Proceedings of the World Congress on Engineering 2008 Vol II

WCE 2008, July 2 - 4, 2008, London, U.K.

Modeling of Banking Profit via Return-on-Assets and

Return-on-Equity
Prof. Mark A. Petersen & Dr. Ilse Schoeman ∗

Abstract—In our contribution, we model bank profitabil- per unit of assets may be represented by
ity via return-on-assets (ROA) and return-on-equity (ROE) in a
stochastic setting. We recall that the ROA is an indication of the
operational efficiency of the bank while the ROE is a measure Net Profit After Taxes
of equity holder returns and the potential growth on their invest- ROA (Ar ) = (1)
Assets.
ment. As regards the ROE, banks hold capital in order to prevent
The ROA provides information about how much profits are
bank failure and meet bank capital requirements set by the reg-
ulatory authorities. However, they do not want to hold too much generated on average by each unit of assets. Therefore the
capital because by doing so they will lower the returns to equity ROA is an indicator on how efficiently a bank is being run.
holders. In order to model the dynamics of the ROA and ROE, Let E r = (Etr , t ≥ 0) be the process representing the return-
we derive stochastic differential equations driven by Lévy pro- on-equity (ROE). Then the net profit after taxes per unit of
cesses that contain information about the value processes of net equity capital may be given by
profit after tax, equity capital and total assets. In particular, we
are able to compare Merton and Black-Scholes type models and
provide simulations for the aforementioned profitability indica- Net Profit After Taxes
tors. ROE (E r ) = (2)
Equity Capital.
Keywords: Stochastic Modeling; Lévy Process; Stochastic Differ-
From this relationship it follows that the lower the equity cap-
ential Equations
ital, the higher the ROE, therefore the owners of the bank (eq-
uity holders) may not want to hold too much equity capital.
1 Introduction However, the equity capital cannot be to low, because the level
of bank capital funds is subjected to capital adequacy regula-
One of the biggest economic considerations in the 21st cen- tion. Currently, this regulation takes the form of the Basel II
tury is the maintenance of a profitable banking system. The Capital Accord (see [1]) that was implemented in 2007 on a
main sources of bank profits originate from transaction fees worldwide basis. Also, from [9] and [8], it follows that there
on financial services and the interest spread on resources that are other measures of the profitability.
are held in trust for clients who, in turn, pay interest on the
asset (see, for instance, [7]). In our discussion, we derive dy- The main problems addressed in this paper can be formulated
namic models for bank profitability via Lévy processes (see, as follows.
for instance, [2], [6] and [13]) appearing in a Merton-type
model (see [10] and [4]). Lévy processes are characterized by Problem 1.1 (Modeling of Return-on-Assets): Can we de-
(almost surely (a.s.)) right-continuous paths with their incre- duce a Lévy process-driven model for the dynamics of the
ments being independent and time-homogeneous. Such pro- ROA? (Proposition 3.1 in Section 3).
cesses have an advantage over Brownian motion in that they
are able to reflect the non-continuous nature of the dynamics
Problem 1.2 (Modeling of Return-on-Equity): Can we de-
of the components of bank profit. In the related Black-Scholes
duce a Lévy process-driven model for the dynamics of the
model, the markets are complete but some risks cannot be
ROE? (Proposition 3.3 in Section 3).
hedged. In addition, the motivation for using Lévy processes
is their flexible (infinitely divisible) distribution which takes
short-term skewness and excess kurtosis into account. The paper is structured in the following way. In Section 2
we present a brief description of the stochastic banking model
There are two main measures of the bank’s profitability (see that we will consider. In the third section, we describe the
[11]). Let Ar = (Art , t ≥ 0) be the process representing the dynamics of two measures of bank profitability, viz., the ROA
return-on-assets (ROA). In this regard, the net profit after taxes and ROE. Section 4 contains numerical examples where we
∗ Department of Mathematics and Applied Mathematics North-West Uni-
compare a Lévy-process driven model with a model driven
versity (Potchefstroom); Private Bag X6001 Potchefstroom, North-West,
by a Brownian motion. In Section 5, we provide concluding
South Africa; The research was supported by: NRF-grant GUN 2074218; remarks and we point out further research problems that may
Email: Ilse.Schoeman@nwu.ac.za; Email: Mark.Petersen@nwu.ac.za be addressed.

