Getting Past the Hype Around Government Innovation
I ran into this topic about five years ago while working on a project that involved public-sector funding for a clean-tech pilot. The person trying to explain the justification kept circling back to something they'd read from Mariana Mazzucato's work, but honestly, most people I talk to about this either haven't read the book or they're using it as a blunt weapon in budget meetings rather than actually engaging with the ideas. So let me lay out what the core argument actually is and how it plays out when you're the one implementing it. Mazzucato's central thesis, laid out in her 2013 book, is straightforward and direct: governments have always been the primary source of high-risk, foundational investment in technology and infrastructure. The private sector then picks up the pieces and commercializes them. This flips the standard neoliberal story that the state is inefficient and the market is the inventive force. She points to things like the internet, GPS, touchscreens, and mRNA vaccine platforms as examples where public money absorbed the risk that no private investor would touch. The more important part of her argument, and the one that gets lost in citations, is that she's not just saying "government spends money on R&D." She's arguing for a specific shift in how governments approach that spending. Instead of just subsidizing private companies or writing blank checks to universities, the state should act like a venture capitalist - taking real equity stakes, demanding performance milestones, and being willing to walk away from projects that aren't working. This is called "mission-oriented" policy, and it's the part that actually matters in practice.
She extends this further in later work, particularly "The Career of a Theorist" and her 2021 book "Forward," where she discusses how the post-2008 financial crisis era demanded a complete rethink of the state's role. The pandemic only reinforced it.
How This Actually Looks When You're Running a Program
I spent about eighteen months embedded with a regional economic development organization that was trying to operationalize mission-oriented investment. The grant writing was one thing. The actual governance mechanics were another matter entirely. Here's what I learned that nobody from the academic side seems to mention. The first problem is that almost nobody in government has venture capital experience. You hand a team of policy people a mandate to "be entrepreneurial" and they still approach it like they're administering a traditional subsidy. They want to distribute widely, minimize political risk, and avoid any appearance of picking winners. This directly contradicts what Mazzucato is asking for. Her framework requires you to bet heavily on specific directions and be willing to lose money on some of them. The second problem is far worse. Mazzucato writes about the state taking equity positions as if it's a normal commercial activity. But in practice, when a government body holds equity in a company, you immediately run into procurement rules, audit requirements, conflicts-of-interest statutes, and in many jurisdictions, outright legal barriers to owning shares. I saw a perfectly viable cleantech project stall for fourteen months because the program manager couldn't figure out which statutory provision allowed the organization to hold even a 0.5 percent stake. The workaround we finally used was routing the equity through a publicly traded special purpose vehicle that the government funded but didn't directly control. It added six months of legal fees and introduced a layer of opacity that made accountability almost impossible. This is a structural flaw in the model, not a solvable implementation detail.
Get the Full Details
The third issue is the measurement problem. In venture capital, you measure success by exit multiples. In government, you measure it by jobs created, tax revenue generated, and a dozen other lagging indicators that don't show up for years. Mazzucato acknowledges this but doesn't give you a practical toolkit for it. What we ended up doing was borrowing SBIR/STTR evaluation frameworks from the U.S. Small Business Administration, which at least had milestone-based disbursement mechanisms built in. It wasn't perfect but it was the closest thing to a VC-style checkpoint system that the regulatory environment allowed.
The Counter-Intuitive Parts Nobody Talks About
Most summaries of Mazzucato's work stop at "government funds innovation." The deeper insight, and the one that actually changes how you'd structure a program, is her concept of the "selectivity of the state." She argues that markets are also selective - they just select based on short-term profitability rather than long-term societal value. The state's advantage is that it can time-horizon beyond electoral cycles if it's structured correctly. But this only works if you have the institutional capacity to enforce consequences when projects fail. Most governments don't. They keep funding failing initiatives because stopping looks like failure politically, even though continuing is the actual failure economically. Another point that gets missed is Mazzucato's distinction between the entrepreneurial state and the entrepreneurial economy. She's not arguing that the state should run companies. She's arguing that the state should create the conditions, de-risk the early stages, and then step back. The mistake a lot of practitioners make is interpreting this as a mandate for permanent state ownership or control. It's not. It's a mandate for smart early-stage intervention with clear exit strategies.
Where the Framework Falls Apart
I need to be straight about the limitations because people who cite this work tend to treat it as settled theory. It isn't. The first major gap is that Mazzucato's examples are almost all large-scale national programs in democratic countries with relatively strong institutions. She doesn't address what happens when you try this in a country with weak regulatory capacity, high corruption, or where the state is the primary actor for reasons unrelated to innovation. The Brazilian attempt at mission-oriented policy in the Amazon region, which I observed briefly, collapsed because the same institutions tasked with green industrial policy were also the ones running environmental enforcement. There's no separating the entrepreneurial function from the rent-seeking function when the state is that small. The second limitation is political sustainability. Even in stable democracies, mission-oriented programs require multi-year commitments that survive election cycles. The U.S. ARPA-E model works partly because it was insulated inside the Department of Energy with a statutory mandate that's hard to dismantle. Most sub-national programs don't have that protection. I watched a regional clean-energy fund get defunded in a single budget cycle because the governor changed and the program couldn't point to a single finished product. Mazzucato's framework doesn't give you an answer for this. It assumes political will as a constant, which is a significant blind spot. The third issue is crowding out. When the state acts as an entrepreneurial investor, it can distort the very market signals it's trying to correct. Companies start tailoring their proposals to what government funders want to hear rather than what the market actually needs. This happened in a UK advanced manufacturing program I reviewed. The pipeline became full of companies that were excellent at writing grant applications and managing government reporting requirements but had no realistic path to commercial viability. The state wasn't de-risking innovation. It was subsidizing application-writing as a business model.
Practical Takeaways If You're Actually Trying to Use This
Don't try to replicate Mazzucato's ideal type exactly. You'll hit legal and political walls that her theory doesn't account for. Instead, focus on the parts that are actually implementable: mission definition, milestone-based funding, and willingness to kill projects. The rest is secondary. Define your mission with the same precision a VC uses for a thesis. "Clean energy" is not a mission. "Reduce grid-scale battery storage costs by forty percent in this jurisdiction within seven years" is. The difference matters because it determines everything downstream - how you pick partners, how you structure disbursements, how you know when to walk away. Build in kill switches from day one. This is the hardest part and the one most organizations skip. You need pre-agreed criteria for when a project stops receiving funding, and those criteria cannot include political protection. I've seen too many programs where the board said "we'll review at year three" and then reviewed it for three more years because no one wanted to be the person who killed the initiative. Write the termination conditions into the founding document before you take a single dollar in.
Accept that equity participation will likely be impossible in your jurisdiction. Find alternatives. Revenue-sharing agreements, conditional grants that convert to loans only upon commercial success, and phased milestone payments can approximate some of what direct equity gives you without the legal headaches. They won't be identical, and you should expect friction from both sides of the table - government lawyers will hate the complexity and startup founders will hate the uncertainty. That's fine. It's better than getting stuck in thirteen months of legal review over a half-percent stake. The broader point is that Mazzucato's work is useful as a diagnostic lens and a rhetorical tool for Justifying government intervention. It's much less useful as an operational manual. The gap between her framework and the institutional reality of how governments actually function is where most people get tripped up. If you're reading this to build something, start with what's legally possible in your jurisdiction, not what the book says should be possible. The rest follows from there.
