PACTA methodology note
Calculation of a company alignment metric
Prepared by Jacob Kastl
The company alignment metric described below is intended to help make sense of alignment patterns in large analyses, e.g. in the case of a supervisory analysis of the banking system, where many loan books are compared. In such situations, high level results are helpful in highlighting the most important patterns that warrant an in-depth analysis of the more granular standard PACTA metrics.
Identifying such patterns means that we need to reduce the information to a more manageable quantity. While this inevitably leads to a loss of granularity, this is justifiable for summarizing main findings, as long as the outcomes are consistent with the main PACTA metrics.
We choose an approach that first aggregates company alignment to a relative metric at the company/sector level for each of the PACTA sectors the company operates in. A methodology to calculating this metric is provided both for the market share approach and the sectoral decarbonization approach.
In this section, we derive the company alignment metric mathematically and provide interpretations of the relevant pieces. While in practice, this metric will most often be calculate at time \(t = 5\), it can be calculated for any point in time, which is why we generalise.
For companies operating in sectors where clear technology transitions should take place, certain technologies need to be built out (\(BO\)) whilst others need to be phased out (\(PO\)). We define the direction \(d\) of technology \(a\) in sector \(b\) as:
\[ d_{a,b} = \begin{cases} BO & \quad \text{if } a \text{ is a low-carbon technology}\\ PO & \quad \text{if } a \text{ is a high-carbon technology} \end{cases}, \]
where low-carbon technology and high-carbon technology are mutually exclusive categories.
The forward-looking production plan \(p\) of company \(c\) for technology \(a\) in sector \(b\) at time \(t\) is:
\[p_{a,b,c,d,t}\]
Which means that at the sector level, we get a net production plan of:
\[p_{b,c,t} = \sum_{\forall a \in b} p_{a,b,c,d,t}\] Correspondingly, the scenario-based production value \(s\) of company \(c\) for technology \(a\) in sector \(b\) at time \(t\) is:
\[s_{a,b,c,d,t}\]
and at the sector level, we therefore get a net scenario-based production value of:
\[s_{b,c,t} = \sum_{\forall a \in b} s_{a,b,c,d,t}\] We define a directional dummy variable \(D\) for technology \(a\), given its direction \(d\) in sector \(b\) as:
\[ D_{a,b,d} = \begin{cases} 1 & \quad \text{if } d_{a,b} = BO\\ -1 & \quad \text{if } d_{a,b} = PO \end{cases} \]
With these definitions, we calculate the total technology deviation (\(\Delta total\)) of company \(c\) for technology \(a\) in sector \(b\) at time \(t\) as:
\[\Delta total_{a,b,c,d,t} = (p_{a,b,c,d,t} - s_{a,b,c,d,t}) \times D_{a,b,d}\] Note that the total technology deviation will be considered aligned when >= 0 and misaligned when < 0. The Directional dummy ensures this is the case both for build out and phase out technologies.
In case a technology trajectory in a transition scenario cannot be clearly classified as a build out or phase out technology over the near to medium term, we treat such a technology as a bridge technology. More specifically, this is the case, when the production trajectory for a technology remains stagnant or the changes vary in direction over a time frame needed to carry out new investments. Usually, a bridge technology will be a high-carbon technology that is less emissions intensive than other high-carbon technologies which need to be phased out first. Whether or not a technology is considered a bridge technology therefore depends largely on scenario assumptions and will sometimes differ between regions even within a scenario.
As a bridge technology is considered high-carbon in principle, but should not yet be phased out, we need to apply a calculation that incentivises production planning as close to the scenario-based value as possible. This ensures that both an overshoot and an undershoot will lead to deteriorating alignment metrics.
We therefore use a two-sided misalignment logic, which adjusts the total technology deviation for bridge technology \(a\) at time \(t\) as:
\[\Delta total_{a,b,c,d,t}^{bridge} = -|p_{a,b,c,d,t} - s_{a,b,c,d,t}|\] Contrary to the standard formulation of the technology alignment deviation, the technology deviation for a bridge technology has a maximum at 0, on the basis that the production forecast exactly meets the scenario-based value at time \(t\). Any deviation from that is considered misalignment. The remainder of the calculation remains unchanged.
Based on the total technology deviations calculated above, we can now derive the net aggregate company alignment metric \(y\). To derive the net aggregate company alignment metric, we first define calculate the total deviation of company \(c\) for sector \(b\) at time \(t\) as:
\[\Delta total_{b,c,t} = \sum_{\forall a \in b} \Delta total_{a,b,c,d,t}\] We then note that the total scenario-based value for company \(c\) in sector \(b\) at time \(t\) is:
\[total_{b,c,t}^{scen} = s_{b,c,t} = \sum_{\forall a \in b} s_{a,b,c,d,t}\] With these definitions, we derive the net aggregate company alignment metric \(y\) for company \(c\) in sector \(b\) at time \(t\) as:
\[y_{b,c,t} = \dfrac{\Delta total_{b,c,t}}{total_{b,c,t}^{scen}}\]
The Net Aggregate Company Alignment Metric is a summary of the alignment of a company in a sector across all relevant technologies within that sector. It is intended as an overall summary metric that shows if a company is aligned with a climate scenario on aggregate. As such, the metric is meant to be a starting point for further analysis. It can be used as a basis to aggregate alignments of individual companies in a loan book to get a high level overview of the alignment in that loan book. It can also be used to further disaggregate the company alignment metric to obtain more granular insights about its underlying drivers.
