Econometric Advisors Services

Our main product is a streamlined retail loan portfolio analysis service, modelled on over a decade of experience in portfolio acquisitions. Relatedly, we provide post-acquisition monitoring technology, turning raw servicer data feeds into intelligible and actionable information.

We also offer our expertise in building data pipelines and bespoke statistical models, and in data-related process automation more generally.

Loan data analysis

Loan data: size matters

Retail loan portfolio transactions come with large data sets on portfolio composition and historical performance, with file sizes sometimes running into the gigabytes. This in itself poses problems for the Excel-based approach to cashflow forecasting many financial analysts are familiar with. Historical performance information in particular is difficult to analyze without relatively sophisticated technical machinery.

Our loan analysis service employs such machinery in the form of purpose-built software libraries, which allow us to deliver results quickly whilst avoiding error-prone spreadsheet calculations. A decade of experience working across the full life-cycle of loan portfolio acquisitions helps guide valuations when data alone do not provide the answer.

Features

Data cleaning and organization.
Detailed portfolio composition analysis, in graphical and tabular format.
Flexible specification of loan-level prepayment and default forecasts.
Optional econometric model-driven forecasts where appropriate historical data for model calibration are available.
Clear pdf-format report with data, forecasts and hand-written commentary.
Use as an add-on to your deal team for fast interaction, or stand-alone e.g. as a back-up for your own computations.

Contextual information

Loan data, and mortgage loan data in particular, often include a wide set of attributes. This creates the possibility to augment the analysis with contextual data. For example, when precise property location information is at hand, one can link this to regional macro-economic, demographic and property transaction data in order to paint a more detailed picture of the portfolio than is possible with the loan tape alone.

We continue to invest in regional macro and housing research to build this capability. Data availability varies widely across jurisdictions; where possible the results are integrated into our portfolio valuation reports.

Portfolio monitoring

Overseeing a portfolio of loan pools managed by different servicers quickly becomes a major drag on internal resources when done manually. Numerous files in different formats and with different naming conventions are sent across every month, emails are overlooked, copy / paste errors are made and people go on holiday. Fortunately, every data-driven business faces this type of problem. The solution is straightforward at a conceptual level: a central database puts data on a common denominator, and is the source for all reporting. A data normalisation ("ETL") engine handles the ingestion of raw servicer data.

Our offering in this space adheres to this general layout. It consists of ready-made reusable software components tailored to the loan reporting use case and can be customized to fit your needs. The data ingestion / normalisation engine is fully configurable via human-readable files to facilitate onboarding new loan pools. You build reporting directly via tools such as Tableau, Qlikview, PowerBI or Excel, or we provide a Flask app serving up a custom-built dashboard.

Process automation

Automate the boring stuff

Management of loan portfolios through 3rd party services often entails a range of back-end processes that go beyond standard monitoring and reporting tasks.

Whenever you find yourself repeating similar data-related tasks in Excel over and over again, we can probably help. We develop custom solutions that are ready to use, or train your staff in essential programming skills and assist you in developing an in-house solution.

Because we already understand the problems you face, we can deliver tools far faster than general-purpose software firms and free up your time to focus on what matters.

Bespoke statistical modelling and forecasting

Let your intuition guide you where hard data cannot, but only from there on. Our bespoke modelling and forecasting service pushes that frontier as far as it should go. We help you frame the problem, organize your proprietary data and augment them using publicly available information - typically this will be macro-economic or financial time-series. We then construct a model and set up automated forecasting machinery so that your predictions are always up-to-date.

The right tool for the job

We are able to do what we do thanks to a truly amazing open-source eco-system of system development and data analysis software, centered around the Python language and the Postgresql relational database management system.

There will be instances where your needs cannot be fully met by this toolkit, particularly when your company has already committed to a different stack. If we are not able to accommodate this, we will say so and help you find an alternative partner if possible.