Improving energy productivity is key for competitiveness and reducing emissions. Energy is a crucial input to industrial production. The energy price crisis in 2022 showed how much firms remain vulnerable to energy price shocks – an issue which is now resurfacing, with global energy prices rising 52% in a single month in March 2026. Raising the value added produced per unit of energy can lower costs, strengthen resilience, and cut emissions, allowing countries to progress towards environmental targets whilst also boosting economic competitiveness.
There are huge differences between firms’ energy productivity, with major implications for aggregate energy consumption. On average, within the same narrow industry, frontier firms generate around 20 times more value added per unit of energy than laggards. If firms in the bottom quartile of the distribution in each industry caught up to the 25th percentile, total industrial energy use and emissions could fall by nearly half while producing the same output.
There is no evidence of a trade-off between economic and energy productivity, and policies can improve both at the same time. At the firm level, there is a strong relationship between energy productivity and labour productivity, which also tend to move together. Furthermore, industries with large differences in energy productivity across firms also have wide gaps in labour productivity. These differences are linked to the same factors – such as competition, access to finance and innovation – that help improve overall economic performance.
Energy and carbon pricing are strongly related to energy productivity. In many countries, large energy users, who are less energy productive on average, benefit from lower effective energy prices and tax exemptions, making them even less energy productive. These policies weaken incentives to invest in more efficient technologies and disadvantage smaller, more energy-productive firms, raising concerns about fairness.
A coherent policy package can boost both energy productivity and economic performance. Supporting investment and innovation by the least energy-efficient firms, reforming energy pricing, and strengthening competition can help laggards catch up and allow more productive firms to grow. Such policies can align the productivity, economic resilience and environmental agendas and deliver lower emissions, reduced energy use and stronger economic competitiveness together.
How improving industrial energy productivity can cut emissions and boost competitiveness
Why is it important to improve energy productivity?
Energy is a crucial input to industrial production and a key driver of competitiveness. The recent surge in energy prices – which tripled from 2020 to 2022 and remained around twice their 2020 level in 2024 – has underscored how vulnerable firms are to energy cost shocks. Improving energy productivity – the value added generated per unit of energy consumed – is therefore a key lever for strengthening firms’ competitiveness and resilience to energy price shocks.
Increasing energy productivity alongside labour productivity is essential to reduce emissions and improve resilience and energy security while maintaining economic growth. Reducing the energy intensity of production is essential to keep on track for environmental goals and reduce dependencies on critical inputs. Yet the growth rate of energy productivity in the industrial sector in IEA countries has roughly halved in recent years compared with the early 2000s.
There are huge differences between firms’ energy productivity. Looking only at the aggregate performance of industries hides important differences in performance between firms and overlooks a major potential policy lever: helping less energy-productive firms catch up. Using official firm-level data for nine countries, a recent OECD study compares the energy productivity of firms in the same narrowly defined industries (De Lyon and Dechezleprêtre, 2025[1]). Averaging across all nine countries, frontier firms at the 90th percentile of the energy productivity distribution generate roughly 20 times more value added per unit of energy than laggards at the 10th percentile (Figure 1). Put simply, if these two firms used the same amount of energy, the frontier firm could produce 20 times more output than the laggard firm. These gaps are around four times larger than the differences observed in labour productivity. This energy productivity dispersion appears in each of the nine countries and the gaps between firms have not narrowed on average over time.
Percentiles and dispersion in energy productivity across countries
Note: The chart plots the 10th, 50th, and 90th percentiles of energy productivity for each country, averaged across industries, in the latest year of available data for each country up to 2019. The labels show the ratio between the 90th percentile and 10th percentile firm. The units are thousands of dollars per tera joule of energy (or dollars per mega joule). The sample is restricted to manufacturing industries with at least 10 firms and firms with at least 20 employees.
Source: OECD calculations.
Why are some firms significantly more energy productive than others? The data show that younger companies and those using more recent vintages of capital are typically more energy productive. Strikingly, larger firms are, on average, less energy productive, even though they are usually more productive in terms of labour productivity.
One important reason why large firms may be less energy productive is that they pay significantly lower prices for energy. Large firms can use their bargaining power to negotiate better contracts with suppliers. On top of this, they often benefit from reduced energy taxes or exemptions designed for energy-intensive industries (Figure 2). These lower prices weaken their incentives to improve energy efficiency, while putting smaller, more energy-productive firms at a relative disadvantage. This has important implications for how energy price and tax policies are designed.
Share of all taxes, fees, levies and charges in electricity costs by firms’ consumption
Note: The chart shows the share of firms’ electricity cost accounted for by taxes, fees, levies, and charges for OECD Member Countries with available data. Small firms are those with consumption below 20 MWh and large firms are those with consumption over 150 000 MWh. Data is from Eurostat.
Source: OECD calculations
There is no trade-off between energy and labour productivity. The strongest predictor of a firm having high energy productivity is high labour productivity (Figure 3). Furthermore, as firms improve their labour productivity over time, they also tend to improve their energy productivity. In practice, this means that policies supporting better management, skills and technology adoption could deliver a “double dividend”: higher output per worker and more output per unit of energy.
