Table of Contents
A System Dynamics Model of Bank Geographic Diversification
John J. Voyer
Professor of Business Administration
voyer@usm.maine.edu
207-780-4597
Presenting author
Contact author
Herbert Smoluk
Assistant Professor of Finance
hsmoluk@usm.maine.edu
207-780-4407
Bruce H. Andrews
Professor of Business Administration
bandrews@usm.maine.edu
207-780-4185
Fax: 207-780-4662
University of Southern Maine
School of Business
Portland, ME 04104-9300
March 15, 2003
A System Dynamics Model Of Bank Geographic Diversification
After seventy years of placing restrictions on banks that wished to operate in more than
one state, in 1994 the federal government enacted the Riegle-Neal Interstate Banking and
Branching Efficiency Act, which went into effect in 1997. With this one move, the government
opened up the arena for banks wishing to operate across state lines. This was supposed to help
banks in at least two ways. First, allowing bank holding companies to convert other-state
subsidiaries into branches would allow for the elimination of duplicated overhead, increasing
bank efficiency (Carow and Heron, 1997; Berger and DeYoung, 2001). Second, since different
regions of the country have different economic profiles, banks would more easily be able to
diversify their assets and liabilities, reducing risk (Hughes, Lang, Mester, & Moon, 1996;
McLaughlin, 1995; Spiegel & Gart, 1996). The hoped-for policy effect of this change was a
system of stronger banks better able to compete against domestic non-bank financial institutions
and foreign banks in the U.S., and foreign banks in their home countries (Matasar & Heiney,
2000).
Rather than focus on the policy aspects of this change, the present paper will examine its
effects on the cognitive map of one large regional bank holding company. It will examine how
the change was incorporated into this bank’s cognitive map, but will be even more concerned
with the implications of the bank's mental model, testing some of those implications with output
from a system dynamics model.
COLLECTIVE MENTAL MODELS AND SYSTEMS THINKING
Weick (1979) suggested that, through a process of enactment, selection and retention, an
organization’s members create a cognitive schema, or map, of the most important aspects of their
collective experience. This schema channels future action, leading to further refinements of the
map, leading to future action, and so on. Senge (1990) describes the phenomenon of cognitive
maps through the term “mental models,” which can apply to either individuals, groups or
organizations. The present paper will use the terms “cognitive map,” “cognitive schema,” and
“mental model” interchangeably.
RESEARCH OBJECTIVES
The present paper examines the geographical expansion of one bank, Aquilo Bank (a
pseudonym). More precisely, it examines the implications of the cognitive map of the
geographical expansion strategy articulated by Aquilo's upper management. The paper examines
these implications by using a system dynamics model of Aquilo's cognitive map.
DESCRIPTION OF SETTING
In the early 1990s Aquilo Bank, on the verge of bankruptcy, was placed into receivership
by a federal court. It had its major operations (including about 20 branches) in one relatively
small state, a small number of branches in an adjacent state, and about $2 billion in assets. At
that point, Aquilo’s management set two targets: one million customers and $10 billion in
assets. Today, Aquilo is a $21.1 billion banking and financial services company with over two
million customers and banking divisions in six states. Aquilo is the third largest banking
company in its region and is one of the 100 largest commercial banking companies in the United
States. It operates over 300 branches and 400 ATMs throughout its region. Aquilo Bank is a
good example of a banking organization that availed itself of a geographical expansion strategy.
The question is, what mental model lay behind this remarkable performance? And what are the
implications of this mental model?
METHOD
The method chosen to address this phenomenon was system dynamics (Forrester, 1961;
Richardson & Pugh, 1981). System dynamics is a quantitative simulation method that attempts
to assist decision makers in thinking systemically about problems they confront. System
dynamics projects are roughly divided into two phases—qualitative and quantitative. The
qualitative phase develops the mental model that depicts the decision makers’ understanding of
what is creating the phenomenon. The quantitative phase attempts to create a model whereby
assumptions and relationships in the mental model can be tested and “played with.”
Given the time constraints on Aquilo Bank management, the present study used a data
gathering approach more streamlined than is typical for system dynamics. The CEO of the bank
was interviewed for approximately one and one-half hours. Three other executives from the
bank (two in strategic planning and one from retail operations) were also interviewed for
approximately the same amount of time. From these interviews, the researchers were able (with
feedback from Aquilo’s managers) to create a causal loop diagram of the cognitive map of the
geographical diversification strategy of Aquilo Bank’s top management, as shown in Figure 1.
(See the figure's caption for a brief explanation of links and loops.)
