Conference Agenda

How is productivity affected by increasing energy prices? To answer this question we exploit a comprehensive firm dataset comprising nearly all French manufacturing firms. We find differential productivity growth biased towards energy. This finding informs our modelling strategy in the second step where we introduce factor specific productivity growth into a scale invariant growth model with endogenous market structure. Due to a continuum of steady states, the level of the equilibrium growth path is affected by energy taxes and the initial technology mix. Differential growth of factor-specific productivities offers a complementary perspective for understanding sluggish labour productivity growth.

Electrification is a key lever for decarbonization. Achieving it requires not only abundant, low-cost clean electricity, but also new technologies that can convert electricity into valuable economic services. While the Directed Technical Change (DTC) literature emphasizes that innovation responds to relative profit opportunities, these profits are shaped by market structure and competition intensity. Building a novel model of directed technical change with endogenous markups across energy generating and consuming sectors, we analyze how imperfect competition and technological linkages jointly shape the pace and direction of electrification. We show that competition effects, in particular free exit and entry of firms, can weaken lock-in effects that typically favor fossil technologies. Furthermore, the competition effects dampen the speed of transition, especially under strong climate policy. Policies that combine carbon pricing with research subsidies and competition policies that reduce fossil-sector markups can direct innovation incentives, achieving faster and more cost-effective electrification.

With the expansion in both the level and scope of carbon pricing in Europe and elsewhere, the use of carbon price revenues has become a central question in climate policy design. This paper examines a fundamental trade-off: whether to recycle revenues to reduce distortionary labor taxes—generating a static efficiency gain—or to subsidize R&D for clean technologies to lower future decarbonization costs. We first develop a general-equilibrium model with distortionary taxation and endogenous directed technical change, showing that recycling revenues into clean R&D can, in principle, generate a dynamic fiscal dividend by expanding the future tax base. We then quantify these mechanisms using a calibrated extension of the model featuring heterogeneous households and a detailled representation of energy technologies. The numerical results show that recycling revenues through labour-tax reductions delivers robust aggregate welfare gains. In contrast, channeling revenues into electricity-sector R&D improves technologies but yields lower aggregate welfare than simple lump-sum transfers. This arises from induced sectoral misallocations when subsidies are not precisely targeted toward the underlying knowledge externalities.

Can growth and emissions be decoupled, and if so, through which mechanism?

This paper develops a framework in which three forces drive the energy transition:

(i)~capital installation, where investment in cleaner vintages reduces emission

intensity; (ii)~Directed Disembodied Technical Change (DDTC), where R\&D shifts the

production frontier; and (iii)~post-installation adjustment, where already-installed

capital adapts its emission intensity over time.

The third force is typically omitted from the literature, which treats capital as

either fully rigid or fully flexible after installation.

I capture it through a parameter $\theta\in[0,1]$ measuring the intensity of

post-installation adjustment frictions.

Studying the decentralized economy, I show that the transition is shaped by three

market failures. Each requires its own instrument: a monopoly markup on R\&D

intermediates, an intertemporal knowledge spillover, and a novel

\emph{Scrapping Delay Effect} (SDE), by which directed innovation endogenously

delays the scrapping of emission-intensive vintages.

A scrapping subsidy $\sigma_t$ is necessary to correct the SDE; no direction

subsidy can replace it, and it vanishes in the putty-putty limit.

Quantitatively, Bayesian estimation on U.S. data (1949--2024) shows that the

assumed value of $\theta$ alone drives the estimated elasticity of substitution from

near zero to near unity, illustrating how the modelling assumption determines which

mechanism appears as the primary driver of the transition.