Empirical Inference of Numerical Information into Causal Strategy Models by Means of Artificial Intelligence - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 295

edges into the structural model. By analyzing the conditional behavior and the graphical model it
becomes possible to direct some or all of these edges. The most common of these causal search algo-
rithms—the PC-algorithm—has first been proposed by Pearl and Verma (1991). Spirtes et al. (2000)
used these principles for their causal proving software application TETRAD. Within this chapter, this
technique is employed to rule out third variable effects and is therefore explained in more detail in the
respective section.

A SYnThe TIC Con CePT of CAuSALITY BeTWeen BuSIne SS VAr IABLeS

Applied to managerial cause-and-effect relations, an appropriate concept of causality must restrict itself
to observational studies in terms of empirical data as a consequence. Therefore we can summarize the
above disquisition of definitions for causality as follows: A cause should provide information which
can be used to (partly) explain its effects. In case of linear cause-and-effect relations this property of
informational redundancy is also known as correlation or covariance. However, as it has been shown
in the introduction these concepts do not succeed to fully explain causality. According to Hume, there
has to be a temporal precedence relation between a cause and its effects additionally to informational
redundancy as for example Pearl and Verma (1991) state:

Temporal precedence is normally assumed essential for defining causation, and it is undoubtedly one
of the most important clues that people use to distinguish causal from other types of associations. (p.
442)

Regardless of the ability of these two necessary properties to fully explain many causal relations
there still remains the problem of an exogenous common cause to induce spurious associations between
presumably causal variables. Therefore the definition of causality has to be enhanced by the postulate
to control for this type of association. This leads to a notion similar to the one of Suppes (1970):

X and Y must covary, X must precede Y, and other causes of Y must be controlled.

Based on these foundations the following definition of an appropriate concept of causality to analyze
associations between managerial variables can be derived:

Theorem 1 (managerial causal relation): A causal relation between variables of a managerial system
exists if and only if there exist appropriate nomothetic (i.e. unproven) cause-and-effect hypotheses based
on causal a priori knowledge where the following conditions are fulfilled:

• The empirical observations of a potential cause provide informational redundancy regarding its

potential effect.

• The variation within the time series of the potential cause must always precede the response of

this variation within the time series of the potential effect.

• The three causality properties as defined above (causal a priori knowledge, informational redun-

dancy and temporal precedence) must not originate from the influence of a known or unknown
cause, common to the potential cause and the potential effect.