APG Asset Management is a fiduciary asset manager for pension funds, managing a total of €544 billion as of January 2020. As a leading long-term responsible investor, we regard robust management of climate risks and opportunities as essential to our mission of providing good pensions in a livable world.

Considering that methods for climate-related risk management are still in development, we have established a climate steering group at APG. This group is tasked with overseeing various initiatives to monitor and manage climate-related risks and opportunities within APG, including the prioritization of research. Its members come from various parts of the organization: portfolio management, risk management, and fiduciary management.

We include climate factors in the analysis used to determine the strategic investment plan for ABP, the pension fund for government and education sector employees in the Netherlands. We began with the central path scenario, based on stochastic modelling. To perform a stress test of this scenario, we constructed four additional deterministic scenarios. Climate change was one of the key factors in constructing these scenarios, along with such other elements as the role of central banks and the strength of international collaboration. One scenario (the “climate pit”) reflects a 4°C global temperature increase scenario, and another (“good globalization”) is comparable to a < 2°C scenario. We used these deterministic scenarios to map potential effects of climate change for economic growth, inflation, and impact on various asset classes. Also, for each asset class, we developed ESG scores based on a methodology that ranks asset classes primarily for their upside potential for responsible investing. Underlying factors include, for example, involvement by asset owners with investee companies, market transparency within that asset class on ESG-related topics, and percentage of Sustainable Development Investments. We integrated the analysis into ABP’s strategic investment plan for 2019–2021.

To map climate-related risks and opportunities in portfolio construction, we conducted scenario analysis at the economic sector level (26 sectors) in collaboration with an external consultant, Environmental Resources Management. To analyze transition risks and opportunities, we followed both a business-as-usual scenario (International Energy Agency Current Policies Scenario [IEA CPS], 3.7°C) and a 2°C scenario (IEA Sustainable Development Scenario [IEA SDS], 2°C), supplemented with specific information from the IEA Energy Technology Perspectives (ETP). Because the IEA scenarios do not cover physical risks and opportunities, we used the RCP 4.5 and RCP 8.5 scenarios to analyze the physical dimension.¹ For both transition and physical analysis, we looked ahead to 2022, 2030, and 2040. We chose the year 2022 as the short-term horizon because it is far enough into the future to observe climate impact but also falls within a relevant investment horizon for investments in liquid capital markets (~five years). The years 2030 and 2040 are common intermediate- and longer-term horizons for climate analysis.

Our climate scenario analysis insights are captured in a “traffic light” model. For each economic sector at each time horizon, the model designates both transition and physical climate-related risks as high, moderate, or low. We determined each score using the difference in the value of the key climate factor (selected for the specific sector) between the business-as-usual and the 2°C scenario for the specific time horizon: The larger the difference between those values, the larger the risk or opportunity. For example, in the oil sector, a key risk factor is declining demand for oil. We take the projected demand in IEA CPS and the projected demand in IEA SDS. The larger that difference, the larger the reduction, hence the risk.

The climate factors are defined as key drivers of global climate-related risk and opportunity that may impact the economic sectors in which we invest. Our taxonomy of climate factors (44 in total) is informed by the TCFD list of climate factors, which include carbon pricing, oil demand, litigation risk, and flooding risk. For every sector, a key climate factor has been proposed by the external consultant and validated by APG.

Beyond the traffic light model, we have developed a climate dashboard that offers supplementary analysis aiming to track the speed of the transition to a low-carbon economy. The dashboard consists of 20 indicators and is updated annually (versus every two years for the traffic light model). The most prominent changes in indicators and the overall score are analyzed. Therefore, this dashboard supports the assessment and management of climate-related risks and opportunities in a shorter time frame.

We also conduct similar analysis for sovereign bonds at the country level. For each country, we look at physical risk (based on the Notre Dame GAIN database) and transition risk (based on HSBC indicators), resulting in a low-medium-high risk profiling of the sovereign bond portfolios.

Together, the traffic light models for sectors and countries and the climate dashboard are our primary instruments for monitoring climate-related risks and opportunities of our portfolio at a high level. For all investments in areas denoted ‘high risk’ within the investment horizon, explicit attention should be paid to climate risk in the investment case, including a rationale about why we are prepared to take the risk, as well as the impact of the specific investment on both our and our clients’ climate goals.

The scenario analysis shows that by 2040, the effects of climate change are large and comprehensive. In the run-up to 2040, climate change transition risk increases gradually for a global and diversified portfolio, such as ours. The transition can be accompanied by disruptive changes and unexpected inflection points, however, that will require close monitoring.

Before 2030, we anticipate major transitions already taking place in the 2°C scenarios, with corresponding risks and opportunities for the following sectors in particular: utilities, real estate, cement, oil and gas, aerospace, food and consumer goods, automotive, semiconductors and electrical equipment, agriculture, chemicals, and construction.