Average importance of each factor in predicting energy productivity and direction of marginal effect
Note: The chart plots the importance of each factor in predicting firms’ energy productivity in a machine learning prediction model. The bar size shows the overall importance of each predictor variable and the colour shows whether the relationship with energy productivity tends to be positive (orange) or negative (blue). The importance of each variable is estimated using the mean of the absolute SHAP value. The response variable is energy productivity at the firm-level. Both the predictor and response variables are measured subtracting the country-industry-year average value, computed separately within the training and test datasets. The prediction model is estimated using XGBoost and the SHAP analysis is carried out on a random sample of 5 000 firm-year observations. The sample is restricted to firms which have at least 20 employees in the manufacturing industry in Chile, Indonesia, and Portugal (where this analysis can be performed).
Source: OECD calculations.
How improving energy productivity can unlock energy and emissions reductions while maintaining competitiveness
Improving the energy productivity of the least efficient firms could deliver large gains for both competitiveness and the environment. If firms in the bottom quartile of the energy-productivity distribution in each industry caught up only to the 25th-percentile firm, total energy use would fall by nearly a half (42%) for the same level of output (Figure 4). The potential reduction in energy-related emissions is even larger (45%).
These results show that reducing energy use and emissions does not need to rely on speculative technology breakthroughs. Huge gains can be made by helping lagging firms adopt technologies and practices that already exist in their own country and industry today. In other words, a large share of the potential reduction in industrial energy use and emissions can be achieved simply by spreading current best practices more widely. Policy can play an essential role in unlocking this catch-up among lagging firms.
Counterfactual exercises showing implied reduction in energy to produce the same aggregate value added if all firms were at least as productive as the 25th percentile firms in their industry
Note: The chart shows the implied percentage reduction in energy that would be achieved if all firms had at least an energy productivity equal to the 25th percentile firm in their 4-digit industry. The data are for the latest year available in each country up to 2019 and is restricted to manufacturing sector firms with at least 20 employees.
Source: OECD calculations
Why do some industries have larger gaps in energy productivity than others? Higher energy prices are associated with smaller differences in energy productivity between firms, but other industry characteristics beyond energy prices are also important. Industries with more dispersion in energy productivity between firms also tend to be more dispersed in other economic characteristics, especially in labour productivity. This suggests that the same underlying factors – such as differences in management quality, skills and technology adoption – may shape how efficiently both labour and energy are used.
Dispersion in energy productivity is higher in industries with weaker competition, less access to external finance and more concentrated innovation. Industries where employment is more concentrated among a small number of firms – a proxy for a lack of competition – have greater dispersion in energy productivity. By contrast, industries where firms have stronger cashflow relative to their capital stock – indicating less reliance on external financing – have smaller gaps, consistent with the idea that easier access to finance helps lagging firms invest in more efficient machinery and processes. Industries where a greater share of firms file patents, suggesting more widespread innovation activity, also have lower dispersion in energy productivity.
These results imply that policy levers to improve the energy productivity of lagging firms go well beyond energy pricing. Competition policy, business dynamism and access to finance all matter for how efficiently energy is used. This provides further evidence that there may be synergies, not trade-offs, between improving energy productivity and economic competitiveness.
Support investment by the least energy-efficient firms and accelerate diffusion of existing technologies. There are large potential gains to be made by helping the least productive firms to catch up with others in their industry and by ensuring that the most productive firms can grow. Crucially, this does not rely on speculative future technological breakthroughs but on encouraging the adoption of already existing technologies. For example, Germany’s EEW programme, which offers technology-open grants and concessional loans for adopting efficient equipment, has supported over 40 000 firm projects, with measurable energy savings.
Make energy pricing reward efficiency rather than size. Reform energy and carbon pricing so that large, energy-inefficient firms no longer benefit from systematically lower effective prices and tax exemptions. While these energy price subsidies may help maintain competitiveness of energy-intensive companies in the short term, they dampen the incentives for efficiency-boosting investments. Instead, these funds could be used to encourage firms to adopt more efficient technologies and production methods. Evidence from the UK Climate Change Levy shows that firms paying the full tax reduced their energy use significantly without losing output or employment compared with energy-intensive firms that benefited from an 80% discount on the tax.
Use competition, finance and innovation policies to close energy-productivity gaps. Strengthen competition and business dynamism in energy-intensive sectors, for example by reducing barriers to entry and reviewing measures that protect large incumbents. Ease financing constraints for energy efficiency investments and broaden support for innovation and technology diffusion. Access to finance is especially important for SMEs: the Netherlands' Energy Investment Allowance, which lets firms deduct around 40% of qualifying energy-saving investments from taxable profit, has experienced a substantial free-rider problem among larger firms, leading its evaluator to recommend a sharper focus on SMEs.
Harness potential synergies between economic competitiveness and energy productivity. Policies to boost business dynamism, management quality, skills and technology adoption could pay a double dividend, raising both labour and energy productivity. Some policies targeted at energy efficiency investments, such as Italy's White Certificates, have been shown to also improve economic productivity, while policies aimed at improving economic competitiveness, such as trade liberalisation episodes in India and Mexico, also improved energy efficiency. Such policies can at the same time reduce emissions, decrease reliance on energy, and boost economic competitiveness.
References
[1] De Lyon, J. and A. Dechezleprêtre (2025), “The great dispersion in energy productivity between firms”, OECD Science, Technology and Industry Working Papers, No. 2025/24, OECD Publishing, Paris, https://doi.org/10.1787/0fa543ce-en.
Contact
Josh DE LYON (josh.delyon@oecd.org)
Antoine DECHEZLEPRETRE (antoine.dechezlepretre@oecd.org)