In general, \(y_{b,c,t} >= 0\) means that a company is - on aggregate - aligned with a climate scenario, whereas \(y_{b,c,t} < 0\) means that it is misaligned. For each of the technologies in the sector that a company operates in, we measure the directional deviation from its corresponding scenario-based value at time \(t\). A positive value of this deviation indicates alignment and a negative value indicates misalignment, regardless of the direction of the technology in the given scenario. The sum of those deviations across all technologies is then divided by the company’s sector level scenario value at time \(t\). This returns a percent deviation of the company production from its scenario value at \(t\), where \(y = 0\) means that there is no deviation, \(y = 1\) means that the company is doing one hundred percent more in terms of buildout and phaseout than it needs to and \(y = -0.5\) means that it is doing fifty percent less than it needs to according to the scenario values. Note that this metric does not show exactly where a (mis)alignment comes from. Hence, following up this analysis on a more granular level is always a good idea.
One particular strength of the PACTA metrics for sectors with technology pathways is that it will surface misalignment on the technology level, which matters particularly in sectors that include both build out and phase out technologies. Transition scenarios often assume that particular technologies must be phased out in order to remain within a carbon budget. Therefore, alignment metrics should not obscure the need to phase out or build out technologies by simply adding up bidirectional technology deviations. To account for this, we define a disaggregation of the net aggregate company alignment metric into its build out and phase out drivers. While this falls short of the maximum granularity by design, it ensures that opposite trends in build out and phase out are not obscured by the analysis.
The disaggregation of the net aggregate company alignment metric into its \(BO\) and \(PO\) drivers is designed such that the resulting sum of the disaggregated metrics equals the net metric, because build out and phase out technologies are mutually exclusive sets. To calculate the \(BO\) and \(PO\) company alignment metrics, we need to derive the total deviation of company \(c\) by technology direction \(d\) in sector \(b\) at time \(t\) as:
\[\Delta total_{b,c,d,t} = \sum_{\forall a \in b,d} \Delta total_{a,b,c,d,t}\] The \(BO\)(\(PO\)) company alignment metric \(y\) of company \(c\) by technology direction \(d\) in sector \(b\) at time \(t\) is then defined as:
\[y_{b,c,d,t} = \dfrac{\Delta total_{b,c,d,t}}{total_{b,c,t}^{scen}}\] Note that the denominator remains at the net level so that the sum of the disaggregated metrics is the net metric:
\[y_{b,c,d=BO,t} + y_{b,c,d=PO,t} = \dfrac{\Delta total_{b,c,d=BO,t} + \Delta total_{b,c,d=PO,t}}{total_{b,c,t}^{scen}} = \dfrac{\Delta total_{b,c,t}}{total_{b,c,t}^{scen}} = y_{b,c,t}\] This must be true because every technology \(a\) must have one and only one direction \(d \in \left\{BO,PO\right\}\).
Previously, it was mentioned that the aggregated value of \(y_{b,c,t}\) does not indicate where the sources of (mis)alignment can be found. The disaggregation into \(BO\) and \(PO\) drivers of the company aggregate alignment metric aims to provide a high level overview of just that. In essence, \(y_{b,c,d,t}\) only adds up the technology deviations for either buildout or phaseout technologies, while still dividing by the sector level scenario value of the company (including all technologies). This means that we get a simple disaggregation of the overall alignment metric into its sum parts. Hence, it should not be read as an indicator in isolation. However, the disaggregation can be very helpful in determining if the overall alignment is driven by a lack of buildout or by slow phaseout. Additionally, what may look like overall alignment near the target with \(y_{b,c,t}\) close to zero may in theory be composed of significantly overaligned buildout and significantly misaligned phaseout or vice versa. This goes to show that, companies active in sectors with both buidlout and phaseout technologies, it is important to understand the drivers behind that overall alignment metric.
For sectors where alignment is calculated using emissions intensity metrics there is no distinction between build out and phase out technologies. This means there is only a net aggregate company alignment metric in emissions intensity sectors \(y^*\) and the calculation of that metric is much more straight forward.
Let emissions intensity based on forward-looking production plan \(p^*\) of company \(c\) for sector \(b\) at time \(t\):
\[p_{b,c,t}^*\] We also define scenario-based emissions intensity value \(s^*\) of company \(c\) for sector \(b\) at time \(t\) as:
\[s_{b,c,t}^*\] We then derive the total emissions intensity deviation (\(\Delta total^*\)) of company \(c\) in sector \(b\) at time \(t\):
\[\Delta total_{b,c,t}^* = p_{b,c,t}^* - s_{b,c,t}^*\] It follows that the alignment measure \(y^*\) of company \(c\) based on emissions intensities in sector \(b\) at time \(t\) is:
\[y_{b,c,t}^* = \dfrac{\Delta total_{b,c,t}^*}{s_{b,c,t}}\]
The company alignment metric for emissions intensity sectors \(y_{b,c,t}^*\) follows a similar logic as the one for sectors with technology pathways, in that it measures the percent deviation of the emissions intensity in time \(t\) for a company from its scenario-based emissions intensity at time \(t\). Again, this means that \(y_{b,c,t}^* = 0\) describes a company that is right on track with with its emissions intensity relative to the scenario-based value, whereas \(y_{b,c,t}^* = 1\) means the emissions intensity is one hundred percent better (lower) than required by the scenario value and \(y_{b,c,t}^* = -0.5\) means that the emissions intensity is fifty percent worse (higher) than implied by the scenario value. There is no disaggregation of emissions intensity based metrics, as these are calculated on the sector level to begin with.