Equity Research interview preparation

Questions

100

Traced to a firm

72

Firms

45

Updated

September 2026

Asked at

All firmsMorningstar12Man Group6Balyasny Asset Management5BLBlackRock5FTFranklin Templeton5MSCI5Jefferies4CSCredit Suisse3Fidelity Investments3Moody's3Perella Weinberg Partners3Point723S&P Global3The Vanguard Group3WMWellington Management3Advent International2Apollo Global Management2Bank of America2Carlyle Group2DED.E. Shaw2Houlihan Lokey2HSBC2Piper Sandler2Sequoia Capital2SSState Street2Viking Global Investors2WBWilliam Blair2ACAQR Capital Management1BGBaupost Group1BMBNY Mellon1Centerview Partners1Coatue Management1Goldman Sachs1GSGuggenheim Securities1HWHarris Williams1Insight Partners1Invesco1Mizuho1Moelis & Company1MSMorgan Stanley1PIMCO1SCSchroders1Scotiabank1T. Rowe Price1TSTruist Securities1

Topic

All topicsResearch process9Stock pitch6Company analysis8Investment philosophy5Valuation14Modelling2Portfolio and risk8Macro8Sector knowledge2Accounting8Career and fit12Industry knowledge6Quantitative research1Sector: technology3Sector: consumer1Sector: healthcare1Sector: energy1Sector: financials2Sector: industrials1Case and estimation2

Level

AnyCoreIntermediateHard

Type

AnyTechnicalCaseFitBrainteaserMarket view

1. 081How would you initiate coverage on a new company?Research processIntermediatetechnicalSell-side research

   Say this

   Read the last three years of filings and transcripts, build the model from drivers, map the competitive landscape, talk to the company and the channel, then decide what your differentiated view is before writing a word.

   Then walk it

   1. Primary documents first: three years of annual filings, the last eight quarterly transcripts, the investor day materials and the proxy for incentives. The transcripts tell you what management has promised and how the questions have changed.
   2. Build the model from drivers and reconcile it to reported history. If you cannot rebuild the last two years from your drivers, your model is wrong.
   3. Map the industry: who competes, what share each has, how the value chain splits economics, what the customers care about. Read the competitors' filings, because they describe your company from the outside.
   4. Channel work: customers, distributors, former employees, industry consultants. This is where a differentiated view most often comes from.
   5. Then form the thesis. An initiation with no variant view is a description, and nobody reads it. Decide what you believe that consensus does not, and structure the note around defending it.
   6. Then the deliverable: rating, target, earnings forecasts that differ from consensus in a specific place, the key debates set out fairly, and the risks. And a clear statement of what would change your mind.

   Where candidates lose it

   Describing a document-gathering exercise with no thesis. An initiation is judged on whether it says something. Leading with the variant view rather than the process is the answer that sounds like an analyst.

   Expect next

   • How long would that take you?
   • Where does the differentiated view usually come from?
   • How would you handle initiating with a sell rating?
2. 082How would you handle publishing a sell rating on a company your bank has a relationship with?Career and fitHardsuperdaySell-side research

   Say this

   Publish it, and rely on the structures that exist for exactly this: research is separated from banking, the rating is based on documented analysis, and compliance reviews it. The answer is process, not courage.

   Then walk it

   1. Start with the structural point: research and investment banking are separated by information barriers, and the analyst's rating is not subject to banking sign-off. That separation exists because of past abuses and is enforced by regulation.
   2. Then the professional standard: the rating must follow the analysis, and the analysis must be documented so it can be defended. If the work supports a sell, the rating is a sell.
   3. Then the practical handling: make sure the note is factually impeccable, put the reasoning in the open, and give the company the chance to correct factual errors, not conclusions.
   4. Acknowledge the real cost honestly, because pretending there is none is naive: you may lose management access, which degrades your product. That is a genuine professional cost and it is why the skew exists.
   5. Then the escalation path: if you were pressured, you raise it with compliance and supervisory analysts. Knowing there is a channel is the answer they want.
   6. And the credibility argument: an analyst who never publishes a sell has a rating scale worth nothing. The value of your buy recommendations depends on your willingness to say sell.

   Where candidates lose it

   Either an idealistic 'I would just publish it' with no awareness of the structures, or a suggestion that you would soften the view. Name the information barrier and compliance explicitly; this is partly a regulatory-awareness question.

