ABSTRACT The following concepts are brought together to prove that - - PDF document

1

COMPLEXITY AND ENDOGENEITY IN ECONOMIC MODELING

M d l Al Ch dh Masudul Alam Choudhury Visiting Professor Summer 2012 Department of Economics Carleton University Professor of Economics College of Economics and Political College of Economics and Political Science (CEPS) Sultan Qaboos University Muscat, Sultanate of Oman masudc@squ.edu.om

ABSTRACT

The following concepts are brought together to prove that although economic models ought to be socially embedded in an extended and extendable field of interactions with diverse world-systems. Such interactions yet lead to predictability and

• controllability. Thereby a degree of nonlinearity exists by way

y y g y y y

• f evolutionary learning in such systemic reality. Yet anomie is

logically avoided. The resulting model configuration requires an epistemological approach that we refer to as unity of knowledge by virtue of pervasive complementarities between selected variables as normatively required. The episteme is contrary to that of rationalism upon which much of economic (socio-scientific) modeling is otherwise premised. 1. Interrelatedness between the concepts of Complexity, Endogeneity, and Circular Causation g y 2. The model formalism with (1) 3. Knowledge-induced social embedding phenomenon 4. From positivism to normative reconstruction and simulacra of repeated evolutionary learning processes 5. From ‘estimation’ to ‘simulation’ and repeated processes 6. Empirical confirmation of wellbeing simulation by the method

• f circular causation

2

OBJECTIVE

• Paper introduces a methodology for economic modeling

that establishes extensive inter-causal organic relations by evolutionary learning in the variables and their embedded systems. Such inter-causal organic relations are able to analyze the systemic complexity by treating y y p y y g completely endogenous inter-relationship between the variables.

• The paper considers the nature of complexity in socio-

scientific methodology, particularizing to economics. it introduces the meaning of pervasively endogenous inter- causal relationships between the variables that arise from the complexity of endogenous relations.

• We then mesh up the Complexity and Endogeneity

p p y g y characteristics of social realism by the method of circular causation.

• An empirical example is provided to explain the nature

and role of circular causation in evolutionary learning model of unity of knowledge as the underlying episteme.

COMPLEXITY

The semblance of uncontrollability of interactions among variables in economic models that defies predictability is formally models that defies predictability is formally replaced with order out of chaos (self- governing). A positivistic solution (estimation) is replaced by the simulative nature of interactions caused by ever- increasing linkages between causally increasing linkages between causally related variables. Complexity is a property

• f systems, yet with controllability and

predictability in our case, contrary to models with anomie (disorder or entropy).

3

FORMAL DEFINITION OF COMPLEXITY

• “The fundamental characteristic of complexity is the

fact that, in the study of the evolution of dynamical t th t l it d l ith th t f th systems that complexity deals with, the nature of the system in question is usually irrelevant. In the vision proposed by complexity, we can identify forms and evolutive characteristics common to all, or almost all, systems that are made up of numerous elements, between which there are reciprocals, nonlinear between which there are reciprocals, nonlinear interactions and positive feedback mechanisms. These systems, precisely for this reason, are generally called complex systems.”

• COMPLEXITY IN SYSTEM VIEW

Separately provided

4

COMPLEXITY IS CAUSED BY INCOMPLETENESS OF INFORMATION FLOW: EVOLUTIONARY LEARNING SYSTEM ANALYSIS: PHILOSOPHY OF SCIENCE G d l’ i l t f th ith ti l Godel’s incompleteness of the arithmetical sentences: undecidability and logical incompleteness Popper’s open conjectural deductive-inductive dialectical system Kant’s rationalism in modeling heteronomy History of the future: past or consciousness Pervasive system-embedded interdependency

• II. ENDOGENEITY

Endogenous variables are system-variables. In economic theory such variables arise from interactions in market exchange process interactions in market exchange process. Exogenous variables are determined independently of the market process; system- relations do not exist for exogenous in terms of endogenous variables and between them Circular Causation: Endo variables are inter Circular Causation: Endo-variables are inter- related in a circular way in generating and reinforcing the embedded effects between the variables in the state of complex interrelations. .

