Equity Research Report: Cover the Rating, Read the Rest

An equity research report is a study of one listed company with a rating and a price target on the cover, usually pinned to the next twelve months. That cover is the only part of the document nobody can verify today. The remaining thirty pages you can check line by line, and not because the author was generous, but because the regulator made him print them.

Take an equity research report example apart#

The rating is the last paragraph of the analyst's work, even though it gets printed at the top. Underneath it sit three sections, and they take up most of the page count.

The first is financial. Several years of revenue, margins and operating cash flow taken from the company's own filings, alongside the shape of the balance sheet, the debt maturity ladder, and the gap between accounting profit and cash that actually arrived. That last gap is often the most interesting number in the whole report. Say a company shows $200m of net income and $120m of operating cash flow. Sixty cents of cash for every dollar of profit, for the third year running, stops being an accounting detail and becomes a question about whether the profit exists outside the table.

A good report asks that on page eight, not in the rating.

The second section is management. What they promised on the last eight earnings calls, how much of it arrived, and how their pay is put together. Eight quarters is the minimum, because only over that stretch can you tell a management team that adjusts its plan from one that rewrites it every quarter. A single CEO quote proves nothing, since speaking well is part of the job description.

The third is competitive position. Who the company competes with for the same customer spend, and whether its margin is higher because the firm can do something others cannot, or because the industry is having a good year. That distinction only shows up against peers and across a full cycle, so a competition chapter that names rivals without comparing a single number is decoration.

Beyond those three, a report may add segment economics, customer concentration, regulatory exposure or a bond maturity schedule. Those are extras, sometimes very good ones.

Without the three, you have an opinion with a ticker in it.

You ask the same set of questions when you analyze a company yourself. The difference is that in a report somebody has already done the work and left a trail you can follow to see whether they did it well.

What the word "buy" actually means#

The rating is not an empty word. Every house sets its own scale and writes down what it means by it. At one firm "buy" means we expect somewhere in the mid teens over twelve months. At another the same label means only that the company will do better than its sector, even if the whole sector falls.

Then there are the vocabularies. One house runs five steps with "accumulate" and "reduce" in the middle, another runs three, a third replaces the sell rating with "underweight", which is a statement about portfolio weight rather than about the company. Comparing bare labels across houses without opening their definitions is comparing grades from two different schools.

The advantage is real, though. A rating compresses a multi-page model into one word you can hold against the current quote.

Say the stock trades at $40 and the target in the report is $52. Twelve dollars of difference, thirty percent above the quote, and against a mid-teens threshold the label fills itself in, even if the analyst privately thinks his own assumptions are stretched.

Fair enough. Except that this number is the only sentence in the report that talks about the future.

The rest of the document makes claims about things that have already happened: this is what the margin was two years ago, this is what the CEO said in the second quarter, this much debt matures in two years. Every one of them can be verified today, with no waiting. The price target is the output of a model whose inputs concern years that have not happened yet, so it cannot be checked now, and once the horizon closes, almost nobody goes back to it.

A rating has an expiry date that hardly anyone looks at.

Looking takes two seconds, because the rules require the date and time the recommendation was completed to be stated on it. A report from five months ago, with two earnings calls in between, describes a company that no longer exists in that form. The business chapters age slowly and the valuation ages fast, which is why the same thirty pages can still be useful when the front page is already stale.

What this document is not#

Research written for a wide audience sits in a different legal category from advice given to one client, and the law draws that line sharply. Advice takes a single person into account: their goals, horizon, tax position, tolerance for loss, and what they already hold.

The report knows nothing about you.

So it cannot know whether a twelve-month horizon suits you, or whether you can afford to be wrong alongside the author for that long. Those two questions stay on your side of the table, and no number on the cover closes them for you.

It is also not a tip. A tip runs on conviction and gets shorter the less there is behind it. A report works the other way round: it shows its steps and exposes itself to being checked, because that is the only way it can argue with a reader at all.

Four things the regulator forces onto the page#

How do we know the rest of the report is checkable? Because somebody wrote it down. The EU's Delegated Regulation 2016/958 lists what a published recommendation must contain, and each of the four items on that list is a tool for the reader rather than a burden on the author.

The second row is worth the most, because a valuation model looks precise while standing on three or four assumptions that actually move it: revenue growth, the target margin, the discount rate, and whatever the model assumes after the last forecast year. Substitute your own numbers and run it again. If cutting growth from eight percent to four halves the target, this is not a report about the company, it is a report about one assumption dressed as a report about the company.

