Practicing economics without a licence
17/07/2012 § 21 Comments
In the June issue of AgScience, Prof Shaun Hendy has an article entitled, ‘New Zealand’s voyage of economic self-discovery’. He also has a post with the same title over a Sciblogs. Before I get too wound up, I should give Prof Hendy his dues. He does do fieldwork amongst economists in their native habitat. But in the end, it is dilettantism.
The article sounds impressive — we have a new approach! we have pretty pictures! Nokia! But really, what he is able to tell us is:
- New Zealand is small and distant
- its economy is based on what has worked in the past
- scale is important
- we should be more productive.
He isn’t telling us anything new. No, really, there is nothing new there. And what is there is either useless or confused.
Let’s start with useless. A large part of the article is about connections, about the networks that drive patenting and thus invention and innovation. The article points to several firms and countries where we see economic performance, innovation, and networks. Okay, great. Now, can you tell us why it worked in those places? Is there some description of how our situation is different? What can we generalise from these findings?
But that’s not what we get. Instead, we hear that networks are important and drive scale, so we should have networks. This sounds like the Underpants Gnomes theory of economic development:
- Build networks
- ???
- Growth!
But even worse than useless is confused. See, Prof Hendy argues both sides, and doesn’t even realise it. First, he states:
Economic development is path-dependent. The investment decisions we make now will lead to specialisations that determine the types of products we can invent and produce later.
Then, he goes on to maintain that, rather than picking winners, we need to fund all science:
Because of the variety and unpredictable nature of the science we will need, we must subsidise the generation of new knowledge across the board and not pick winners.
The problem is that path dependence and product space analysis point in exactly the other direction. Roger Procter has done the heavy lifting to explain the economics. The logical conclusion is that certain innovations are more likely to be successful, only certain types of science should be funded, and focusing efforts on those areas is more likely to produce economic returns.
I’m a knowledge fanboy. Let’s have more, let’s make as much as we can. But let’s not fool ourselves that activity = productivity. If the point of science is knowledge, then yay! let’s have as much as our budget constraint will allow. If the point is economic growth, well, that’s different. Then we need to think about playing the odds that are in our favour.
But then I’m just an economist. Maybe Prof Hendy would like to hear my thoughts on the Higgs boson. After all, I’ve been watching The Big Bang Theory.
Groping towards Bethlehem 
Classic work Bill – laugh out loud funny and a an excellent lesson to be learnt. It seems very common for those in the physical sciences to comfortably and without too much thought wander into disciplines of which they have very little knowledge. We see this all the time when the biologists and epidemiologists start talking about the ‘stigma’ of weight. I spent quite a few years trying to get my head around the psychology of stigma and yet they just blah on about it as if it’s simple… Grrr.
The wider issue of course is that we treat those in the physical sciences as if they have some special access to the Truth (with a capital T) rather than what they actually have which is normal access to an often contested version of truth (with a small t)!
You’re very kind, Andrew. The whole ‘access to Truth’ posture is the discourse of the University, is it not?
Yeah, that is certainly how I would read it in this circumstance. Of course hidden beneath the facade of the university is the Master, in this case the discourse of ‘science’. It is interesting that not all of us in the university have the same access to the power the discourse offers, depending on the relationship to the master some scholars are treated more as hysterics – which is of course a privilege as well as a burden.
Sheldon has me pretty convinced about string theory. Maybe I should write something about it for NZEP. Except the editor there would rightly call me out for talking crap. What’s up with the science mags that they don’t do the same?
Higgs Boson/Field…. here you go…. http://profmattstrassler.com
The good news is that he doesn’t bang on about economics….
Thanks for sharing — I’ll have a look.
[…] Update: Bill discussed this here. […]
Crikey guys, steady on – even that heavy-lifter Roger Proctor started out in physics!
Regarding networks, if you haven’t had much to do with the New Zealand science system you won’t know about the difficulties we’ve had with large-scale collaboration in a funding environment that incentivises competition between small teams. That scale is important in innovation and that inter-institutional, multi-disciplinary collaborative networks can deliver scale are messages that our science policy makers need to hear. Nonetheless, the article in AgScience was a transcript of a 20 minute talk I gave at a forum – I do give a longer version of it (sometimes I dare to give it to licensed economists!) where I consider how we can encourage such networks to form in New Zealand.
My point about path dependence is that by picking the obvious winners with the best short term returns (e.g. AgScience), I think we have driven ourselves into a cul-de-sac with fewer opportunities for future economic development. If it helps, here is how Daron Acemoglu puts it: ”innovations are made for current gain – the future benefits from these innovations are not fully internalized. This externality discourages research towards technologies that will bear fruit in the future because, in these research lines, current innovations are likely to be followed by further innovations before these technologies can be profitably marketed” (full article here –http://economics.mit.edu/files/4373). We have one of the most narrowly specialised public science portfolios in the world and I wanted to make the case for greater diversity in the types of science we do.
Would enjoy discussing this (or the Higgs) further!
Shaun, unlicensed economics-practitioner.
First, Shaun, thanks for stopping by and commenting.
Your comment can be taken apart both factually and rhetorically.
