Risk Management interview preparation

Questions

100

Traced to a firm

37

Firms

12

Updated

September 2026

Asked at

All firmsUBS14MSCI7BLBlackRock5FTFranklin Templeton3Oaktree Capital Management2Scotiabank2Jane Street1Moody's1Neuberger Berman1PIMCO1SSState Street1TSTruist Securities1

Topic

All topicsMarket risk and VaR14Tail risk and stress testing5Greeks and sensitivities5Credit risk11Counterparty risk and CVA6Operational risk5Model risk and validation6Regulatory capital7Liquidity risk and ALM6Statistics and quant foundations7Indian regulation7Risk governance and appetite4Markets and macro9Fit and career8

Level

AnyCoreIntermediateHard

Type

AnyTechnicalCaseBrainteaserMarket viewFit

1. 051You are validating a gradient boosting credit model that beats the existing logistic scorecard by eight Gini points. Do you approve it?Model risk and validationHardsuperdayModel validationBank credit risk

   Say this

   Not on the Gini alone. Eight points of discrimination is worth having, but I'd need calibration, stability, explainability and fair-lending testing before approving it, and I'd want to know whether the gain survives out of time rather than just out of sample.

   Then walk it

   1. First question: is the eight points real? Check for leakage, which is the most common cause of a suspiciously strong challenger. Any feature that encodes the outcome, a post-application field, a collections flag, a date artefact, and the gain evaporates.
   2. Second: out of time, not just out of sample. Boosted models overfit to the period as well as to the sample. If the gain is eight points on a random split and two points on a later year, the story changes completely.
   3. Third: calibration. Gradient boosting ranks well and is often badly calibrated in the extremes, which is where pricing and provisioning live. Check the predicted-versus-observed curve by decile and consider isotonic or Platt scaling.
   4. Fourth: monotonicity and explainability. A credit model has to survive being explained to a customer who was declined and to a regulator. Unconstrained boosting can learn that higher income increases risk in some segment, which is a spurious interaction you can't defend. Monotonic constraints usually cost very little Gini and buy a lot of defensibility.
   5. Fifth: fairness. Test outcomes across protected characteristics and proxies for them. Complex models find proxies more efficiently than simple ones, so this risk genuinely rises with model power.
   6. Sixth: operational reality. Feature pipeline stability, retraining cadence, latency, reproducibility, version control, and whether anyone can support it in three years when the builder has left. Model risk includes the risk that nobody understands the production model.
   7. So my recommendation would be conditional approval with constraints: monotonic constraints on the key variables, calibration layer on top, capped score-level overrides, tightened monitoring thresholds, and the logistic model retained as a live benchmark. That is a real validation outcome rather than a yes or a no.
   8. And the commercial framing to say out loud: eight Gini points on a large retail book is worth real money, so the answer isn't to refuse complexity. It's to price the governance cost and decide deliberately.

   Where candidates lose it

   Picking a side. Reflexively rejecting machine learning makes you look like an obstacle; approving it on Gini alone makes you look like you've never validated anything. The answer is conditional approval with named conditions, and leakage plus out-of-time degradation are the two checks that must come first.

   Expect next

   • How would you test for leakage?
   • What would you tell a declined customer?
   • How much Gini would you give up for monotonicity?
2. 052How do you detect overfitting in a risk model?Model risk and validationIntermediatetechnicalModel validationBuy-side risk

   Say this

   The signature is a large gap between in-sample and out-of-sample performance, and for financial models specifically between in-sample and out-of-time. If accuracy collapses on data from a later period, you've fitted the era, not the relationship.

