Predicting the unpredictable: how economic models can shape our understanding uncertainty in economic scenario modelling

Econometric scenarios, just like a road trip, are subject to uncertainty. Even with careful planning, unexpected events can occur: a detour due to road construction, a sudden rainstorm, or maybe even a flat tire. These unpredictable events can throw your plans into disarray and make it difficult to reach your destination on time.

Just like road trips, economic scenarios are also subject to uncertainty. Economists use complex models to predict how the economy will behave in the future, but these models can’t account for every possible event. Unexpected “shocks”, like a global pandemic or a sudden change in government policy, can disrupt even the most carefully crafted scenarios.

Our newly published article on building renovation policy in Norway (Perez-Valdes et al, 2024) showed how economic modelling can provide insights into decision- and policy-making through macroeconomic analysis. Economic future is full of uncertainty, which can significantly influence a model’s ability to deliver reliable results. Specifically in the context of input-output modelling, uncertainty in data estimation of econometric interactions over many simulated periods can lead to large effects in the final results. This uncertainty can originated from many sources, like a shifting policy landscape, inaccuracies in econometric prognoses, or missing data points. To handle these challenges, our article showed how different types of uncertainty affecting the systems can be methodologically treated, in an illustrative case on building renovation for energy efficiency. The goal for this methodology is to make macroeconomic analysis more robust.

One major source of uncertainty is data limitations. Economic models rely on vast amounts of data to represent how different industries produce goods, how consumers spend their money, and how the government influences the economy. However, this data can be incomplete, inaccurate, or out of date, which can lead to errors in the model’s predictions.

Another source of uncertainty is the inherent unpredictability of the future. Even with the best data and the most sophisticated models, economists can’t foresee every event that might impact the economy. Unexpected events, like natural disasters, political upheavals, or technological breakthroughs, can significantly alter the economic landscape and render even the most accurate forecasts obsolete.

Finally, the structure of the economic model itself can also introduce uncertainty. We make choices about how to represent different parts of the economy and how they interact with each other. These choices can influence the model’s predictions and introduce uncertainty, especially when dealing with complex or poorly understood phenomena.

Different Types of Uncertainty

In the context of dynamic econometric input-output models, we have identified four broad types of uncertainty:

• Pathway Uncertainty: This type of uncertainty deals with the big-picture direction of the economy. This is equivalent of choosing a route for our car trip. Major trends, like globalization, technological change, or demographic shifts, can significantly impact the economy’s long-term trajectory, but their exact course is often uncertain, but we have a rought idea of it.
• Regression Uncertainty: This arises from the use of statistical methods to forecast future economic variables, like consumer spending or government investment. Returning to the car trip analogy, this includes external factors liks the state of the roads, detours, amount of trafic, and so on. They can be helpful guides, but they’re not perfect, and the margin of error can vary depending on the data available and the forecasting method used.
• Coefficient-Matrix Uncertainty: This type of uncertainty focuses on the relationships between different sectors of the economy. It is more internal, akin to our car’s condition, variable fuel efficiency, number of stops to rest and use the toilet. In economic models, these relationships are represented by a matrix of coefficients that show how changes in one industry, like a decrease in car production, can affect other industries, like steel manufacturing or transportation. However, these relationships can be complex and subject to change, which introduces uncertainty into the model.
• Qualitative-Estimation Uncertainty: This arises when economists have to make educated guesses about certain aspects of the economy due to a lack of reliable data. We can think of it as similar to Coefficient-Matrix Uncertainty but even harder to quantify, like the mood of a child in the backseat. In economic models, this can happen when dealing with new or rapidly evolving industries, where historical data is scarce or difficult to characterise.

Applied Case: Bulding  Renovation in Norway

The methods for dealing with uncertainty in economic models have applications in various fields, from urban planning and energy policy to environmental management and public health.

For example, we presented a case study on building renovation in Norway used stochastic modeling to account for uncertainty in factors like energy prices and construction costs. The results highlight the importance of understanding uncertainty in building renovation policies for policymakers to make informed decisions.

Moreover, the study’s findings can help policymakers understand the trade-offs between different policy options. For instance, a policy that mandates energy-efficient renovations in all new buildings may be more expensive in the short term but could lead to greater long-term benefits in terms of reduced energy consumption and emissions. By using economic models to quantify these trade-offs, policymakers can make more informed decisions that balance short-term costs with long-term benefits.

Our case study on building renovation in Norway demonstrates how our economic models can be used to inform policy decisions and guide investments in critical areas, even in the face of uncertainty. By embracing uncertainty and developing tools to manage it, we can make more informed decisions and create a more sustainable and prosperous future.

This content is based on Perez-Valdes, G. A., Wiebe, K. S., & Werner, A. T. (2024). Uncertainty in dynamic econometric input-output models: a Norwegian case study. Economic Systems Research, 1–21. https://doi.org/10.1080/09535314.2024.2413552. Research for the original article was financed through SINTEF Industry – Department of Sustainable Energy Technology’s easi-System project (102023679).