Estimating Power Law Exponents For Extreme Capital Losses Without Probability Scoring
Power law tail exponents measure drawdown severity scaling directly from ordered loss ratios, bypassing subjective probability scoring in fat-tailed domains.
log log regression
Meaning
Mathematical techniques involve transforming both the independent and dependent variables into natural logarithms before performing a linear analysis. Using log log regression allows researchers to identify power law relationships and constant elasticity between two sets of data. This approach is common in economics and biology where growth rates are prioritized over absolute changes.
The resulting coefficients represent the percentage change in the output for a one percent change in the input.
Elasticity Measurement
Constant proportional relationship is the primary output of this type of model. When a researcher applies log log regression, the slope of the line indicates how sensitive one variable is to another across different scales. For instance, if the coefficient is two, a small increase in the input leads to a doubling of the rate of change in the output.
This measurement remains consistent whether the values are small or large.
Scaling Law
Natural systems and market behaviors often follow patterns that are invisible in a standard linear view. By using log log regression, a practitioner can see if a process scales according to a specific exponent. This helps in predicting how a business or a biological population will behave as it grows by several orders of magnitude.
The transformation flattens the exponential growth into a straight line that is easier to quantify and project.
Transformation Requirement
Data used in this method must be positive because the logarithm of zero or a negative number is undefined. If a dataset contains zeros, log log regression cannot be applied without adding a constant or using a different model. This limitation means the technique is best suited for quantities like price, volume, or population size.