So, by changing the reference point relatively, if even absolute figures get changed, then it's not so linear transformation...It's likely higher order non-linear issue !
So, if absolute figure = f(reference point) then things might not be so straightforward linear ! The selection of reference point may not be so arbitrary!
Even Growth % can also be relative... 5% growth rate may not be same in every reference frame in higher order ?!
Statistical figures such as GDP are not directly observed; they are constructed through a measurement methodology and a chosen reference system, such as a base year, price structure, weights, deflators, or aggregation method. If changing the reference system merely rescales the original series linearly, the percentage growth rate remains unchanged. However, when the transformation is non-linear, non-proportional, or changes the underlying weights and methodology, the resulting statistical series can change in a way that also changes the measured growth rate.
Therefore, the robustness of a statistic should not be judged only by its estimation error. It is also important to test its reference-point and model fragility—that is, how much the reported figure changes under reasonable alternative reference points or measurement methodologies.
The broader principle is:
A statistical conclusion may depend not only on the underlying data, but also on the measurement model and reference system used to transform that data into the reported statistic.
Thus, for measures such as GDP, the relevant question is not only “What is the growth rate?”, but also “How robust is the measured growth rate to the choice of reference point and measurement methodology?” That's a deeper question to analyse statistically
Just for educational purpose
