Earlier quoted context omitted.
In many cases it could truly be that programmers aren't worth enough for them to pony up. It isn't that they're being cheap. It's that it doesn't make financial sense. There are plenty of contexts where programmers are useful but still only worth "normal" salaries.
Or, perhaps many companies do not know how to leverage software to make money. I suspect there are many of those sorts out there, and if they do have developers on staff, it makes sense to let them go.
A great example is my friend who works for a provincial political party. It's clear that they would be able to compete with other parties on many levels if they were more technically capable, and someone on staff being paid a lot to facilitate that would be well worth it. They'd almost certainly secure more donations, learn to reach their demographics better by utilizing data properly, work far more efficiently, and so on. My friend is acutely aware of this and how technology is currently impeding them, but the rest of the organization is more like vaguely aware of this yet almost entirely uncertain of what to do about it.
If they were to pay someone with significant software experience around $200,000 per year to untangle and revamp the technology they're depending on, they'd certainly see this paying dividends within a couple of years. Instead they will carry on encumbered. They currently spend around $38,000 per year on software that's serving them partially and otherwise weighing them down substantially.
Part of the problem they face too is that they'd need to hire someone who truly gets the politics and people involved. You couldn't bring someone into that scenario who loves the IC role. There would be a lot of white boarding and explaining to stakeholders. A lot of very mundane untangling of existing services accomplished by digging through accounts and billing, talking to the people who use it, documenting how they use it, why, figuring out what they forgot to tell you, etc. Then the very dissatisfying step of determining solutions which can't be totally bespoke or custom, but would inevitably involve some degree of compromise. It wouldn't be a glamorous role. It could transform the capability and velocity of the organization, though.
I often think this is exactly the type of role which software developers will tend to move into as AI takes on more low-level IC tasks. I don't expect that to be tomorrow, but very likely within the next decade. Software developers will need to be far more people-oriented than they tend to be today in order to allow their skills to shine in ways an AI's can't.