Opinion

If data really is the new oil, the raw materials need a different economic approach

The International Monetary Fund is the latest to say that data is the new oil – the change agent in the economy – but striking it rich is a complicated matter

Tim Worstall
By
Published: 27 Sep 2019

Data really is the new oil – so says the International Monetary Fund (IMF) – and why not believe those who apparently know everything about money? However, it is important to understand the details of what they are saying – which isn’t that those who have a lake of data in the back garden are now as rich as the Beverly Hillbillies.

Rather, they mean there are certain economic similarities between the place of oil in the economy and that of data in the new economy. These are matters to consider as we ponder how to structure the new data economy.

One thing to understand is that the IMF experts don’t really mean data – they mean information. We in the business of economics know the difference.

Data is unstructured, unprocessed; information is what comes from performing those tasks. It is information that is valuable. To continue the analogy with oil – the crude itself, what comes out of the ground, has almost no value in actual use; it is the refined products that are valuable. Refineries are the value adding part of the process. Actually drilling for, and sucking up, oil only collects the other, extra, value between the value in use and the costs of refining.

When there is a restriction on refining capacity and a flood of crude oil, then refining margins do indeed rise. We have seen this recently, when fracking increased the amount of US-produced oil. Crude was substantially cheaper in the US than elsewhere, gasoline the same price – refineries were making out like bandits.

We still have something similar, in fact – crude oil in parts of Canada is one-third of the world price, but there are no pipelines to get it to market.

This is why the profit margins in transforming data into information – what Facebook, Google, et al are actually doing – are so large while the price of data is minimal. This is simply because there are few able to produce the information, so the profits are flowing to those with that scarce resource – ability.

What the IMF is really trying to point to is that data – again, they mean information – is the change agent in the economy, just as oil was, and perhaps coal was before that.

The economist Deepak Lal says fossil fuels allow Promethean growth – we are no longer limited by the energy output of human and animal muscles. The other two types of growth are Smithian – from the division and specialisation of labour – and increased resource use.

Without going too far into Greek mythology, given that Prometheus gave us fire and is regarded as the god of human wisdom and the striving for it, perhaps data really is the new oil in this sense. What the IMF is getting at is that data is that new grease of the economy that allows us to do everything better. We can become more efficient in our division and specialisation.

Of course, they could just be saying that data is important, and that would work too.

Economic peculiarities

They go further, though, and point out that data has certain economic peculiarities. It is non-rivalrous, by which they mean exactly the opposite of common parlance. My having and using it doesn’t diminish the supply available to you – we usually use rivalry to mean something quite different. It is also excludable – while data doesn’t run out the more people that use it, it is entirely possible to stop people from having it.

Something that is non-rivalrous and non-excludable is also known as a public good and there are good reasons why we want government action here. Simple economic goods, something like an apple, are rivalrous and excludable. My having a banana means you can’t have this one, and it’s easy enough to stop you having one – as the Soviets proved for decades. Being one and not the other similarly means that there is a correct method of management.

If something is non-rivalrous and also beneficial, then we will, in aggregate, benefit from the widest possible access to it. We don’t have to worry about running out of it, after all. Therefore we have a wish to make this data – information – easy to transmit and share.

That something is excludable means that we’ve got to get property rights correct. For who get to determine who uses it, this is pretty much the definition of what is the ownership of property. This is where that difference between data and information becomes so important. Those who process from one to the other should clearly own the results of their own efforts – even the results of their effort to collect that raw material. But not that raw material in its original form.

That means we do all own our own data, but not the processed form of it being spat out by someone else’s computers. Their value-add belongs to them as the people adding the value.

This combination of both being the grease of the new order plus greater benefit from wider dissimulation also answers the data nationality question for us.

Read more about the data economy

We want the widest dissemination – so the restriction of data and/or its processing to one geographic location is counterindicated. In other words, if something is known to be true, we want as many as possible to have that truth, distance be damned. We benefit from more truth being widely known and we don’t wear it out by it being so.

Of course, these economic points don’t determine exactly what we must do, they just give us some guidelines to how we would best benefit in a purely economic sense. If we really do want to say that Americans can’t process European data because, well, they’re Americans, then we can do that, as people are trying to. As long as we grasp that, in doing so, we are missing some of the economic gains of this new world we are in.

Perhaps the biggest lesson here, though, is that the specifics of data, information, the whole world of processing the one into the other – sure, they’re new. But they are examples of known and identified economic phenomena. As such, we already have decent rules for how to treat the new to best effect. The output of the data processing technology might be entirely new, but the rules we should use to manage it are not in the slightest.

Get property rights properly assigned, insist on as global a marketplace as possible, and then sit back and see what happens. We’ll all be richer for it. 

Read more on Big data analytics