Under the Keynesian model of government investment, government investments are a means to lead the way and break the ice, in a stalled market, where private investors have become squirrelly about investing.  However, this Keynesian model assumes that government is applying the same criteria for investment that the market would have otherwise applied, so what happens when government applies criteria different from market demands?  Easy: investors loose money, and companies that should have flourished fail or never start.

Over the last two decades, private investors have become increasingly dependent on government "investment," particularly in technology.  As technologies have become more complex and multidisciplinary, they have become more difficult to evaluate.  Investors were particularly roughed up in the IT and biotech booms, and subsequent busts, where they lacked the tools to evaluate those early-stage technologies.  In many cases, they were simply the victims of shysters, while in others, the possible rewards were so great that the investors were impelled to take ridiculous risks.  So government grants and earmarks (aka investments) have become a way to test the waters, because if a company got government money to research its technology, then that technology must have some substance, right?

By the letter of the law and regulations, that government money (apart from the earmarks) is supposed to be going toward meritorious winners of competitive grants, but it rarely, if ever, happens that way.  Assuming that everything goes proper and above-board, reviewers tend to favor hot, new technologies, which may never deliver, over minor improvements on existing technologies.  One example is the carbon nanotube transistor (CNT).  Nanotubes are hot (they include the term "nano"), and making such small things is really neat, but CNTs will never outperform our 1950s silicon technology.  This was fairly well-known and accepted, until Jan Hendrik Schön published results that CNTs could work better than current technologies, and then the money flowed after "nano."  Even after it was discovered that Schön lied, money still flowed after CNTs and other nanotechnology.  Yes, scientists make mistakes (or worse), and the reasonable voices can be quashed by the herd, so even if everything is fair and proper, government investment in technology can be a craps shoot.

Unfortunately, the letter of the law is rarely followed (except in the case of earmarks), and technologies are often chosen based on politics, rather than merit.  This politics can take multiple forms at multiple levels.  The most common thing to see is topic authors creating topics pre-targeted toward a specific technology and submitter.  Lettermarking is a common way for congressmen to inject their influence into the evaluation of grant proposals, while earmarks completely bypass the entire process.  From the administrative side, top bureaucrats have a lot of influence over the low-level government employees reviewing the proposals; you don't turn something down when the big boss says he likes it.

Over the past couple decades, and particularly the last few years, the amount of politics involved in those granting decisions has increased substantially.  Following government investment in technology has turned from a craps shoot into following politics and hype down a hole.  Solar energy makes a great example.  Every month, there is one or more announcement about some new solar company or project obtaining hundreds of millions in government loan guarantees (one page of news alone totals to $5.9 billion), millions in government research grants, and tens of millions in investor money.  Knowing the technologies that these companies are using, they will never realize their claims, and they will fail.  It is a wonder that 90% of them received money, because only 1 or 2 of the dozen technologies might actually do what they claim.  Sure enough, the number of failed, government-backed solar companies grows by the month.

Many investors seem to be naive to all of this, or worse, in denial.  I have asked several investors, "What will you do when the government money disappears, or becomes unreliable?"  Every single one of them has denied those possibilities.  Government money will always be there, and will always be reliable.  Looking at the situation in the DoD and DoE, the situation at government labs, and the situations in the university labs, one thing has become quite clear: the grants are vanishing.  Despite the budgetary offices' and politicians' claims that more money is spent on research, less money is getting to the researchers.  The budget isn't lying, and the researchers aren't lying, so where is the money going?  Into under-the-table politics -- loan guarantees, non-competitive contracts, administrative influence, and earmarks.  The only thing government investment reliably indicates is political connectivity.

When investors send all of their money after government money, they run out of money to send after the things government missed.  When the miraculous new solar technology loses out to presidential donors, those donors get their payback, solar panels continue to fall short of promises, and investors have missed out on the next Apple or Microsoft.  Government investment has become bad investment, and it is time for it to end, or be massively reformed.