taxes, as if listing assumptions were the same as criticising the theory. The criticism that matters is estimation error in expected returns. Say that and you sound like someone who has actually run an optimiser.

   Expect next

   • Which input is the optimiser most sensitive to?
   • So do you use mean-variance optimisation at all?
   • What happens to the frontier when you add a no-shorting constraint?

   Reported by candidates at Northern Trust (Asset Management, Chicago, 2025). Source: Wall Street Oasis.

2. 003Why does diversification work, and where does it stop working?Portfolio theoryIntermediatetechnicalNorthern TrustAsset Management · Chicago · 2025

   Say this

   It works because idiosyncratic risks partly cancel, so portfolio variance falls faster than expected return does. It stops working once you have removed the diversifiable part, because what is left is systematic risk that every holding shares.

   Then walk it

   1. The arithmetic: portfolio variance depends on the average variance divided by the number of holdings, plus the average covariance. The first term shrinks toward zero as you add names, the second does not.
   2. So the average covariance is the floor. In a single equity market, roughly 25 to 30 reasonably spread names get you most of the way, and the marginal benefit after that is small.
   3. That is why the next step is diversifying across things with genuinely different drivers: other markets, other asset classes, duration, real assets. Thirty Indian banks are not a diversified portfolio.
   4. Where it stops working: in a liquidity event, everything correlated to the same funding conditions moves together. In March 2020 credit, equities, EM, even gold for a few days, all fell at once because people were selling what they could.
   5. It also stops working when the diversification is only nominal. Three funds that all own the same quality-growth factor are one position with three fee loads, and that only shows up in a factor decomposition.
   6. So my summary: diversification removes stock-specific risk cheaply and reliably, and does almost nothing about the systematic risk you are actually paid for. Anyone who claims a portfolio is safe because it is diversified has confused the two.

   Where candidates lose it

   Saying more names is always better. Past roughly 30 well-spread holdings you are adding cost, monitoring burden and closet indexing, not risk reduction. The interviewer wants the covariance floor, and wants you to name the crisis case where correlations converge.

   Expect next

   • How many stocks do you actually need?
   • What happened to correlations in March 2020?
   • How would you check whether two funds are really diversifying each other?

   Reported by candidates at Northern Trust (Asset Management, Chicago, 2025). Source: Wall Street Oasis.

3. 006Explain CAPM, and then tell me where it fails empirically.Factor modelsIntermediatetechnicalAsset managementQuantitative research

   Say this

   CAPM says the only risk you get paid for is covariance with the market, so expected return is the risk-free rate plus beta times the equity risk premium. Empirically the beta and return relationship is far too flat, and several characteristics that should not matter clearly do.

   Then walk it

   1. The logic is clean: idiosyncratic risk diversifies away for free, so nobody pays you for it, and the only priced risk is the part that moves with the market.
   2. Failure one, the security market line is too flat. High beta stocks earn less than CAPM predicts and low beta stocks earn more. That is the betting-against-beta result, and it is the foundation of low-volatility investing.
   3. Failure two, characteristics predict returns after controlling for beta. Small size, cheap valuation, high profitability and recent momentum all carry return premia that beta does not explain. Fama and French built their models on exactly this failure.
   4. Failure three, the market portfolio is unobservable. Roll's critique is that you cannot test CAPM at all, because the proxy you use, usually a cap-weighted equity index, is not the true market portfolio that includes human capital, housing and private assets.
   5. Failure four, the assumptions that break in practice: unlimited borrowing at the risk-free rate, no constraints, no taxes. Leverage constraints are actually the leading explanation for why the line is flat, because investors who cannot lever buy high beta instead.
   6. Where it still earns its place: as the discipline that says only undiversifiable risk is compensated, and as the cost of equity in every DCF ever built. Nobody has replaced it for that job, which is worth saying out loud.

   Where candidates lose it

   Stopping at the formula. The question has two halves and the second is the interesting one. Candidates who cannot name at least the flat security market line and the size or value anomaly sound like they learned CAPM from a textbook and never looked at data.

   Expect next

   • If the line is flat, how would you exploit that?
   • What is Roll's critique?
   • Do you still use CAPM for cost of equity, and why?
4. 007Walk me through the Fama-French three factor model, and tell me what Carhart added.Factor modelsIntermediatetechnicalQuantitative asset managementFactor investing

   Say this

   Fama-French keeps the market factor and adds size, small minus big, and value, high minus low book to market. Carhart added a fourth, momentum, winners minus losers, because the three factor model could not explain the persistence of past winners.

