Because those people wield the power in the relationship, by virtue of the fact that they can stop working and disrupt the system economically.
This is, I believe, the author's point: political power follows economic power, over time. The two can diverge temporarily, but over the long run this tends not to be stable, or has to be maintained by force with much attendant unpleasantness.
Historically, I think this checks out. The combination of market(ish) economies and representative(ish) democracy that exist in most Western states evolved alongside the increased economic power of industrial workers, starting in the 18th century. Marxist theory takes the same ideas and runs with them further, hypothesizing inevitable conflict when worker-labor is alienated from the economic gains it produces. Neither of these two systems, which basically dominated the 20th century, would have emerged under conditions of agrarian peasantry. (And in the early Soviet Union, which basically blundered into Communism by mistake -- Marx had always thought it would emerge in more-industrialized Germany first! -- the peasantry was largely either forcibly transitioned into industrial laborers or exterminated, since they had no place in that political-economic system.)
But even in a democratic context, a "vote" is fundamentally meaningless; it is at best merely a suggestion. Underlying the idea of democracy is always the question of what the citizenry will do if the government decides to ignore the vote. Governments that depend on economic activity which citizens can disrupt through simple nonparticipation have a vested interest in keeping those same people happy and productive. Governments that only need to worry about armed insurrection can employ more draconian methods.
In a widely-diverse economy, it's impossible to have huge voting blocks like that. There's just too many vastly different economic models represented. But if you start heading down the road of UBI, you're basically splitting the population into two groups. Something's gotta give somewhere. You either get debt default or oligarchy.
It currently works in the U.S. -- assuming you believe the U.S. is not dysfunctional -- because while we have retired folks, they're just a small part of a vastly more diverse electorate. And even then, you'd imagine they would bring some of their experience in the economy to bear on their current decisions. A nation full of people who have never economically participated would have no data to decide or discuss which voting decisions to make. It'd all just be about more consumption.
In fact, assuming you could make it work, it would end up with folks voting in short-term bumps in UBI increases at the expense of long-term budgets, for much the same reasons as corporate boards vote in CEOs who can increase the share price short-term at the expense of long-term strategy. Only way out of that is tyranny/oligarchy.
Note that I'm not saying it won't work. I want to see more experiments and real data instead of all of this arguing about how many angels can dance on the head of a pin. Real data beats this kinda stuff hands-down. I'm simply saying it's an interesting argument that deserves some attention -- especially when looking at the results of the experiments when they roll in. The last thing I would want to see is something like YC doing an experiment where the only people looking at results were folks who were already true believers. That's total crap. You want results you can throw rocks at.
It's true that UBI is often sold as being a better way of providing support for the permanently unemployable, which it is, but it also reduces the ranks of the unemployable and removes disincentives to work that exist with means-tested safety net programs.