What can governments do to encourage innovation? Well, as we learnt this week, Malcolm Turnbull can think of $1.1 billion-worth of things to do.
His "national innovation and science agenda" involves 24 mainly small spending or tax concession programs, grouped under four headings.
Culture and capital, to help businesses embrace risk and to incentivise early-stage investment in start-ups.
Collaboration, to increase the level of engagement between businesses, universities and the research sector to commercialise ideas and solve problems.
Talent and skills, to train Australian students for the jobs of the future and attract the world's most innovative talent to Australia.
And government as an exemplar, to lead by example in the way government invests in and uses technology and data to deliver better quality services.
Will all those programs prove to be money well spent? Who knows? The safest prediction is that some will and some won't.
At present, the government is spending almost $10 billion a year on research and development. This involves about $2 billion on government research activities (mainly the CSIRO), almost $5 billion on grants to university and other research institutions (including medical research), and about $3 billion on tax breaks to business to encourage them to engage in R&D.
We do know a fair bit about the effectiveness of schemes to subsidise business R&D activity, whether in Oz or other countries.
And last week we saw the Australian Industry Report for 2015, produced by the chief economist of the Department of Industry, Innovation and Science, which reported the results of a new study of the effectiveness of the government's R&D tax concession scheme.
But first things first. This week's innovation statement tells us "innovation and science are critical for Australia to deliver new sources of growth, maintain high-wage jobs and seize the next wave of economic prosperity".
Which is nice, but what exactly is it? "Innovation is about new and existing businesses creating new products, processes and business models."
Ah, so that means innovation is just the latest business buzzword for what economists have always called technological advance. That means we can believe the happy chat about how wonderful innovation is.
Economists have long known that most of the rise in our material standard of living over the decades and centuries has come from advances in technology, which include better knowhow as well as better machines.
R&D, the industry report informs us, is the main vehicle for innovation. You wouldn't know it from the cost-cutting efforts of Treasury and the Department of Finance over the years, but economists have long accepted that there's a good case for government spending on R&D and for government subsidy of business spending on R&D.
A business engages in R&D in the hope that it leads to new or improved products and processes which will allow it make more bucks. They don't do it because they're nice guys but, even so, the rest of us benefit from their contribution to technological advance.
This means R&D has the characteristics of a "public good" – a good (or service) that's "non-excludable and non-rivalrous". You can't exclude me from using it (which means you can't charge me for using it) and my use of it doesn't interfere with other people's use of it.
Trouble is, public goods are a major instance of "market failure". We obviously benefit greatly from public goods – particularly because they're non-rivalrous – and so would benefit from them being produced in large quantity.
But we can't rely on the market – profit-motivated businesses – to produce as much of them as we'd like. Why not? Because they're non-excludable. Because too many people can use them without paying.
Economists call this the "free-rider" problem. They also say public goods generate "positive externalities" – benefits that go to people even though they weren't a party to the original transaction between seller and buyer.
Where market failure can be demonstrated, you've made the case for government intervention in the market to correct the failure by "internalising the externality" – always provided the intervention doesn't end up making matters worse, which these days is called "government failure".
So economists have long accepted the case for government to subsidise private R&D because this will benefit all of us, not just the business that gets the subsidy.
Of course, this is just theory. It's worth checking to see if our government's R&D tax concession really does produce positive externalities. Does the knowledge generated by the subsidised firm really "spill over" to other firms? And, if so, what can we learn about how this works?
To answer these questions the Industry department made available to Dr Sasan Bakhtiari and Professor Robert Breunig, of the Australian National University's Crawford School of Public Policy, data from its administration of the R&D program.
The program began in 1985, but the data used was from 2001 to 2011, during which time the number of participants grew from less than 4000 firms to more than 9000.
The program was open to firms in all industries, but the main industries using it were manufacturing, professional and scientific services, mining, and information media and telecommunications.
The researchers found evidence of significant spillovers of knowledge to particular firms from firms in the same industry, their suppliers, their client firms and from universities. Significantly, these spillovers came from outfits located within 10 km of the receiving firm, except in the case of suppliers, which were located more than 250 km away.
This leads the researchers to conclude, in line with other research, that knowledge spillovers from competitors and client firms mostly occur through face-to-face contacts between the R&D staff of the two firms.
So now you know why firms in the same business tend to cluster together, why that's a good thing and also, perhaps, why more and more of the nation's economic activity happens in or near the central business districts of our capital cities.