Decomposition Analysis Flashcards
What is a decomposition analysis
The disaggregation of a number into >2 components
Disaggregation of energy consumption
into volume/activity and intensity effect
Disaggregation of CO2 emissions into
energy demand and carbon intensity
Three reasons for decomposition
- getting insights in past developments
- understanding scenario outcomes
- supporting policy analysis (trough top-down analysis)
Energy consumption can be disaggregated in three effects
- Volume (V)
- Structure (S)
- Energy savings/intensity/efficiency (I)
CO2 emissions can be disaggregated in 5 effects
- Volume (V)
- Structure (S)
- Energy intensity (I)
- Fuel mix (EF)
- Emission factor (EF)
2 factor decomposition methods
- successively
- one-by-one
Successively approach
Decompose: b-a
Look at Delta V with I constant. Thus Delta I = b-a- Detla V
Do also for Delta I with V constant.
One-by-One approach
- Determine Delta V with constant I
- Determine Delta I with constant V
- Add up Delta V and Delta I.
-Leaves a residual and is therefore not the same as b-a
2 factor decomposition without residual
Laspeyres method
When to use Laspeyres method
With 2 factors since it is easy to communicate and explains the factor you want to decompose.
-Does work for 3 or more factors but gets really complicated
Laspeyres method
½ (V0 +Vt) x ΔCI + ½ (CI0+CI1) x ΔV
Advantage of LMD appraoch
no residuals
What does LMD stand for
Log Mean Divisia
Logarithmic mean formula
(a-b)/ln(a/b)
Formula for LMD approach
L * ln(Xt/X0)
X should be replaced by the appropriate effect
Determining the decomposition trough LMD
Apply the LMD formula to all factors you use and add up the results
Difference between Laspeyres method and one-by-one
In the Laspeyres method instead of freezing one factor you take the average of it and multiply it with the Delta of the other factor.
Downside of LMD approach
Does not work for 0 or negative data
How to determine which factors to use
Make it into a formula, every factor has to be in their twice, either as a whole or in a (breuk) with another factor