Economic variables and the lies we tell ourselves

When regulators propose scenarios for use in climate stress tests, they have particular narratives in mind.  They will imagine a story unfolding in the intended manner and then infer the shape of economic pathways that, in their view, are consistent with the narrative.

But economic variables – like gross domestic product (GDP), median house prices, and the consumer price index (CPI) – are statistics that must be constructed; they do not simply fall from the sky.  Measuring the underlying concepts – the amount of income, the value of housing stock, and the general price level – is a difficult technical challenge that economists have grappled with for many decades.  The economic variables we know and love all have technical quirks, and their own strengths and weaknesses, which can have a material impact on the nature and quality of the analysis we attempt to undertake.

In simple terms, when it comes to GDP et al, we think we are measuring a particular concept – but in reality, the data is measuring something different entirely.

Here I want to talk about how the data for these three concepts might behave if the climate scenarios described by regulators actually start to play out.  I’m going to assume that the scenarios are specified in terms of the core underlying concepts.  In other words, if the scenario describes a 10% decline in GDP, they actually mean a 10% decline in economic income. 

This distinction will become clear – I hope – when we get into the actual concepts.