ISBN:978-988-17012-3-7 WCE 2008

Proceedings of the World Congress on Engineering 2008 Vol II
WCE 2008, July 2 - 4, 2008, London, U.K.

2 The Banking Model

In our model, we consider the filtered probability space Value of Assets (A) = Value of Liabilities (Γ)
(Ω, F, (Ft )0≤t≤τ , P). As usual, we assume that P is a + Value of Bank Capital (K). (5)
real probability measure, F = (Ft )0≤t≤τ is the natural fil-
tration, F0 is trivial and Fτ = F. The jump process ∆L =
2.1 Assets
(∆Lt , t ≥ 0) associated with a Lévy process, L, is defined by
∆Lt = Lt − Lt− , for each t ≥ 0, where Lt− = lims↑t Ls is In this subsection, we discuss bank asset price processes. The
the left limit at t. Let L = (Lt )0≤t≤τ with L0 = 0 a.s. be the bank’s investment portfolio is constituted by m + 1 assets in-
cádlág version of a Lévy process. Also, we assume that the cluding loans, advances and intangible assets (all risky assets)
Lévy measure ν satisfies and Treasuries (riskless asset). We pick the first asset to be
the riskless Treasuries, T, that earns a constant, continuously-
Z Z compounded interest rate of rT . Profit maximizing banks set
|x|2 ν(dx) < ∞, ν(dx) < ∞. (3) their rates of return on assets as a sum of the risk-free Trea-
|x|<1 |x|≥1 suries rate, rT 1, risk premium, µr , and the default premium,
E(d). Here the unitary vector and risk premium are given by
Furthermore, the following definition of the Lévy-Itô decom-
position is important.
1 = (1, 1, . . . , 1)T and µr = (µ1 , µ2 , . . . , µm )T ,
Definition 2.1 (Lévy-Itô Decomposition (see [4])) Let (Lt ) respectively. Also, we have that the default premium is defined
be a Lévy process and ν its Lévy measure, given by equation by
(3). Then there exist a vector γ and a Brownian motion (Bt )
such that
E(d) = (E(d1 ), E(d2 ), . . . , E(dm ))T ,

E(di ) 6= 0 i-th asset is a loan,
Lt = γt + Bt + Lft + lim Lǫt ,
ǫ↓0 E(di ) = 0 i-th asset is not a loan.

where Lft is a compound Poisson process with a finite number The sum rtT 1 + µr covers, for instance, the cost of monitoring
of terms and Lǫt is also a compound Poisson process. However, and screening of loans and cost of capital. The E(d) compo-
there can be infinitely many small jumps. nent corresponds to the amount of provisioning that is needed
to match the average expected losses faced by the loans. The
m assets besides Treasuries are risky and their price process,
An implication of the Lévy-Itô decomposition (see [4]) is that S (reinvested dividends included), follows a geometric Lévy
every Lévy process is a combination of a Brownian motion process with drift vector, rT 1+µr +E(d) and diffusion matrix,
and a sum of independent compound Poisson processes. This σa , as in
imply that every Lévy process can be approximated a jump-
diffusion process, that is by the sum of a Brownian motion
with drift and a compound Poisson process. In this paper, Z 1  
we will consider Merton’s jump-diffusion model (see [10] and St = S0 + Iss rT 1 + µr + E(d) ds +
[4]) of L. Thus 0
Z 1 X
Iss σa dLs + ∆Ss 1{|∆Ss|≥1} , (6)
0 0<s≤t
Nt
X
Lt = at + seBt + Yi , 0 ≤ t ≤ τ, (4) where ItS denotes the m × m diagonal matrix with entries St
i=1 and L is an m-dimensional Lévy process. Also, ∆Ss is the
where (Bt )0≤t≤τ is a Brownian motion with standard devi- jump of the process S at time t > 0 and 1{|∆Ss|≥1} is the
ation se > 0, a = E(L1 ), (Nt )t≥0 is a Poisson process indicator function of {|∆Ss | ≥ 1}. We suppose, without loss
counting the jumps of Lt with jump intensity λ. The Yi (i.i.d. of generality, that rank (σa ) = m and the bank is allowed
variables) are jump sizes, the distribution of the jump sizes is to engage in continuous frictionless trading over the planning
Gaussian with µ the mean jump size and δ the standard devia- horizon, [0, T ]. Next, we suppose that ρ is the m-dimensional
tion of Yi . stochastic process that represents
 the current
 value of risky
assets. Put µa = rT + ρT µr + E(d) , σ A = seσa and
A typical bank balance sheet identity consists of assets (uses
of funds) and liabilities (sources of funds), that are balanced µA = µa + aσa . In this case, the dynamics of the current
by bank capital (see, for instance [5]), according to the well- value of the bank’s entire asset portfolio, A, over any reporting
known relation period may be given by

