[I]n a wealthy society like ours, no individual lacks the physical or material capacity to meet their needs. There are plenty of resources nearby, sitting in shop windows. Anyone is capable of taking those resources. Their problem is that other people in society won't let them. Security guards will interfere, using force to block the individual's access, or to reclaim what they now call "stolen" goods.
(That's not to say we should do away with property, or anything silly like that. Some forms of interference are justified, after all. But we shouldn't let that blind us to its coercive elements, which - upon appreciating - we may seek to mitigate.)
Now, my friend objected that theft is relatively rare, and thus taxation licenses far more interference than do exclusive property rights. But this is irrelevant. Imagine a tyrant whose totalitarian control is so ubiquitous that nobody ever dares step out of line. He thus never actually has to actively exercise his power by interfering with people. That doesn't mean the people are free, of course. The violation of negative liberty instead comes from the threat of interference.
Contrary to my friend's suggestion, then, a relative lack of theft does not mean that property rights are less of an imposition against our negative liberty. Quite the opposite: the imposition is so complete that most of us would not dream of acting against it. We fully internalize the fact that we are not at liberty to take goods that are deemed to be "owned" by another. This societal ordering closes off to us actions that we could otherwise have performed. (Again, it's probably for the best, but one shouldn't pretend it's not a form of ubiquitous interference that has a huge impact on the options available to us every day of our lives.)
It's worth noting that this sort of implicit interference is distinct from Pettit's notion of domination (the modal capacity of one person to interfere with another). If Nora can do whatever she wants, but it is by her husband Torvald's leave - a general permission he could, but let us suppose won't actually, withdraw - then she has negative liberty but suffers domination. Pettit argues for the importance of domination by asking us to imagine that Nora learns of Torvald's power over her, and thus becomes servile, making extra efforts to please him and ensure that he never feels any need to exercise his power by interfering with her. Clearly, as Pettit says, Nora's freedom has not thus increased, even if active interference is now even less likely than before. But my earlier remarks show that this is no argument for non-domination. What Nora suffers here is an increase of implicit interference: the threat of potential interference shapes her behaviour, obstructing her from behaving the way she might prefer or freely choose to, were it not for the shadow of the tyrant.
The difference between negative liberty and non-domination is instead seen in cases where actual non-interference is guaranteed, e.g. by the potential dominator's reliable preference not to interfere. It may be that Torvald could interfere if he wanted, but so long as he actually doesn't want to (and this is certain not to change), who cares? David Braddon-Mitchell asks us to imagine a fantastic neuroscience which allows us to prove to Nora that there is zero probability of Torvald actually interfering with what she wants in life. Why, then, should she care about merely counterfactual coercion? (In practice, of course, there are no such guarantees. So we have very good practical reasons for fearing and opposing domination. But these thought experiments show that the value is instrumental.) Pettit responds that insofar as we think of each other as agents, such deterministic guarantees are ruled out. But that just shows the value of non-domination to be ineliminable, not non-derivative.