In fairness, the US has always done well harvesting the skills of foreign workers in pharmaceutical, chip design and other high-value services. Look at the quantity of patents researched in (say) Switzerland but owned by funding US interests and you'll see what I mean.

The problem for development is that it teetered on the brink of becoming a proper profession but then flopped lazily towards commodity. Professions always protect their customers (and themselves) with quality guilds involving defined standards and peer review. Think of any profession or high-paying trade and you'll find such a guild/association/whatever. But development? Nope. Apart from a few small areas where rigorous standards were demanded, usually by vendors like Cisco, practitioners seemed satisfied with vendor awards and anecdotal expertise. IMHO this made outsourcing and lowest-price contracting almost inevitable.

The other issue is that introducing IT in many industries used to bring a measurable competitive advantage. But as many/most businesses acquired IT, the benefits from constantly spending more $ became less easy to define. Obviously (and especially these days) smart businesses won't spend unless they see a ROI. The recent experience with a certain OS is one example: people chose not to spend $ because they couldn't see a reason. All this actually makes it easier for people with niche expertise who can show a definite ROI from their innovation, but harder for mainstream practitioners.

I don't think this is one to blame MS for, though. Had MS been allowed to abandon the MVP award that some insist is a proxy for guild/peer review, perhaps a proper guild might have taken its place. But we're told that hundreds of thousands of developers sent e-mails of protest to Balmer. QED.