Rhetorically: why do you start with the ‘steady on’? It is purposely belittling — you position yourself as the one with the cool head, as opposed to your interlocutor who is becoming ’emotional’ and therefore needs to ‘calm down’. Very clever, by the way, but also underhanded. The next sentence continues that approach: ‘if you haven’t had much to do with the New Zealand science system…’. Again, you position yourself as the one with access to special knowledge. Of course, I have had a bit to do with the NZ science system, but that fact is powerless against your rhetoric. Again, I bow to your rhetorical skill. It is curious that you resort to such techniques, when you are positioning yourself as a hard-headed man of science. But, y’know, whatever works.
Factually: you don’t actually address the key point, which is that you are trying to have it both ways. If path dependence is true, then the economically efficient thing to do (especially considering the 8% discount rate policy) is to exploit it. The only way a broad base of science — a dollar each way, if you will — is the appropriate betting strategy is if the probabilities of success are identically and independently distributed.
I do agree with what you are saying about networks versus competitive funding. I see it in my own research projects, and I see it at the aggregate level. But…I also wonder about the difference between actual practices of scientists versus the stated policies. In practice, NZ science is based very much on personal relationships; institutional boundaries put an extra cost on collaboration but do not prevent it (except at the margin, but then we need to talk about marginal costs and benefits). To understand the actual behaviours of scientists given the current institutions, we need social science research on scientists and their networks. I have not seen much appetite for funding that research — although it would put us at the international frontiers of understanding science.
If we can get over each other’s rhetoric then this conversation could get really interesting!
I agree that an omniscient social planner could exploit path dependence.
But in Acemoglu’s model, the market fails to do this – the market funds R&D into technologies that generate better short term gains over those that have higher future pay-off at the expense of economic growth. Hence it’s arguable that markets don’t actually do particularly well at exploiting path dependence – the portfolio of research they fund is too narrow. Governments that take an “investment” approach like the one you suggest are also going to fund too narrow a portfolio.
For this reason a key role of government funding should be to encourage diversity in R&D. We do have funds for this purpose (e.g. the Marsden fund) but I don’t think the size of this fund is sufficient – we still seem to have a much narrower portfolio of scientific output than other comparable small countries.
I’m not a fan of the second mover strategy. Knowledge does not travel as easily as it could. This is partly because a lot of knowledge is tacit – you need to be doing your own R&D to be able to exploit that of others. But localised spillover effects also mean that firms benefit a lot more from R&D performed down the road than R&D overseas. A second role of government funding in New Zealand should be to generate spillovers.
On to networks – we have studied NZ scientist networks and we see that they are growing, although we are not as networked as the Australians (there are scale effects here). It is still painfully difficult though to manage inter-institutional or inter-firm IP and I think this probably kills a lot of collaboration dead in it’s tracks. I reckon we have a long way to go.
Sorry for taking a while to get to this.
I’ll have to re-read Acemoglu’s paper to comment on it specifically, but let me comment on path dependence. I think your idea of path dependence is different from mine. You say that an omniscient planner could exploit it. Now, that may be true, but my point is different. It isn’t about selecting an optimal path through research and innovation, starting from some theoretical original position (a la Rawls). Instead, it’s the old Irish joke: if I wanted to go there, I wouldn’t start from here. Given our current position in the research and innovation space, it is extremely costly to moved into some spaces and less costly to move into others. We also need to consider the payoff for moving into each of those new spaces.
An objection might be that those easy, high-payoff spaces are cul-de-sacs, and we are better off paying the cost of moving into wide open research spaces. Well, maybe yes and maybe no. In fact, I think there is an empirical (that is, amenable to research and modelling) question here. What does the research and innovation space look like, and what do the payoffs have to be? What I particularly like about Roger’s work is that it gives some structure (texture?) to the innovation space.
Yes, you’ve anticipated my objection. And yes, I agree that looking at the structure of “innovation” space is the way forward. To get at this, both Roger and I have made use of Cesar Hidalgo’s product space (the first figure in the AgScience article is a rendering of Hidalgo’s product space). My group has also some (mathematically) similar tools that use patents and scientific articles. Hidalgo builds his space using country level data – we are trying (with some success) to take our tools to the industry and firm level.
So my comments on diversity in R&D are based on some of this work – which does suggest we that much of our R&D is corralled in a cul-de-sac – plus my take on what I can find in the literature (with some case studies thrown in for colour). And, as best I can tell, I don’t think I am flying in the face of the empirical evidence – see for example:
http://www.sciencedirect.com/science/article/pii/S0014292198000476
Unfortunately, what we can’t do it the moment is a cost-benefit analysis of the type you would like. Instead, we look at what other successful small economies are up to and they are busy exploring richer areas of Hidalgo’s product space. This is no doubt more expensive – sure enough, these countries spend considerably more than us on R&D.
Our studies of networks is really about trying to make our innovation system more efficient. We think that the agglomeration effects that are seen in innovation data come from innovators exploiting denser networks of skills and expertise. There is a body of literature on micro-level modelling based on this idea, and some of these models reproduce some of the stylised facts that are knocking about. So when I say we need more networks, I am really talking about lowering transaction costs for sharing information and expert knowledge.
