Econometric modelling: Opening the black box

The methodologies we use are all based on the empirical, peer-reviewed work of scholars at the intersection of economics and statistics. Generally, the results we deliver are comprised of several different statistical models that can be compared with a baseline visualization of the data.

Since a statistical model is only as good as the assumptions it relies upon, our ironclad rule is to use only methods that are logically sound and justifiable given the specific context and data. In other words, common sense trumps sophistication for the sake of sophistication.

Our careful, scientific approach results in a model uniquely designed for each and every context. Central to this approach is an emphasis on disentangling causation from correlation. Many services and software packages can provide analyses of the correlations between variables, but uncovering and quantifying true causation requires an understanding of what to do when statistical assumptions fail. It is also what provides clients with the power to make informed, data-driven decisions.

The results of our modeling are subjected to a rigorous battery of checks, a refining process that provides quantitative insights that can be trusted. The end product is a tailored report designed to meet a client’s need by answering data-driven questions in a concise and straightforward way.

To read more about our methodologies, see Publishing.