   Then walk it

   1. Standard test: hold out data, or cross-validate. But for time series you need forward-chaining rather than random k-fold, because random folds leak the future into the training set and make everything look good.
   2. The financial-specific test is out of time. Fit on 2015 to 2019, test on 2021 to 2023. Random splits from the same period share the same regime, so they flatter the model badly.
   3. Structural warning signs: parameter count relative to observations, coefficients with implausible signs or magnitudes, a variable that only works in one sub-period, and results that change materially when you drop a single year.
   4. For strategy or signal work, the killer is multiple-testing bias. If you tried 500 specifications, the best one will look great by construction. The deflated Sharpe ratio and the idea of testing on a truly held-out period exist precisely for that.
   5. Stability diagnostics: rolling-window coefficient estimates. A genuine relationship has coefficients that wobble; an overfitted one has coefficients that flip sign. And bootstrap the performance metric to see how wide the confidence interval actually is.
   6. The cultural control, which matters more than any statistic: limit how many times you look at the holdout. Every peek at the test set converts it into training data, and in practice that's how overfitting enters a model that passed every formal test.
   7. And the counterweight to state: underfitting is also a failure, and a heavily constrained model that misses real non-linearity costs money too. The judgement is whether added complexity buys performance that survives out of time.

   Where candidates lose it

   Naming cross-validation and stopping. For financial models, random cross-validation is itself a source of false confidence because of regime and look-ahead leakage. Out-of-time testing and multiple-testing bias are the two points that show you've done this on real data.

   Expect next

   • Why is random k-fold cross-validation dangerous for time series?
   • How would you account for having tried 200 specifications?
   • How would you tell overfitting apart from a regime change?
3. 053Give me Basel III in one minute.Regulatory capitalCorephone / first roundRegulatory reportingIndian bank risk and treasury

   Say this

   Basel III was the post-2008 response, and it did three things the earlier accords didn't: it raised the quality and quantity of capital, it added a leverage ratio as a non-risk-based backstop, and it introduced liquidity standards for the first time.

   Then walk it

   1. Capital quality: the focus moved to common equity Tier 1, real loss-absorbing equity. Minimum CET1 of 4.5 percent of RWAs, Tier 1 of 6, total capital of 8, plus a 2.5 percent capital conservation buffer, so a functioning bank runs at 7 percent CET1 minimum before any add-ons.
   2. Buffers on top: a countercyclical buffer of zero to 2.5 percent that supervisors raise in a boom, and a surcharge for global and domestic systemically important banks. Breaching the buffers doesn't close the bank, it restricts dividends and bonuses, which is the point.
   3. Leverage ratio: Tier 1 over total unweighted exposure, minimum 3 percent. It exists because risk weights were gamed before 2008 and banks ran 50-to-1 leverage with beautiful risk-based ratios.
   4. Liquidity, entirely new in Basel III. LCR requires 30 days of high quality liquid assets against stressed outflows. NSFR requires stable funding against illiquid assets over a year. Northern Rock was solvent, so capital rules alone were never going to be enough.
   5. Plus counterparty reforms: a CVA capital charge, higher standards for exposure modelling, and incentives to clear centrally.
   6. The Indian version: RBI applies CET1 of 5.5 percent rather than 4.5, plus a 2.5 percent conservation buffer, so minimum CRAR is 11.5 percent against Basel's 10.5. India has been consistently more conservative on the capital ratio and slower on some of the market risk pieces.
   7. The fair criticism to volunteer: complexity. The framework is thousands of pages, RWA calculations are barely comparable across banks, and that opacity is exactly what the leverage ratio and the Basel IV output floor were added to contain.

   Where candidates lose it

   Listing ratios without the three themes. What an interviewer wants is capital quality, a non-risk-based backstop, and liquidity standards. And know the Indian numbers if you're interviewing in India, because CRAR of 11.5 percent versus 10.5 is a detail that immediately places you.

   Expect next

   • Why add a leverage ratio if you already have risk weights?
   • What happens if a bank dips into its conservation buffer?
   • How does RBI's implementation differ?
4. 054What is CET1, and what qualifies as CET1 capital?Regulatory capitalCoretechnicalRegulatory reportingIndian bank risk and treasury

   Say this

   Common equity Tier 1 is the purest loss-absorbing capital: ordinary shares, share premium, retained earnings and disclosed reserves, minus a set of regulatory deductions. It's the numerator regulators actually care about, because it absorbs losses while the bank is still trading.

