Robeco: Re-baselining for a net-zero commitment

Signatory type: Investment manager

HQ country: Netherlands

Operating region: Global

AUM: €171bn (Dec 2022)

Asset class(es): Listed equities and fixed income

 

Robeco is an international asset manager offering an extensive range of active investments. In November 2020, we committed to achieving net-zero greenhouse gas emissions (GHG) across all assets under management by 2050. In October 2021, we published our Net Zero Roadmap laying out interim emissions intensity targets (-30% by 2025 and -50% by 2030) and a strategy to achieve those targets and our commitment. This case study explains how we strive to accurately track our progress by controlling for changes in financial metrics that, if uncorrected, can create misleading impressions of changes in a portfolio’s carbon intensity.

Why is a process for re-baselining necessary for a net-zero commitment?

According to the Institutional Investors Group on Climate Change’s supplementary guidance on target setting, which Robeco helped develop, asset managers committing to decarbonisation should choose a carbon emissions metric, a base year, and a reference point for measuring decarbonisation. This reference point can be the asset manager’s portfolio at the base year or the market index or aggregation of indices to which the asset manager’s portfolio is benchmarked. In the case that a market index is used, decarbonisation targets are then framed as reductions from the carbon emissions level of the reference point in the base year, not as reductions from the carbon emissions level of the manager’s portfolio in the base year. This avoids penalising asset managers whose carbon intensity may be lower than their benchmarks from the outset.

Re-baselining is needed whenever the carbon emissions metric chosen is a ratio whose denominator can change significantly independent of changes in carbon emissions or when there are significant changes to the asset mix, which also do not directly impact real-world emissions. For example, a company’s carbon intensity per dollar of market capitalisation could decline in a year in which its stock price rose, even if its carbon emissions trend was flat. To help investors navigate this technical challenge, the Partnership for Carbon Accounting Financials (PCAF) provides guidance for developing a baseline recalculation policy.

At Robeco, we have chosen 2019 as a base year and tonnes of CO2 equivalent per million euro of enterprise value including cash (EVIC) as our carbon intensity metric.^[1] Our reference point is the aggregated carbon footprint of the indices used as benchmarks for our in-scope portfolios in the base year. As detailed in our Net Zero Roadmap, these in-scope portfolios represent 40% of our AUM and will be referred to for the purposes of this case study as the entity. The carbon intensity of the aggregated benchmark in the base year will be referred to as the baseline carbon footprint. The entity carbon footprint at any point in time will be defined as the weighted average of each of the in-scope portfolios’ footprints. We refer to decarbonisation at any point in time as the difference between the entity carbon footprint at that time and the baseline carbon footprint. In addition to decarbonisation vs the baseline carbon footprint, we also measure performance of the entity against Robeco’s organisational target, which we set by adjusting the baseline footprint each year to create a downward path.

The goal of reporting progress against our net-zero targets is to accurately represent real-world improvements in carbon footprint that have occurred in the in-scope portfolios, either because companies we have continued to hold throughout a given period have become more carbon efficient, or because we have made trades that replaced companies that are less efficient (in terms of GHG emissions) with more efficient ones. We must therefore find ways to adjust for the following factors that could impact carbon footprint calculations independent of actual emissions changes:

1. Changes arising from AUM shifts amongst funds due to client in(out)flows, which we refer to as changes in our asset mix (e.g., due to shifting popularity of funds which, due to their mandates, are dominated by firms with higher or lower carbon intensity);
2. Changes driven by shifts in measures of financial valuation (e.g, rising or falling stock prices, or changes in capital structure that affect EVIC).

Re-baselining, in essence, is the process of calculating what the baseline carbon footprint would have been at the start date, had the entity had the asset mix it does today, and had the financial valuation metrics of the portfolio companies been at today’s levels. Comparing this new baseline to the current entity footprint makes it possible to identify i) real-world improvements in carbon intensity of the companies held and ii) changes in carbon metrics thanks to portfolio manager trading decisions.

It is also worth noting that an element of re-baselining, EVIC inflation correction, is prescribed by the EU climate benchmark regulation^[2] (and advised by the PCAF standard^[3]). With regards to the latter, the new standard released in December 2022 provides additional guidance on the EVIC inflation correction, which is in line with our approach.