Oi! Both of you! Play nice! – says me, rhetorically.
But seriously, I don’t think that path-dependence is a binary factor. For all innovations the path-dependence is a matter of degree, some are, some are not, and most are in-between. Given that uncertainty, I think Shaun is justified in trying to have it both ways.
The path-dependence is also rarely clear. There may be a clear path leading to HTS-110, for instance, but if it exists then that path isn’t widely known. For most of us, that particular innovation seemed to come out of the blue. Should we close off that kind of revolutionary advance simply because of whatever tenuous theory of innovation that we happen to hold to today?
As for the costs that hold back collaboration in NZ, the NZ ‘market’ for collaborative research is sufficiently small that I’m not sure it makes sense to talk about marginal costs and benefits – each possible transaction is situated in a very small market niche, competing with often only a couple of other possible transactions. If we’re going to squeeze discussion of that into a supply-demand market-clearing framework, then at the very least we need to recognise that supply curves are stepped, not smooth. More realistically, I think we need to recognise that a supply-demand framework may be a misleading framework for these kinds of transactions.
I agree with what you’re saying about uncertainty, but I don’t reach the same conclusion. First, we have to remember that innovation funding (PGSF, FRST, MSI, MBIE) isn’t the only source of science funding. Universities and RSNZ also fund science. If we want to fund research and thinking, they are more properly funded through those mechanisms. Innovation funding is explicitly about producing effects that are valuable to NZ. And that means (given, again, the 8% discount rate from Treasury that applies essentially across the board to government spending), that innovation funding should be short- to medium-term and focused on results. The uncertainty in that timeframe is much less than the longer (10 to 20 year) research uncertainty.
The other thing that uncertainty brings me to is an insurance/portfolio model. By way of analogy: my best investment strategy is to mimic the overall market, rather than trying to pick winners. I don’t have the money, time, or inclination to set up investments that would mimic the S&P 500. So, I’m better to find an S&P 500 fund to do the work for me. And then, I can also invest in a few investments that I can control or that I understand: a rental property, stocks in a couple of major companies, and perhaps a local business. Similarly, we can focus some (the bulk?) of our innovation investments on things we understand (and can control), and put the residual into some broad, collaborative, international networks.
Finally, knowledge is expensive to produce and cheap to reproduce. That means that NZ with its small budget is better to be a second mover on science and innovation. That may not bring international scientific kudos, but it will bring more innovation for less money.
And I didn’t even use supply and demand curves! 😉
I just wandered over here somewhat belatedly because of a contemporary link from SciBlogs. I feel constrained to say that the path that led to HTS-110 was pretty obvious, but not necessarily easily achieved.
If you get the patent on the first material used to create long length HTS conductors the path to high field magnets is straightforward (read “Magnetic Venture: The Story of Oxford Instruments”). Helped along the way by the fact (as Shaun has noted) that Buckley Systems was already in the game of high field copper magnet based instruments (and Magritek was working on earth’s field and the instrument side of MR).
The hard journey for NZ will be in HTS based power systems equipment.
There are two points I’d make on the wider debate here. (1) R&D is just one component of economic activity (aka we need to look to all our resource endowments to work out where best to go – sub-optimisation is just that), and (2) Of course we should think about transaction costs (aka networks matter).
Nothing like doing a simple SWOT analysis with these points in mind to help decide what’s likely to work.
That came from me, the computer I was on took control of my life.
The specific technology examples I’ve researched have been interesting for the roles that R&D, law, business management, etc. have all played. As you point out, it isn’t just R&D, and it is my contention that the NZ system doesn’t fully recognise that fact.
If you look at the HVMS medium-term applied R&D investments over the last decade they have not been where the bulk of NZ industry is, but where the science is (hence all the work on materials science, rather than high value equipment).
This reflects a system that has over-weighted current science endowments (and the fact that this science is cheaper to do than high value product engineering). The attempts to get better alignment with industry have been weak and dominated by Wellington views of where the fashionable places to invest are. Providers have had limited ownership of the outcomes of the investments and so the interests of the investigators have naturally won the day by default.
Having said all that I don’t think the answer is necessarily cold turkey on the current programmes (leaving it to the universities to fund perhaps). Rather I think you say given this expertise how do you exploit it in an international economy. A-Star in Singapore gives us one way to do this – recruit global/multi-national companies (or even well funded overseas start-ups) to set up their development activities alongside our centres of R&D. If they aren’t interested of course this tells you something about the value of what you have.
Much better than spending time on under capitalised one person bands spun out by researchers (which seems to be the current solution).
As I said, just do the SWOT.
[…] Economist friend Bill wrote a classic critique of a physicist who had crossed unhelpfully into the discipline of Economics. Bill’s problem was that the scientist in question had no hesitation jumping into an economic […]
[…] takes on the disappointing proclivity of New Zealand’s bench scientists to opine on economics. In the June issue of AgScience, Prof Shaun Hendy has an article entitled, ‘New […]
[…] takes on the disappointing proclivity of New Zealand’s bench scientists to opine on economics. In the June issue of AgScience, Prof Shaun Hendy has an article entitled, ‘New […]