   Then walk it

   1. What's in it: paid-up ordinary share capital, share premium, retained earnings, accumulated other comprehensive income, and statutory reserves. Minority interests only in limited circumstances.
   2. The deductions are where the real work is: goodwill and other intangibles, deferred tax assets arising from losses, defined benefit pension surpluses, own shares held, significant investments in other financial institutions above thresholds, and the IRB shortfall of provisions against Basel expected loss.
   3. Why deductions matter so much: goodwill has no value in a liquidation, and a DTA from past losses is only worth something if you're profitable, which you aren't in the scenario the capital is for. So both get removed.
   4. The tiers above it: Additional Tier 1, which is perpetual and loss-absorbing through conversion or write-down, typically AT1 contingent convertibles that trigger when CET1 falls below 5.125 or 7 percent. Then Tier 2, mostly dated subordinated debt, which only absorbs loss in a gone-concern.
   5. That distinction between going-concern and gone-concern capital is the whole logic of the tiering, and it's the sentence that shows you understand it rather than having memorised a list.
   6. The 2023 reality check: Credit Suisse's AT1 was written down in full while shareholders received value in the UBS transaction. That inverted the expected hierarchy and repriced the whole AT1 market, and it's a live example of how gone-concern capital behaves under political pressure.
   7. Indian specifics: RBI requires CET1 of 5.5 percent, and Indian public sector banks have historically carried large DTAs and government recapitalisation bonds, so the deduction rules have a bigger effect on reported CET1 there than the headline ratio suggests.

   Where candidates lose it

   Listing what's included and skipping the deductions. The deductions are where CET1 differs from book equity, and goodwill plus DTA are the two that matter most. The answer that stands out explains going-concern versus gone-concern capital as the reason for the tiering.

   Expect next

   • Why is goodwill deducted?
   • What is an AT1 CoCo and when does it convert?
   • What did the Credit Suisse AT1 write-down change?
5. 055What are risk-weighted assets, and how are they computed?Regulatory capitalCoretechnicalRegulatory reportingBank credit risk

   Say this

   RWAs are the denominator of the capital ratio: exposures scaled by how risky they are. You compute them separately for credit, market and operational risk and add them up. A sovereign bond might carry a zero weight and an unsecured corporate loan 100 percent, so the same balance sheet size can imply very different capital.

   Then walk it

   1. Credit risk RWA, standardised approach: exposure times a prescribed weight by counterparty type and rating. Cash and most domestic sovereign zero, banks 20 to 100 depending on rating, residential mortgages 35 or lower under the revised rules, unrated corporates 100, and some specialised lending at 150.
   2. Credit risk RWA, internal ratings based: you feed your own PD, LGD and EAD into the Basel formula, which computes a 99.9 percent one-year unexpected loss and multiplies by 12.5. Same idea, but the weight is derived from your models rather than a table.
   3. Market risk RWA covers the trading book, now under FRTB with a sensitivities-based standardised approach or an internal models approach built on expected shortfall.
   4. Operational risk RWA under the standardised measurement approach: a Business Indicator Component from income and balance sheet size, scaled by an internal loss multiplier from your own ten-year loss history.
   5. Then the capital ratio is CET1 divided by total RWA. So there are two ways to improve it: raise capital or shrink RWA. RWA optimisation, shifting to lower-weighted assets, buying protection, improving collateral documentation, is a whole industry and a legitimate one within limits.
   6. The criticism: RWA density varies enormously across banks with similar books, largely because of IRB model differences. A European bank might run RWAs at 30 percent of total assets and a US bank at 60 for comparable risk. That comparability failure is why Basel IV added an output floor.
   7. Rough feel for scale: for a typical commercial bank, total RWA runs 50 to 70 percent of total assets, with credit risk 80 to 90 percent of the RWA total. Market risk is usually small unless there's a real trading book.

   Where candidates lose it

   Explaining the weights and never mentioning that banks can and do manage RWA down. An interviewer wants to hear both that RWA optimisation is a real activity and that its abuse is why the output floor exists. Also know the rough RWA-to-assets ratio, because it makes the number concrete.

