How much bundling?

Traditionally, many industries have been dominated by integrated conglomerates. We would have a "phone company" which did all manner of telecom services. In recent years, we have seen fresh interest in breaking up these markets into smaller areas, where specialised firms compete.

From an economic perspective, there are subtle tradeoffs involved. There may be economies of scale and scope, so that a single phone company which deals with local voice calls, cell phones, Internet services and cable TV enjoys cost efficiencies. For many consumers, it'd be convenient and efficient to have one billing relationship, and one customer support number to call. On the other hand, this integration could also throw up severe entry barriers and generate monopoly rents for the giant company. Today, we have competition in cellular phones, and that has been a source of technological innovation and dropping prices. If a single phone company had been secure in it's monopoly, we might not have obtained such benefits.

How much bundling is optimal? We do not have a general answer. In the real world, we see that automobile engines and automobile transmissions are bundled in cars -- customers do not separately shop from a competitive engine market. We see that computer hard disks are unbundled, and customers shop for hard disks in a competitive market. There is no general recipe which drives these outcomes: in each situation, we need good reasoning, and empirical economics, to guide us on the efficient choices.

Cellular phones. In India, we have been beneficiaries of the excellent European efforts in developing GSM.

GSM handsets
are a competitive market, with multiple vendors, so we can buy handsets without paying monopoly rent.
GSM equipment
that is used by cellular phone companies, is a competitive market, so they do not need to pay monopoly rent.
GSM services
are a competitive market -- a person who buys one particular cell phone is not tied down to any one vendor. By simply switching the SIM, he can switch providers. This helps control monopoly rents in this market also.

In all, GSM is a great success story of a set of open standards around which we have a highly efficient industry with near-zero monopoly rents. Without the standards effort, it would have been extremely difficult for the three industries to coordinate their work.

We are now increasingly moving towards information services that run over mobile phones. It is possible to design these services around open standards, so that a transaction between a consumer and a content provider takes place without paying any tolls to the cellular phone company. This is similar to the idea that when a consumer buys a book from Amazon, his internet service provider (ISP) gets no toll out of the transaction; the ISP is just a competitive provider of bandwidth with no control on how the bandwidth is used.

There are countries which do this differently. In Japan and the US -- the last holdouts where GSM does not rule -- it is common to have a cellular company produce handsets, produce switches, provide cellular phone services, and sign contracts with content providers. The consumer cannot easily switch from one cell company to another owing to this baggage. This hurts competition. On the other hand, it is claimed that this approach can foster faster technological innovation since the process of setting open standards is bypassed.

Computers. In the 1960s and 1970s, IBM was a fully bundled provider of computer systems. They provided hardware, systems software, applications software and (in many cases) custom software development for the client. Over the years, we have seen each of these become separate industries. As with GSM, openness of standards has been essential towards enabling entry. If IBM bundled their own operating system with their hardware, it was not possible for a competing firm to produce operating systems.

In recent years, we have seen a similar problem with the PC industry. Many PC vendors forced users to buy the OS from Microsoft. Microsoft's control of the OS helped Microsoft win on applications software. Any user of Microsoft applications and OS was forced to buy Intel CPUs. This situation generated enormous monopoly rents for Intel and particularly for Microsoft.

Today, Microsoft is moving towards a "rental model" where users are charged roughly Rs.1000 per month per computer. This is an enormous tax, particularly in the Indian context where numerous office workers are paid less than Rs.10,000 per month. Every CEO today should carefully compare their annual payments to Microsoft against their net profit: I know many companies where fees to Microsoft are larger than 50% of the net profit.

This situation is slowly giving way to competition. Intel faces competition from alternative CPU vendors, ranging from Cyrix, AMD and Transmeta to Sun and IBM. Indeed, Intel's profit rates have never been as pronounced as those of Microsoft.

Microsoft's dominant position has been greatly undermined by the rise of the Internet. The Internet originated as a set of open standards, which fostered competition. In the last five years, Microsoft repeatedly tried to subvert Internet protocols. If Microsoft had controlled either the desktop or the server, it could have used these to inject proprietary protocols. So far, it has failed each time; users have carefully avoided using Microsoft standards such as COM/DCOM or Microsoft.NET.

The most interesting frontier is free operating systems and applications, which make it possible to bring the fee of Rs.1000 per seat per month down to 0. This involves moving into free operating systems such as Linux or BSD, and free applications software such as StarOffice, KOffice, etc.

The pension sector. The integrated model for pensions would involve a pension provider who did all the work -- starting from collection of contributions, investment of contributions through the working years of the worker, and then conversion of the asset into a flow of annuities after retirement. This approach is replete with many difficulties.

In recent years, an alternative design has come into favour: to break up this industry into three separate functions: (a) a distribution industry focused on dealing with individuals across the country, (b) a fund management industry focusing on asset management and (c) life insurance companies who sell annuities. This is the design which has been proposed by Project OASIS as a blueprint for India's pension sector.

In this case, it seems that we would have substantial efficiency gains by having three separate industries with competition between multiple providers in each area. The consumer would compare alternative fund managers, and choose one that he likes. Once he reaches retirement, the consumer would shop amongst the annuity providers, and buy an annuity from the lowest-cost competitor. If these functions are bundled, the industry becomes less transparent and less competitive.

Back up to Ajay Shah's media page