ISBN:978-988-17012-3-7 WCE 2008

Proceedings of the World Congress on Engineering 2008 Vol II
WCE 2008, July 2 - 4, 2008, London, U.K.

dAt
 Nt  EtR = Πnt − δe Et − δs Ot (12)
X
= At µa dt + At σa adt + s̃dBtA + d[ Yi − rT Dt dt
i=1
 XNt  We assume that the retained earnings remain constant during
= At µA dt + σ A dBtA + σa d[ Yi ] − rT Dt dt (7) the planning period so that dE R = 0. Therefore, the dynamics
i=1 of the net profit after tax may be expressed as
where the face value of the deposits, D, is described in the
usual way, and rT Dt dt represents the interest paid to deposi-
tors.  Nt 
X
2.2 Capital dΠnt = δe Et− µE dt + σ E dBtE + σe d[ Yi ] +
i=1
The total value of the bank capital, K = (Kt , t ≥ 0), can be δs r exp{rt}dt. (13)
expressed as

3 Dynamics of ROA and ROE

Kt = Kt1 + Kt2 + Kt3 , (8)
In this section, we derive stochastic differential equations for
where Kt1 , Kt2 and Kt3 are Tier 1, Tier 2 and Tier 3 cap- the dynamics of two measures of bank profitability, viz., the
ital, respectively. Tier 1 (T1) capital is the book value of ROA and ROE. The procedure that we use to obtain the said
the bank’s equity, E = (Et , t ≥ 0), plus retained earnings, equations is related to Ito’s general formula (see [13]). An
E R = (EtR , t ≥ 0). Tier 2 (T2) and Tier 3 (T3) capital (col- important observation about our aforegoing description of the
lectively known as supplementary capital) is, in our case, the assets and liabilities of a commercial bank, is that it is sugges-
sum of subordinate debt, O where Ot = exp{rt} and loan- tive of a simple procedure for obtaining a stochastic model for
loss reserves, RL . However, for sake of argument, we suppose the dynamics of the profitability of such a depository institu-
that tion. The solution of the SDE (13) is

Kt = Et + EtR + Ot . (9) 
1
Πnt = EtR + δe E0 exp σ E BtE + (µE − (σ E )2 )t +
For σ E = seσe and µE = (µe +aσe ) we describe the evolution 2
of O and E as XNt 
σe [ Yi ] + δs exp{rt}. (14)
i=1

dOt = r exp{rt}dt, O0 > 0 (10)

and 3.1 Return-on-Assets (ROA)

The dynamics of the ROA (see equation (1)) may be calculated

 Nt  by considering the nonlinear dynamics of the value of total
X
dEt = E E E
Et− µ dt + σ dBt + σe d[ Yi ] (11) assets represented by (7) and the dynamics of the net profit
i=1
after tax given by (13). One can easily check how efficiently
a bank has been managed over a certain past time period by
respectively. Where σe , µe and BtE are the volatility of E, monitoring the fluctuations of the ROA. A stochastic system
the total expected returns on E, and the standard Brownian for the dynamics of the ROA for a commercial bank is given
motion, respectively. below.

