Vega P&L Attribution

External question — Quantitative Finance Stack Exchange Author: Tracy Chen Original post: https://quant.stackexchange.com/questions/85839 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I am trying to decompose Vega PnL into ATM vol, skew, and curvature components. To that end, I am building an SVI model and using the SVI-JW parameterization to attribute Vega PnL via the following formula: $VegaPnL = Vega \times dVol \approx Vega \times (\frac{dVol}{dv_t}dv_t+\frac{dVol}{d\psi_t}d\psi_t + \frac{dVol}{dp_t}dp_t + \frac{dVol}{dc_t}dc_t + \frac{dVol}{d\tilde{v}_{t}}d\tilde{v}_{t})$ However, the results are not satisfactory — the residual is extremely large. What is the correct way to perform this Vega PnL breakdown, and where might the problem lie? Thx