If the Government gets involved in business it is taking risks. That's the nature of business. But should a government take risks with taxpayer money?
Act's Rodney Hide is adamant: it shouldn't. Government is not business. Hide said the best contribution would be lower taxes, lighter regulation and less red tape.
Economic Development Minister Jim Anderton is also adamant: well-targeted government action can make more business for business.
He has sector task forces, regional development initiatives and New Zealand Trade and Enterprise (NZTE). From day one as Minister for Economic Development, Anderton has been toting a bag of cash and looking for ways to apply it to stimulate better business. A fair amount has gone in grants to individual companies.
Anderton has been in business, unlike almost all his Cabinet colleagues - the exceptions are Trade and Agriculture Minister Jim Sutton, who owned a farm, and Social Development Minister Steve Maharey, who managed a shoe shop - so he has known risk at the sharp end.
He has always argued that NZTE won't be doing a good job for the taxpayer unless it is taking some risk and it won't be taking some risk unless some of the risk goes bad.
He has so far been prepared to wear the flak for projects that have gone bad. Contrast that with his colleagues in the social portfolios. For all his past experience in business, Maharey quickly closed down assistance to social entrepreneurs when some grants seemed to have been used more for the recipients' pleasure than New Zealand's advancement, even though others did give value for money.
Such risk-averseness might be good politics but it doesn't get you far in business. So NZTE has been hiring some staff from the private sector and refocusing its working style from individual company welfare to seeking a "net economic benefit" - additional profits and wages, set against wider impacts on the economy and the cost to NZTE. And that means taking calculated risks. "Potential economic benefit means risk and uncertainty - if the future benefit is certain, there is no need for the Government to be involved," says a guidebook for staff about the changes.
The focus on net economic benefit means that just adding to a company's turnover or exports does not qualify for NZTE involvement. Instead, it is to get involved with companies taking risks that potentially - but not definitely - benefit the economy as a whole. The point of the intervention is to reduce risk and bring about the net economic benefit more quickly than if the company was left to its own devices.
And that means taking risks itself. Some projects will fail and lose taxpayers' money or waste staff time and resources. The test will be whether overall NZTE will earn a return on its outlays of time and money.
One guideline in that will be its alignment with the Government's "growth and innovation framework" for lifting productivity growth and real incomes. That means getting away from the production-push mentality that for decades drove much of its predecessor agencies' export and industry assistance.
It means instead trying to be more market-led: trade representatives overseas doing more to spot niche opportunities for New Zealand companies to develop and putting less time into introductions for any exporter who turns up. At home, it means looking for infrastructural and whole-of-sector solutions, not just company ones.
There is also accent in the new approach on spreading scarce resources: some trade positions overseas are being cut because they don't promise enough; at home companies are to be prodded to look for solutions that don't involve NZTE or go outside NZTE to other agencies and private sector companies.
It also means being clear in advance about getting out of involvement: the exit strategy must be built in at the beginning.
And it means commitment of time and resources and risk-taking by the companies it gets involved with. This is not about gently upward-sloping gains - it is about "significantly accelerated growth".
All of this is encapsulated in a decision chart that requires NZTE staff to explore alternative options and be clear that there is at least potentially a net economic benefit from intervening.
But how will we know if it works? National's John Key has caned NZTE for past programmes that shovelled grants out to businesses without any monitoring to find out how those businesses then fared. Part of the answer is in a performance-based measure NZTE is developing for its interventions. It expects to have it operating in six months.
Key favours shifting the focus from individual company grants - picking winners - to developing clusters (of IT, marine, fashion companies), building capability and so on - picking priorities. "It is unlikely one business grant will be game-changing," he says. But broader-based assistance can be.
So, fashion weeks are fine for the fashion industry because they work on the principle that the whole is more than the sum of the parts. Grants to individual designers are not because "if they are strong enough they will make it anyway". Applying for grants take a company's eye off the ball, he says and locks NZTE into "the business of making grants".
And, Key says, what NZTE does needs to be tied more tightly to what other agencies, notably the Foundation for Research, Science and Technology (which contracts for research), are doing. The aim should be to "build serious capability".
There is little in that that is not consonant with NZTE's revised market-led, productivity-focused, exit-conscious approach.
But NZTE does intend to continue to work with individual companies. And that is where the risk principally lies. Anderton and his Cabinet colleagues will continue to need stiff nerves in Parliament when Key and Hide bark and bite - and some hard figures from NZTE's audits about offsetting benefits.
Because the obverse of risk is opportunity. If NZTE's approach works, there will be more opportunity and potentially a sharper economy. At best the jury is out. At most, we won't know for several years. By which time, Anderton will be preparing for his political retirement.