Look at enough congressional disclosures and a pattern appears that feels like a discovery. The members from a state keep showing up in the same handful of companies, and those companies happen to be headquartered in that state or to run the biggest plant in the district. It looks like an inside track. Most of the time it is the least surprising thing in the dataset.
People everywhere overweight what is nearby and familiar. That is one of the sturdier findings in the household finance literature, it shows up in ordinary brokerage accounts with no political dimension at all, and members of Congress are ordinary investors in this specific respect. So the question is never whether home-state concentration exists. It does. The question is whether a particular instance of it carries information, and there are two checks that answer that.
The view the module does not give you
Start with an honest statement of what you have to work with, because the shortcut does not exist. The Heatmaps tab carries three views, Sector, Ticker and Committee Correlation, with a window selector reading 30 days. There is no state or delegation view. You cannot press a button and see what the members from one state did.
That means this analysis is manual, and it should be, because the manual version forces you to look at the individual rows rather than at an aggregate that hides them. The workflow is: use the Ticker view to see which names carried the month's activity, take the handful you care about, and then look up who filed them and where those members are from. It is fifteen minutes for one ticker and it does not scale, which is a feature. You are not going to run this on fifty names, so run it on the two or three where the answer would change what you do.
The three view buttons above the tiles are worth knowing individually. Sector aggregates the window by industry. Ticker regroups the same window by name, which is where you start when you want to know who filed a specific company. Committee Correlation is the one that speaks to the question of whether a member's assignments line up with what they trade, and it is a better tool for the informational question than any geographic slice would be.
Why utilities and regional industrials dominate the pattern
The sectors where home bias shows up hardest are the ones whose businesses are geographically defined, and the tiles show it. Utilities on my capture listed AEE, CMS, AWK, NEE and PEG. A regulated utility serves a state or part of one. If a member holds one, the overwhelmingly likely explanation is that it is the utility that serves their house, that it has been in the family account for twenty years, and that the local broker recommended it in the nineties.
The Industrials tile listed OC, CTAS, ITW, NVT and JCI, and industrials carry the same flavour for a different reason. These are companies with large physical footprints, which means a district with a plant, which means a member who knows the name because it employs their constituents and hosts their factory tours.
Neither of those is a trading signal. Both of them will produce exactly the same disclosure pattern that a genuinely informed position would produce, which is why looking at the pattern alone gets you nowhere. What separates them is behaviour over time and behaviour around events, and that is what the two checks measure.
Check one, is it a position or a habit
Pull the filer's history on the ticker and ask a single question. Is this a first purchase, or is it the twentieth entry in a long series?
A member who has bought the same home-state utility in small amounts every few months for six years is running a dividend reinvestment habit. There is no decision embedded in the most recent one. A member who has never filed a company before and suddenly initiates a position in it is making a decision, and a decision is the only thing worth reading. The distinction between an initiation and a top-up is invisible on a single row and obvious across a history, and it is the cheapest filter in this entire exercise.
Two things make this harder than it sounds. The disclosed amounts are ranges, and the module reports average trade size at $40K under the label Midpoint USD, so you are comparing midpoint estimates rather than actual dollars and a small top-up can look like a meaningful add. And disclosure covers a household, not a person, so a spouse's employer stock plan will file under the member's name and will produce a metronomic series of purchases that has nothing to do with the member's views at all. If the cadence looks like a payroll schedule, it probably is one.
Check two, does the timing sit near anything
The second check asks whether the trade landed near a company or policy event, and it has to be run in the strict order or it produces a story every time.
Write down the events first. Before you look at the trade date, list what was publicly scheduled for that company and that industry in the relevant period, from the company's own calendar and from committee schedules. Then compare. If the purchase sits in an unremarkable stretch with nothing on either side of it, you have a habit, and you should say so and stop. If it sits immediately before something that was on the calendar, that is more interesting, though still not proof of anything, because the whole market could see the same calendar.
The reason the order matters is that if you look at the trade date first and then go searching, you will find a plausible event near every single date. Something always happened. Building the list first is what makes the check falsifiable, and a check that cannot come back negative is not a check.
One structural caution about timing generally. The module reports average delay from trade to filing at 32.5 days, so you are reading about a decision taken roughly a month ago. Any event you line up against the trade date is an event whose market impact has long since happened. This check tells you whether a member's home-state position looks informed. It does not hand you a tradable entry.
What to do when the pattern survives both checks
Suppose a name clears both. First purchase rather than a habit, in a company with a real presence in the member's state, dated near something you wrote down in advance. What is that worth?
Less than it feels like. You are looking at one household's decision, disclosed a month late, in a stock that your own analysis may say nothing about. The correct response is to add the company to a watchlist and do normal work on it, then size it as an ordinary position if the normal work supports it. The disclosure earned the company a place in your reading queue and nothing beyond that.
The thing I would actively avoid is the version of this that turns into a narrative about a particular person. It is tempting, because the pattern is vivid and the names are public, and it is also the fastest route to being wrong in an embarrassing way. The rows are structural facts about a household portfolio. They are not a story about a legislator, and treating them as one will lead you to hold a position for reasons that have nothing to do with the company.