Political Capital

Credibility is a budget. Every ask beyond your formal authority is a withdrawal, and the account is funded only by past delivery.

The budget model

Your job title gives you formal authority: you can merge code you own, approve reviews in your area, run your own sprint. Everything else — pushing back on a decision that's already been made, requesting an exception to a process, proposing a rewrite, escalating past your manager, asking another team to change how they work — is an ask beyond that authority. Each one costs something.

The currency is credibility, and it's finite. When you object to a plan, you're asking people to trust your judgment over the judgment already applied. Whether they do depends almost entirely on your track record with them, not on the quality of your argument in the moment. Arguments are evaluated through the lens of who's making them. This isn't irrational — it's how people compress the impossible task of evaluating every claim from scratch.

Engineers who miss this treat every disagreement as a fresh, isolated debate to be won on technical merit. There is no fresh debate. There is one long running account, and every interaction posts to it.

How it's earned

Deposits are boring and slow:

Shipping. Delivery on commitments is the base rate of the whole system. Nothing else works without it.
Being right on the record. Not right in your head — right where people saw you call it. "You flagged that race condition in the design review and it bit us in prod" is worth ten confident opinions.
Doing unglamorous things well. Taking the on-call cleanup, fixing the flaky test suite nobody wanted, writing the migration runbook. These earn disproportionately because everyone knows nobody wanted to do them.
Making other people successful. Unblocking the adjacent team, giving the junior engineer the win, making your manager look prepared. People remember who made their job easier.

One structural rule: capital is domain-scoped. Deep credibility from five years of backend work funds asks about backend architecture. It does not fund opinions on org design, hiring, or the frontend team's framework choice (see Political Capital and status-doesn't-transfer-across-domains). Spending outside your domain is spending at a punishing exchange rate.

New joiners get starter credit: roughly ninety days of benefit-of-the-doubt, extended on the assumption you were hired for a reason. It expires whether you use it or not. Spend it on learning the system, asking questions that would be embarrassing later, and banking small visible wins. The classic misspend is the day-one rewrite proposal — burning unearned capital to declare that everything the current team built is wrong, before you understand why it's built that way.

How it's spent — and overdrawn

Every ask spends. But some things drain the account far faster than a normal withdrawal:

Being wrong loudly. A confident public prediction that fails costs several multiples of a quiet one.
Missing commitments after winning the argument. You spent capital to get your approach approved; the delivery was the collateral. Fail to ship and you lose the spend plus a penalty — next time your objection is remembered as "the thing that cost us a quarter."
Burning people publicly. Making someone look bad in a meeting is the most expensive transaction available, because they and everyone watching reprice you at once.

Engineers mismanage the budget in two symmetric ways:

Hoarding. Never spending — staying silent on the design flaw, the doomed deadline, the bad hire — feels safe, but unspent capital does nothing, and silence gets read as agreement and hardens into the org's default (silence-becomes-structural). The hoarder ends the year with a pristine reputation and zero influence on anything that mattered.

Overdrawing. Dying on every hill. Object to the linter config, the sprint format, the naming convention, the meeting time — and the org applies the boy-who-cried-wolf discount: your objections get repriced toward zero. By the time the hill that actually matters arrives, you're "the person who always objects," and the one pushback that deserved full weight gets the same eye-roll as the other forty.

The compounding spend

Spending well is not just non-destructive — it can return more than it cost. You spend real capital to push a controversial storage migration past a skeptical team. It ships, latency drops 40% visibly, the pages stop. The result: you didn't just recoup the spend, you're now the person whose controversial calls pay off, and your next controversial call starts with a subsidy. Credibility spent on bets that land compounds; that's how staff-level influence is actually built — a chain of well-chosen spends, each funded by the last one's return.

The corollary is that a spend is really a leveraged bet. The moment you win the argument, you hold outstanding debt until delivery clears it.

Concrete cases

The hoarder. Priya sees the new service design will fall over at 10x load. She's the strongest engineer in the room and says nothing — "not my decision." It falls over eight months later. Cost of speaking: one moderate withdrawal. Cost of silence: the outage, plus the quiet knowledge among peers that she saw it coming. Counter-move: one written, specific, on-the-record concern with a proposed mitigation. Cheap to post, and it converts into a large deposit when proven right.

The overdrawer. Marcus objects in every design review — style, tooling, process, architecture, all at equal volume. In month nine, he flags a genuine data-loss risk. It's triaged as "Marcus being Marcus" and shipped anyway. Counter-move: an explicit personal quota — pick at most one or two real objections per quarter, concede everything else out loud ("I'd do it differently, but this works — ship it"), which itself earns capital.

The new joiner. Dana joins from a company with a slicker deploy system and spends week two proposing to replace the CI pipeline. Nobody adopts it; her starter credit is gone by week six. Counter-move: spend the first ninety days shipping two small fixes, asking why the pipeline is the way it is, and writing down the answer. Propose the change in month four with the org's own incident history as evidence.

The overdrawn account. After a failed argument and a slipped date, Wei notices nobody engages with his proposals. Counter-move: stop proposing. Ship, quietly and repeatedly, for a quarter or two. Arguing from a zero balance is not persuasion, it's noise; delivery is the only deposit that clears an overdrawn account.

Diagnostic test

Score yourself honestly:

In the last quarter, name one thing you pushed back on that mattered. If you can't — you're hoarding.
Name the last three things you objected to. Would a colleague say all three were worth it? If not — you're overdrawing.
List your open spends: arguments you won where the result hasn't shipped yet. Each one is outstanding debt. More than two or three, and you're over-leveraged.
Was your last strong opinion inside your domain of demonstrated delivery? If not, you spent at the bad exchange rate.
When you spoke up last, did people lean in or brace? Their reflex is your live account balance.

If the balance is low: stop arguing, start shipping, and let the next argument be funded before you make it.