How will our electricity system change in the future? Will we consume more electricity over time or less? And how will we meet our energy needs?

The answers to these questions have considerable public policy implications but our ability to forecast these changes has so far been limited.

Electricity forecasts just four years ago predicted strong, uninterrupted growth in electricity demand. In reality, demand has fallen for the past four years.

Electricity forecasts over timeadapted from AEMO data

Why does that matter? These overestimates have contributed to the over-investment in electricity network infrastructure, commonly referred to as “gold plating”. When the demand didn’t materialise as forecast, the result was a much-publicised electricity price shock.

Energy forecasts also underpin global scenarios of greenhouse gas emissions. In fact a new report from market analyst RepuTex suggests Australia has been overestimating future greenhouse gas emissions, in part due to poor energy demand forecasts.

A matter of perspective

There are two broad perspectives for analysing the energy system: the “top-down” perspective which views energy systems as a set of of markets, and the “bottom-up” perspective which views energy systems as a collection of appliances, equipment and infrastructure.

Both are critically important perspectives. The top-down perspective contributes important insights into how changes in markets impact the energy sector, including shifts in world fuel prices and economic activity. The bottom-up perspective contributes important insights into the impacts of changes in behaviour and technology.

Yet when it comes to energy policy making, the top-down perspective has dominated in Australia and internationally to our collective detriment, leading to unreliable forecasting, poor policy choices and billions in infrastructure spending that turned out to be unnecessary.

Conventional electricity forecasting

The dominant method for energy forecasting is a top-down modelling approach called econometric forecasting. This involves forecasting future energy requirements based on historical trends and correlations with a range of factors including weather, income, electricity prices and economic activity.

So why have the models overestimated demand? Subsequent research has revealed that the econometric forecasts failed to account for improvements in the energy efficiency of appliances and buildings, the rapid uptake of solar PV and structural shifts in industrial energy consumption.

In other words, the forecasts failed because they falsely assumed that historical correlations between energy demand and economic activity would remain consistent.

However, the efficiency of appliances and buildings has improved markedly in recent years thanks to a suite of regulated minimum performance standards, while many energy intensive industries have scaled back considerably. The result has been an incremental decoupling of energy demand from economic activity, making old models redundant at best, and misleading at worst.

A different approach

Instead of solely using statistical correlations to predict energy use, we can model energy demand using end-use forecasting.

This approach, currently being developed by the UTS Institute for Sustainable Futures and others, forecasts energy demand from the bottom up based on how energy is consumed by appliances, buildings, and industrial equipment.

The result is a much more detailed picture of how energy is consumed, including how changes in behaviour and improvements in technology will impact the energy system in the future.

Household electricity demand simulation

Biased forecasts to biased policy

A top-down bias also results in a bias towards top down policy. This is because the models we use to understand a policy problem inevitably frame the proposed solutions.

Energy policy makers in Australia and internationally have focused much of their attention on carbon pricing as a means to decarbonise our energy system. Although carbon pricing is desirable, it is far from a silver bullet.

Complementary measures are needed to overcome widespread market failures such as imperfect information, inefficient pricing and split incentives.

A top-down bias also leads to a bias toward supply-side solutions such as new electricity generators and network augmentations.

Energy policy makers and climate change campaigners alike often ignore the dominant role that energy efficiency will play in the decarbonisation of our economy. Energy efficiency is the largest and cheapest source of energy and greenhouse gas abatement.

Much like solar energy, energy efficiency is available in every home and business. It also reduces energy bills and improves industrial competitiveness.

A more integrated policy response therefore combines top-down carbon pricing policies with a tailored combination of policies including information, regulation and incentives.

For example, it is widely recognised that households and businesses often don’t take into account energy costs when purchasing buildings, appliances and vehicles. In response consumers need a tailored combination of clear energy labeling, minimum performance standards and in some cases rebates or finance, particularly for low income households.

The silver lining

The upshot of our historical top-down bias is that decarbonising our economy will be easier and cheaper than we previously imagined.

Mandatory performance standards in appliances and buildings are already driving significant reductions in the energy consumption of households and businesses. For example, a fridge purchased today consumes around half the electricity of a fridge purchased 20 years ago while air conditioners are now 50% more efficient.

Over a similar period solar PV and other renewable energy technologies have experienced cost reductions far beyond forecasts made even three years ago, leading to a dramatic uptake of renewable generation across Australia and internationally.

These shifts weren’t incorporated in electricity demand forecasts or climate change emission projections. This means we can achieve much more ambitious emission reduction targets at a lower cost than we previously thought possible.

But to get there, we need to raise our collective ambition and make smart decisions, based on a more nuanced understanding of how our society uses and generates energy today.

John McKibbin works for and owns shares in Ecologic Apps, a startup company that provides energy efficiency and solar PV assessment tools. He also receives funding from utilities and government agencies to provide energy policy advice and tools.

The Institute for Sustainable Futures at UTS undertakes contract research for electricity network businesses, governments, community groups and NGOs on economic, technical and regulatory matters regarding distributed energy. Ed Langham is also a member of Pingala, a Sydney based not for profit community energy association.