MADDALA-NELSON EFFECTIVE DEMAND PRICE ADJUSTING DISEQUILIBRIUM MODEL     
          
OCZKOWSKI(1997, AUSTRALIAN ECONOMIC PAPERS,36, 283-307). 
MODEL DEFINED BY EQN (1)-(4) & (9)-(10) PP287-8.


QD(I) = A(1)*XD(I,1) +........ + A(K1)*XD(I,K1) + A(K1+K2+5)*P(I)

QS(I) = A(K1+1)*XS(I,1) +..... + A(K1+K2)*XS(I,K2) + A(K1+K2+6)*P(I)

EQD(I) = QD(I) + A(K1+K2+3)*MAX[0,(QD(I)-QSE(I))] +UD(I)

EQS(I) = QS(I) + A(K1+K2+4)*MIN[0,(QDE(I)-QS(I))] +US(I)

Q(I) = MIN[EQD(I),EQS(I)] 

P(I) - P(I-1) = (1/A(K1+K2+7))*(EQD(I)-EQS(I))

UD(I) AND US(I) ARE  NIID ERROR TERMS 

QD(I) AND QS(I) DEFINE NOTIONAL DEMAND AND SUPPLY
EQD(I) AND EQS(I) DEFINE EFFECTIVE DEMAND AND SUPPLY

K1   = NUMBER OF COEFFICIENTS IN DEMAND EQUATION                  
K2   = NUMBER OF COEFFICIENTS IN SUPPLY EQUATION                  
A( ) = COEFFICIENTS TO BE ESTIMATED; A(1),...,A(K1) ARE DEMAND    
       EQUATION COEFFICIENTS, A(K1+1),...,A(K1+K2) ARE SUPPLY     
       EQUATION COEFFICIENTS (EXCLUDING PRICE) 
       A(K1+K2+1) IS THE DEMAND ERROR TERM VARIANCE
       A(K1+K2+2) IS THE SUPPLY ERROR TERM VARIANCE 
       A(K1+K2+3) IS THE DEMAND MANIPULATION COEFFICIENT
       A(K1+K2+4) IS THE SUPPLY MANIPULATION COEFFICIENT
       A(K1+K2+5) IS THE DEMAND PRICE COEFFICIENT
       A(K1+K2+6) IS THE SUPPLY PRICE COEFFICIENT
       A(K1+K2+7) IS THE PRICE ADJUSTMENT COEFFICIENT
Q( ) = DATA OBSERVATIONS ON TRANSACTED AMOUNT
XD( )= DATA OBSERVATIONS ON RIGHT HAND VARIABLES IN DEMAND (EXCLUDE PRICE)
XS( )= DATA OBSERVATIONS ON RIGHT HAND VARIABLES IN SUPPLY (EXCLUDE PRICE)
QSE( )=DATA OBSERVATIONS ON SUPPLY EXPECTATIONS BY DEMANDERS 
QDE( )=DATA OBSERVATIONS ON DEMAND EXPECTATIONS BY SUPPILERS
P( ) = DATA OBSERVATIONS ON PRICE 

THE MAIN PROGRAM SHOULD ALSO CONTAIN THE FOLLOWING STATEMENTS

	COMMON/USER1/Q,XD,XS,QSE,QDE,P
	COMMON/USER2/N,K1,K2
	COMMON/USER3/APP
	EXTERNAL MNEDP
	EXTERNAL SPLINE
      	N = ...
	K1= ...
	K2= ...
	NP= K1+K2+7
	APP=...
           .
           .
           .

	CALL OPT(X,NP,F,METHOD,ITERL,MAX,IER,ACC,MNEDP,ALABEL)


        

APP IS THE APPROXIMATION COEFFICENT FOR THE SPLINE APPROXIMATION                     
THE MNED SUBROUTINE SHOULD BE CALLED AT LEAST TWICE BY CHANGING APP.


N.B.  THE EXPECTATION VARIABLES QDE(I) AND QSE(I) CAN BE DEPEND ON PARAMETERS
BY INSERTING ADDTIONAL CODE IN MNEDP AND DEFINING ADDTIONAL PARAMETERS
STARTING AT A(K1+K2+8) AND REDFINING NP.
SET A(K1+K2+6) = 0 IF SUPPLY DOES NOT DEPEND ON CURRENT PRICE.