Keeping things very simple, all of our economic and policy analysis and the statistics on which that analysis is based is boundary focused.
Take economics as an example. Modern economic analysis was formed during a period that saw the rise of the nation state. It is state focused. The simple equations underpinning much economic policy focus on the state entity, relegating the international relations between that state entity and the rest of the world as a residual, an add-on. Important, but secondary.
You can see this in the way the old theory of comparative advantage is applied. As a theory, it attempts to explain why people might choose to specialise in particular activities. I am a good doctor but a great gardener. Comparative advantage dictates, however, that I should focus on being a doctor if that provides a greater return, buying in gardening services even if those services are not as good as those I might provide myself. This approach was then generalised to countries seen as economic entities, dictating that trade and benefits between countries would be maximised if those countries followed the route dictated by comparative advantage. I, Australia, am great at primary production and minerals. I should specialise there. That will maximise local benefits in part because it also maximises the return to others.
While the theory of comparative advantage is still popular, it dropped out of favour because of its inability to properly explain international trade. Comparative advantage is based on difference, yet trade seemed to be dictated by commonality and overlap. Similar countries with apparently similar factor endowments traded more with each other than countries with varying factor endowments.
All this analysis was based on geographic national boundaries. It was focused on trade between entities. But what if those boundaries have reducing meaning? What happens then? How does it affect the analysis?
In similar vein, Australians treat the states within the Federation as though they are economic entities. In one sense they are. But what does it actually mean to speak of the NSW economy? Is there such a thing as compared to a series of regional economies of which Sydney is the largest and therefore dominates the statistics? Mind you, the definitions of Sydney itself have changed over time. What do we mean by Sydney? Indeed, what do we actually mean by the word economy?
At a different level, Australian policy makers make considerable use of ARIA, a measure of remoteness from major centres with its sub-classifications of major city, inner regional, outer regional, remote and very remote. The Australian Bureau of Statistics describes the intent in this way:
The concept of remoteness is an important dimension of policy development in Australia. The provision of many government services are influenced by the typically long distances that people are required to travel outside the major metropolitan areas. The purpose of the Remoteness Structure is to provide a classification for the release of statistics that inform policy development by classifying Australia into large regions that share common characteristics of remoteness.The intent of ARIA is summarised in the last sentence: ARIA provides a statistical classification intended to "inform policy development by classifying Australia into large regions that that share common characteristics of remoteness." Sound reasonable? Well, what does it actually mean? In 2006, Darwin and the small NSW town of Balranald were both classified as outer regional, while Armidale was classified as inner regional. Under what conditions did it make sense for Darwin and Balranald to be treated as identical in policy terms, while Armidale was different? Do social or economic policy measures based on ARIA classifications make any sense at all? Why should an Aboriginal community in the Northern Territory be treated differently in policy terms from an equivalent Aboriginal community in NSW simply because the first is classified as remote or very remote, the second as outer regional?
I first came across the interacting problems created by boundaries and statistics based on boundaries many years ago and have been struggling with it ever since. Problems have become more intense with the growing importance of measurement and of performance indicators based on measurement.
It's become an absolute pain from my viewpoint, for now you have to constantly ask two questions: are the boundaries correct for the purposes they intend to serve; and then given the answer to one, are the indicators used appropriate? In both cases, the answers are often no.
A further difficulty lies in our sometimes inability to actually recognise that boundaries and boundary conditions are deeply embedded in the analytical structures we use.We just take them for granted. As an aside, I wonder to what degree the decline in geography as a discipline has reduced our ability to recognise spatially distributed difference?
All this is just a further gripe at this point. However, I am pretty sure that I will be returning to the issue!