   Then walk it

   1. Construction is the same idea each time. Rank the universe on the characteristic, form portfolios from the extremes, and the factor return is the long-short spread, rebalanced annually for value and size and monthly for momentum.
   2. SMB is small cap minus large cap, HML is cheap minus expensive on book to price. Together with the market they explain something like 90 percent of the variation in diversified US portfolio returns, against about 70 percent for CAPM alone.
   3. Carhart's momentum factor, usually called UMD or WML, is built on 12 month returns skipping the most recent month. It matters because momentum is the most statistically robust anomaly and the hardest to explain as compensation for risk.
   4. Fama and French later went to five factors, adding profitability and investment, at which point HML becomes close to redundant. That is a useful thing to know because it tells you value is partly a profitability story.
   5. The practical use is attribution rather than prediction. Regress a fund on these factors and the intercept is what the manager added beyond cheap systematic exposures, which is the number an allocator actually pays for.
   6. The honest limitation: these are in-sample constructions on US data, the value premium has been weak for long stretches including most of 2010 to 2020, and factor timing has an awful record. So I would use them to explain returns, not to promise them.

   Where candidates lose it

   Getting the sign convention wrong or describing HML as growth minus value. Also, do not present the model as a return forecast. Its main professional use is attribution, and saying that shows you know how these models get used in a real seat rather than in a paper.

   Expect next

   • Why did they add profitability and investment?
   • What does momentum being unexplainable by risk imply?
   • How would you use these factors on Indian equities?
5. 014What is strategic asset allocation, and how much of the outcome does it really explain?Asset allocationIntermediatetechnicalMulti-assetAsset management

   Say this

   Strategic asset allocation is the long-run policy mix you would hold if you had no view, set from the objective and the constraints, and it explains most of the variation in a portfolio's returns over time. What it does not explain is the difference between two funds with the same policy mix.

   Then walk it

   1. Mechanically, it is a set of target weights and permitted ranges per asset class, agreed in an investment policy statement, with a benchmark for each sleeve and a rebalancing rule.
   2. It comes from the liability or the objective: required return, horizon, drawdown tolerance, liquidity needs, tax status, regulatory constraints. Not from a market view. The market view lives in the tactical overlay.
   3. The famous number is Brinson's, that about 90 percent of the variability of a fund's returns over time comes from the policy mix. That is routinely misquoted as 90 percent of the return level, which is wrong.
   4. Ibbotson and Kaplan cleaned this up: policy explains roughly 90 percent of the variation over time within a fund, about 40 percent of the variation across funds at a point in time, and slightly more than 100 percent of the level of return, because active management and costs net out negative on average.
   5. So the honest framing is: allocation dominates the risk profile and the path, and manager selection determines whether you beat your peers. Both matter, for different questions.
   6. Practically that is why governance time is best spent on the policy mix and the rebalancing rule, not on the monthly manager review. The decision with the largest effect is made once and revisited every three years.

   Where candidates lose it

   Quoting 'asset allocation explains 90 percent of returns'. It explains 90 percent of the variability over time, not the level, and the distinction is a standard trap in asset management interviews. Getting it right separates people who read the Brinson paper from people who read a marketing deck.

   Expect next

   • So does manager selection matter at all?
   • How often would you revisit the strategic allocation?
   • What inputs would you use for long-run expected returns?
6. 015A new institutional client hands you a mandate. How do you set the strategic asset allocation?Asset allocationIntermediatecase studyVanguardInvestment Research · Malvern · 2024

   Say this

   Start from the obligation, not the assets. What has to be paid, when, in what currency, and what shortfall is intolerable. Then build capital market assumptions, then solve for the cheapest mix that meets the obligation with acceptable risk, then write down the rules.

   Then walk it

   1. Define the objective precisely. A pension has a liability with a duration and an inflation linkage. An insurer has regulatory capital. An endowment has a spending rule. Each of those implies a different portfolio even at the same risk tolerance.
   2. Separate risk capacity from risk tolerance. Capacity is what the balance sheet or the funding position can absorb; tolerance is what the trustees will actually sit through. Build to the lower of the two, because a policy abandoned in a drawdown is worse than a more modest one that survives.
   3. Set capital market assumptions for each asset class: expected return, volatility, correlation. I would build expected returns from building blocks, real yields plus inflation for bonds, earnings yield plus growth for equities, rather than extrapolating history, because historical equity returns include a valuation re-rating that cannot repeat.
   4. Then optimise, but with a heavy hand on the inputs. Constrain sensible ranges, use resampling or shrinkage, and test the candidate mixes against the objective in a scenario framework rather than trusting one frontier.
   5. Then stress it. What does a 1970s inflation path, a 2008 correlation shock, or a decade of 2 percent real yields do to the funding position? A policy chosen on a single expected return path is untested.
   6. Then write the governance: target weights and ranges, the rebalancing rule, hedging policy for currency, liquidity budget, and review triggers. The document is the deliverable, because it is what stops the committee changing course at the worst moment.