Robeco’s process for re-baselining

Once emission reduction targets have been set, at any point in time there are two main numbers to monitor:

1. The entity carbon footprint at that time
2. The organisation-wide target footprint at that time, equal to the baseline carbon footprint multiplied by the targeted reduction of 7% per annum at that time

The graph in Figure 1 illustrates what could be observed about a hypothetical portfolio over the years 2019-2025 before any adjustments. For illustrative purposes, the entity carbon footprint (assuming we are in year 2025) is the green dot and the organisation-wide target at that time (as of 2025) is the red dot on the downward target path that starts from the the baseline carbon footprint in 2019 (shown by the purple marker on the left). In the absence of any re-baselining, it would appear that the in-scope portfolios are doing well, being slightly ahead of the target.

 

Figure 1: Unadjusted figures for entity, baseline and target carbon footprints

Note that Figure 1 is referred to as unadjusted because the baseline carbon footprint has been calculated using the 2019 carbon footprint and weights, with no adjustments applied. This will be referred to as the unadjusted baseline. The target reduction percentages which are applied to the unadjusted baseline are then referred to as unadjusted targets.

As explained in the previous section, in addition to changes in the actual GHG emissions released by the investee companies, the entity-level carbon footprint is influenced by changes in allocations between funds (asset mix), within funds (changes in asset prices, trading) and EVIC changes. At this stage we want to neutralise changes in asset mix and EVIC. In future, we may also think about neutralising asset price changes but this is complex and requires further research.

To adjust the baseline carbon footprint (referred to going forward as the adjusted baseline), we do the following:

1. Adjust for changes in asset mix (allocations between funds):

To account for changes in asset mix (i.e. changes in the value of the portfolios in scope for the target), we apply the asset mix at time t to the portfolio level EVIC-adjusted baseline carbon footprints in the base year. The result is the adjusted baseline at time t.

2. Adjust for changes in EVIC:

We calculate an EVIC inflation factor for each company:

t is the current time, and b is the base year.

The inflation-adjustment factor for each portfolio and reference index combination can then be calculated as the market-value-weighted average EVIC using the reference index weights at time t.

For each portfolio, the unadjusted baseline is then multiplied by the EVIC inflation factor to obtain the EVIC-adjusted baseline.

Figure 2 shows the adjusted baseline with black marker on the left. This is then used to recalculate the target line (shown in grey). This is so that we can compare “apples with apples” when calculating the year-on-year decarbonisation of the portfolio vs the baseline and Robeco’s target. In this illustration, as of 2025, the entity-level emissions reduction (green marker) lags behind the re-baselined target (the grey line).

 

Figure 2: Entity lags adjusted target

Theoretical example

We constructed a simplified example to illustrate what is happening in Figure 2.

Consider this is for an asset manager managing three portfolios: A, B and C, with a baseline year of 2019 and current year of 2025. Table 1 shows the AUMs in 2019 and 2025 as well as the footprint of the portfolios’ reference indices in 2019 and the actual portfolio footprints in 2025. The aggregate of the three portfolios is referred to as the entity.

Table 1

	AUM 2019 (€m)	AUM 2025 (€m)	Reference index footprint 2019 (tCO2e/€m invested)	Portfolio footprint 2025 (tCO2e/€m invested)
Portfolio A	100	120	270	220
Portfolio B	200	500	40	35
Portfolio C	30	40	150	130
Total AUM	330	660
Index footprint calculated using the 2019 AUM weights:			120
Entity footprint calculated using the 2025 AUM weights:				74

Based on this information alone, the reader would conclude that the portfolio decarbonisation is 38% (74/120-1). But asset managers commonly set an organisation-wide target of a 7% year-on-year reduction from the 2019 baseline (as implied by the IPCC’s 1.5°C scenarios). So the organisation-wide target would be 77 tCO2e/€m over the six years to 2025 (120*(1-7%)^6). Based on this, the reader would conclude that the asset manager has outperformed its target by 4% (74/77-1). But is this the full story?

Looking at the individual portfolios, compared to the reference index in 2019, 2025 emissions for Portfolios A, B and C are 19% (i.e. 220 vs 270), 13%, and 13%, lower, respectively. Yet, at the entity level, the decrease was much higher, at 38% (74 vs 120). This is because the lower-emissions portfolio B gained significantly more in AUM over the six years compared to the higher-emissions portfolios A and C. This is not a reflection of the real change in carbon emissions and hence we want to correct for it. In Table 2 we do that by re-calculating the baseline footprint in 2019 using the 2025 asset mix. Each portfolio’s 2019 reference index footprint figure is multiplied by that portfolio’s 2025 weight in the new AUM €660m total, and the products are then aggregated to arrive at 88 tCO2e/€m.

