In fact, when things are tight is the very moment you should go all in for R&D.
Let's clear up something simple people forget: when things are tight for you, they are normally also tight for your customers. In fact, that's generally how they become tight for you.
So you can forget investing in production and sales when things are tight. Things are tight because the world is in a period where your customers aren't buying. You won't change that by trying to sell harder.
However, small research projects are an incredibly cheap way to learn things and invent new technologies. When you're not encumbered by turning things into products, you can discover at a frightening speed. So you spend on research when the times are tough, and then when things go better you have a huge backlog of technologies you can apply right away.
And recessions are the time when this research is at its cheapest.
But that's how it's meant when people use it like that. They mean they want to do the work somewhere where it's (considered by the accounting department) a profit centre; not cost. That doesn't mean it could be another way for that company or that nobody should work there, much the same way as OP isn't saying nobody staff vs. line is right vs. wrong nor vice versa.
Second, as we move to more open source for infrastructure software (the tools used by staff) the above becomes more clear. SAP is charging for something with zero marginal cost, therefore it's perfectly valid to say they are renting software and hopefully using the money to develop the next version.
As the world moves to more open source I think you'll see fewer line software people, as the staff will do both support of the business and push some changes upstream (equivalent to what the line guys at sofware companies do). Once software is mature the rent model is really obsolete, but maintenance still has to be done by someone.