Implications from literature

According to previous literature, efficiency barriers and information disadvantage are two major characteristics for foreign bank entry. However, most studies don’t consider the costs from efficiency barriers. So, it seems important to incorporate the concept of

"efficiency barriers" into an analytical model. Besides, like most studies, I assume that domestic banks possess more information than foreign banks do. Faced with two major disadvantages, foreign bank may try to reduce costs from efficiency barriers and gain more information by choosing alternative entry modes. Joint venture and acquisition can reduce efficiency barriers by accessing local resources (Petrou, 2009). However, Greenfield investment can’t reduce the costs. So, Greenfield investment may be preferred if costs from efficiency barriers is lower. Also, acquisition can bring the information held by acquired domestic bank to foreign bank. The benefits of information may play an important role in entry mode choices. In the following sections, I compare three entry mode choices of Greenfield investment, joint venture and acquisition. The model set-ups are introduced in next section.

3 The Model

In this thesis, I follow Lehner (2009) by using similar model settings. Consider a banking market and two kinds of players, borrowers and banks. Both borrowers and banks are risk neutral and try to maximize profits. Banks serve as the only source of capital. The market has two domestic banks, Bank 1 and Bank 2. Meanwhile, Bank F from another country is making entry into the market. Foreign bank and two domestic banks are located equidistantly along a circular city (Salop, 1979). In the market, the number of borrowers is m and all the borrowers are uniformly distributed along a circle with circumference 1.

Salop’s circular city model is capable of analyzing competition between more than two banks while Hotelling’s linear city model focuses on only two competitors.⁶ In addition, along circular city, locations of competitors are located equidistantly from one another.⁷The choice of location is exogenously imposed. This allows the model to focus on the choice of entry mode in a simple and tractable way.

Every borrower can undertake a project with initial investment of 1. With no cash at hand, borrower must apply for a loan at a bank to finance the project. Borrower can either take a good project, which generates a return R with certainty, or a bad project, which always yields a return of zero. A borrower must incur a linear form of travel cost t x , where x is the distance between borrower and bank and t is a parameter of travel cost, when applying for a loan.⁸ Borrower can only apply for a loan at one bank. I assume that banks can observe the location of borrowers so that borrower whose application is denied by one bank can’t apply at another bank.⁹ Comparing the interest rates offered by different banks and the travel cost determined by the relative distance, borrower will decide at which bank to apply for financing. In order to satisfy the individual-rationality condition, I assume that the return R is always big enough to cover the repayment and travel cost.

6In Dell’Ariccia (2001), the number of banks is N along a circular city and Lehner (2009) considers two domestic banks and one foreign bank in Salop’s model.

7Mueller (2007) only considers one foreign firm and one domestic firm in Hotelling’s model with two firms located at two endpoints.

8In spatial competition models, the cost usually refers to travel cost or transportation cost because of the distance between consumers and firms. However, in this paper, the cost can be regarded as application cost for borrower. Also it can serve as an indicator of development of the host country.

9This assumption is also used by Dell’Ariccia (2001) and Lehner (2009) to simplify analysis in a circular market model.

In the market, borrowers consist of old borrowers with a fraction ofθ and new borrowers with a fraction of 1− θ . The fraction of borrowers with good projects is λ and that with bad projects is 1− λ . A portion of borrowers with bad projects will be screened out based on banks’ screening technologies and the application of loan is rejected by banks. Through long-term relationships with clients, only domestic banks have access to soft information of the old borrowers. So, with the help of soft information, domestic banks can screen out all the "old" borrowers with "bad" projects. As for the new borrowers, the screening technology help domestic banks screen out a fraction 0≤ δD≤ 1 of borrowers with bad projects. On the other hand, since foreign banks just enter this market and haven’t established any long-term relationship with local borrowers, soft information is not available for foreign bank. All the borrowers are new to foreign bank. So, foreign bank can screen out a fraction 0≤ δF ≤ 1 of borrowers with bad projects. The costs of fund for domestic and foreign banks are i_Dand i_F. If foreign bank enters the host country via Greenfield investment, it incurs "efficiency barriers" due to less local knowledge to the host country. I define C as the costs which are caused by efficiency barriers. These costs undermine the profits of foreign bank. On the other hand, entry via joint venture gets the local knowledge by introducing a local partner and giving up part of profits. So, the problem of efficiency barrier can be alleviated. Also, entry via acquisition can reduce efficiency barrier by acquiring the knowledge of acquired domestic bank. Acquisition can increase the efficiency of foreign bank because foreign bank can distinguish between old borrowers and new borrowers by obtain the soft information held by acquired banks. However, an acquisition price P^AC is paid to acquired bank if foreign bank choose acquisition as its entry mode.

