Epistemic Risk and Requisite Variety
Epistemic risk relates to the risk of being wrong. Requisite variety relates to every possible way of being wrong. To manage epistemic risk, your epistemic "inspection process" or methods of checking, validating, and governing risk must have enough variety to match the possible errors which could occur.
As was pointed out in this blog post, "completeness" is not produced by reasoning; what is considered "complete" must be declared or governed in some way. That is one thing that the Seattle Method provides, a definition or declaration of completeness. What authority do I have to define what is complete? Well, I am not really defining what is complete, I am defining eight indisputable things that are necessary for a financial statement to be considered correct. What I am prescribing is necessary to be considered correct, but it might not be considered sufficient.
The ways one can be wrong when creating something like an XBRL-based report when a report model can be customized is greater than when a standardized "form" is used where the report model cannot be changed.
So, if "customization" increases the risk of error; then why not always use "standardized" forms? Well, life is always a trade off. When using standardized forms you run the risk of loss of idiosyncratic details of information which is being collected. Permitting reporting economic entities to customize their report models preserves idiosyncratic detail when the customizations are created correctly.
If you only prepare for one type of mistake; you will very likely be blindsided by other types of mistakes. Understanding the Law of Requisite Variety helps you better understand the dynamics of what is going on.
Effectively managing and controlling requisite variety requires that there be a balance or "matching" between the potential "perturbance variety" or potential possible disturbances which may occur within a system and the "control variety" which is the information such as rules available to the system to make sure the "residual variety" is as close to 0 as possible, preferably equal to zero. Basically, requisite variety manages your risk control process.
To manage epistemic risk, your epistemic “inspection process” which is your methods of checking, validating, governing that risk must have enough variety to match the possible errors.
Key to managing epistemic risk is for stakeholders, usually within an area of knowledge, is for the stakeholders to agree on the goals and objectives of a system; governance. Governance mechanisms are used to agree on those stakeholder goals and objectives, which has an impact on possible perturbances and therefore on perturbance variety, which impacts the necessary control variety, which impacts residual variety, which then impacts epistemic risk and therefore the risks which must be managed.
The point is to create a virtuous cycle. The feedback loop of this virtuous cycle continually improves both the knowledge representation of the area of knowledge and the system as a whole. The virtuous cycle makes the following core characteristics possible:
- Clarity. Clear thought processes and communication processes with an understanding of the difference between objective facts and subjective opinions and beliefs.
- Theory: A sound approach grounded in proven theory which demonstrates the validity of the approach which drives work.
- Discipline: Standard procedures and processes that are grounded in good practices and best practices drive all work. There is no room for spontaneity or letting people simply do work "their own way". Accountancy is not all "art" or unconscious "wild behavior" or "anarchy". There is some art in accountancy.
- Rigor: Best practice standards-based processes and procedures are followed precisely. Rigor is not about inflexibility. Rigor is about understanding the reason for the best practices.
- Feedback: Processes and procedures are not pulled from thin air. Rather, they are carefully determined using quantitative feedback from measurements and thorough testing to make adjustments to systems, techniques, processes, and procedures.
Consider the following metaphor; imagine building your own house:
- Epistemic risk: What is the chance the house isn’t safe (bad materials, poor design, faulty construction, inferior craftsmanship).
- Requisite variety: The inspection team must have enough different skills (structural engineer, electrician, plumber, code inspector) to catch all possible faults.
- Governance: Creating proper building codes that inspectors, engineers, builders, craftsmen like plumbers all need to adhere to enables effective communications.
If the inspection only covers one dimension (say, just the roof), epistemic risk remains high.
Epistemic risk is the possibility of error in knowing. The Law of Requisite Variety is the principle that only a sufficiently varied set of controls matched with controls can absorb those errors. You can’t eliminate epistemic risk, but you can manage it if your system’s variety matches the variety of possible mistakes which could occur.
Additional Information:
Comments
Post a Comment