   Expect next

   • What if a senior banker called you about it?
   • Why do so few sell ratings get published?
   • How do you maintain company access after a downgrade?
3. 083How would you value a company with a large stake in a listed subsidiary?ValuationHardtechnicalIndian research desks

   Say this

   Sum of the parts. Value the core business on its own operating metrics, then add the market value of the listed stake, usually at a holding company discount, and subtract net debt at the parent.

   Then walk it

   1. Value the core operating business separately, using only its own earnings. This means stripping out any consolidated contribution from the subsidiary, which is the step people get wrong.
   2. Value the stake at its observable market value. That is the cleanest input in the whole exercise, so use it rather than modelling the subsidiary again.
   3. Apply a holding company discount, typically 20 to 50 percent, to reflect tax on disposal, the fact that the parent will probably never sell, and the governance discount investors apply to conglomerate structures.
   4. Subtract parent-level net debt and any other claims to get to equity value.
   5. Watch the consolidation treatment carefully. If the subsidiary is consolidated, its revenue and EBITDA are in the group numbers, so applying a group multiple double-counts the stake. Either deconsolidate or do not add the stake.
   6. This is a very common structure in India and Korea, where promoter-led holding companies own listed operating subsidiaries. The persistent discount is one of the most reliable features of those markets and also one of the most persistent value traps, because the discount rarely closes without a structural event.

   Where candidates lose it

   Double-counting by applying a group multiple to consolidated earnings and then adding the market value of the stake. That is the classic error and it is why this question gets asked.

   Expect next

   • What discount would you apply and why?
   • What would cause the discount to close?
   • How does minority interest affect your bridge?
4. 084What is minority interest and why does it appear in enterprise value?ValuationIntermediatetechnicalBulge bracket IB

   Say this

   It is the portion of a consolidated subsidiary the parent does not own. You add it to enterprise value because the consolidated EBITDA includes 100 percent of that subsidiary, so the numerator must reflect 100 percent too.

   Then walk it

   1. Accounting: if a parent owns more than 50 percent it consolidates the whole subsidiary, taking all of its revenue and EBITDA, then deducts the minority's share of profit below the line.
   2. So consolidated EBITDA overstates what belongs to the parent's shareholders.
   3. To keep the multiple consistent, you add minority interest to enterprise value. Both numerator and denominator then represent the whole enterprise, including the part owned by others.
   4. Use the market value of the minority if the subsidiary is listed. Book value is the fallback and is usually a poor estimate.
   5. The alternative approach is to deconsolidate: strip the subsidiary's EBITDA out and value the parent's stake separately. Cleaner conceptually, more work, and it is what you do when the subsidiary is very different from the core business.
   6. The error to avoid is forgetting it entirely. If you ignore minority interest, a company that consolidates a large partly-owned subsidiary will look artificially cheap on EV/EBITDA, and that appears constantly in emerging market comp sets.

   Where candidates lose it

   Knowing the rule but not the reason. The reason is consistency between numerator and denominator, and being able to state that is what shows you understand enterprise value rather than having memorised the bridge.

   Expect next

   • Should you use book or market value for it?
   • When would you deconsolidate instead?
   • How does this distort a comp set?
5. 085How would you analyse a company with related-party transactions or a controlling shareholder?Company analysisHardtechnicalIndian research desksEmerging market research

   Say this

   Treat governance as a valuation input rather than a footnote. Map every related-party flow, ask whether value is leaking out of the listed entity, and apply a discount if minority shareholders are not being treated equally.

   Then walk it

   1. Read the related-party transactions note in full and map the flows: who sells what to whom, at what price, and is there an independent benchmark for that price.
   2. The question is always whether the listed entity is transacting at arm's length. Buying raw materials from a promoter-owned entity above market, or selling output below it, transfers value out of the listed company.
   3. Watch for the classic structures: royalty or brand fees paid to the parent, shared services agreements, loans and guarantees to group companies, and asset purchases from related entities.
   4. Check pledged promoter shares. A promoter who has pledged a large proportion of their holding has an incentive problem and a forced-selling risk that can hit the stock independently of fundamentals.
   5. Look at the board: how many genuinely independent directors, who the auditor is, and whether auditors have resigned. An auditor resignation is one of the strongest negative signals available.
   6. Then price it. A governance discount is real and persistent, so the honest output is a lower multiple rather than a refusal to cover. But if you cannot verify that the cash belongs to minority shareholders, the correct answer is to avoid it, and saying that is a legitimate conclusion.

