I am running an options book containing listed options across multiple products. I trade mostly equity and index related options - with a preference for European expiration products. I trade products listed across multiple continents - so there is an FX risk component as well.

I have lots of spreadsheets etc which I use for adhoc risk management controls - typically, looking at the sensitivities of a single product (e.g. SPY book). Typically, I have a single workbook detailing the positions for a single product, and I use the greeks for risk management purposes. However, as my portfolio increases, I have decided it is time to implement a more consolidated approach to risk management.

I would like some guidance on how to:

  1. Consolidate option sensitivities (i.e. across the entire portfolio as opposed to single products)
  2. Get a handle on FX exposure risks in the portfolio

Ideally, I would like to create a single report that outlines the entire portfolio risk (i.e. option sensitivities and FX risk).

I would be grateful for any guidance or links to information that will help me extract data that I am already generating, and put that data together in a framework as the one described above, with the aim of obtaining a more global view of my portfolio risk.

Note: I am not looking for third party system, I simply want to know how to aggregate the data I already have, in a meaningful and practical way (hopefully with some theoretical/empirical underpinning), in order to manage my portfolio risk better.


2 Answers 2


The most successful market risk management frameworks separate calculations from products. This makes them centralized, scalable and allows easy aggregation. Please note I'm referring to the aggregation of security-level exposures; not computing new portfolio-level metrics. Speaking very generally, such an infrastructure requires the following:

  1. A library of risk calculations.

  2. A separate product database for tracking the various indicatives of each security.

  3. A layer of aggregation logic to group similar results across different products:

    • Calculations X and Y both yield vega, but for different products
    • Calculations X and Y both yield JPY exposure, but for different products
  4. A report that displays the (optionally aggregated) results of some set of calculations on another set of products.

The key idea is to produce your exposures in some centralized infrastructure and pull only what you need into a report. This way you let the reports just filter and aggregate and avoid the redundancy of making each one perform calculations as well (additionally it minimizes the operational risk of having a calculation defined differently in two spreadsheets).

Given your current setup, the simplest way to start might be by putting all your securities and calculations on a single spreadsheet, making sure that calculations with the same type of result are in the same column, and then use pivot tables to filter and aggregate the information. So one table shows all risk calculations broken out across a certain product, reproducing your current set up; another shows a single calculation (say, vega) summed across all products, giving you a portfolio-level view. This may not be ideal, but you could implement it almost immediately.

Depending on your level of development experience, you can build more sophisticated infrastructures that don't require jamming everything into a spreadsheet -- or using spreadsheets at all!


I was running a very similar book just a month ago. On the market/option risks, your best bet is to convert all sensitivities into something "familiar". For example, I converted all deltas into SPX delta using index/stock beta, converted all vega to SPX vega using vol-beta etc. FX is the unpleasant one, you end up looking at a list of pair exposures and usually end up doing nothing.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.