You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When we do trajectory projection (mean or median or however we calculate the trend), there is a final trajectory trend rate which is applied to the last reported data and projected forward. Each year we get new data, we get both a new trend calculation (which doesn't change too dramatically, because it weighs past behavior) and a new starting point (which can cause dramatic changes to the trajectory estimation, because a projection from a lower number lowers ALL the future carbon projections based on that new lower starting point).
When we do target projection, the original code treated the forward-looking target budget a bit like a bank loan: the projection was done from the target start year, filling in all future values to the target end year, and then actual numbers sit on top of those numbers (like payments against a bank loan). If a company emits less than projected, it "banks" the benefit (its actual used budget goes down), but the forward-looking targets remain unchanged. So updating data changes target estimates very little until all the years are ultimately calculated, or until new targets are set that change the forward-looking data (like refinancing a loan).
An alternative implementation for targets could be that yes, there is a base year and yes there is a target reduction amount, but each year we get fresh data, we could change all the unrealized target projections to start from the most recent realized data and then compute the intercept based on the target reduction aspiration from that point. Good early progress would dramatically change the forward-looking target estimates because the target reduction would be starting from a lower point, giving a lower overall carbon budget.
Is it better for the tool to make ITR scores more sensitive to data updates, by recalculating forward-looking data based on trends and/or target reductions from the newest baseline? Or better to make it less sensitive to such by anchoring against the base year and just showing the small changes that each new year data brings to the forward-looking budget year by year?
Hi Michael,
Ultimately, I think it is better to also use the latest data available in the target projection. Which would logically also imply that, when realized reductions have been larger (smaller) than according to the required annual reduction rate to get from base year to target year emission intensity, the target EI reduction rate from the actual start year used in the calculations will be smaller (larger) than the reduction rate derived from what was required to get from base year to target year EI.
Bert
Thanks for that, Bert. Adjusting the projections to hit the target from the perspective of the latest data is actually the easiest to implement from my perspective. It will make target and trajectory handling more consistent.
When we do trajectory projection (mean or median or however we calculate the trend), there is a final trajectory trend rate which is applied to the last reported data and projected forward. Each year we get new data, we get both a new trend calculation (which doesn't change too dramatically, because it weighs past behavior) and a new starting point (which can cause dramatic changes to the trajectory estimation, because a projection from a lower number lowers ALL the future carbon projections based on that new lower starting point).
When we do target projection, the original code treated the forward-looking target budget a bit like a bank loan: the projection was done from the target start year, filling in all future values to the target end year, and then actual numbers sit on top of those numbers (like payments against a bank loan). If a company emits less than projected, it "banks" the benefit (its actual used budget goes down), but the forward-looking targets remain unchanged. So updating data changes target estimates very little until all the years are ultimately calculated, or until new targets are set that change the forward-looking data (like refinancing a loan).
An alternative implementation for targets could be that yes, there is a base year and yes there is a target reduction amount, but each year we get fresh data, we could change all the unrealized target projections to start from the most recent realized data and then compute the intercept based on the target reduction aspiration from that point. Good early progress would dramatically change the forward-looking target estimates because the target reduction would be starting from a lower point, giving a lower overall carbon budget.
Is it better for the tool to make ITR scores more sensitive to data updates, by recalculating forward-looking data based on trends and/or target reductions from the newest baseline? Or better to make it less sensitive to such by anchoring against the base year and just showing the small changes that each new year data brings to the forward-looking budget year by year?
Is the question clear @LeylaJavadova @ImkeHorten @BertKramer? What are your thoughts?
The text was updated successfully, but these errors were encountered: