Effects of Shared ATM Networks on Efficiency -Turkish Banks

The effects of shared ATM networks on the efficiency of Turkish banks

H. Evren Damar

• This study investigates whether forming shared ATM networks has yielded positive benefits for banks in Turkey by increasing their productive efficiency.

• The performance of the banking sectors of developing countries has recently become a topic of interest in the literature. Most of this attention has been focused on the impact of financial liberalization on the performance efficiency of banks in a variety of countries.

• An aspect of financial liberalization that has not been addressed in this literature is the impact of new technology adoption and sharing that usually accompanies the liberalization and opening up of the banking sector.

• The usage of technologies such as Automated Teller Machines (ATMs) in developing countries has increased significantly during the past 15 years or so.

• Although the theory behind the benefits associated with the adoption and sharing of such technologies is well-understood, empirical studies that look at the actual realization of these benefits are relatively few.

• The goals of this study are to investigate whether ATM sharing has benefited Turkish banks by increasing their productive efficiency and to contribute to the literature on bank efficiency in developing countries through focusing on aspects of bank behaviour that have yet to be fully examined.

• The ideas behind ATM sharing and its benefits are based on the development of shared ATM networks in the USA during the 1970s and 1980s.

• In broad terms, there are two opposing effects associated with shared ATM networks. The benefits of ATM sharing are called ‘network’ and ‘economies of scale’ effects (Prager, 1999).

• Network effects suggest that the value customers attach to ATM services offered by a bank goes up as the size of the ATM network increases. In other words, the addition of a new bank or a new ATM to the network increases the attractiveness of all banks within the network to their customers. This is an important issue because it allows for banks to capture more business without having to increase the size of their branch or ATM networks.

• For example Vesala (2000) finds that after the start of ATM sharing in Western Europe, banks have opened fewer new branches and deployed fewer new ATMs.

• Economies of scale imply that the cost per transaction at an ATM declines as the number of transactions increases.

• Each ATM location has a variable cost and a fixed cost associated with it.

• Although variable costs (film, paper, etc.) are directly proportional to the number of transactions conducted at the particular ATM, fixed costs (such as the cost of purchasing or leasing the ATMs) decline as the number transactions increase (Saloner and Shepard, 1995).

• Therefore, by increasing the number of transactions, a shared ATM network can turn an unprofitable ATM into a profitable one.

• On the other hand, the presence of shared-ATM networks has also been shown to have negative effects on participating banks. These effects arise because ATM sharing reduces the level of product differentiation between banks and allows depositors to switch banks without incurring high costs. In their study of ATM network compatibility, Matutes and Padilla (1994) refer to this as the ‘substitution effect’ and show that its presence can be an impediment to achieving full ATM compatibility within the banking sector.

• In simple terms, whether a bank can benefit from a shared-ATM network will depend on which one of the effects described above dominates. If the network and economies of scale effects dominate, then the bank will be able to offer a more convenient product, collect more deposits and potentially increase profits. On the other hand, if the substitution effect dominates, then ATM sharing may result in a loss of depositors and profits.

• This problem can easily be framed within the concept of productive efficiency of banks.

• If ATMs are considered an input in the ‘production’ of deposits, then the presence of any benefits from ATM sharing would be reflected in the efficiency scores of banks. If indeed the network and economies of scale effects dominate, then banks that are engaged in ATM sharing will have relatively higher efficiency scores. On the other hand, if the substitution effect dominates, this would result in lower productive efficiency.

• Evolution of shared ATM networks in Turkey
• Similar to other developing countries, ATM technology was introduced in Turkey during the late 1980s. As the level of competition in the banking sector increased in the 1990s, there was a widespread increase in ATM usage. By 1995, there were 5000 ATM locations in Turkey and this number doubled by the end of 1999 (Isik and Hassan, 2002). By this date 27 out of 62 deposit collecting institutions had adopted ATM technology and another seven had issued ATM cards to their customers, although these banks themselves did not own or operate their own ATMs.

• The first shared ATM network in Turkey (referred to as the ‘Pamukbank-YKB Network’) was formed in 1993, and was soon followed by a shared ATM arrangement between four banks, named ‘Golden Points’.

• Unlike the USA, shared ATM networks in Turkey did not start as regional networks between local competitors. Since Turkey is significantly smaller than the USA, most banks operate in all major cities and some rural provinces.

