Physical commodity trading quantitative risk return model

To Anyone else looking for a base model to go off on, I was not able to incorporate the more high level (market risk, political risk), etc I was looking for. However, Trafigura does publish their value at risk model (VaR) for a single transaction: https://www.trafigura.com/media/1490/2014_trafigura_the_economics_of_commodity_trading_firms_section_iii_english.pdf This was the single most useful article I found. To incorporate the additional risks, I would theorize using this risk as a baseline against which to go, and adding constants/variables in as blackbox variables (whatever metric you chose as political risk), and with your own company/public historical data, reverse identifying a good weight for the given variable. I have moved no from Petrochemical risk and am now a Data Scientist, so have not tested this theory

Besides taking the suggested reading materials above and skimming them for help, I have been in commodity markets probably 15 years. I can tell you that there are NO commodity risk programs I've found that will answer your questions we all have as quants. The typical quant will invent their own method combining concepts from the financial markets which are typically covered in school. I've seen some of the strangest modeling in commodities including a model combining interest rate models with volatility surface models and models never combined in any research paper. All I can suggest is that computing power today is so high most problems can be "solved" i.e. estimated via Monte Carlo. Whether that leads you to a Python 3D array (correlations - what time period - EWMA - Garch - up to you - vol surface lookups, 1 week or a year model etc.). Just backtest whatever creation you come up with in and out of the sample used to create it before using it and see how it performs before using it. I'm also looking to boost my quant skills as a risk quant I have to examine / replicate / replace trader desk models. If I knew of a good course or certificate I'd tell you, but I can't find any. I work with some super creative quant traders no one can understand their mix of financial market models with commodity models either. Apparently the are elite PhD graduates a.k.a. genius mathematically

I think you might benefit from reading a few resources.

• Hélyette Geman's book Commodities and Commodity Derivatives is an excellent overview from a quantitative perspective.
• Morgan Downey's Oil 101 is a good explanation from an oil trader's perspective of the oil market.
• For petrochemicals, pipelines, refinining or other issues, William Leffler's "in Nontechnical Language books are great. Petrochemicals in Nontechnical Language is written by Leffler and Donald Burdick and could prove useful for you.
• If you find yourself getting into natural gas or power, Bob Shively and John Ferrare's Understanding Todays {Natural Gas, Electricity, Global LNG} Business books are informative and usually up-to-date with recent practice.

Risk appetite or aversion are often absent from commodity models because the contracts are often used to hedge risks. That said, if investors became a large presence in a certain market, a factor model with a commodity or multi-asset-class index (similar to the CAPM) might be useful. I suspect you might do better just trying to model the range of suppliers and their GPMs so that you can infer the supply function (similar to the "powerstack" function in power markets).

Commodity prices depend on global supply and demand but not on the perception of the market regarding an adequate risk premium for a specific asset class.

Have a look on this Claude Erb and Campbell R. Harvey, "The Tactical and Strategic Value of Commodity Futures"

Your question piqued my interest. While not specific to commodities, this looks like a good starting point for quantifying political risk:

Practically, this means taking the following steps (according to Rigobon 2003):

Step 1: Define the treatment group, or a set of “event” days on which the variance of the unobservable factor is high, such as the Italian referendum on December 4.

Step 2: Choose a set of “non-event” days to serve as a control or comparison group. Common practice suggests choosing non-event days one or a few days before the event days, so as to minimize the influence of risk factors other than the political one. Political risk may also change on these days, but (by assumption) they change less than on event days. This assumption is a leap of faith, but hopefully a reasonable one.

Step 3: Apply a standard econometric technique known as instrumental variable (IV) regression. Consider regressing changes in one variable of interest (e.g., the CDS spread of Spanish sovereign debt) on changes in a second variable (e.g., Italian CDS spreads), that is then instrumented by a proxy.3 This proxy is the same variable but with opposite sign on non-event days.

Unlike a traditional event study, this approach does not apply the unrealistic assumption that political risk only changes on event-days. In situations like the Italian referendum, in which periodic polls revealed changes in voter preferences, the traditional event-study method may underestimate the magnitude of the political risk.

source: Quantifying Political Risk on Financial Markets—Italian Case Study