2.3 Profit
Let Πn be the bank’s net profit after tax which is used to meet Proposition 3.1 (Dynamics of ROA using Merton’s
obligations such as dividend payments on bank equity, δe , and Model): Suppose that the dynamics of the value of total
interest and principal payments on subordinate debt, (1+r)O. assets and the net profit after tax are represented by (7) and
Put δs = (1 + r). In this case, we may compute the retained (13), respectively. Then a stochastic system for the ROA of a
earnings, E R = (EtR , t ≥ 0), as bank may be expressed as

ISBN:978-988-17012-3-7 WCE 2008

Proceedings of the World Congress on Engineering 2008 Vol II
WCE 2008, July 2 - 4, 2008, London, U.K.

 
n −1
 dEtr = Etr [Πt ] δe Et µE + δs rOt
dArt = r
At δe Et (σ E )2 {(σ A )2 σa2 dBtA − σa2 } + σa2
 + δe Et (σ E )2 {2(σ E )2 + σe2 dBtE − σe2 }
 
A 2 A n −1 E
+ (σ ) − µ + [Πt ] {δe µ Et + δe rOt } dt −δe Et (σ E )2 + [σ E ]2 − µe + σe2 dt
 X Nt  
+ d[ Yi ]δe Et σ E {σ A σa dBtA − σa } + [Πnt ]−1 δe σ E Et − σe dBtE
i=1
   X Nt
+ [Πnt ]−1 δe σ E Et dBtE + [Πnt ]−1 σ E δe Et − σe + 2σ E σe dBtE d[ Yi ]
 
i=1

+ [Πnt ]−1 σ E δe Et + σ A σa dBtA − σa + δe Et σ E {2σ E σe dBtE − σe }

+ δe Et σ E [Πnt ]−1 dBtE {σ A σa dBtA − σa }  XNt 
 X Nt − δe Et (σ E )2 dBtE d[ Yi ] . (17)
−δe Et [Πnt ]−1 σ E σ A dBtA d[ Yi ] i=1
i=1

− σa dBtA . (15) PNt we consider the special case where Lt = Bt i.e.
Next,
i=1 Yi + at = 0.

Corollary 3.4 (Dynamics of ROE using Black-Scholes

Next, we consider the special case where Lt = Bt , i.e., Model): Suppose that Lt = Bt in equations (11) and (13).
P Nt
i=1 Yi + at = 0.
Then a stochastic system for the ROE (using Black-Scholes
model) of a bank may be expressed as

Corollary 3.2 (Dynamics of ROA using Black-Scholes  

n −1
Model): Suppose that Lt = Bt in equations (7) and (13). dEtr = Etr 2
[σe ] − µe + [Πt ] δs rOt + δe Et µe
Then a stochastic system for the ROA (using Black-Scholes 
model) of a bank may be expressed as −δe Et (σe )2 dt
  
+ [Πnt ]−1 σe δe Et − σe dBtE . (18)

dArt = Art {σa2 − µr + [Πnt ]−1 (δe µe Et + δs rOt )}dt

4 Numerical Examples
n −1 E A
+ [Πt ] δe σe Et dBt − σa dBt . (16) In this section, we simulate the ROA and the ROE of the SA
Reserve Bank (see [12]) over a two year period. There are
a few methods for simulating stochastic differential equation
(SDE). First we assume that the ROA and the ROE do have
3.2 Return-on-Equity (ROE) jumps. We therefore simulate the stochastic differential equa-
tions (15) and (17) by using Merton’s model. Note that in
The dynamics of the ROE (see equation (2)) may be calculated Merton’s model the driving Lévy process is a compounded
by considering the equation for the dynamics of the equity Poisson process.
capital given by (11) and the net profit after tax represented by
(13) and using Ito’s formula (see [13]). A stochastic system Although Lévy based models are structurally superior, the es-
for the dynamics of the return on equities for a commercial timation procedures are complicated. For comparative pur-
bank is given below. poses (see [14]), we compute the average absolute error (APE)
as a percentage of the mean ROA (or ROE) as

Proposition 3.3 (Dynamics of ROE using Merton’s 1

Model): Suppose that the dynamics of the value of total AP E =
mean ROA (or ROE) value
assets and the net profit after tax are represented by (11) and 24
X |Data value − M odel value|
(13), respectively. Then a stochastic system for the ROE of a ∗ .
(19)
bank may be expressed as i=1
number of ROA (or ROE) values

ISBN:978-988-17012-3-7 WCE 2008

Proceedings of the World Congress on Engineering 2008 Vol II
WCE 2008, July 2 - 4, 2008, London, U.K.

