Financial Analysis, Planning and
Forecasting
Theory and Application
Chapter 26
Econometric Approach to Financial Analysis,
Planning, and Forecasting
By
Cheng F. Lee
Rutgers University, USA
John Lee
Center for PBBEF Research, USA
Outline










26.1 Introduction
26.2 Simultaneous nature of financial analysis, planning,
and forecasting
26.3 The simultaneity and dynamics of corporate-budgeting
decisions
26.4 Applications of SUR estimation method in financial analysis
and planning
26.5 Applications of structural econometric models in financial
analysis and planning
26.6 Programming vs. simultaneous vs. econometric financial
models
26.7 Financial analysis and business policy decisions
26.8 Summary
Appendix 26A. Instrumental variables and two-stages least
squares
Appendix 26B. Johnson & Johnson as a case study
26.1 Introduction
26.2 Simultaneous nature of financial
analysis, planning, and forecasting
 Basic
concepts of simultaneous
econometric models
 Interrelationship
of accounting
information
 Interrelationship
policies
of financial
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
 Definitions
of endogenous and
exogenous variables
 Model
specification and
applications
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
TABLE 26.1 Endogenous and exogenous variables
1. The endogenous variables are:
a) X1,t = DIVt = Cash dividends paid in period t;
b) X2,t = ISTt = Net investment in short-term assets
during period t;
c) X3,t = ILTt = Gross investment in long-term assets
during period t;
d) X4,t = -DFt = Minus the net proceeds from the
new debt issues during period t;
e) X5,t = -EQFt = Minus the net proceeds from new
equity issues during period t.
2. The exogenous variables are:
5
5
a) Yt = Σ Xi,t = Σ X*i,t, where Y = net profits +
i=1
i=1
depreciation allowance;
a reformulation of the sources = uses identity.
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
TABLE 26.1 Endogenous and exogenous variables (Cont.)
b) RCB = Corporate Bond Rate (which corresponds to the
weighted-average cost of long-term debt in the FR
Model [Eqs. (20), (23), and (24) in Table 23.10],
and the parameter for average interest rate in the
WS Model [Eq. (7) in Table 23.1].
c) RDPt = Average Dividend-Price Ratio (or dividend yield,
related to the P/E ratio used by WS as well as the
Gordon cost-of-capital model, discussed in Chapter 8).
The dividend-price ratio represents the yield expected by
investors in a no-growth, no-dividend firm.
d) DELt = Debt Equity Ratio (parameter used by WS in Eq. (18)
of Table 23.1).
e) Rt = The rates-of-return the corporation could expect to
earn on its future long-term investment (or the
internal rate-of-return discussed in Chapter 12).
f) CUt = Rates of Capacity Utilization (used by FR to lag
capital requirements behind changes in percent sales;
