In your paper you point out that, in order to analyze the relation between economic growth and environmental impact, it is necessary to focus on the complex interactions between several aspects of structural change which may have an influence on economic growth and climate change. What approach do you take in addressing the complexity of the relations between the different aspects of structural and climate change within the modeling framework?

I have been working on structural change and its impact on economic performance (productivity, growth, levels and composition of employment) for several years now. As economists, we need to be bold in tackling complexity and dispelling assumptions that oversimplify analyses and/or are unrealistic.

Both the several dimensions of structural change taken into account in this paper and their interactions require models equipped to tackle complexity. This is particularly so when we aim to provide a reliable and exhaustive representation of the impact of economic growth on different aspects of climate change.

In the paper we assess several families of models, based on: their different theoretical approaches, the number of dimensions of (endogenous) structural change included, on how their interaction is modeled and whether they are micro-founded.

This choice is dictated by the belief that by assuming away the complexity of structural change in macro and policy models this may lead to wrong outcomes and predictions. For example, how do changes in the composition of industry affect changes in firm size, employment, wages, income distribution, and thus in consumption preferences? Different industries have different impacts on GHG, and therefore the specialization of countries matters. Consumers will purchase green goods, on the basis of income distribution and preferences. All these aspects will matter for GHG. Given that they influence each other, modelling their interaction is crucial.

You single out six different aspects of structural change, which can be unbundled into different components, in order to understand their relations with climate change. Can you explain the main findings arising from your analysis?

We start from six aspects of structural change and discuss the extent to which they are integrated and) their interactions (endogenously) modeled within different families of models: Integrated Assessment Models (IAM), Computable General Equilibrium (CGE) models, Structural Change Models (SCM), Ecological Macroeconomic models (in the Keynesian tradition) (EMK) and Multi agent and Evolutionary models (EABM). We highlight the strengths and weaknesses of each modelling approach, based on the criteria mentioned above. Results are summarised in the table below. 

EABM are comparatively best equipped to capture the interactions between several aspects of structural change and climate change, as they are micro-founded and are able to represent out-of-equilibrium scenarios, which are essential when modelling qualitative aspects of structural change.

What kind of limitations derive from the application of the main modeling traditions that investigate the relation between the economy and climate change, including aspects of structural change?

As summarised in the Table above, the IAM model accounts for some sectoral differences, distinguishing between sources of energy use, but the relative growth of these sectors is not endogenized; industrial organization is not relevant; technical change is modeled as marginal changes in an aggregate production function through LBD; the main demand changes are limited to time preferences. CGE models include a larger number of industries, related through I/O coefficients, with different learning curves and contributions to emissions, but these relations are fixed; they also include different consumption patterns of the average representative consumer, which does not allow for consumer heterogeneity. SCM models focus on the I/O interaction between industries and how they may change due to climate change. Nevertheless, we only found a couple of models that also introduce changes in the I/O coefficients and model the emergence of new intermediate industries. Differences on the demand side are not more sophisticated than those already included in CGE models. EMK models integrate ecological unbalances and monetary unbalances within a unique framework to study the relations between ecological and macroeconomic balance constraints. Although models are demand driven, there is no structural change on the demand side. The main contribution is in the analysis of the nexus between (un)employment dynamics, investment, environmental impact and economic growth.

In addition, three main assumptions underpinning all of the above families of models render their use problematic in accounting for the most significant components of structural change (and their relation with climate change). Firstly, perfect rationality (even with limited information) does not allow for true uncertainty, which characterizes the impact of economic activities on the environment, e.g. typically in terms of unintended consequences. Secondly, in the absence of heterogeneity of actors, there is no real scope for endogenous structural change. Thirdly, under the market clearing framework in which there is one unique equilibrium along which an economy grows, it is not possible to capture “the emergence of qualitatively different entities”. Even with multiple equilibria, it is not possible to model the transitional adjustments that crucially underpin structural change.

In your view, which family of models best suits the interaction between the analyzed aspects of structural and climate change? And how can it be applied to better address the different aspects of structural change?

EABMs relax the three basic assumptions of the models mentioned above: aggregate behaviour (or average behaviour in the case of microfounded CGE models); perfect rationality of heterogeneous interacting agents; out-of-equilibrium dynamics. As a result, they are comparatively best equipped to model evolving complex systems, which account for a higher number of aspects of structural change and model their interactions. For instance, they are able to consider economies in which: (1) the significance of industries changes across time and space; firms in different industries behave differently and have different incentives; (2) the relation between industries also changes constantly, affecting industrial dynamics, size, trade relationships; (3) technical change is a complex process per se, which involves non-measurable risks, investment, radical shifts and non-reversible choices, and determines future technical advances; (4) shifts in industries, industrial organization and technical change determine radical changes in the demand for labour, and therefore in wages and income distribution; (5) changes in income distribution and technologies induce changes in consumption preferences and behaviors; (6) all above changes depend on how institutions change; (7) last but not least, climate change has heterogeneous impacts on each of the above aspects of structural change, which is not predictable without considering how these impacts are distributed.

Given the restrictive assumptions of IAM, CGE models, most SCM and the aggregate construction of EMK models, to date only EABM seems able to take up the challenge of studying the interactions between several aspects of structural change and the environment.

We hope that this modelling tradition, albeit demanding, can inform policymaking in a more systematic way, particularly within the framework of the EU Green New Deal.