The present paper will discuss both the loops in Aquilo's mental model and the
implications of the quantitative model that has been developed from the data, turning to an
examination of the results of test runs done using the system dynamics model that was developed
from Aquilo's mental model.
CAUSAL LOOP RESULTS
Causal loops
Causal loop diagrams are composed of system elements whose causal linkages are
symbolized by arrows. Arrows indicate the direction of causality. Signs indicate the polarity of
relationships. A “+” sign means that, all else equal, if the cause increases (decreases), the effect
increases above (decreases below) what it would otherwise have been. A “-” sign means that, all
else equal, if the cause increases (decreases), the effect decreases below (increases above) what it
would otherwise have been. Loops are designated “R” for “reinforcing” or “B” for “balancing.”
Reinforcing loops signify that increases or decreases in the value of any variable in a series of
connected variables are ultimately amplified (“reinforced”) as their effects are traced through
that connected series. In balancing loops, increases in the value of any variable in a series of
connected variables are ultimately decreased (and decreases are ultimately increased) in a way
that “balances” or dampens these changes. Causal structures dominated by reinforcing feedback
loops generate behavior that moves component values progressively away from initial or
equilibrium values. In contrast, structures dominated by balancing feedback loops generate
behavior that resists continued change in a given direction and that directs system behavior
toward a goal or equilibrium condition (Goodman, 1974). Balancing feedback loops and rein-
forcing feedback loops combine to produce much of the behavioral change observed over time in
organizations (Forrester, 1961; Senge, 1990).
Reinforcing loops
Aquilo Bank’s cognitive map contains five reinforcing loops and nine balancing loops.
Reinforcing Loop R1, “Revenue Drives Acquisition,” (see Figure 2) is straightforward. Aquilo
makes an acquisition, which leads to more customers, more revenue, and more profit. The
higher profit raises the confidence level of Aquilo’s management, which, to close the loop, leads
to even more acquisition. Loop R2, “Confidence Drives Acquisition,” (Figure 2) says simply
that as the amount of acquisition rises, so does the management team’s confidence, which further
drives up acquisition.
Loop R3, “Pressure to Expand Geographically,” (Figure 3) comes from Aquilo’s
beginning position in a smallish state. To reach the asset and customer targets mentioned earlier,
it felt the need to acquire, which necessitated geographical expansion (because it could not reach
those targets from in-state operations only). This reduced the bank’s traditional preference for
operating in contiguous states, which in turn increased the number of regions available, leading
to more acquisition opportunities and, ultimately, even more pressure to acquire. Loop R4,
“Culture-Based Expansion,” (also in Figure 3) depicts how geographical expansion increased the
opportunities to acquire banks with similar culture, something that was important to Aquilo’s
management.
In Loop RS, "Attractiveness From Convenience," (shown in Figure 4) the bank begins by
increasing its acquisitions. This leads to more banks and an increase in automatic teller
machines (ATM). These in turn increase the convenience factor and the bank's attractiveness,
leading to more customers. This increases revenue, profit, and earnings per share, which raises
management's confidence level, leading to even more acquisition, closing the loop.
Balancing loops
Balancing Loop B1, "Cost Brake on Acquisition," (Figure 5) shows one of the systemic
negative feedback processes in this cognitive map. As acquisition increases, so does the number
of banks owned by the corporation. This leads to higher costs and lower profits and earnings per
share (the primary profit target monitored by management). Lower EPS reduces management's
confidence, leading to less acquisition, closing the balancing loop.
Loop B2, "Get More Assets," (also shown in Figure 5) is technically a balancing loop,
but it also depicts a major engine of growth for Aquilo Bank. Another term often used in system
dynamics for balancing loop is "goal seeking loop," and Loop B2 is a classic example. In this
instance, the goal it is seeking is more assets. As desired assets increase, the asset shortfall
increases, leading to pressure to acquire. The resulting acquisition increases the number of banks
and assets, closing the asset shortfall. This reduces the pressure to acquire and shuts down this
loop, but only after the bank has grown considerably larger than it was at the outset.
A similar dynamic can be seen in Loops B3, B4 and BS, shown in Figure 6. All three are
seeking the goal of more customers, and each responds to the resulting customer shortfall with a
different action. In Loop B3, "Community Bank Strategy," the bank responds to the customer
shortfall with efforts to improve its community bank image, which increases its attractiveness
and leads to more customers, closing the customer shortfall. Loop B4, "Customer Growth
Through Selling," shows how the bank reacts to the customer shortfall by increasing its cross-
selling, which leads to more customers (since a person with two accounts is considered two
customers). Loop BS, "Customer Retention," builds off the cross selling by reducing the churn
rate. Churn rate is the rate at which customers close out accounts and move their business to
another bank. A customer with multiple accounts at Aquilo Bank is less likely to do this, so
increased cross selling reduces churn, which increases customers, closing the customer shortfall.