   Where candidates lose it

   Going straight to weights, '60 percent equities, 40 percent bonds, done'. The sequence is objective, then capacity and tolerance, then capital market assumptions, then mix, then stress, then written policy. Also do not extrapolate historical equity returns as your expected return input; build it up from yield and growth and say so.

   Expect next

   • How would you build a long-run expected return for equities?
   • How does the answer change for a closed pension scheme?
   • What ranges would you set around the targets?

   Reported by candidates at Vanguard (Investment Research, Malvern, 2024). Source: Wall Street Oasis.

7. 018How would you invest ten million pounds?Asset allocationIntermediatetechnicalSCSchrodersAsset Management · London · 2023

   Say this

   My first move is to ask whose money it is and what it has to do, because the same ten million belongs in completely different portfolios depending on the answer. Then I would build a low-cost core, add satellites only where I can justify an edge, and write down the rebalancing rule.

   Then walk it

   1. Ask four questions: what is the money for, when is it needed, what loss would force a change of plan, and what tax wrapper and jurisdiction are we in. Volunteering those questions is most of the marks on this question.
   2. Assume a long-horizon investor with no near-term call on the money. I would run something like 55 to 65 percent global equities, broadly market weighted with a modest home bias for currency reasons, 20 to 25 percent high quality duration, 5 to 10 percent inflation-linked or real assets, and a working cash buffer.
   3. Build the core passively. At ten million, total cost matters more than cleverness: a global tracker at under 10 basis points versus an active fund at 80 basis points is a certain 70 basis points a year of advantage, which compounds to real money over twenty years.
   4. Use satellites sparingly and only where there is a reason: small cap and emerging market inefficiency, credit where the manager can hold to maturity, trend following as a diversifier. Cap the total satellite sleeve so a bad manager choice cannot break the plan.
   5. Then the practicalities, which is where candidates win this question: tax wrappers first, staged entry over a few months if the money arrived as a lump sum, currency hedging policy on the bond sleeve, and a rebalancing rule with 5 percentage point bands.
   6. And the caveat: if the money is earmarked for something in three years, most of this is wrong and the answer is short-dated bonds and cash. Say that, because it shows the horizon is driving the portfolio rather than your product preferences.

   Where candidates lose it

   Launching into a product list before asking what the money is for. This is a test of process, and the specific distinction that separates good answers is horizon and purpose driving the mix. Also, give real numbers. A candidate who cannot commit to approximate weights sounds like they have never built a portfolio.

   Expect next

   • How would that change if the client needs the money in three years?
   • Would you invest it all at once or phase it in?
   • Where would you actually use an active manager?

   Reported by candidates at Schroders (Asset Management, London, 2023). Source: Wall Street Oasis.

8. 021How do private assets fit into a strategic asset allocation when they are only valued quarterly?Asset allocationIntermediatetechnicalMulti-assetInstitutional asset management

   Say this

   You have to unsmooth the returns before they go anywhere near an optimiser, otherwise the quarterly marks make private assets look like low-volatility, low-correlation magic and the optimiser puts everything there. Then size them off the liquidity budget, not off the expected return.

   Then walk it

   1. The measurement problem: appraisal-based or model-based marks are stale and averaged, so reported volatility is understated and correlation to listed markets is understated too. Private equity marked quarterly shows maybe 10 percent volatility when the underlying economic exposure is levered equity at 25 percent or more.
   2. So unsmooth. The standard approach is to regress reported returns on current and lagged public returns and sum the betas, or to use a listed proxy plus leverage as the economic exposure. You will usually find a private equity sleeve behaves like 1.2 to 1.4 times small cap equity.
   3. Then the optimiser gives sensible answers. Feed it raw private marks and it will recommend 60 percent private assets every time, which is the single most common abuse of mean-variance in institutional investing.
   4. Size by liquidity. Work out committed capital, the drawdown schedule, expected distributions and the worst case where distributions stop for two years while calls continue. That denominator problem is what forced endowments into secondaries at discounts in 2009 and again in 2022.
   5. Then decide what you are actually buying: an illiquidity premium, access to companies not available publicly, and manager selection dispersion that is genuinely wide in private markets. Not diversification, because the economic exposure is the same cycle.
   6. And the honest caveat about the reported numbers: IRRs are money-weighted and subscription lines flatter them, so I would compare a private fund on a public market equivalent basis rather than on headline IRR.

