Tuesday, May 17, 2011

The economic value of broadband

Is it worth investing in the NBN? What are the benefits? These big questions are largely impenetrable. But so is any decision about the future that has to be made. We employ a number of techniques to do so. Anticipating the "wisdom of crowds" we use a thing called "democracy" to try to reflect the communal will or opinion) of the people.

But there are other more scientific tools one can use. One of these is an econometric model, whereby the correlation between certain past events is measured, and controlled for other factors, and a relationship established.

A new paper in The Economic Journal by researchers at the University of Munich (partially funded by Deutsche Telecom) provides a convincing case that higher broadband penetration indeed raises economic growth rates.

The paper "Broadband Infrastructure and Economic Growth" uses OECD Broadband penetration data and a number of models to determine a 10 percentage point increase in broadband penetration raises annual per capita growth by 0.9 to 1.5 percent.

The study is incredibly robust and controls for workforce size, workforce capability and rate of capital formation. They discuss a favourite topic of mine, that is broadband as a General Purpose Technology (or GPT). They summarise this nicely;

High-speed internet via broadband infrastructure facilitates the spatial distribution of large batches of information that previously had to be collocated, which in turn allows new business models, entrepreneurial activities and collaboration of firms producing specialised inputs. This can lead to lower entry barriers and higher market transparency, increasing both labour productivity (also through better job matching) and market competition and ultimately economic growth.

They note that their first simple association between growth and penetration could be due to reverse causation (growing economies can spend more on broadband through higher disposable income or investment by the state0) and omitted variables (the diffusion of broadband has occurred at the same time as other diffusions).

In their modelling to overcome these issues the researchers provide the first model I've seen that matches the individual country broadband diffusion curves with a predicted (logistic) diffusion curve. The paper is worth looking at for this (Figure 1) alone.

Of course, for the NBN the case needs to be made that even faster broadband means even faster growth. There are two ways to think of this. The first is that "even faster broadband" is a new GPT that should be expected to have similar benefits - after all Broadband penetration growth is slowing and so therefore is its contribution to growth.

Another way would be to add the OECD's data series on speeds as another variable to the model. Unfortunately that data set doesn't go back quite so far.

This is, however, an important paper. It is worth reading and building on.

Novae Meridianae Demetae Dexter delenda est


Richard Hayes said...

I can't immediately access the paper - is the 0.9-1.5 percent figure a 1 off increase or in perpetuity? If in perpetuity then it seems much too high.

David Havyatt said...

In perpetuity for each 10% is the claim as I read it.

Richard Hayes said...

Well that does seem too high. For example that implies going from 40% coverage to 90% coverage would give 4.5-7.5 % extra growth per annum in perpetuity. Perhaps I feel suspicious of things that are claimed to make such a large difference to growth rates in perpetuity.

David Havyatt said...

Well part of the issue is that you never get to 90% because it is a penetration rate per 100 inhabitants, not a population coverage measure.

I'm sure you can pay the price to buy the online copy .....