A robust data-driven macro-socioeconomic-energy model

Abstract

In a resource-constrained world with growing population and demand for energy, goods, and services with commensurate environmental impacts, we need to understand how these trends relate to aspects of economic activity. We present a computational model that links energy demand through to final economic consumption, illustrated by application to UK data. Our model fits within a whole-economy framework which harmonises multiple national accounting procedures. Our model minimises both the number of exogenous aspects and tuning factors by using historical data to calibrate relationships. We divide economic activity into a number of distinct but interdependent outputs that are non-substitutable in the short-term. The dynamic aspects assume that supply follows demand, but are constrained in the short-term by physical infrastructure. At the same time, capital formation grows the physical infrastructure. Our model regenerates historical data dynamically as a basis for projecting forward scenarios to discuss pathways to a lower carbon future.

Modelling socio-economic and energy data to generate business-as-usual scenarios for carbon emissions

“…The investment loop is crossed by a dashed arrow of shortfall from a shortfall circle. The manu loop applies control (Roberts et al, 2016) where the shortfall circle denotes the point in the loop where a mismatch of supply to demand is derived at each time-step. This mismatch gives rise to a change in FC4 to resolve the mismatch.…”

“…The 7see model (Roberts et al, 2016) operationalises this framework. The purpose of this approach is to express constraints between the component stocks and flows of the macroeconomy, for example, capital stocks and capital formation, specifically gross fixed capital formation, GFCF.…”

“…The evolution of relationships between stocks and data flows in a 7see-type model are captured by timevarying coefficients (Roberts et al 2016). To regenerate historical stocks and flows, these coefficients are ratios linearly interpolated between annualised historical data.…”

“…In Table 1. we disaggregate an economy into six industries along with other types of stocks which we number by j (Roberts et al, 2016). An issue with any model of this nature is distinguishing between endogenous and exogenous variables.…”

“…Furthermore, timeseries data should be used for as many parameters as possible, and be calibrated and tested using historical data. Fixed capital formation can be treated as an explicit flow within a model (Roberts et al, 2016), which is overlooked by most energy system model developers.…”