   Expect next

   • How would a bank legitimately reduce its RWAs?
   • Why do RWA densities differ so much between banks?
   • What proportion of RWAs is credit risk for a typical bank?
6. 056Standardised versus internal ratings based approach. Which would you rather run?Regulatory capitalIntermediatetechnicalRegulatory reportingBank credit risk

   Say this

   IRB usually gives lower capital and better risk sensitivity, so a large bank wants it. But it costs a great deal to build, maintain and defend, and after the Basel IV output floor the capital saving is capped, so for many portfolios the honest answer is now standardised.

   Then walk it

   1. Standardised: prescribed risk weights by exposure class and external rating. Cheap, comparable across banks, transparent, and completely insensitive to whether your borrowers are good or bad within a bucket. Every unrated corporate gets 100 percent whether it's excellent or nearly insolvent.
   2. Foundation IRB: you estimate PD, the supervisor sets LGD and EAD. Advanced IRB: you estimate all three. Both need supervisory approval, years of clean data, validated models and demonstrated use in actual credit decisions, the use test.
   3. The capital saving is real, often 20 to 40 percent lower RWA on a good-quality retail or mortgage book, because the prescribed weights are calibrated conservatively for the average bank.
   4. The costs: model development and validation teams, data infrastructure with long histories, annual supervisory review, and the risk that a supervisor imposes a multiplier or pulls approval, which produces a sudden capital hit. Several European banks have taken exactly that.
   5. Then Basel IV changes the calculus. The output floor requires total RWA to be at least 72.5 percent of what the standardised approach would give, phased in. Advanced IRB has been removed for exposures to large corporates and banks, and equity IRB is gone. So the saving is bounded and the cheapest portfolios to model no longer qualify.
   6. My answer: IRB where the portfolio is large, homogeneous, data-rich and where internal models genuinely discriminate, so retail mortgages and retail lending. Standardised for low-default wholesale portfolios where you were never going to estimate PD credibly anyway. Running IRB for its own sake is a large cost for capped benefit.
   7. The point worth making that goes beyond capital: the real value of IRB was never the capital saving, it was that building the models forces a bank to understand its own credit risk. Banks that adopted IRB seriously ended up with better credit decisions, and that survives the output floor.

   Where candidates lose it

   Answering 'IRB because it's lower capital'. Post-Basel IV that's only partly true, and a candidate who hasn't registered the output floor and the withdrawal of advanced IRB for large corporates is working from pre-2017 knowledge. Mention the use test too; supervisors care more about it than about the maths.

   Expect next

   • What is the use test?
   • Which portfolios can no longer use advanced IRB?
   • What happens if a supervisor withdraws IRB permission?
7. 057What is Basel IV, and what changed for market risk?Regulatory capitalHardsuperdayRegulatory reportingBank market risk

   Say this

   Basel IV, formally the finalisation of Basel III, is about comparability rather than more capital. The headline is the 72.5 percent output floor on internally modelled RWA. For market risk it's FRTB, which replaced VaR with expected shortfall and drew a much harder line between trading and banking book.

   Then walk it

   1. The output floor: total RWA can't fall below 72.5 percent of the standardised calculation, phased in over several years. It caps the benefit of internal models and restores comparability between banks, which was the central complaint after 2008.
   2. Credit risk: advanced IRB removed for large corporates and financial institutions, IRB removed for equities, input floors on PD and LGD, and a more granular standardised approach with real loan-to-value sensitivity on mortgages.
   3. Operational risk: internal models abolished entirely, replaced by the standardised measurement approach driven by business indicators and your own loss history.
   4. FRTB for market risk, and the four things that changed. Expected shortfall at 97.5 percent replaces 99 percent VaR, so tail depth is captured. Liquidity horizons vary by risk factor from 10 to 120 days, so illiquid risk costs more capital.
   5. Third, non-modellable risk factors. If a factor lacks enough real price observations, you can't model it and it attracts a stress-based add-on. That was a large and unwelcome surprise for exotic and emerging market desks.
   6. Fourth, the trading and banking book boundary became prescriptive with restrictions on reclassification, ending the pre-crisis practice of moving positions to whichever book carried less capital. And the internal models approval is now at desk level with P&L attribution tests, so one desk can fail and lose modelled treatment while others keep it.
   7. Implementation dates have slipped repeatedly and differ by jurisdiction, with the US, UK and EU all on different timelines and different versions, especially for FRTB internal models. That fragmentation is itself a live commercial issue for global banks.
   8. The fair criticism: the aggregate capital impact is modest but very unevenly distributed, falling hardest on European banks with big IRB books and on trading desks in illiquid products. And the complexity it adds runs against the original goal of simplicity.

