Sign up for free newsletter


Randeep Grewal

The network effect

Randeep Grewal (pictured), our regular columnist, ends the year commenting on the power of networks.

Robert Metcalfe is perhaps not well known in financial circles. In 1973, whilst working at Xerox PARC (the legendary research centre of Xerox at Palto Alto) he, together with David Boggs, invented Ethernet. Starting as a local area networking technology it is now used throughout networking in both local area and wide area networks. Indeed, nowadays it is fairly difficult to find a wired network device not using Ethernet. Over time a whole family of electrical standards (IEEE 802.3) have been defined for different variants of Ethernet

Subsequently Robert went on to found 3COM which at one stage was one of the networking hardware majors.

In technology circles Robert Metcalfe is best known for Metcalfe’s Law which was formulated around 1980 and states: “The value of a network is proportionate to the square of the number of nodes”. In economics the corollary is often described as ‘The Network Effect’. Numerous ‘social media’ companies have used variants of this to explain their high valuations.

In its original form, Metcalfe’s law implicitly assumes each node communicates equally with each other node and has equal value to the network. Real life is rarely quite so equitable; for instance, I am likely to find it of great value to join a social network which contains family, friends and business counterparts; but less likely to join a network with whom I have no common interests. Of course the interests and connections of each member of the network can be different and also asymmetric e.g. ‘John’ might find it desirable to communicate with ‘Rosie’, but the latter might not ascribe the same value to the relationship.

We are used to technology companies, particularly in the social media space, growing rapidly as the network effect takes hold. If the network also has a low cost of entry (typically free) and a cheap method of promotion (e.g. access to the users’ contact lists) then the ‘viral effect’ helps fast take-off. 

Perhaps the greatest network effect the world has ever seen has been due to a combination of an international treaty, technology, urbanisation and one of the oldest modes of transportation – i.e. shipping. 

On 11 December 2001 China entered the World Trade Organisation (WTO). In one swoop, Chinese businesses and those in the rest of the world became part of the same network. (Actually a pedant might argue it merely improved the connectivity between the nodes).

One might argue whether mass urbanisation was a consequence of entry into the WTO or was inevitable. Either way it lead to an available workforce in cities willing to work hard in the workshops that would produce the goods that the West wanted. At the same time mobile phones and the internet were entering the mass market allowing Chinese suppliers to communicate with their Western customers and their own supply chains. (Perhaps more subtle, but modern communications also help allow families to stay in contact when dispersed between the countryside and the city.) Combine all of this with the shipping industry, in particular container ships but also bulk carriers increasing their trips to and from China. In summary several different drivers of the ‘network effect’ coincided and drove GDP growth at a rate that has rarely been seen in a large economy.

Now however, there is a growing consensus that China’s growth is slowing – indeed some commentators suggest that China may presently be growing at a low single digit percentage.

Though the network effect is often discussed in the context of growth, it is rarely considered how fast a business dependent on the network effect can lose its way. Generally not all users (or all nodes) have equal value – and losing the wrong 10-20 per cent of the user base can dramatically alter the value of the whole network. 

Those of us with a tad of grey hair will recall how quickly the network effect reversed during and after the dotcom boom. Few younger readers will recall that once AOL (America Online) and Yahoo were considered insurmountable as the global number one and two most visited web sites. Fewer still will remember Geocities which at one stage was the third most visited website globally. Much of Geocities was shut down in 2009, at which time it was owned by Yahoo. Perhaps one might be surprised that it actually held on for so long, but the whole Geocities experience shows that traffic (or ‘eyeballs’) do not always equate to commercial viability. And lack of investment and innovation can lead to rapid obsolesce.

In the case of China it is almost inconceivable to imagine that it will not grow to become the world’s largest economy. However the present downturn and the wage inflation that China has faced in the last few years raise significant questions as to what happens after that. At the price sensitive end of the market there has been a trend for a few years now for Vietnam, Bangladesh and other countries to undercut China. At the high end some companies have onshored back to Japan, Europe or the USA. Of course the first reaction of local players will be to petition for a cut in the exchange rate to make them ‘competitive’ again. 

But often, business executives use currency as a scapegoat rather than accept that they have a structural problem. Management all too often belatedly acknowledges the absence of commercial viability, suppressing the opportunity to modernise by a risk-off appetite and by running out of prudence free investors. But it is often too late to deleverage the balance sheets and improve operational performance.

It is doubtful that China will go the way of Geocities or AOL. But it is worth remembering that the network effect can be just as potent in reverse – and an ageing population leading to a fall in those of working age and increasing wages of those actually in work could be the driver of the reverse. The key question to ask as an investor is surely what industries are susceptible to being supplanted by foreign competition, and how will the loss of relatively small percentages of some high impact industries affect the economy? 

As with any network, not all industries and sectors have equal impact on the economy. Many commentators have been impressed by the volumes of products being churned out by Chinese industries – few have asked how commercially viable the businesses inherently are? And in a downturn how quickly will China’s competitive position erode?

Perhaps it is worth remembering that, at one stage, the cotton mills of Yorkshire and Lancashire dominated the world but now many are derelict shells. Similarly in many mining villages in the north of England, the closure of the mines impacted whole communities. Even now, a quarter of a century later, many mining villages have not recovered from pit closures. 

For China, the risk of segments of its industrial base becoming uncompetitive could have significant social consequences that perhaps are being overlooked by many in the markets. It is not just the amount of its industrial base that is under threat from competing economies, but the relative importance of those parts to the wider economy. 

However investors in the West should not be too sanguine about unwinding network effects. It is clear that a number of ‘Unicorn’ start-ups have based their valuations on growth rates driven by network effects. But how many of them really understand what value they are providing their customers, and who are their high value customers? Growth due to the network effect is enticing – but is it profitable? How many Unicorns will become the Geocities of this generation

other gfm publications