used here to define the Rt expected).
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
5
5
Σ Xi,t = Σ X*i,1 = Yt,
i=1
(26.1)
i=1
where X1,t, X2,t, X3,t, X4,t, X5,t, X*1,t and Yt are identical to those defined in Table
25.1.
Expanding Eq. (25.1) we obtain
X1,t + X2,t + X3,t + X4,t + X5,t = X*1,t + X*2,t + X*3,t + X*4,t + X*5,t = Yt.
(26.1')
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
X*t = AZt,
(26.2)
where
X*' = (DIV*IST*ILT* - DF* - EQF*),
Z' = (1 Q1 Q2 Q3 Y RCB RDP DEL R CU),
┌ a10 a11 ... a19 ┐
│ .
. │
A=│ .
. │.
│ .
. │
│ a50 a51 ... a59 │
└
┘
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
DIV*t = a10 + a11 Q1 + a12 Q2 + a13 Q3 + a14 Yt
+ a5 RCBt + a16 RDPt + a17 DELt
+ a18 Rt + a19 CUt,
IST*t = a20 + a21 Q1 + a22 Q2 + a23 Q3 + a24 Yt
+ a25 RCBt + a26 RDPt + a27 DELt
+ a28 Rt + a29 CUt,
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
ILT*t = a30 + a31 Q1 + a32 Q2 + a33 Q3 + a34 Yt
+ a35 RCBt + a36 RDPt + a37 DELt
+ a38 Rt + a39 CUt,
-DF*t = a40 + a41 Q1 + a42 Q2 + a43 Q3 + a44 Yt
+ a45 RCBt + a46 RDPt + a47 DELt
+ a48 Rt + a49 CUt,
-EQF*t = a50 + a51 Q1 + a52 Q2 + a53 Q3 + a54 Yt
+ a55 RCBt + a56 RDPt + a57 DELt
+ a58 Rt + a59 CUt.
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
Xi,t = Xi,t-1 + δi(X*i,t - Xi,t-1)
(26.3)
or
(a) X1,t = X1,t-1 + δ1(X*1,t - X1,t-1),
(b) X2,t = X2,t-1 + δ2(X*2,t - X2,t-1),
(c) X3,t = X3,t-1 + δ3(X*3,t - X3,t-1),
(d) X4,t = X4,t-1 + δ(X*4,t - X4,t-1),
(e) X5,t = X5,t-1 + δ5(X*5,t - X5,t-1).
25.3 The simultaneity and dynamics of
corporate-budgeting decisions
5
5
Σ X*i,t = Σ Xi,t = Yt.
i=1
i=1
X2,t = X2,t-1 + (1 - δ1)(X*1,t - X1,t-1).
5
Xi,t = Xi,t-1 + Σ δij(X*j,t - Xj,t-1)
j=1
5
Σ δij = 1.
i=1
(i = 1, 2, 3, 4, 5),
(26.4)
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
Xt = Xt-1 + D(X*t - Xt-1)
= Xt-1 + D(AZt - Xt-1)
= Xt-1 + DAZt - DXt-1
= (I - D)Xt-1 + DAZt,
┌ δ11 δ12 ... δ15 ┐
│ .
.
│ .
.
D=│ .
.
│ δ51 δ52 ... δ55 │
└
(26.5)
│
│.
│
┘
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
TABLE 26.2 An expanded version of Eq. (26.5)
X 1  DIVt  DIVt 1  11 ( DIVt*  DIVt 1 )  12 ( ISTt*  ISTt 1 )  13 ( ILTt*  ILTt 1 )  14 ( DFt*  DFt 1 )  15 ( EQFt*  EQFt 1 )
X 2  ISTt  IDTt 1   21 ( DIVt*  DIVt 1 )   22 ( ISTt*  ISTt 1 )   23 ( ILTt*  ILTt 1 )   24 ( DFt*  DFt 1 )   25 ( EQFt*  EQFt 1 )
X 3  ILTt  ILTt 1   31 ( DIVt*  DIVt 1 )   312 ( ISTt*  ISTt 1 )   33 ( ILTt*  ILTt 1 )   34 ( DFt*  DFt 1 )   35 ( EQFt*  EQFt 1 )
X 4   DFt   DFt 1   41 ( DIVt*  DIVt 1 )   52 ( ISTt*  ISTt 1 )   43 ( ILTt*  ILTt 1 )   44 ( DFt*  DFt 1 )   45 ( EQFt*  EQFt 1 )
X 5   EQFt   EQFt 1   51 ( DIVt*  DIVt 1 )   52 ( ISTt*  ISTt 1 )   53 ( ILTt*  ILTt 1 )   54 ( DFt*  DFt 1 )   515 ( EQFt*  EQFt 1 )
X = BXt-1 + CZt + Ut,
(26.6)
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
 DIV t 1 
 STI 
t 1 
TABLE 26.3 An expanded form of Eq. (26.6)

