The Climate Actions and Policies Measurement Framework (CAPMF) - FAQ

TOPICS

1. The CAPMF
2. Policy stringency
3. Structure and aggregation 
4. Data and applications
5. Key resources 

The OECD Climate Actions and Policies Measurement Framework (CAPMF) is an internationally harmonised climate mitigation policy database that covers the principal climate mitigation policies (targets, market-based and non-market-based policy instruments).

The CAPMF covers 130 policy variables, grouped into 56 policies for 50 countries (OECD + G20 + OECD accession countries) between the period 1990 to 2023. The CAPMF focuses on government action only.

1.     To track the evolution of climate action defined both in terms of policy adoption and policy stringency.

2.     To provide the foundational policy data to enable analytical work (see use cases).

The CAPMF fills a significant gap in the current global climate policy data landscape since:

1. It provides consistent and coherent climate policy data to the extent that it tracks the same policies across a large number of countries and a long time period.
2. It is the only database that provides harmonised information on policy stringency in addition to the adoption of policies.
3. Policy data is validated by countries.

The CAPMF covers the following types of policy: 

The CAPMF has different aggregation levels. Each of the levels that are higher in the hierarchy consists of several lower-hierarchy elements (see chart below). For example, the building block ’International policies’ (Level 1) comprises 3 modules (Level 2); the module ‘International co-operation’ (Level 2) comprises 3 policies (Level 3); and the policy ‘Climate treaties’ (Level 3) comprises 6 policy variables (Level 4). Some policies (e.g. Climate initiatives) only comprise one policy variable.

There are multiple ways of defining and operationalising policy stringency. The CAPMF defines policy stringency as ‘the degree to which climate actions and policies prescribe, incentivise or enable GHG emissions mitigation’. The CAPMF bases its stringency metric on government action only. It focusses on the underlying climate policy data, which reflect the key characteristics of the policy intensity as defined in law or policy documents.

1.     To track countries’ climate action. The stringency scores used in the CAPMF recognise that countries can increase their climate action not only by adopting policies but also by increasing their respective stringency. Only looking at policy adoption would severely understate countries’ efforts on climate action by failing to capture the strengthening of existing policies. Consider, for example, the European Union emissions trading scheme (EU ETS). Only looking at policy adoption would not signal any progress after EU countries implemented the EU ETS in 2005. However, the ETS price increased sharply from below EUR 5 to EUR 80 between 2005 and 2022 thanks to government interventions. The CAPMF's policy stringency metric can reflect an increase in the ETS price.

2.     To aggregate policies into categories. The stringency metric used in the CAPMF can translate different policy variables (e.g. emissions limit values, tax rates, government expenditure) into a common unit (i.e. the stringency score) using the same methodology. This is important because it facilitates the aggregation across different policy types, thus permitting to track progress in climate action across different policy instruments and sectors (e.g. transport, electricity) or policy types (e.g. market-based instruments).  

For each policy variable, the CAPMF operationalises stringency as a relative concept by calculating a stringency value between 0 (not stringent) and 10 (very stringent) based on the in-sample distribution across all countries and years of the policy variables’ level (e.g. tax rate, emission limit value, government expenditure) following the methodology of the OECD Environmental Policy Stringency Index. A stringency value of 0 indicates that a policy is not adopted. The steps would be:

1.     Order policy variables (e.g. carbon tax rate) across all countries and years from lowest to highest.

2.     Assign a 10 for the 10% highest values, a 9 for the next 10%, etc.

3.     Assign a 0 if the policy is not in place.

4.     Repeat for all other policy variables (e.g. standards, ETS).

The figure below illustrates the calculation of policy stringency taking the example of a carbon tax.

For binary variables, a stringency value of 10 is assigned if the policy is in place. Categorical variables are linearly mapped into the space from 1 to 10. For example, a categorical variable with 5 elements would get the stringency values 2, 4, 6, 8, and 10, respectively.

A missing value implies that the CAPMF does not have information about whether or not the policy exists. A zero is assigned when the policy is not in place.

The policy stringency is determined based on all observed data until 2022, meaning that a country’s policy stringency will not decrease if other countries strengthen their efforts in the future. Adding new countries or years to the sample may require a re-attribution of bin thresholds (i.e. the orange vertical lines in the figure above) and therefore lead to a re-attribution of stringency scores. To address this concern, the bin thresholds will remain fixed for the coming years based on data until 2022. Bin thresholds will be updated less frequently (e.g. every five years). Once updated, the entire time series of all stringency scores will also be updated.

