Some of the most frequent questions we get here at Crest Capital have to do with the finer points of assessing credit risk. Which is a fancy way of saying: "what exactly do we look at in making an equipment financing decision, and how do we reach these decisions?"

The answer is actually more complex than anyone thinks, and it's especially multifaceted with equipment financing, where a transaction could be in the hundreds of thousands of dollars (or more.)

To give an example, when a lender finances a car for someone, they look at the credit score, and that's generally that. But when an equipment financing company lends money to a company for equipment, other things come into play. These can include (but are not limited to) the following: age and nature of the business applying for credit, comparable borrowing history, the equipment's useful life and aftermarket value, soft costs, why we don't like to finance anything that requires ongoing vendor reliance, risk-based pricing, etc. The list is long and varied. Hence, this series of articles designed to give you a better idea regarding the inner workings of our business.

Let's start with an answer to a question that's asked often: how much "human" decision is factored into credit decisions?

The answer may surprise you – not that much. In the old days, of course, humans made ALL of the credit decisions. But today, computer models and reams and reams of data have led to the creation of sophisticated algorithms that are – we must admit – incredibly accurate.

Let's look at a few points regarding what we call "credit risk models".

• In this day and age, credit risk models are used by almost every lender. Prior to this, credit decisions were based on a loan officer's ability to make a judgment using available data. In other words, it was a subjective decision. Credit Risk Models made this decision objective. Almost a complete reversal.

• However, it's not like these Credit Risk Models appeared out of the blue one day. It was a slow evolution, beginning with credit managers and loan officers making their own "scorecards" based on their own findings. If you looked in a loan officer's desk circa 1980, you likely would have found some type of notebook with data and findings regarding credit risk factors. However, since these were personal, they could vary drastically from one lender to the next (and could be silly, as well. For example, say a loan officer got burned by someone named "Smith" - you can bet the next "Smith" applying for a loan will start off on the wrong foot in the loan officer's mind, even if there is no relation whatsoever.)

• Over the years, however, it was inevitable that the best pieces of these formulas (not the "Smith" part) were similar. Then, models developed by Fair Isaac Corporation (FICO) became commonplace. We work off of these today.

• These models were/are used because they have taken the best of the best, and have enabled lenders to make complex, high-volume decisions quicker. They also utilize large-scale optimizations involving dozens of interconnected action and effect models, and enables exploration and simulation of many optimized "what if" scenarios.

• The predictive power of these models continuously improves, and also adapts to current economic conditions. Lender loss typically decreases with every revision.

• Of course, one model does not work for everyone, so most firms will customize proprietary risk models using the company's own data and past transactions.

• Head to head comparison testing between older and revised models is sometimes done, using past accounts. This allows a lender to see if they can expect an improvement in risk assessment (e.g., "Would the new model have prevented that bad loan last year?")

• Despite all of this, there still exist lenders who go by the old-school "gut feeling" approach. However, the last recession weeded out almost all of them. In plain terms, the "outside-the-model exceptions" these firms gave credit to typically ended up defaulting much more frequently than anticipated.

In the end, it's been proven that decision management utilizing credit risk models is definitely superior to just "gut feeling". Scoring automation provides faster decisions, cuts fraud losses, minimizes credit risk, complies with regulatory requirements, and meets competitive demands.

Now you may ask "ok, so why do we even need humans?" The answer is the models do have a "line in the sand", and many transactions fall right at or around that line. For transactions like that, humans are still needed to make the final call. Plus, someone needs to enter in all of the data!!!

Thanks for reading. Next time, we'll look at what data elements are considered in these scoring models.

Crest Capital is an EquipmentFinance and Equipment Leasing company that provides businesses with the funds they require to obtain the equipment they need. Regardless of the strength of the economy or the current economic climate, Crest understands that solid businesses still wish to grow, and strives to provide easy financing at great rates, with the fastest approval time in the industry.