 LTI t 1 


DF
t 1 

 EQFt 1 


 DIV t   b11b12b13b14 b15 c10 c11c12 c13c14 c15 c16 c17 c18 c19   I 
 IST  b b b b b c c c c c c c c c c   Q 
t 
1

 21 22 23 24 25 20 21 22 23 24 25 26 27 28 29  

 ILTt    b31b32b33b34b35 c30 c31c32 c33c34 c35 c36 c37 c38 c39   Q2 

 



DF
b
b
b
b
b
c
c
c
c
c
c
c
c
c
c
t 

 41 42 43 44 45 40 41 42 43 44 45 46 47 48 49   Q3 
 EQFt   b51b52b53b54b55 c50 c51c52 c53c54 c55 c56 c57 c58 c59   Y 


RCB


 RDP 


DEL


 R 


 CU 
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
D = i - B,
(26.7)
A = D-1 C.
(26.8)
5
Σ Xit = Yt
for every period t.
i=1
5
5
i=1
i=1
Σ bij = Σ ĉik = 0 for all j and all k≠4,
and that
5
Σĉ = 1.
i=1
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
TABLE 26.4 Adjustment coefficients of
DIVt  DIVt 1
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
TABLE 26.4 Adjustment coefficients of
DIVt  DIVt 1 (Cont.)
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
TABLE 26.5 Summary of results
26.3 The simultaneity and dynamics of
corporate-budgeting decisions
Fig. 26.1 (From Spies, R. R., “The dynamics of corporate capital budging,”
Journal of Finance 29 (September 1974): Fig. 1. Reprinted by permission.)
26.4 Applications of SUR estimation method
in financial analysis and planning
 The
role of firm-related
variables in capital-asset
pricing
 The
role of capital structure in
corporate-financing decisions
26.4 Applications of SUR estimation method
in financial analysis and planning
R1t = α1 + ß1Rmt + γ11X11 + γ12X12 + γ13X13 + E1t,
R2t = α2 + ß2Rmt + γ21X21 + γ22X22 + γ23X23 + E2t,
.
.
.
Rnt = αn + ßnRmt + γn1Xn1 + γn2Xn2 + γn3Xn3 + Ent,
(26.9)
26.4 Applications of SUR estimation method
in financial analysis and planning
where
Rjt = Return on the jth security over time interval t
(j = 1, 2, ..., n),
Rmt = Return on a market index over time interval t,
Xj1t = Profitability index of jth firm over time interval
t (j = 1, 2, ..., n),
Xj2t = Leverage index of jth firm over time period t
(j = 1, 2, ..., n),
Xj3t = Dividend policy index of jth firm over time
period t (j = 1, 2, ..., n),
γjk = Coefficient of the kth firm-related variable in
the jth equation (k = 1, 2, 3),
ßj = Coefficient of market rate-of-return in the jth equation
Ejt = Disturbance term for the jth equation, and
aj's are intercepts ( j = 1, 2, ..., n).
Rjt = α′ + ß ′ + Ejt.
(26.10)
26.4 Applications of SUR estimation method
in financial analysis and planning
TABLE 26.6 OLS and SUR estimates of oil industry
26.4 Applications of SUR estimation method
in financial analysis and planning
TABLE 26.6 OLS and SUR estimates of oil industry (Cont.)
*t-values appear in parentheses beneath the corresponding coefficients.
†Denotes significant at 0.10 level of significant or better for two-tailed test.
‡Denotes significant at 0.05 level of significant or better for two-tailed test.
From Lee, C. F., and J. D. Vinso, “Single vs. simultaneous-equation models in capital-asset pricing: The role of firm-related variables,”
Journal of Business Research (1980): Table 3. Copyright 1980 by Elsevier Science Publishing Co., Inc. Reprinted by permission of
the publisher.
26.4 Applications of SUR estimation method
in financial analysis and planning
TABLE 26.7 OLS parameter estimates of oil industry-Sharpe Model*
* t-values appear in parenthesis beneath the corresponding coefficients
From Lee, C.F., and J.D. Vinso, “Single vs. simultaneous-equation models in capital-asset pricing: The role of firm-related variables.”
Journal of Business Research (1980): Table 2. Copyright 1980 by Elsevire Science Publishing Co., Inc. Reprinted by permission of the
26.4 Applications of SUR estimation method
in financial analysis and planning
TABLE 26.8 Residual correlation coefficient matrix after OLS
estimate
26.4 Applications of SUR estimation method
in financial analysis and planning
ΔLDBT = α1(LDBT* - LDBTt-1) + α2(PCB* - PCBt-1 - RE)
