If I sell 5M of an IG corporate and buy 5M of another with similar duration. I then drag and drops the ISIN into the model and the system shows the new exposures, sector weights, duration impact, spreads. All looks good, run through compliance and off to the trading desk.
What the system did not model for is whether the buy is in fact available for me to purchase at the current liquidity today. It has no way of accounting for various compliance breaches at the time of execution: cash concentration, various settlement conventions, pricing updates affecting MV%.
These execution problems are handled by traders, and the trade is re-iterated throughout the day with no guarantee that the price hasn't moved by the time something that passes all the checks is ready to be executed on.
What if scenarios have always been built too narrowly. the job to be done is very simple: take the current portfolio, apply a hypothetical trade, recalc MV% and contributions to portfolio level analytics. This was enough in slow moving markets with slow blipping prices. Nowadays the markets move fast and T-1 data is only part of the risks that can derail a portfolio manager performance.
What if scenarios should account for pre-trade compliance, broker eligibility, cash & settlement availability, any active limits on the account (issuer, country, etc.) and ideally a model of the dependent trades that will need to follow. This is what most trading tools fail to capture: If my 5M corporate bond switch from earlier tips a sector limit and forces a trim on another position to make the trade work, the portfolio management system should surface that dependent trade and let the PM stage the two trades synchronously. Otherwise the dependency surfaces post the second compliance run at which stage the price may have already moved and the trade may no longer make sense in the way it did a few hours ago. back to square one.
Arithmetic recalculates a state. Modelling simulates the state plus its dependencies. Front-office tools that claim to do the second mostly do the first, and leave the dependencies for PM's to discover.
This becomes harder, not easier, at scale. Modelling a single switch against a single mandate is tractable. Modelling a rebalance across 30 portfolios with mandate-specific constraints, broker restrictions, and overlapping liquidity buckets is a different problem entirely, and one that no vendor has fully solved.
The framing worth challenging is "impact". Impact on what? Most tools mean impact on the snapshot of the portfolio. The more useful question, and the one buy-side desks are increasingly asking, is impact on the probability that this trade completes today, at a price the PM would still choose.