   Where candidates lose it

   Treating governance as a qualitative aside. In emerging markets it is frequently the dominant driver of returns. Naming pledged shares and auditor resignations shows real familiarity with how these situations actually unfold.

   Expect next

   • What is a promoter pledge and why does it matter?
   • How would you size a governance discount?
   • Would you ever refuse to cover a company?
6. 086Give me a two-line thesis on a company you would short.Stock pitchHardsuperdayBalyasny Asset ManagementGeneralist · New York · 2020

   Say this

   One line on the structural problem, one line on the catalyst that forces the market to see it. Something like: the company's growth is funded by discounting that is destroying its unit economics, and the covenant test at the next refinancing will expose it.

   Then walk it

   1. Line one has to be a mechanism, not a valuation. 'Expensive' is not a thesis. 'Gross margin has fallen 600 basis points over six quarters while revenue growth held up, because they are buying volume' is a mechanism.
   2. Line two is the clock: the specific event that makes the market agree. A refinancing, a covenant test, a contract renewal, a patent expiry, a competitor launch, a change in the disclosure that removes the cover.
   3. Then the numbers that support it, in one breath: the trend in the metric, and the gap between what management guides and what the trend implies.
   4. Then the risk: what would squeeze you. A cheap balance sheet, a possible takeover, heavy existing short interest, or a founder who could take it private.
   5. And the practicalities: borrow cost and days to cover, because a 20 percent borrow makes a slow thesis unprofitable even if you are right.
   6. Prepare a real one before you walk in. Being unable to construct a short is a common failure in buy-side interviews, and it reveals that you have only ever thought about why things go up.

   Where candidates lose it

   Not having one prepared. Long-short interviews ask for both sides, and candidates almost always have three longs and no shorts. Prepare one short properly, including the borrow cost and the squeeze risk.

   Expect next

   • What is the borrow?
   • What would squeeze you?
   • How would you size it?

   Reported by candidates at Balyasny Asset Management (Generalist, New York, 2020). Source: Wall Street Oasis.

7. 087Tell me about a time you were wrong about a stock.Career and fitIntermediatesuperdayMorningstarEquity Research · Chicago · 2023Point72Investment Research · New York · 2026

   Say this

   Give a real position, state the thesis you held, say what actually happened, and identify the specific analytical error rather than blaming the market. Then the process change it caused.

   Then walk it

   1. State the original thesis in one sentence, exactly as you held it at the time. Reconstructing it charitably in hindsight is obvious and undermines the whole answer.
   2. Then what happened and what you missed. Be specific about the type of error: you overestimated pricing power, you trusted a management forecast, you ignored the balance sheet, you anchored on the purchase price.
   3. Distinguish a bad decision from a bad outcome. Some losses come from good process and bad luck; others from process failure. Showing you can tell them apart is the highest-value part of the answer.
   4. Say what you did when the evidence turned. Did you cut, add, or freeze? Freezing is the honest answer for most people and admitting it is fine if you explain what you now do instead.
   5. Then the process change, concretely. 'I now write the falsifier down when I initiate and check it every quarter' is better than 'I learned to be more careful'.
   6. Avoid stories where you were secretly right and the market was wrong. That is not an answer about being wrong.

   Where candidates lose it

   Choosing a loss you can blame on an external shock. That avoids the question. Pick one where the error was yours and where the lesson changed a specific habit.

   Expect next

   • Was that a bad decision or a bad outcome?
   • What do you do differently now?
   • How long did it take you to change your mind?

   Reported by candidates at Morningstar (Equity Research, Chicago, 2023); Point72 (Investment Research, New York, 2026). Source: Wall Street Oasis.

8. 088How do you avoid confirmation bias in your research?Research processHardsuperdayHedge fundsAsset management

   Say this

   Build the disconfirming case deliberately rather than waiting to encounter it. Write the bear case before you buy, define in advance what would falsify the thesis, and seek out the best argument against you.