The dynamics of ROA using the Black−Scholes model The dynamics of ROE using the Black−Scholes model
2.4 35
Brownian motion Brownian motion
2.2 Data Data

30
2

1.8
25
1.6
dA 1.4 dE 20

1.2
15
1

0.8
10
0.6

0.4 5
J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D
Time Time

Figure 1: A solution of the SDEs (16) and (18) with a fitted linear trend line

The dynamics of ROA using Mertons model The dynamics of ROE using Mertons model
2.4 40
Merton Merton
2.2 Data Data
35
2

1.8 30

1.6
25
dA 1.4 dE
20
1.2

1 15

0.8
10
0.6

0.4 5
J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D
Time Time

Figure 2: A solution of the SDEs (15) and (17) with a fitted linear trend line

Another measure which also gives an estimate of the goodness the Euler-Maruyama method takes the form
or quality of fit is the root-mean-square error (RMSE) given by
Aj = Aj−1 + f (Aj−1 )∆t + g(Aj−1 )(B(tj ) − B(tj−1 )),
8
where j = 1, 2, · · · , 2R .
v
u 24
uX (Data value − M odel value)2 The following data on the ROA and ROE from the SA Reserve
RM SE = t . (20) Bank was used in our simulation.
i=1
number of ROA (or ROE) values
ROA/ROE ROA/ROE
We estimate the model parameters by minimizing the APE and Jan-2005 0.9/11.2 Jan-2006 1.3/16.4
the RMSE errors. In Table 2 we give the relevant values of Feb-2005 1.8/22.0 Feb-2006 1.3/16.9
APE and RMSE. The calibrated Lévy model is very sensi- Mrt-2005 1.0/12.0 Mrt-2006 1.2/14.9
tive to the numerical starting point in the minimization algo- Apr-2005 0.5/6.2 Apr-2006 0.8/9.8
rithm or small changes in the input data. In our case, we use May-2005 1.2/14.2 May-2006 1.0/13.0
Merton’s model with intensity λ = 2 (ROA case) or λ = 16 Jun-2005 1.2/13.9 Jun-2006 1.5/20
(ROE case) and the average jump size as µ = 0.06 (ROA) Jul-2005 1.6/19.5 Jul-2006 1.4/17.9
or µ = 0.01 (ROE). For another intensity the results of the Aug-2005 1.2/15.0 Aug-2006 1.8/23.4
minimization will be different. Sep-2005 0.7/8.7 Sep-2006 1.2/15.5
Oct-2005 1.1/13.3 Oct-2006 1.4/18.1
Secondly, we assume that the ROA and the ROE do not have Nov-2005 1.4/16.4 Nov-2006 1.1/14.5
jumps. In this case our SDEs (16) and (18) are driven by Dec-2005 1.5/18.2 Dec-2006 2.2/27.5
Brownian motions. We apply Euler-Maruyama Method to
simulate these SDEs over [0, T ] discretized Brownian path us- Table 1: Source SA Reserve Bank
ing time steps of size Dt = R ∗ dt for some positive integer R
and dt = 2T8 . For a SDE of the form
Using SA Reserve Bank’s data we get the following parameter
dAt = f (At )dt + g(At )dBt , 0 ≤ t ≤ T, choices σe = 0.69, µe = 0.06, σa = 0.01, µr = 0.003.

ISBN:978-988-17012-3-7 WCE 2008

Proceedings of the World Congress on Engineering 2008 Vol II
WCE 2008, July 2 - 4, 2008, London, U.K.

