Thoughts on PolicyHack

Fri, 09 Oct 2015 09:31:59 +0000

Yesterday a Policy Hackathon event was launched by the Assistant Minister for Innovation Wyatt Roy. This kind of outreach to the community is a great thing to see from politicians, and I look forward to seeing more examples of it. However, I'm concerned that the way in which this particular event has been framed won't lead to successful outcomes.

The first problem with Policy Hack is that it focuses on the solution space (i.e.: policy ideas) and completely fails to explore the problem space. Ironically, this is the same kind of thinking that leads to the demise of many technology companies who focus on the cool technology instead of the problem their customers need them to solve.

The Policy Hack website has, at the time of writing, received over 60 submissions (of varying quality). All of them are policy suggestions, but few, if any, provide any kind of detail on the specific problem they are trying to solve. In some cases it is implied (e.g. a policy suggestion to lower taxation implies that high taxation is in some way impeding innovation), but even then there is little to convince the reader that the implied problem is actually a problem. (e.g.: where is the evidence that the current level of taxation is actually impeding innovation?)

Before we get into policy ideas we really need to have some kind of problem definition. What is the problem that Policy Hack is trying to solve?

The closest we get to that on the website is policy ideas designed to foster the growth of innovation industries including tech startups, biotech, agtech, fintech, renewables and resources..

This isn't at all a bad starting point, but it's a very high-level aspirational kind of statement, rather than an actionable definition of the problem. A lot more work needs to be done to fully flesh out the problem before we even start thinking about the solution.

An example of a more actionable problem statement could be: The growth of the tech startups, biotech, agtech, fintech, renewables and resources industries as measured by percentage of GDP is too low for Australia's future economic prosperity. Of course this raises further questions. Maybe percentage of GDP isn't the best measure, and we should be using a different metric such as tax revenue, or employment. And the actual terms should be defined a bit more clearly as well; what even is a tech startup? I don't think that is one of the business classifications used by the Bureau of Statistics, which raises even more questions!

Of course, once we have that problem statement, some kind of data would be nice. Like, what is the current growth rate? Off what base? What does it need to be to ensure economic prosperity? If we can't get those numbers directly, what are reasonable proxies for those numbers that we can measure? How reliable are those proxies?

With a problem definition, trying some kind of root cause analysis might lead to specific problems that we can solve with policies. One example could be: there aren't enough people creating startups, because there aren't enough STEM graduates, because high-school students don't choose STEM courses, because they are worried about the HECS/HELP debt. Or it could be: we have plenty of startups but once they get initial traction they move to the US, because there isn't enough capital available in Australia, because investors are taxed too heavily on capital gains. I'm sure there are hundreds different chains of thought like this that should be explored.

Of course, those are just hypotheses! We should be applying some kind of scientific method here. Do startups really move to the US once they get success? Maybe they don't. Do they move because of funding reasons, or maybe it is access to customers? Is capital not available here because of capital gains tax, or maybe it is because other asset classes provide better returns?

None of that is really easy, and certainly can't be done collaboratively during a hackathon (despite the excellent opening up of government data sources to make the underlying data available). This needs patient analysis, by diligent people, spending time to find the right data sources.

Without access to this information how could you possibly come up with a useful policy suggestion? Of course, once this work has been done the actual policies probably become somewhat obvious!

Ideally this is work that has already been done, either within government policy units, or external consultants. If so, it should be directly available on the Policy Hack website, and it should be required reading for anyone wanting to put forward a policy idea.

The second problem that I see with the Policy Hack event, or at least some of the suggestions put forward to date is that many of them are simply reinventing the wheel. (Which, ironically, is the cause of the downfall of many a startup!)

Many of the suggestions currently put forward are basically the same as policies that already exist in some shape or form. This isn't meant to be a criticism about the person putting forward the idea; it isn't exactly simple to find what policies are currently out there. But surely, if we are going to have an event that looks at putting forward new policy ideas on innovation we should at the very least have a list of the current policies the government has in place with the aim of fostering innovation! Ideally, we have also have an analysis of each policy with some commentary on what works well about it, what doesn't work so well, and some form of cost-benefit analysis.

With this information in place we may find that a simple tweak to an existing policy could have a large impact. Or we could find ineffective policies that can be scrapped to make funds available for new policies (because any new policies will need to be funded somehow, right?). Or we could find that we already have policies addressing many of the things that we think are problems, and we need to go back to the root cause analysis to determine other possible things that need addressing.

Again, maybe this information is available somewhere, but it should be made available to participants of Policy Hack in advance. And, if you are going to propose a new policy, that is similar in some way to an existing policy, then should explain what's different about your proposal. For example, if you put forward an idea such as allow special visas to hire technical immigrants, without any reference at all to the current 457 visa scheme you're really just wasting everybody's time!

Finally, there are a lot of Country X does Y, let's do Y policy suggestions without any critical analysis of why Y works for X (or even if Y works for X!). (And to continue the theme, this kind of me too thinking seems to plague startups as well.)

Ideally an input to the Policy Hack should be a comprehensive review of other innovation policies that have been put in place by other jurisdictions. This review should explain how these have, or have not, been effective in the jurisdictions in which they have been implemented, along with some commentary as to how they compare to similar policies in Australia. For example, holding up the US low capital gains tax and long-term dividend tax rate without explaining that the US has no mechanism for relief from double taxation at the corporate and individual level is not providing a fair evaluation.

And again, if you are going to put forward policy ideas based on ideas from other jurisdictions, it better actually reference that research document and justify why that policy would work in Australia. For example, if you think that Australia should have lower corporate tax rates because Ireland has lower corporate tax rates, you need to at least give some kind of backup that this is actually working for Ireland.

If we want to have an effective discussion on policy, we first need to agree on the problem we are solving, and have a clear understanding of what has been tried before both locally and overseas.

Instead of policies, maybe the Policy Hack could instead work through the problem space. A morning session that presented a summary of all existing data to the participants, followed by an afternoon brainstorming session on identifying possible reasons why we don't see the levels of innovation that we'd like. With this information in hand the policy experts could perform the relevant research to determine which of the identified reasons actually has an evidence base to support it. The results of this could feed in to a future hackathon that starts to look at policy from an informed base.

blog comments powered by Disqus