The first row does something less obvious. Since the author has to name sources, every strong claim in the report can be traced back to the original, and a claim that cannot be traced stands out from the rest like an answer with no working shown.

Source discipline is a property of the document, then, not of the author's conscience.

The fourth row is the one authors like least. The regulation makes an investment firm publish, every quarter, the proportion of ratings it has issued over the previous twelve months, and next to it the share of issuers in each category to which it supplied material investment services. That one table says more about a rating's credibility than any statement of independence.

Who pays for the other thirty pages#

A ratings distribution is not a statistical curiosity; it is a variable that predicts how good the ratings are. Barber, Lehavy, McNichols and Trueman compared the directions in a 2006 paper in the Journal of Accounting and Economics. Upgrades issued by brokers with the smallest share of buy ratings went on to do better than the same upgrades from brokers where buy is the default. Which means the identical label carries different information depending on how freely a firm hands it out, and without the distribution you do not know which of the two you are holding.

So where does the money for those thirty pages come from? Until 3 January 2018, research travelled inside trading commissions: the fund paid for execution and the reports arrived alongside. From that date Delegated Directive 2017/593, article 13 requires the two to be separated. Research has to be paid for on its own: out of the investment firm's own resources, or from a ring-fenced account funded by a charge to the client that is not linked to trading volume.

The intent was sound and partly worked: once research carries a price, you can see what it costs and whether it earns that price. The side effect was equally predictable. Once it carries a price, it gets bought where it pays for itself, and it pays for itself where a lot of stock changes hands.

Smaller companies lost coverage first.

Hence a third funding model, which the Warsaw exchange has been running for several years. Under its analytical coverage support programme, reports on companies the exchange itself describes as medium-sized and less liquid are commissioned and paid for by the exchange, and the fifth edition covers 65 main-market companies. Which means 65 companies get written up at all, on research that commissions on trading their shares would not fund for a year.

Before that sounds like an accusation, in fairness: a report paid for by an exchange or an issuer is not automatically worse than one paid for by a fund. A sell-side analyst has something you do not. He goes to investor days, calls the company's suppliers, has sat in one industry for ten years and knows which question on the call was the genuinely hard one. Who pays does not settle whether a report is good. It settles which way the author will be wrong, if he is wrong.

I do not know whether any particular analyst gave in to that pressure, and from the outside it cannot be established. I do know that his firm's ratings distribution takes fifteen minutes to look up.

Where to find an equity research report for free#

Reports from the exchange programme sit publicly on the exchange's site and on the sites of the firms that write them, which makes them the cheapest place to start. Brokerage research usually reaches clients first, and what reaches the press is one sentence and one number, which is precisely the part with the least information in it.

The material those reports are built from is free to everyone. US filings sit in the SEC's EDGAR database, current and periodic reports for European issuers sit in exchange announcements and on investor relations pages, and earnings call transcripts sit with the companies themselves. If you would rather start there than with someone else's summary, primary sources in stock research sets out what each document is and the order to read them in.

Does that mean the reports are not worth reading? The opposite. It means you read them the way you read somebody else's homework next to your own working, rather than instead of it.

Ten minutes and three moves#

You have a document and ten minutes, so do not read it front to back. Do three things with it:

1. Cover the front page. If, with the rating and the target hidden, you cannot say how this company makes money and what would have to break for it to stop, the report has not done its job.
2. Find the assumptions page and change one input. Half the revenue growth, two points off the margin. See how much of the valuation survives.
3. Pick the three strongest claims and follow them to a document. Land on a filing or a transcript and you are reading research. Land on an article citing another article and you are reading commentary in a report's cover.

If the second move is impossible because there is no assumptions page, there is no point going further either. A valuation you cannot recompute is a number to be taken on faith, and faith is not what you came to an equity research report for.

The third move settles it most often, which is why I treat citations in AI-written research as a precondition rather than a garnish. And a bias I will admit to: I trust a report that shows me an assumption I can disagree with more than one that shows me a conclusion I am meant to agree with.

A Full report in Taufolio runs in that order: financials, management and competitive position, a thesis broken into assumptions you can test, every meaningful number carrying a link to the document it came from, and no trading call on the cover, because there isn't one. Full report pro adds the Three models and a judge mechanism: three models write the same analysis independently, a fourth compares the versions and keeps only what a source can carry. A free account gets 100 credits a month and a Full report costs around 100 credits, so you can run those three moves on our document before you run them on anyone else's. If you are unsure which depth to start at, the two levels of company report differ in depth rather than in source discipline, and you can see how a Full report is put together on the product page.