P&L
Common Causes of P&L Differences
When two P&L numbers that should match don't, the cause almost always falls into one of a small number of well-understood categories. Here's the reference list.
When two P&L figures that should agree - a desk's own estimate versus the ETRM's number, or two systems that are both supposed to be showing the same book - don't match, the cause almost always falls into a small number of recurring categories. This is a practical reference list for the most common ones, roughly in order of how often we encounter them.
1. Market data timing or source mismatch
The two numbers were calculated using prices from different points in time, or from different sources for the same instrument. This is by far the most common cause we see, and the fastest to check: confirm both calculations used the identical price snapshot.
2. Position timing differences
One system's position reflects trades booked slightly later (or earlier) than the other. A trade booked at 4:58pm might be included in one end-of-day snapshot and excluded from another, depending on exact cutoff timing.
3. Valuation methodology differences
Different discounting conventions, interpolation methods for curve construction, or volatility surface assumptions between two systems will produce genuinely different - and both individually defensible - valuations. This is a methodology reconciliation, not a bug.
4. FX conversion differences
For any cross-currency book, using a different FX rate or a different conversion timing between two calculations will produce a P&L difference that has nothing to do with the underlying commodity position.
5. Fee and cost treatment
One calculation includes transaction fees, broker commissions or transportation costs; the other doesn't, or includes them in a different period. Small individually, but a common source of a persistent, low-magnitude discrepancy.
6. Realized vs unrealized boundary differences
Disagreement about exactly when P&L moves from unrealized to realized (at trade date, delivery date, or settlement date) will shift how much of a given movement shows up in each category, even when the total is correct.
7. Silent fallback or default data
As covered in Why Market Data Quality Matters, one calculation may have used a stale fallback price without anyone realizing it, while the other used a live, current price.
A practical checklist
| Check | What to compare |
|---|---|
| Market data | Same source, same timestamp, same snapshot for both calculations |
| Position | Same cutoff time, same trade set included |
| Valuation method | Same discounting, interpolation and volatility assumptions |
| FX | Same rate source and conversion timing |
| Fees/costs | Same inclusion and period treatment |
| Realized/unrealized boundary | Same definition of when P&L crystallizes |
Key Takeaways
- Most P&L differences trace back to a small set of recurring causes: market data timing, position timing, valuation methodology, FX, fee treatment, or the realized/unrealized boundary.
- A methodology difference between two systems isn't necessarily a bug - it needs to be reconciled and understood, not just 'fixed.'
- Working through a consistent checklist, rather than starting from scratch each time, dramatically speeds up P&L reconciliation.
Related Services & Reading
Working through this in your own environment?
Tell us about your platform and the specific issue you're facing - we'll suggest the right next step.