What Actually Buys You the Power to Say No

Published August 21, 2026

TL;DR

One employer is a single point of failure. But a handful of clients isn't a diversified portfolio either. It's correlated risk. The actual lever behind "walk away" power is savings and marketable skills, not your employment status.

In February 2026, Dave Kline posted that going independent in his mid-40s looked like the riskiest move he had made. Then he ran the numbers. "Now, if my biggest customer fires me, I lose 5%. If my manager had fired me? 100%." Simran Malhotra replied:

your negotiating position flips too
with a job, you negotiate from weakness - they hold 100% of your income. so you take shit you shouldn't have to
with clients, losing one hurts but doesn't kill you. so you can actually say no to bad deals or difficult ppl. that's when you start making real money - when you can walk away

The claim worth testing is the mechanism. Employment puts 100% of your income behind one relationship. Losing it is total loss. A client base, in theory, spreads that risk. And spread risk is supposed to change who can say no to a bad deal.

The concentration part holds up

Justin Skycak's Jenga-tower framing sharpens why "stable company" and "personal security" aren't the same thing. A stable employer runs fine without any one contributor. That's what stability means, so your individual leverage there is low by design. The employees with real security are the ones whose removal would visibly break something. The same test applies to clients who'd notice if you vanished.

Employment also concentrates risk invisibly. George Pu put it in three sentences. Your manager has a career plan, and it isn't yours. Read that forward: the manager's incentives point at their own promotion. Nothing about "stable job" implies anyone there is managing your risk on your behalf.

Where the popular version overclaims

The "5% loss" framing borrows the language of portfolio diversification, and diversification has a requirement the tweet skips. It only reduces risk when the things you're diversifying across don't move together.

That's the exact mechanism behind the 2008 mortgage-bond collapse, as documented in the Big Short story. Tranches built from the same subprime loans got priced and insured as independent risks, when a macro shock would sour them together. The insurance was cheap because the market treated correlated exposure as diversified exposure. It wasn't. The transfer to career risk is an analogy, not a study of freelance clients, but the math holds regardless of asset class.

Freelance clients in the same industry, hit by the same recession or budget cuts, aren't independent either. If they lose demand together, "five separate clients" behaves like "one client." Ray Dalio states that about fifteen uncorrelated bets improve the return-to-risk ratio 4.3x against one concentrated bet.

Most solo freelancers never get near fifteen clients, and the ones they have often share a sector. Five invoices in the same downturn are one hit with extra paperwork. Fifteen clients in the same industry is the same problem at a larger headcount.

Employment is still the 100% case. Spreading income across clients still beats that. The tweet oversold how much. You get a smaller catastrophe, and those clients can still fail together.

What does the work

If the portfolio-theory framing doesn't hold precisely, something else has to explain why operators with distributed risk negotiate better. The clearest answer already has a name. BATNA is your best alternative if the current deal falls apart. Itamar Turner-Trauring, in The Programmer's Guide to a Sane Workweek (v1.3.0, chapter 4), puts the two halves plainly. Skills let you find other work, savings let you wait, and together they let you walk away. Your negotiating power is proportional to your ability to survive the relationship ending, not to how many relationships you have.

Freelancing isn't the actual lever. It's one way to get one. An employee with real savings and a marketable skill set can build the same walk-away power without leaving a job at all. A credible competing offer functions as an improved BATNA whether or not you take it. Freelancing without that financial and skill foundation doesn't automatically produce leverage either. It just swaps one dependency (one employer) for a thinner, correlated version of the same problem (a few clients in the same market).

The corrected version

Employment concentrates career risk into one relationship. That part of the claim is directionally right.
Freelancing spreads that risk, but rarely as far as "portfolio" implies. Most freelancers don't have enough clients, or clients uncorrelated enough, for the analogy to be rigorous.
The actual source of negotiating power is savings plus marketable skills. That is the capacity to survive the relationship ending, whichever relationship that is.
That capacity is buildable inside employment. Freelancing is neither necessary nor sufficient for it.

Going independent replaces one relationship that can zero your income with several that usually cannot. That is an improvement. The walk-away still comes from cash that lasts if the work stops, and from a skill a stranger will pay for. You can start both this year without quitting.

Once you can survive losing a client, going deep with one of them can still be the right call for the work. Treat that as a quality decision. Do not treat it as diversification. One demanding client is still one basket.