Except the developer takes a hit there, too. Given that not everyone needs every feature, users end up subsidizing each other's desired improvements as part of a major upgrade. On top of that, not all features have clear lines that can be drawn around them, technically or marketing-wise.
It's a great model for games and certain specific types of apps (Paper is very clever). But to push all apps onto this model is highly unrealistic.
I'm not sure if the developer necessarily takes a hit. Say I have $15 to spend on a video editing app. $10 for the initial app and $5 for the only premium feature I want results in me giving $15 to the developer (less Apple's cut etc). However if the developer adds 5 new features to the app and I have to choose between buying features that I mostly don't want and can't afford or not spending the money I'll most likey give $0 to the developer.
Similar situation with luxury features on cars: it makes business sense to sell them individually.
If you wanted/needed the feature badly enough (ie, it was a mission-critical tool), you might pay much more for an update that included your desired feature in addition to stuff you don't need (or even stuff you might not yet know that you want). Also, the extra "wasted" revenue helps pay the developer for the thankless but necessary task of under-the-hood improvements, whose costs have to absorbed elsewhere.
I'm not saying the model can never work, but it's not a good fit for every product. The developer should be able to create the relationship with their customers that they think fits best; sometimes that will be IAP, sometimes subscriptions, sometimes upgrades. There will never be "one business model to rule them all", or else we'd see all marketplaces naturally converge in that direction.
It's a great model for games and certain specific types of apps (Paper is very clever). But to push all apps onto this model is highly unrealistic.