There are many articles and advice columns on how to maximise your mystical credit score yet few of them deal with the psychology of how credit-worthiness is assessed. This was something that fascinated me the first time I was rejected for a mortgage (while the second rejection, coming directly after the first rejection, evoked emotions a little stronger than fascination).
A great proportion of landlords and first-time buyers have good stable jobs and buy property as an investment. Many of them are buying for the first time. Many of them believe that having a “clean” record of credit is enough to make them attractive to a lender and secure that mortgage.
Many of them don’t have a credit card or only sparingly use one thinking that this shows responsibility with credit. Many of them believe they are financially savvy by swapping and changing bank accounts or phone providers securing the best cashback deals or other incentives.
They are wrong!
How can that be, you ask? How can someone, on a very good salary, with great career prospects, who uses mainly debit cards and has never missed a payment in their life be rejected for credit while a person who is on an inferior salary, has never been promoted, has worked in the same office all their career, has three credit cards, and still lives with Mum is welcomed with open arms by the same bank for the same mortgage?
The answer is in what’s called an algorithm. Almost every mainstream lending decision today is decided by a computer using an algorithm. In fact, many mortgages are applied for, processed, assessed, and provisionally given without a single pair of human eyes even glancing at the application.
Do you give good algorithm?
An algorithm does not make value judgements. It does not care that your CV is exceptional. It does not care that you have had a wide variety of positions and had the good fortune to be promoted and relocated for your very impressive job.
Instead, it assesses you purely on data that you submit and if that data matches the data that it has been programmed to match you with. In each area, it simply assesses whether you comply or not and gives a weighted score. An algorithm is a process that is concerned only with cold, hard, unemotional data.
An algorithm, in this instance, can be seen as a way to simply “profile” an applicant. It profiles you against what it sees as an ideal customer. It creates this profile of an ideal customer based on years and years of data collated from all its countless customers, from the model customer who successfully paid their debt to the delinquent customer who did a “runner”, as well as all the various types of customers in-between.
You see, this is their business. Lenders are experts at identifying the (sometimes seemingly irrelevant) characteristics of good customers based upon historical data even if the data they are looking at make little sense to you or me.
In fact, (and it may not make sense to senior management at the bank either) if the algorithm says that people with green eyes are more likely to be highly represented in the list of customers who are stable with credit than those with blue eyes, then having green eyes makes you are more likely candidate.
While this last example is obviously nonsensical as you cannot show a correlation between eye colour and credit, (could be a great topic for a PhD, though) it is not that far from what actually happens with a credit assessing algorithm. It is about understanding the difference between “correlation” and “causation”.
Good correlation = Looking good to a computer and not another human
A correlation is a relationship between two sets of data. This data may have absolutely no connection at all. For example, people who are left-handed may be better at a certain sport than people who are right handed. Scientists may have vague theories but really have no conclusive evidence as to why this is true. It just “is”.
On the other hand, causation includes a specific reason why one characteristic causes something to happen. For example, teenagers who start smoking before they have fully developed are more likely to be shorter than those that do not smoke at the same age because the chemicals contained in cigarettes. In other words, smoking is “causes” stunted growth whereas a left-handed person is more likely to be good at sports for reasons unknown and possibly unrelated.
It’s more about looking good on paper
Lenders deal better with correlation as it is quantitative (numbers and such) data that can be processed by an algorithm to spit out a result (this is what computers mainly do). Causation deals with qualitative (the “quality” of something) data that typically needs be interpreted and is multi-dimensional.
As such, correlations are much easier to locate and assess than causations. A lender is less concerned about evidence or causation (as it’s difficult to quantify and prove), and more interested in something easier to measure, being data matching or correlation.
Essentially, lenders want an easy life. They want a homogeneous customer base with each customer having the save characteristics as the next. This is because banks don’t like risks. Correlations and algorithms show them where the risks are. Your job is to mitigate this risk for them and make their job giving you credit that much easier.
For daily motivational advice, wealth creation articles, and a place to ask any and all questions you may have about becoming financially free, head over to our free Financial Freedom coaching group.