Also, for the Euler-Maruyama method we chose the value of should invest. This means that banking decisions and equity
net profit after tax as Πnt = 16878, the dividend payments on policy have to be simultaneously addressed by bank managers.
E as δe = 0.05, the interest and principal payments on O as Further investigations will include descriptions of the dynam-
δs = 1.06, the interest rate as r = 0.06, the subordinate debt ics of the other measures of bank probability.
O = 135 and the bank equity E = 1164.
References
Model APE(%) RMSE [1] Basel Committee on Banking Supervision, International
Black-Scholes (ROA) 27.58 0.3472 Convergence on Capital Measurement and Capital Stan-
Black-Scholes (ROE) 38.7145 6.2475 dards; A Revised Framework, Bank of International Settle-
Merton (ROA) 1.2588 0.06 ments, http://www.bis.org/publ/bcbs107.pdf, 2004.
Merton (ROE) 0.6124 0.3687
[2] Bingham, N.H., Kiesel, R.,Risk-Neutral Valuation: Pric-
ing and Hedging of Financial Derivatives, Second Edition,
Table 2: Lévy models: APE and RMSE Springer Finance, 2004.
[3] Cannella, A., Fraser, D., Lee, S., “Firm failure and man-
agerial labor markets: evidence from Texas banking.” Jour-
In Figure 1, we plotted the actual ROA (or ROE) values versus nal of Economics and Finance, V38, (1995), pp. 185210.
the Black-Scholes model values for the two years 2005 and
2006. From Figure 1 and Table 2 where we used the model [4] Cont, R., Tankov, p., Financial Modelling with Jump Pro-
without jumps it follows that the root-mean-square error for cesses, Chapman & Hall/Crc Financial Mathematics Se-
the ROA is 0.3472 and for the ROE it is 6.2475. ries, 2004.
[5] Diamond, D.W., Rajan, R.G., “A theory of bank capital,”
In Figure 2, we plotted the actual ROA (or ROE) values Journal of Finance, V55, (2000), 2431-2465.
versus the Merton’s model values for the two years 2005 and
2006. From Figure 2 and Table 2 where we used the model [6] Eberlein, E., Application of generalized hyperbolic Lévy
with jumps it follows that the RMSE for the ROA is 0.06 and motions to finance, Handbooks in Mathematical Finance:
0.3687 for the ROE. Option Pricing, Interest Rates and Risk Management,
(Cambridge University Press, Cambridge), 2001, pp. 319-
Note that the APE (%) decreases from 27.58 % to 1.2588 % 336.
and the RMSE value decreases from 0.3472 to 0.06 for the
ROA. Furthermore, in the ROE case the APE (%) decreases [7] Freixas, X., Rochet, J.-C., Microeconomics of Banking,
from 38.71 % to 0.6124 % and the RMSE value decreases Cambridge MA, London, 1997.
from 6.2475 to 0.3687. We therefore conclude that in the [8] Goldberg, L.G., Rai, A., “The structure-performance re-
ROA case and even more for the ROE case the Black-Scholes lationship for European banking.” Journal of Banking and
model performs worse than Merton’s model. However, we Finance V20, (1996), pp. 745-771.
still observe a significant difference from the data values. Note
that calibrations to other datasets can favor the Black-Scholes [9] Kosmidou, K., Pasiouras, F., Zopounidis, C., Doumpos,
model more. M., “A multivariate analysis of the financial characteristics
of foreign and domestic banks in the UK,” The Interna-
tional Journal of Management Science, V34, (2006), pp.
5 Conclusions and Future Directions 189-195.
Although the Black-Scholes model is powerful and simple to [10] Merton, R., “Option pricing when underlying stock re-
use, most profit indicators exhibit jumps rather than contin- turns are discontinuous,” J. Financial Economics, V3,
uous changes. Therefore, we have constructed asset-liability (1976), pp. 125-144.
models in a stochastic framework driven by a Lévy process for
[11] Mishkin, F.S., The Economics of Money, Banking and Fi-
two measures of commercial bank profitability. In this regard,
nancial Markets, Seventh Edition, Addison-Wesley Series,
the ROA, that is intended to measure the operational efficiency
Boston, 2004.
of the bank and the ROE that involves the consideration of the
bank owner’s returns on their investment was central to our [12] The South African Reserve Bank, www.resbank.co.za.,
discussion. These stochastic models arose from a considera- 2008.
tion of the bank’s balance sheet and income statements asso-
ciated with off-balance sheet items. [13] Protter, P., Stochastic Integration and Differential Equa-
tions, Second Edition, Springer, Berlin, 2004.
Discussions on the profitability and solvency of banking sys- [14] Schoutens, W., Lévy Processes in Finance, Pricing Fi-
tems are intimately related (see, for instance, [3]). In par- nancial Derivatives, Wiley Series in Probability and Statis-
ticular, asset-liability management by banks cannot be sepa- tics, 2003.
rated from the decision about how much equity the bank owner

ISBN:978-988-17012-3-7 WCE 2008