• Therefore, the Turkish shared ATM networks started from a ‘national’ and not a ‘regional’ network stage. By 1999, ATM sharing had become a widespread phenomenon, with three more banks joining the Golden Points network and 16 smaller banks forming another network in 1998, named ‘Common Points.’ However, three of the five biggest banks in Turkey continued to operate proprietary networks.

Variables:

• ATM transactions
• Total deposits
• Total loans
• Fees and commissions
• no. of ATMs
• no. of shared NW ATMs
• no. of branches
• no. of employees
• Interest on deposit
• Operating expenses

Conclusion (READ IT AGAIN)

This study has looked at the evolution of shared ATM networks in Turkey and has attempted to see whether banks have been able to realize net positive network and scale effects through ATM sharing.

The main finding of this study is that participation in shared ATM networks has failed to increase the efficiency of small and medium size banks. The fact that most of these banks tend to share their ATMs with each other (and not with big banks) can be an important factor in their relatively lower efficiency scores.

The lack of significant positive benefits for many medium and small banks fits the conclusions reached by Matutes and Padilla (1994). Their results suggest that ATM compatibility is easier and more effective if shared-ATM networks are formed by banks that operate in separate locations because of regulatory reasons or due to geographical factors.

On the other hand if banks that compete for deposits within a market decide to share their ATMs, this may decrease the level of product differentiation between these banks, causing the sharing arrangement to become costly and ineffective.

Similarly, Holden and El-Bannany’s (2004) conclusion that there was no relationship between ATM sharing and bank profitability in the UK may be due to the fact that banks in their sample are not differentiated according to size and geographic concentration.

The findings of this study also support Carbo et al. (2003), who argue that technology adoption and sharing do not always yield cost savings for small banks. The results discussed above take this finding one step further by arguing that such technology adoption by small banks can turn into costly idle capacity. For the case of Turkey, there is ample evidence of such idle capacity. For example, Table 5 shows that many medium size banks exhibit DRS.

It is likely that some of this excess capacity is caused by ATMs that are deployed in urban areas, but infrequently used by depositors. A similar observation has been made by the Banks Association of Turkey, which has concluded that ‘some ATMs operated by banks are located too close to each other and this is a waste of resources.’ They suggested that banks should try ‘to increase the sharing of existing ATMs before deploying new ones’.

One possible solution to this problem of low efficiency among small and medium banks would be for these banks to form sharing arrangements with bigger banks. This would allow them to truly expand the services they can offer and gain an advantage over their competitors. Recently, strides have been made towards such arrangements, with one small and one medium bank from the Common Points network signing an ATM sharing agreement with the Pamukbank-YKB network in early 2003.

However, it is also possible that big banks would not be eager to allow smaller banks to join their shared ATM networks, as this would make it harder for bigger banks to differentiate themselves. Similar worries have been echoed by big banks in Turkey, who have stated that ‘the possible effect of increased sharing on the banks with extensive branch and ATM networks is an important issue’.

Another possibility has been put forward by Carbo et al. (2003), who argue that the lack of uniform benefits from technology-sharing arrangements should promote consolidation in the banking sector. In the context of the Turkish banking sector, the consolidation argument would suggest that the large number of small and medium banks that offer similar products can be consolidated into a few big banks that would be able to offer differentiated products and compete with other banks. One of the consequences of the 1999–2001 banking crisis in Turkey has been a government-encouraged wave of consolidation, the efficiency impact of which remains to be seen.

Other than being the first study to look at the effects of ATM sharing on productive efficiency, the period covered in the analysis is also significant.

The period 2000–2003 corresponds to one of the worst banking crises in Turkish history and the beginning of the best planned and executed rehabilitation program of the banking sector.

Although the number of branches and depository institutions decreased significantly between 2000– 2003, the growth rate of ATMs has remained high and positive. This may suggest that banks and regulatory authorities do not consider the build-up of ATMs as a serious overcapacity issue. This study, however, makes a point of caution that more ATM sharing does not automatically mean efficiency gains for banks. On the contrary, an increase in ATM sharing solely within the urban markets will be likely to increase the costs of banks without generating any additional benefits to their customers.

Although ATMs are still significantly cheaper than branches, operating and sharing unproductive ATMs can possibly contribute to another build-up of overcapacity. The only two options for preventing this potential problem are either further consolidation of the banking sector or a carefully planned restructuring of the existing sharing arrangements.