   Where candidates lose it

   Treating Basel IV as just higher capital requirements. The theme is comparability and constraining internal models, not level. And for market risk you need FRTB's specifics: expected shortfall, liquidity horizons, non-modellable risk factors and the desk-level P&L attribution test. Naming only the first shows shallow reading.

   Expect next

   • Why did FRTB move to expected shortfall?
   • What is a non-modellable risk factor and what does it cost?
   • What happens when a desk fails P&L attribution?
8. 058Why does a leverage ratio exist alongside risk-weighted capital?Regulatory capitalIntermediatetechnicalRegulatory reportingBank market risk

   Say this

   Because risk weights are model outputs and models can be wrong or gamed. The leverage ratio is a non-risk-based backstop: Tier 1 over total exposure, minimum 3 percent, and it doesn't care what you think the risk is.

   Then walk it

   1. The pre-crisis evidence is the whole argument. Banks entered 2008 with comfortable risk-based ratios and leverage of 30 or 50 to one, because sovereign debt, AAA tranches and repo books all carried tiny weights and turned out not to be riskless.
   2. So the design intent is a floor that survives being wrong about risk. It binds when a bank holds a lot of assets it has judged safe, which is exactly the situation that has historically preceded trouble.
   3. The exposure measure is deliberately broad: on-balance-sheet assets, derivative exposures including a potential future exposure add-on, securities financing transactions, and off-balance-sheet commitments converted at credit conversion factors. You can't shrink it by netting the way you can for RWA.
   4. Who it binds: banks with large low-risk-weight books. Custodians, repo intermediaries, and banks holding large government bond portfolios. For those, leverage rather than RWA is the constraint that drives the business decision.
   5. Its own weakness, and you should say it: it's risk-insensitive by construction, so it treats a treasury bill and an unsecured emerging market loan identically. That creates an incentive to shift toward higher-yielding, higher-risk assets once the ratio binds, which is the opposite of what you want.
   6. So the two measures are deliberate complements. Risk weights give you sensitivity and can be gamed; leverage gives you robustness and rewards risk-taking at the margin. Neither alone is adequate, which is the point of having both.
   7. Live example: in 2020 several jurisdictions temporarily excluded central bank reserves from the exposure measure, because deposit inflows and QE were inflating the denominator and constraining lending. That's a good illustration of the ratio binding for reasons unrelated to risk.

   Where candidates lose it

   Stating the definition without the pre-crisis motivation, and without the downside. A complete answer says the leverage ratio pushes banks toward riskier assets at the margin, because the interviewer wants to see you can criticise a rule you also support.

   Expect next

   • Which kinds of bank does the leverage ratio bind?
   • What perverse incentive does it create?
   • Why did supervisors exclude central bank reserves in 2020?
9. 059A bank reports CET1 of 11 percent against a 9 percent requirement. Is it safe?Regulatory capitalHardsuperdayRegulatory reportingBank credit risk

   Say this

   Not from that number alone. A capital ratio tells you about solvency under the RWA model, and banks fail from liquidity and from concentration, not from a ratio. I'd want to know the composition of the denominator, the funding profile and the trajectory before answering.