Generally, foreign bank’s profit is given by πF= mSF[λ(rF− iF) − (1 − λ)(1 − δF)(1 + iF)],

where SF is the market share for foreign bank. In a circular city model, market share for foreign bank S_F can be solved by finding marginal borrower who is indifferent to applying loan from foreign or domestic banks (Tirole, 1986). When entering via Greenfield investment, costs from efficiency barriers C must be subtracted fromπF. As for joint venture, C is discounted because of the help of local partner. So, costs from efficiency barriers via

joint venture becomeβC , where 0 < β < 1. Moreover, a fraction 1 − α of the total profits are distributed to local partner. So, foreign bank can only getα(πF− βC ). Acquisition price and costs from efficiency barriers lower the profits for acquisition. Foreign bank gets different discounts when it acquires different domestic banks. So, costs from efficiency barriers becomeρjC , where j= 1,2. The final profits for acquisition is πF− P^AC− ρjC .

4 Analysis of Entry Modes

4.1 Greenfield investment

The fraction λ of good borrowers make repayment so that the foreign bank can get margin r_F−iF. Since foreign bank doesn’t have any access to soft information, it can’t screen out old borrowers with bad projects. Foreign bank needs to evaluates all the borrowers, including old and new borrowers. So, foreign bank can only screen out the fractionδF of all borrowers with bad projects. The fraction 1− δF of bad borrowers get financed and don’t make any repayment. That results in default loss, (1 − λ)(1 − δF)(1 + iF). The total market share is mS^{G R}_F , where m is the number of total borrowers in the market and S^{G R}_F is the demand for foreign bank. The marginal borrower who is indifferent between getting loan from foreign bank or domestic bank is located at ˆx, where ˆx is the distance between foreign bank and marginal borrower and ¹₃− ˆx is the distance between domestic bank and marginal borrower. Comparing the travel cost or application cost and interest rates offered by foreign bank and domestic bank, marginal borrower will be indifferent to getting loans from foreign or domestic bank. The equality is ˆxt + rF = (¹₃− ˆx)t + rD. ˆx = ¹₆−^rD_2t^−rF. Since foreign bank is located equidistantly from two domestic bank The market share S^{G R}_F equal 2ˆx. And market share S^{G R}_D for domestic bank is¹₃+ ˆx. Foreign bank incurs efficiency barriers, or information costs when entering another country. Efficiency barriers decrease profit of Greenfield investment by the amount of C . So, the profit for the foreign bank is given by

π^{G R}_F = mS^{G R}_F [λ(rF− iF) − (1 − λ)(1 − δF)(1 + iF)] −C , (1)

with

S^{G R}_F =1

3+r_D1+ rD2− 2rF

2t .

On the other hand, domestic banks can screen out all the old borrowers with bad projects with soft information and don’t incur efficiency barriers. Profits for two domestic banks are the same, the profit is given by

π^{G R}_Dj = mS^{G R}_D [λ(rD− iD) − (1 − θ )(1 − λ)(1 − δD)(1 + iD)], j = 1,2 (2)

with

Since the locations for both foreign and domestic banks are assumed to be equidistant, banks compete in interest rates,rF, rD, to maximize their profits. Lemma 1 shows the equilibrium profits when foreign bank selects Greenfield investment.

Lemma 1. If foreign bank enters another market via Greenfield investment, the equilibrium profits for foreign bank and domestic banks are given by

π^{G R}_F = mλt 1

Proof. The marginal borrower between domestic bank and foreign bank is give by

r_F+ t ˆx = rD+ t 1

whereˆx is the distance between foreign bank and marginal borrower and¹₃− ˆx is the distance between marginal borrower and domestic bank. Borrowers within the range of ˆx will apply for loans at foreign bank. So, multiplying ˆx by 2 yields the market share for foreign bank.

Since two domestic banks are symmetric, their market share are equal to ¹₃− ˆx+ ¹₆. The market shares are given by

The loss of profits resulted from efficiency barriers is C . So, the profit functions are given by π^{G R}_F = mS^{G R}_F [λ(rF− iF) − (1 − λ)(1 − δF)(1 + iF)] −C and

π^{G R}_Dj = mS^{G R}_D [λ(rD− iD) − (1 − θ )(1 − λ)(1 − δD)(1 + iD)] j = 1,2.

Differentiating the profit functions with respect to interest rates and letting the F.O.C. equal

The lowest interest rates that banks can bid are defined by

r˜F≡ iF+(1 − λ)(1 − δF)(1 + iF)

λ and (5)

˜

r_D≡ iD+(1 − θ )(1 − λ)(1 − δF)(1 + iD)

λ . (6)

Solving the three reaction functions yields equilibrium interest rates:

r_D^{G R}

The equilibrium profits are given by π^{G R}_F = mλt 1

I defineΦ as an indicator for the relative efficiency between foreign bank and domestic bank. Technically, from the profit function,Φ is the difference between the lowest interest rates at which foreign bank and domestic bank can bid when efficiency barrier is not taken into consideration. In the profit functions, efficiency barrier results in the reduction in profit margin for foreign bank and increase in the profit margin for domestic banks. The coefficient of relative efficiency is positive for foreign bank. That’s because ifΦ positive, it means that

foreign bank is more efficient than domestic banks and foreign bank will gain more market share and higher interest margin in the profit function. On the contrary, if Φ is negative, it means that foreign bank is less efficient than domestic banks and will lose some market share and interest margin.