The stringency scores should be interpreted in an informative, not in a normative or prescriptive way. It is not a performance metric and the CAPMF does not suggest that countries should aim to reach the highest score for all policies. A score of 10 only implies that the intensity of a particular policy is among the top 10% values observed across all countries and years. A score of 10 neither implies that it is sufficient to reach countries’ mitigation targets nor that the policy is more effective than a lower score in another country. For example, a high score for a high carbon tax in a developed country may be less effective than a low score for a relatively low carbon tax in an emerging country that still has many low-cost abatement options available. Similarly, a high score for one policy instrument does not imply that this policy is more effective than another policy instrument with a lower score. This is expected because the selected stringency metric is not designed to measure effectiveness. Nevertheless, there is likely a positive relationship between the CAPMF’s stringency metric and effectiveness or compliance costs. For example, higher carbon taxes or tougher emission standards can be reasonably expected to lead to more emissions reductions and higher compliance costs in the same country context.

The primary structure of the CAPMF is shown in the figure below:

The starting point of the CAPMF are the 130 Level 4 variables. These are aggregated into 56 Level 3 policies. Based on the 56 policies, the CAPMF proposes different classifications. The primary classification is shown above. In this classification, the 56 policies are aggregated into 15 Level 2 Modules (e.g. electricity market-based), which are then aggregated into 3 Level 1 Building blocks (e.g. sectoral policies). The correspondence between Level 4, Level 3, Level 2 and Level 1 variables can be found in the correspondence table.

The stringency values for policies (level 3), modules (level 2) and building blocks (level 1) are calculated using the unweighted average of the stringency values of the elements in the underlying hierarchy level. For example, the stringency value of the policy ‘Feed-in-Tariffs’ is the unweighted average of all four elements that underpin the policy ‘Feed-in-Tariff’ (table below). The stringency value of the module ‘Electricity – market-based instruments’ is the unweighted average of the seven policies included in this module.

For the calculation of stringency values for level 1, 2 and 3 variables, missing values of variables in the lower hierarchy are generally considered to have a stringency value of 0. However, for the last five years, any missing policy variable data is replaced with the last available value from that period. For example, if a policy variable has missing data for the years 2021 and 2022, then the stringency value of the year 2020 will be used for these two years in the calculation of the stringency values of the higher hierarchy levels. Note that this rule is only used for the aggregation to higher hierarchy levels. The CAPMF still reports the policy variables in 2021 and 2022 as missing. In cases where extrapolated values and/or missing values counted as 0s are used for the aggregation, the calculated observation includes a flag "E" for "Estimated". If all sub-level policy instruments are missing in a given country-year, then the aggregated values are also reported as missing and no flag is assigned. 

Adopted policies count the number of adopted policies (i.e. the number of level 3 variables). Hence, showing this number on the level 3 would result in exactly either a 0 or a 1. As level 3 variables are aggregates of level 4, it is not possible to show this number on the level 4.

Yes, users can also calculate other aggregates. The example below shows how the user could create an index for market-based instruments.

The primary classification is shown above. In addition to the primary classification, the correspondence table offers other classifications:

·       By policy instrument type (e.g. market-based instruments, non-market-based instruments, and climate targets, governance and climate data)

·       More granular classification by instrument type (e.g. split of market-based instruments into ‘taxes and fees’, ‘trading systems’ and ‘subsidies’)

·       Climate and climate-relevant policy instruments.

Absolutely. If the proposed classifications do not work for the task at hand, users can define their own classification system. The only thing that needs to be done is to assign a tag to each of the 56 policies (or 130 policy variables if the assignment of policy variables need to be regrouped into policies).

Yes, but country-level indices should be interpreted carefully. In particular, they should not be used as performance metrics (see Interpretation of stringency scores).

There are at least two ways of calculating country-level indices. First, by calculating the unweighted average of the stringency of the three building blocks. Second, by calculating the unweighted average of the 56 policies. In principle, one could also calculate a country-level index by averaging the 130 policy variables, but this would put more weight on some policies that have more information underpinned. For example, this aggregation method would implicitly assign four times the weight to the policy feed-in-tariffs (which have four policy variables) compared to the policy Renewable Portfolio Standard (which has only one policy variable). It is not advisable to use a combination of Level 3 and level 4 variables to create any aggregate as this will likely result in double counting. 

The two aggregation methods implicitly assign different weights to the policies. The unweighted average across all 56 policies assigns equal weight to each policy (1/56). Consequently, sectoral policies, which account for 36 of the 56 policies, implicitly get a higher weight than under the unweighted average across the 3 building blocks aggregation, where sectoral policies receive a weight of 1/3. Note that any aggregation inevitably requires the assignment of weights. Using unweighted averages is conservative as it does not make any judgement call on the relative importance of some policies over others.

The data is available either on the IPAC Dashboard or on the OECD Data Explorer. The data can be downloaded in bulk or for selected countries, indicators or years. The OECD Data Explorer also features an API. 

• Find out different applications of the CAPMF.
• Explore the latest Climate action data explainers.
• Check the citations of the CAPMF working paper to see further use cases.