   Then walk it

   1. First, what's the 9 percent made of? Pillar 1 minimum, plus the conservation buffer, plus Pillar 2, plus any systemic surcharge. If the 9 percent is mostly buffer, breaching it restricts dividends rather than triggering resolution, which is a different kind of 2 percent of headroom.
   2. Second, interrogate the denominator. RWA density against total assets, how much is IRB-modelled, and single-name and sector concentration. An 11 percent ratio on a book with 25 percent in one sector is far weaker than the same ratio on a granular one, and the IRB formula won't show it.
   3. Third, the trajectory, which is what actually matters. Was it 13 percent two years ago? Is it falling through loan growth, buybacks or rising provisions? The direction and the stress path matter more than the level.
   4. Fourth, and this is the real answer, liquidity. SVB had a capital ratio comfortably above requirement the week it failed. Look at LCR, NSFR, deposit concentration, the uninsured deposit share, and unrealised losses in held-to-maturity securities that don't touch CET1 until they're sold.
   5. Fifth, asset quality and provision adequacy. Coverage ratio, NPL ratio, Stage 2 share, and whether provisioning looks light relative to peers. A thin provision stock means the capital ratio is borrowing from the future.
   6. Sixth, the stress result. What does CET1 do in the adverse ICAAP or supervisory scenario? If it drops to 8 percent, the 2 percent buffer is already spoken for and the bank is effectively at its constraint.
   7. So my answer would be: 11 against 9 is adequate headroom on a granular, well-funded, well-provisioned book with a stable trajectory, and thin on a concentrated book with a volatile funding base. And I'd say what I'd need to see rather than guess, because the interviewer is testing whether I'll commit to a number without the information.

   Where candidates lose it

   Answering yes or no. There isn't enough information, and the interviewer is testing whether you know that solvency ratios don't capture liquidity or concentration. SVB is the example that proves it, and naming unrealised held-to-maturity losses is the detail that lands.

   Expect next

   • What would you want to see to be comfortable?
   • SVB had a fine capital ratio. Why did it fail?
   • Does breaching the buffer requirement mean the bank fails?
10. 060Distinguish funding liquidity risk from market liquidity risk.Liquidity risk and ALMCoretechnicalTreasury and ALMBank market risk

    Say this

    Funding liquidity risk is not being able to meet your obligations as they fall due. Market liquidity risk is not being able to sell an asset at anything near its marked price. They're different risks, and the danger is that each one triggers the other.

    Then walk it

    1. Funding liquidity is a balance sheet and cash flow problem: deposits leave, a wholesale line isn't rolled, a margin call lands, and you need cash today. It's binary and it's fatal. You are either able to pay or you are not.
    2. Market liquidity is a price problem: bid-offer, depth, and how far the price moves against you when you try to sell size. It's continuous, and it shows up as a haircut on what your book is really worth.
    3. The spiral is the real answer. You need funding, so you sell assets. Selling into a thin market depresses the price. The lower mark reduces your collateral value and your capital, which makes funding harder, so you sell more. That's the liquidity spiral, and it's what turned 2008 from a credit event into a systemic one.
    4. Measurement differs completely. Funding liquidity: contractual and behavioural cash flow ladders, survival horizon, LCR and NSFR, and a stress test on deposit outflow. Market liquidity: bid-offer spreads, days of average daily volume to exit, and a liquidity-adjusted VaR or an exit-cost haircut.
    5. Worked example: a bond book marked at 100 crore, where the position is ten days of average volume. In a stress you might realise 92, so the honest liquidity-adjusted value is 92, not 100. The mark is not the exit price, and that 8 crore is the market liquidity risk in money.
    6. FRTB codified this by making liquidity horizons vary from 10 to 120 days by risk factor, so illiquid risk now costs more capital. That's the regulatory acknowledgement that a mark is not a price you can get.
    7. The thing to say without prompting: almost every bank failure is ultimately a funding liquidity failure. Solvency problems kill banks slowly and liquidity kills them in a week.

    Where candidates lose it

    Conflating the two, or giving definitions without the interaction. The answer that earns respect explains the spiral in both directions and says that a marked price is not an exit price. And naming that banks fail from liquidity, not capital, frames everything else you say.

    Expect next

    • How would you measure market liquidity risk in a bond book?
    • Which one killed more institutions in 2008?
    • How does FRTB handle illiquidity?

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.