It is interesting to note that two of the previously mentioned reinforcing loops, R1,
"Revenue Drives Acquisition," and R5, "Attractiveness From Convenience," also close the
customer shortfall by increasing the number of customers. In the case of R1, as banks are
acquired, they automatically bring new customers with them, closing the customer shortfall. In
R5, newly acquired banks bring ATMs with them, increasing Aquilo's attractiveness and
drawing new customers. This is an interesting example of balancing and reinforcing loops
working in concert to drive corporate growth. Raising the goals in the balancing loops leads to
actions within those loops that will close the resulting gaps, but also leads to the kicking in of
growth-oriented reinforcing loops.
By contrast, the four remaining balancing loops, B6 through B8 (shown in Figure 7) and
B9 (shown in Figure 4), are all examples of "brakes" on growth. In Loop B6, "Size Brake on
Acquisition," as pressure to acquire increases, acquisitions increase, which lower the bank's
ability to control its organization. This motivates a reduction in the number of regions to which
the bank might expand, thereby reducing the opportunities for acquisition and the pressure to
acquire, closing the balancing loop. In Loop B7, "Geographic Brake on Acquisition,"
acquisitions lead to an increase in regions served, which reduce the bank's ability to control its
organization and its motivation for seeking new regions in which to acquire, which ultimately
lowers the opportunities for acquisition and the pressure to acquire. The reduction in the number
of regions served reduces, in Loop B8, "Brakes on Culture Based Expansion," the number of
banks with cultures compatible with Aquilo's, which reduces the number of acquisition
opportunities and therefore the pressure for acquisition. Lastly, in Loop B9, "Image Brake on
Acquisition," we see the dark side of ATMs. While they provide convenience, they detract from
Aquilo's community bank image, reducing its attractiveness and leading to a reduction in the
growth rate of new customers, revenue, profit, acquisition, and, ultimately, in ATMs, thereby
closing this balancing loop.
POLICY TESTING RESULTS
There are dozens of possible policy choices that can be simulated in the model. The
present paper will discuss only a few, and will concentrate on the effects of particular policy
choices on the bank’s actual strategy. As Senge (1990) pointed out, the key to change and
growth in systems like Aquilo's is to find the leverage points and exert pressure on them. Aquilo
Bank’s cognitive map suggests that if it wants to continue to grow, it should consider, among
other things, these policies:
1. Implement systems that will improve control, even over longer distances or larger
organizational size. Or, learn to live with less-than-perfect control.
2. Let go of the idea that its managers must be able to drive to contiguous states to
perform their jobs.
3. Be willing to look at potential acquisitions where the cultural match is less than
perfect.
4. Let go of the idea that it is a community bank and simply go for growth and size.
5. Expand outside of home area, where returns might be higher.
Each of these policies was tested in the model. The results of these tests on Net Income and
Earnings per Share (EPS) are shown in Figures 8 and 9, respectively.
Improving control. Aquilo Bank’s management very much wants to maintain control
over its operations. It feels that it must contain its expansion for that to happen. But what if the
bank were to improve its systems or in other ways boost its capacity to control its operations?
The model shows that there would be only a negligible increase in net income (Figure 8), and
actually a decrease in EPS (Figure 9), compared to its actual strategy. What happens is a "fix
that fails." Increasing management's ability to control leads to a larger number of banks being
acquired, but this leads to lower ability to control, so the policy of increasing control proves to be
self-defeating. This is probably not a high leverage policy for Aquilo Bank, although it probably
needs to work on increasing the potency of its control systems merely as good practice and
continuous improvement, and to enable it to control its ever-expanding empire.
Letting go of preference for contiguous-state operations. Aquilo Bank’s management
has a preference for expanding from state to contiguous state. This, in their judgment, allows for
greater, and easier, control. It also appears to be a part of the company's somewhat cautious
culture. Were Aquilo to relax this preference, the result would be an increase in net income (see
Figure 8), but little effect on earnings per share (Figure 9). This is a "fix that succeeds a little
bit." By relaxing this cultural preference, Aquilo Bank allows itself to expand to a greater
number of regions, and acquires a greater number of banks. This yields more assets on which to
earn a return. This is a bit more promising than the increasing-control policy.