+ α3 STOCKT + α4RT + ε1,
(26.11)
ΔGSTK = ß1(LDBT* - LDBTt-1) + ß2(PCB* - PCBt-1 - RE)
+ ß3 STOCKT + ß4RT + ε2,
(26.12)
STRET = η1(LDBT* - LDBTt-1) + η2(PCB* - PCBt-1 - RE)
+ η4RT + ε3,
(26.13)
ΔLIQ = LIQ* + γ2(TC* - TCt-1) + γ3(ΔA - RE) + γ4RT
+ ε4,
(26.14)
26.4 Applications of SUR estimation method
in financial analysis and planning
ΔSDBT = ΔLIQ* + λ2(TC* - TCt-1) + λ3 (ΔA - RE) + λ4RT + ε5,
(26.15)
where
LDBT* = bSTOCK (i/i) = A target for the book value of long-term debt,
STOCK = Market value of equity,
b = LDM/STOCK = Desired debt-equity ratio,
LDM = Market value of debt = (LDBT)(i/i),
i/i = Ratio between the average contractual interest rate on long-term debt outstanding and the current new-issue rate
on long-term debt,
LDBTt-1 = Book value of long-term debt in previous period,
PCB* = Permanent capital (book value) = net capital stock (NK) + the permanent portion of working assets (NWA),
PCBt-1 = Permanent capital in the previous period,
RE = Stock retirements,
STOCKT = Stock-market timing variable = average short-term market value of equity divided by average long-term
market value of equity,
RT = Interest timing variable, weighted average (with weight 0.67 and 0.33) of two most recent quarters' changes
in the commercial paper rate,
TC* = Target short-term capital = short-term asset-liquid assets,
TCt-1 = Short-term debt in the previous period,
ΔA = Changes in total assets,
ΔLIQ* = Change of target liquidity assets.
26.5 Applications of structural econometric
models in financial analysis and planning
A
brief review
 AT&T’s
econometric
planning model
26.3 Applications of structural econometric
models in financial analysis and planning
Fig. 26.2 Tripartite structure of FORECYT.
26.6
Programming vs. simultaneous vs.
econometric financial models
26.2
Applications
of structural
econometric
models in
financial
analysis and
planning
Fig. 26.3
Flow chart of FORECYT.
(From Davis, B. E., G. C. Caccappolo,
and M. A. Chaudry, “An
econometric planning model for
American Telephone and
Telegraph Company,” The Bell
Journal of Economics and
Management Science 4 (Spring
1973): Fig. 2. Copyright © 1973,
The American Telephone and
Telegraph Company. Reprinted
with permission.
26.7
Financial analysis and business
policy decisions
26.8 Summary
Based upon the information, theory, and methods
discussed in previous chapters, we discussed how
the econometrics approach can be used as
alternative to both the programming approach and
simultaneous-equation approach to financial
planning and forecasting. Both the SUR method
and the structural simultaneous-equation method
were used to show how the interrelationships
among different financial-policy variables can be
more effectively taken into account. In addition, it
is also shown that financial planning and
forecasting models can also be incorporated with
the environment model and the management
model to perform business-policy decisions.
Appendix 26A. Instrumental variables and
two-stages least squares
Appendix 26A. Instrumental variables and
two-stages least squares
Appendix 26A. Instrumental variables and
two-stages least squares
Appendix 26B. Johnson & Johnson as a
case study
TABLE 25.A.1
Sales in Different Segment
2003
Division
Sales
2004
Profits
Sales
%
2005
Profits
Sales
%
2006
Profits
Sales
%
Profits
%
Consumer
18%
13%
18%
11%
18%
12%
18%
10%
Pharmaceuticals
47%
56%
47%
58%
44%
48%
44%
48%
Medical Devices
and Diagnostics
36%
31%
36%
31%
38%
40%
38%
43%
Total
100%
100%
100%
100%
100%
100%
100%
100%
Domestic
60%
59%
56%
56%
International
40%
41%
44%
44%
Total
100%
100%
100%
100%
Appendix 26B. Johnson & Johnson as a
case study
Table 26.B.2 Balance sheet
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Xt = BXt-1 + CZ + Ut
(26.B.1)
Appendix 26B. Johnson & Johnson as a
case study
Appendix 26B. Johnson & Johnson as a
case study
Xt-1 = BXt + CZt + Ut
= B2Xt-1 + BCZt + CZt+1 + BUt + Ut-1
(26.B.2)