Table 2

	AUM 2025 (€m)	Reference index footprint 2019 (tCO2e/€m invested)	Reference index footprint x 2025 AUM weight	Adjusted footprint calculation:
Portfolio A	120	270	270 x 0.18 =	49
Portfolio B	500	40	40 x 0.76 =	30
Portfolio C	40	150	150 x 0.06 =	9
Total AUM	660
Total (2019 baseline footprint adjusted for 2025 AUM weights)				88

*Unadjusted reference index footprint

The baseline footprint has now decreased from 120 to 88. If we compare this to the 2025 entity footprint of 74 we see a decrease of 16% instead of a decrease of 38%.

The next step is to adjust for the asset price inflation or deflation that influences EVIC. We will demonstrate this for Portfolio A only, assuming its reference index has 5 holdings.

Table 3

	Weight 2019 (MV%)	Weight 2025 (MV%)	EVIC 2019 (€m)	EVIC 2025 (€m)	Inflation Factor
Holding 1	20%	25%	1000000	600000	1.67
Holding 2	30%	25%	140000	120000	1.17
Holding 3	15%	20%	200000	190000	1.05
Holding 4	10%	15%	100000	120000	0.83
Holding 5	25%	15%	800000	640000	1.25
Total	100%	100%			1.23

 

The last column in Table 3 shows the inflation factor, which is calculated by dividing EVIC 2019 by EVIC 2025. We use the formulas outlined earlier to calculate the weighted average EVIC inflation factor for the reference index using 2025 as the reference year. For portfolio A, the EVIC has decreased by such an amount that we need to scale the reference index’s 2019 footprint by 1.23 to correct for the EVIC change. The baseline value for Portfolio A is then adjusted from 270 to 332 (by multiplying 270 by the inflation factor of 1.23). Table 4 shows the results for the baseline for all three in-scope portfolios.

Table 4

	Unadjusted baseline 2019 (tCO2e/€m invested)	Inflation factor	Adjusted baseline 2019(tCO2e/€m invested)
Portfolio A	270	1.23	332
Portfolio B	40	1.18	47
Portfolio C	150	1.12	168

Looking at portfolio A, without applying an EVIC adjustment, the 2025 carbon footprint of the portfolio is 19% (220/270-1) lower than the unadjusted baseline, i.e. a decarbonisation of 19%. By correcting for EVIC inflation, we get an adjusted 2019 baseline of 332, which results in a decarbonisation of 34% (220/332-1).

Note that by calculating the weighted average inflation factor this way at the level of the reference index (1.23 in example above – Table 3) and using it to adjust the baseline overcomes the problem of potentially missing EVIC inflation factors amongst benchmark holdings (if we were to calculate the inflation factors for each benchmark constituent).

Table 5 shows the results required for our decarbonisation calculation and target setting. The step we have just carried out captures both the adjustment for the asset mix change and the EVIC inflation adjustment. The fourth column reflects the inflation-factor-adjusted reference index footprints and the calculation of 106 is based on the 2025 AUM weights.

Table 5

	AUM 2025(€m)	Reference index footprint 2019(tCO2e/€m invested)	Adjusted baseline 2019(tCO2e/€m invested)	Entity 2025(tCO2e/€m invested)
Portfolio A	120	270	332	220
Portfolio B	500	40	47	35
Portfolio C	40	150	168	130
Total	660	120	106	74

We can now calculate decarbonisation of the 2025 entity level carbon intensity vs. the 2019 adjusted baseline. For the entity, after correcting for the asset mix and EVIC inflation, it is equal to 30% (74/106-1).

But how does this compare to the asset manager’s target? Applying the 7% annual reduction to the adjusted baseline of 106 for 6 years implies a 2025 target of 69 (106*(100%-7%)^6). So as of 2025, the asset manager is lagging behind its target by 8% (74/69-1). This compares to an initial assessment that the asset manager was performing better than its target by 4% so indeed, the initial assessment was not the full story.

Concluding remarks

This may appear a complex topic but it is one that asset managers and owners are likely to face as they set net-zero targets and start monitoring them. Whilst the fundamentals of our approach to re-baselining will remain the same, there are likely to be adjustments made to the methodology to adapt to lessons learnt as we continuously monitor our decarbonisation trajectory.