Less-than-perfect cultural match. Aquilo Bank is very concerned that any bank it
acquire be compatible with its culture. This obviously restricts the number of possible
acquisition or merger candidates. What if the bank became less choosy about the culture of its
prospective partners? Figures 8 and 9 show that the effects on profits and EPS are negligible.
By being less "picky" about its targets, Aquilo Bank increases the number of banks it might
acquire, but this has only a small effect on its goal for number of banks, and as a result very little
changes. This is a very low leverage policy.
Abandon community bank image. Aquilo Bank, like many banks, wants it both
ways—it wants to be large, but it wants customers to think of it as a friendly community bank.
What if it abandoned this image-building stance by simply going for larger size? The model
shows that this is a high leverage policy, showing a 33% increase in net income and a 19%
increase in the earnings per share. In essence, this is the strategy that Aquilo Bank chose in the
first place, i.e., to grow. But there is an even higher leverage strategy available to the company.
Expand outside of home area. In a previous study (citation withheld for blind review
purposes), it was shown that Aquilo Bank’s home area was actually a low one for bank profits.
Other areas, some far away from its base but others very much within striking distance, had the
potential for much higher profits. When that is entered into the model, along with the desire to
grow and a loosening of the contiguous-state preference (both prerequisites for this strategy), net
income and earnings per share rise the most of all—49% for net income and 25% for earnings
per share. Clearly this is the strategy that Aquilo Bank should pursue, but to do so it would have
to let go of two of the most subtle aspects of its collective mental model—its preference for
contiguous-state operations and its desire to remain a “community” bank.
CONCLUSION AND NEXT STEPS
Many other policy test could be conducted with this model. The effects of increased
cross-selling effort, greater emphasis on customer service, and product attractiveness could all be
tested. The model, if set at equilibrium, could assess the effects of single and combination
policies on a small bank (i.e., Aquilo when they came out of receivership) to see which would
have been the highest leverage. Lastly, by moving away from Aquilo Bank itself and trying to
simulate a more “generic” bank, many loops could be added to the model. The model also needs
to refine its profit loop to see those effects more clearly. Many other tests and changes are
possible. Much could be done with, and learned from, this model. But all of these lessons would
require a far longer manuscript than is currently possible. Suffice it to say here that the present
exercise has illuminated a clear course of high-leverage action for Aquilo Bank. Whether it
chooses to act on this learning is its decision.
10
REFERENCES
Berger, A.N. & R. DeYoung. 2001 " The effects of geographic expansion on bank efficiency."
Journal of Financial Services Research 19, n2 (Apr/Jun): 163-184.
Carow K.A. & R.A. Heron. 1998 " The interstate banking and branching efficiency act of 1994:
A wealth event for acquisition targets." Journal Of Banking & Finance 22, n2 (Feb): 175-
196.
Forrester, J. 1961. Industrial dynamics. Cambridge: MIT Press.
Hughes, J. P., W. Lang, L. J. Mester, and C. Moon. 1996 "Efficient Banking under Interstate
Branching." Journal of Money, Credit, and Banking v28, n4 (Nov.): 1045-71
Matasar, A. B. & J.N. Heiney. 2000 "Lemonade or Lemon? Riegel-Neal and the Consolidation
of American Banking." International Advances in Economic Research, May, Vol. 6 Issue
2, pp. 249-258
McLaughlin, S. 1995 "The impact of interstate banking and branching reform: Evidence from
the states." Current Issues in Economics & Finance 1, Issue 2 (May): 1-5.
Richardson, G., & A. Pugh III. 1981. Introduction to system dynamics modeling with dynamo.
Cambridge: MIT Press.
Senge, P. 1990. The fifth discipline: The art and practice of the learning organization. New
York: Doubleday Currency.
Spiegel, J. W. & A. Gart. 1996 "What Lies Behind the Bank Merger and Acquisition Frenzy?"
Business Economics v31, n2 (April): 47-52
Weick, K. 1979. The social psychology of organizing, 2nd ed. Reading, MA: Addison-Wesley.
1979.
11
al
ATMs per Image brake on
Bank ‘sctulatton .