   Where candidates lose it

   Treating reported private asset volatility as real, which makes the optimiser allocate everything to them. The two things that must appear are unsmoothing the returns and sizing off the liquidity budget including the denominator effect. Calling private assets a diversifier without qualification is the tell that a candidate has only seen marketing material.

   Expect next

   • What is the denominator effect?
   • How would you compare a private fund to public equity?
   • How much of a portfolio can be illiquid?
9. 023Would you allocate to retail real estate today, and why?Asset allocationIntermediatetechnicalNUNuveenInvestment Management · New York · 2021

   Say this

   Selectively yes, and the reason is that the sector already took its pain, so pricing reflects the structural problem in a way it does not in some other property types. But only the dominant assets, and only with a view on the capex and tenant mix.

   Then walk it

   1. State the structural case against first, because the interviewer is testing whether you will be honest: e-commerce took share, rents on secondary centres are still resetting downwards, and retail needs continuous capital to stay relevant.
   2. The investable case is bifurcation. Grocery-anchored and necessity retail has proved resilient through both the pandemic and the rate shock, because footfall is non-discretionary and leases are short enough to reprice with inflation.
   3. Then the pricing argument, which is the whole point. Retail derated from 2016 onwards, years before offices did, so entry yields already price a bad outcome. Buying a repriced asset class with a known problem is often better risk-adjusted than buying one whose problem has not been marked yet.
   4. What I would underwrite: tenant sales productivity and occupancy cost ratio, because that tells you whether the rent is actually affordable to the tenant, the capex required per square foot to keep the asset trading, and the covenant quality of the top ten tenants.
   5. What I would avoid: secondary and tertiary centres in declining catchments, and anything where the exit assumes a cap rate tighter than entry. If the return depends on yield compression rather than on income, it is a rates bet wearing a property costume.
   6. So the position: a modest allocation to dominant grocery-anchored and outlet formats bought on income, funded out of the office allocation rather than out of logistics, with the debt maturity profile matched so I am never a forced seller.

   Where candidates lose it

   Answering with a sector narrative and no price. Everyone knows e-commerce hurt retail; the question is whether that is already in the entry yield. A candidate who cannot say what they would underwrite at the asset level, occupancy cost ratio and capex per square foot, is giving a newspaper answer.

   Expect next

   • What yield would you need to buy a secondary centre?
   • How would you fund that allocation?
   • Does listed retail REIT pricing tell you anything useful here?

   Reported by candidates at Nuveen (Investment Management, New York, 2021). Source: Wall Street Oasis.

10. 024Which sector would you overweight today, and what would you fund it from?Asset allocationIntermediatetechnicalWMWellington ManagementGeneralist · Hong Kong · 2022

    Say this

    The 'funded from' half is the real question. A sector view is only a portfolio decision once you have said what it displaces, and the pair determines what risk you have actually taken, because an overweight funded from cash is a beta increase and one funded from a correlated sector is a relative value trade.

    Then walk it

    1. Pick a sector where you can state the mechanism in one sentence, name the two or three numbers, and say what the market is assuming that you think is wrong. Specificity beats breadth here.
    2. Then be explicit about the funding leg. Funding an overweight from cash adds market beta. Funding it from a defensive sector adds beta and cyclicality. Funding it from a sector with the same driver, say industrials out of materials, isolates the idiosyncratic view, which is usually what you intend.
    3. Check what else comes with it. Sector bets carry factor exposures whether you want them or not: financials are a rate and curve bet, staples are duration, technology is long growth and long the multiple. Say which factor you are unintentionally buying.
    4. Size it against a tracking error budget. A 3 point sector overweight in a portfolio with a 3 percent tracking error budget is a substantial use of that budget, and I would say how much of the budget I am spending.
    5. Give the falsifier and horizon: the data point that would tell me I am wrong, and when I would review. Sector rotations often need two to three quarters to work, so a one month judgement is noise.
    6. And I would be honest that sector allocation has a weaker evidence base than stock selection within sectors, so I would keep the tilt modest unless the mispricing is unusually clear.

    Where candidates lose it

    Giving a sector view and never saying what you sell. Interviewers in allocation seats are specifically listening for the funding leg and for the factor exposure that comes attached. Naming the unintended factor bet, rates in financials or duration in staples, is what makes it sound like a real portfolio decision.

    Expect next

    • What factor exposure does that pair give you?
    • How much of your tracking error budget does it use?
    • What would make you close it?

    Reported by candidates at Wellington Management (Generalist, Hong Kong, 2022). Source: Wall Street Oasis.

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.