Sectors that are especially vulnerable to, but also show opportunities for, the physical impact of climate change include agriculture, forestry, real estate, oil and gas, food processing, road and rail transport, mining, utilities, health care, construction, and water utilities.

From the analysis on climate risks in sovereign debt, we conclude that our exposure to countries with high climate risk (physical and transition) is limited. Countries with lower sovereign credit ratings (emerging economies) are more exposed to climate risk than higher-rated countries, and there is evidence that this exposure is already being priced into investments.

Our clients gain insight into our scenario analysis results and the climate dashboard through a digital client reporting tool. We update this report twice a year. Furthermore, we hold deep-dive sessions to brief clients on the monitoring and management of climate-related risks and opportunities. Finally, our fiduciary management department reviews the various asset classes (at least annually) as part of granting and evaluation of mandates. Climate risk is part of this analysis, and we report the results to clients.

We have gained the following insights from our climate-related work thus far:

• Scenario analysis provides insights on a generic level about the most prominent impacts of climate change on the overall portfolio. A true understanding of climate-related risk and opportunity for individual investments (including financial impact), however, requires more granular analysis. For this reason, within the APG governance model, the investment teams are primarily responsible for managing climate-related risks and opportunities, whereas overall exposures are monitored and reported across the portfolio. In addition, on a portfolio-wide level, attention needs to be paid to second-order and network effects of climate change, which affect the entire portfolio.
• The scenario analysis conducted in 2018 highlighted the importance of having insight into the macroeconomic spillover effects of climate change. The analysis was based on IEA scenarios that are linear by nature, but in reality, the changes are most probably dynamic and non-linear. Climate change can be seen as not a risk by itself but rather a risk multiplier, with impacts on conflicts, migration, and scarcity that might materialize via such economic variables as economic growth, interest rates, and inflation. For the next iteration of our scenario analysis, therefore, we are considering the addition of more disruptive scenarios (akin to the Inevitable Policy Response [IPR] scenario developed by the Principles for Responsible Investment [the PRI]). To inform our scenario analysis, we are collecting insights on the underlying patterns of impacts on economies and financial markets, including their speed of recovery, through analogies with historical cases in which physical destruction and major government interventions took place (e.g., natural disasters and wars).
• For the real estate asset class, we conducted a pilot on measuring physical risks. As part of this pilot, we tested six different methods for measuring physical risk on a single asset. The results showed large differences among the six methods, and no single model successfully accounted for all physical risks. We concluded from this analysis that careful interpretation is required in evaluating results from off-the-shelf products. A combination of insights and analysis is essential for complete understanding.

Measurement, monitoring, and management of climate-related risks and opportunities is in an incipient phase. A number of critical limitations are therefore important to consider.

First, robust quantitative metrics to measure climate risk in portfolios, as well as to integrate this process into regular risk management, are missing. Many (semi-) quantitative metrics and methodologies are becoming available. In practice however, we observe that these metrics depend strongly on models and assumptions, and therefore we do not yet consider these suitable for setting explicit limits on the portfolio with regard to climate risk. Thus far, we have opted to work with a semi-quantitative approach (the traffic light model). We are closely monitoring developments in this field and are looking to strengthen our approach in the future using more quantitative risk management metrics.

Second, for such asset classes as sovereign debt and such sectors as finance, we observe that climate risks are not direct but rather indirect, based on the underlying economy and financial relationships. Our analysis for the sovereign debt asset class proxies climate risk for countries based on their underlying economies. The next step will be to also make this methodology applicable to financials and proxy climate for companies operating in this sector.

Third, we have developed a dashboard to track the speed of the transition to a low-carbon economy. Currently, this dashboard relies on relatively conventional indicators that use globally available data, such as oil demand and capacity of renewable energy versus fossil fuels. Because these indicators are all backward-looking in nature, we aim to supplement them with some more disruptive forward-looking indicators to enhance our understanding of strong changes in the speed of the transition to a low-carbon economy (e.g., policy developments and social sentiment). We are looking into the possibility of whether innovative technology and data sources, such as unstructured data, can help us enhance the dashboard.

Finally, we have concluded that our scenario analysis does not sufficiently account for the physical risks of climate change. It has merely touched on physical vulnerabilities rather than physical risks. Because we conducted our scenario analysis at a global level, the information is too general to map physical risks for individual investments. We need more detailed information on the physical risks of climate change at a local level for the specific locations where each investment has a footprint. As a follow-up analysis, therefore, we aim to map local physical risks for specific sectors, starting with the real estate asset class.

© 2020 APG Asset Management N.V. All rights reserved.