5

EPISTEMOLOGICAL UNDERLINING OF ENDOGENEITY TO SOLVE COMPLEXITY

the methodological application of the circular causation interrelations between the variables reveals the degree of entangled

• complexity. CC converts this into solvability,

predictability, and controllability over the evolutionary learning processes. Endogeneity is thereby epistemologically driven by a normative worldview based on driven by a normative worldview based on correcting and reformulating a given problem situation by ethical construction.

EPISTEMOLOGICAL QUESTIONS IN ECONOMIC MODELING Contra episteme of unity of knowledge: Rationalism is the epistemological Rationalism is the epistemological underlining of mainstream economic modeling due to its exogenous content. Rationalism is understood as dichotomy between the Kantian type a priori reasoning and a posteriori reasoning (e.g. monetary and fiscal policies; monetary neutrality of real output at full-employment rare of output).

6

COMPLEXITY, ENDOGENEITY AND CIRCULAR CAUSATION This system of organic inter-relations proves the theorem that, disentangling of g g complexity, endogeneity of relations between the variables, and the circular causation relations between them are coterminous in economic modeling with extensive endogeneity extensive endogeneity. Pervasive complementarities prevail between the selected variables in normative epistemological sense of unity

• f knowledge.

INTEGRATED FORMALISM

• Simulate

W(X()); X() = {x1, x2, …, xn}[] (1)

• Subject to, xi() = fi(xj(),);

(2)

• where xj() = {X()}ji i,j = 1,2,..,n
•  = F(X())  W(X())

(3)

The revolutionary scientific contribution: would be to include inter-causal embedding. The emergent discipline of economic modeling will then enable methodological convergence of the varied disciplines towards an explanatory whole. Such a convergence of systemic embedding establishing interrelations means endogeneity establishing interrelations means endogeneity that disentangles complexity by means of the circular causation interrelations. The complete phenomenological model is given later.

7

THE NEW EPISTEMOLOGICAL FOUNDATION IN ECONOMIC MODELING

The new epistemic methodology for attaining inter- variable and inter-system integration is premised on t i it f k l d Th d f t i systemic unity of knowledge. The order of systemic differentiation reflected in the prevalent nature of economic modeling, arising ultimately from the epistemology of rationalism, is replaced by the episteme of unity of knowledge intra- and inter-

• systems. The resulting theory of economic modeling

causes every variable to be endogenously interrelated with the other selected variables through the process of embedding in and across systems. There is Only One Exogenous Variable: The Epistemic

• foundation. There can be lagged knowledge-flows

across processes over the dimensions of knowledge, space and time.

THE PHENOMENOLOGICAL (ECONOMIC) MODEL PREMISED ON UNITY OF KNOWLEDGE

• Event: {, X(),t}, with (,S)

(4)

• X() = (x1, x2, …,xn)[]
• The total phenomenology of the evolutionary learning system as the unique
• The total phenomenology of the evolutionary learning system as the unique

and universal methodology of all systems by virtue of the episteme of unity of knowledge as follows (Figure 2):

• Process 1(in time period 1) Process 2 (in time period 2)
• [S (,S)] 

{} X() (5)

• 
• {,X(),t} W(X();t)
• Simulate by circular causation

i fi( j( ) ) i j 1 2

• xi = fi(xj(;t), ), ij = 1,2,..,n
•  = F(X();t)  continuity across
• 

evolutionary learning epistemic recalling  processes of

• simulacra.
• Figure 2: The phenomenological model of unity of knowledge in ‘everything’: Economics

as embedded evolutionary learning system in unity of knowledge across systems with endogenously knowledge induced variables in knowledge, space and time dimensions

8

RECURSIVE CIRCULAR CAUSATION RELATIONS BETWEEN GROWTH AND POVERTY VARIABLES AND THE WELLBEING INDEX

• lnG = -2.632 + 1.564lnP + 1.737lnTHETA

(A1)

• t-stats:
• 7.735 6.554

24.362

• R2 = 0.986
• lnP = 1.618 + 0.482lnG – 0.895lnTHETA

(A2)

• t-stats

44.109 6.554

• 8.615
• R2 = 0.904
• lnG = 1.573 + 0.562lnG -0.940lnP

(A3)

• t-stats:

10 860 24 362 -8 615

• t-stats:

10.860 24.362 -8.615

• R2 = 0.991

COMPUTER GENERATED DIAGRAMS: CIRCULAR CAUSATION RESULTS 1 SPSS

• 1. SPSS
• 2. Simulation: Spatial Domain Analysis

separately provided