   Then walk it

   1. Pre-commit the falsifier. When you initiate, write down the specific observable outcome that would prove you wrong, with a date. Then you cannot rationalise it later.
   2. Write the opposing case yourself, properly, not a straw man. If you cannot write a persuasive bear case, you do not understand the stock well enough to own it.
   3. Actively read the other side: the short reports, the bearish sell-side note, the competitor's investor day. Seek out the smartest person who disagrees.
   4. Structure the review so it starts from the evidence rather than from your note. Re-underwriting the position from scratch once a year, ignoring what you previously wrote, is the most effective single habit.
   5. Use a team process: have someone else argue the other side, or present the bear case yourself to the portfolio manager. Making the disconfirming work someone's explicit job is how good funds handle it.
   6. And watch for the behavioural tell: noticing that you are discounting bad news because it is inconvenient. Catching that in yourself is most of the battle.

   Where candidates lose it

   Giving a generic 'I try to stay objective'. Everyone believes that; that is what makes the bias work. The answer needs specific mechanisms, pre-commitment and structured disconfirmation, not intentions.

   Expect next

   • What would falsify your current best idea?
   • How do you handle it when a position moves against you?
   • Who do you go to for the other side?
9. 089What is the case study process here, and how would you approach a two-week modelling test?Career and fitIntermediatecase studyPoint72Investment Banking · London · 2026DED.E. ShawGeneralist · New York · 2025

   Say this

   Treat it as a recommendation, not a model. Spend the first day deciding what the investment question is, then build only the model you need to answer it, and reserve the last quarter of the time for the write-up.

   Then walk it

   1. Identify the decision first. What is the one variable this investment turns on? Everything you build should serve answering that.
   2. Build a model sized to the question. A beautiful 20-tab model that does not resolve the debate scores worse than a three-tab model that does. Graders are checking judgement about materiality.
   3. Do primary work. Transcripts, competitor filings, industry data, and anything you can verify externally. This is what separates entries, because everyone can build a model.
   4. Structure the output like a note: recommendation and target up front, the variant view, the evidence, the valuation, the risks and the falsifier. Never make the reader hunt for your conclusion.
   5. Sensitise honestly. Show the bear case with a number attached, and say what you are least confident about. Overclaiming certainty is the most common way candidates lose credibility.
   6. Then rehearse defending it out loud, because the presentation is usually where the decision is made. Expect them to attack your weakest assumption, and have the downside quantified before they ask.

   Where candidates lose it

   Spending all the time on the model and writing the conclusion in the last hour. The model is the working; the recommendation is the product. Budget time backwards from the write-up.

   Expect next

   • How did you get your assumptions?
   • What is the bear case worth?
   • What would you have done with another week?

   Reported by candidates at Point72 (Investment Banking, London, 2026); D.E. Shaw (Generalist, New York, 2025). Source: Wall Street Oasis.

10. 090How did you arrive at the assumptions in your case study?Research processHardcase studyDED.E. ShawGeneralist · New York · 2025Houlihan LokeyInvestment Banking · Richmond · 2025

    Say this

    Each assumption should trace to something external: a historical rate, a disclosed contract, an industry data point, a peer's experience. Name the source for each, and say which ones you are least confident about.

    Then walk it

    1. Go through them in order of importance to the answer, not in model order. The interviewer cares about the two that drive the result.
    2. For each, give the anchor: 'I used 6 percent price growth because that is what they have taken in each of the last four years and the contracts reprice annually to an index.'
    3. Where you had no data, say so explicitly and explain the logic you substituted. Inventing a source is fatal; reasoning openly from a gap is respected.
    4. Distinguish the assumptions that matter from the ones that do not. 'The tax rate assumption is immaterial; the retention assumption drives 70 percent of the value' shows you understand your own model.
    5. Present the sensitivity around the critical ones rather than defending a point estimate. The honest position is a range with a most likely case.
    6. And volunteer your least confident assumption before they find it. Doing so converts a vulnerability into evidence of self-awareness.

    Where candidates lose it

    Defending every assumption equally, or saying 'that is what management guided'. Guidance is an input to be tested, not a source of truth. Trace assumptions to independent evidence wherever possible.

    Expect next

    • Which assumption are you least confident about?
    • What if that assumption is 20 percent wrong?
    • Where did you disagree with management's guidance?

    Reported by candidates at D.E. Shaw (Generalist, New York, 2025); Houlihan Lokey (Investment Banking, Richmond, 2025). 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.