+, Convenience —
Community - Effors to Improve
factor Bank Image"
“ota Arig Desired — Communiy Bank
Attractiveness Assets + ac uy
rom convenience | ce ans vin (28h
+ Attractiveness
{ Assets —™ Asset culturebaved
\ ‘Shortfall — \ ‘ext pansion \
— Investment in eS \
i i + 4 } Opportunities for
poo Organization | eres or Gap @ |
Bank Costs + Pressure for Investment in culture Beaed |
rc + ‘Acquistion Technology (gy ere es Community |
Bt | | Bank Strategy
/
Geographic brake
Cost bake on
ccaulttion Number of cnacquistion |
i Shares Peculiar Desire for Small
: +” + Avaliable —— Town Locations
ae caringe ee nets
+ + Share Gaocrephical Control | Desire for
eorephis . —simarty of
me Nas Customers
. ‘ares Regions Pressure to Desire to
acquisition Seas th Dominate Market
Management red
se enenerent, Geog Bentaisn geographically
acquisition Desire for Operations
Revenue + pin Contiguous States
PS sine \ :
Fee Income per "sean _
Customer ______chstomers a
a “a Desired
42) Susi Customers
so ‘Shortall mma
chum Kate through seling
customer cross #7
fetenton es
Figure 1Aquilo Bank Geographical Diversification Cognitive Map
Note: Arrows indicate the direction of causality. Signs indicate the polarity of relationships, a “+” sign meaning that, all else equal, if the cause
increases (decreases), the effect increases above (decreases below) what it would otherwise have been. Loops are labelled “R” for “reinforcing,
signifying that increases or decreases in the value of any variable in a series of connected variables are ultimately amplified (“reinforced”) as their
effects are traced through that connected series.
ye
Reveme™
Customers
a
; es
Profit
Revenue
acquisitio
(map Acquisitio:
> Confiden ah
Earnings per drive
Share acquisitio
#
Management
+ Team
Confidence
Figure 2 Reinforcing Loops 1 and 2
Banks With
Similar Culture. +
+
a Opportunities for (Rf
Acquisition + «
Pressure for ulture-based
Acquisition expansion
Om
* Regions
Available
Need for geographically
Geographical -
Expansion
Desire for
Operations in
- Contiguous States
Figure 3 Reinforcing Loops 3 and 4
13
+
Number of Total
+ Banks ATMs
Convenience
Acquisition R5 factor
+
C ifsttr™' iveness from
community convenience
Bank Image +
Management Team ae Bank
Confidence i) Attractivenes|
a +
Image brake on
acquisition +
Customers
Earnings per
Share
+
Revenue +
Profit<¢————
+
Figure 4 Reinforcing Loop 5 and Balancing Loop 9
Desired Assets
Bank Costs gf
‘Assets
* Asset
a Shortfall
Number of
- Banks
. t=) Pressure for
Bt Get more assets Acquisition
Cost brake on
acquisition
‘Acquisition _*
‘Eamings per
Share
ae Management
+ Team Confidence
Figure 5 Balancing Loops 1 and 2
14
Community _+
{ Bank Image i
Efforts to Improve
Community Bank
Image
Bank
Attractiveness
Community bank
strategy
Customer
___ Shortfall d
Coe Cwiomes ees
Churn Rate Customer growth
through selling
- Customer
retention
‘Cross Selling *
Figure 6 Balancing Loops 3, 4 and 5
B8
Banks with similar
Brakes on
i culture-based culture
Pressure for .q———— Opportunities expansion +
Acquisition for Acquisition (2 loops)
Las BT Regions
Acquisition (@} Available
Geographic brake ‘
on acquisition
+ Ability to
Regions =, “Control
Served Meg
acquisition
Figure 7 Balancing Loops 6, 7 and 8
15
Effects of Various Strategies on Net Income
400M
298.5 M
197M
95.5M
-6M
1992 1994 1996 1998 2000 2002 2004 2006 2008
Time (Year)
Actual Strategy 1 1 2 1 2 2 $/Year
More Control 2 2 2 2 $/Year
Relax Contiguous-State Preference 2 2 2 2 $/Year
Loosen Grip of Culture 4 4 4 4 4 4 $iveae:
Relax Community Bank Strategy ——5- 5 5 5 5 5— s/Year
Expand Outside Home Arca 6 6 A 6 s/Year
Figure 8 Leverage-Testing Results for Net Income
Effects of Various Strategies on Earnings per Share
4
2.9
—<—
1.8 &
0.7
-0.4
1992 1994 1996 1998 2000 2002 2004 2006 2008
Time (Year)
Actual Strategy 4 4 3 3 4 4 $/(Year* share)
More Control 2. 2: 2 2 2— $/(Year*share)
Relax Contiguous-State Preference ES z 3 $/(Year* share)
Loosen Grip of Culture 4 4 4 4 4 $/(Y cart share)
Relax Community Bank Strategy —5- 5 5 5 5 $/(¥ear* share)
Expand Outside Home Arca 6 6 6 6 $/(Year* share)
Figure 9 Leverage-Testing Results for Earnings per Share
16
Back to the
CC BY-NC-SA 4.0