→ 🇫🇷 French version here

It has been exactly one year since I immigrated to Canada. I adapted quite easily to almost every change that came with it, but one specific thing still leaves me puzzled.

April 2025. I arrive in Calgary with two huge suitcases, a hiking backpack, plus a smaller bag strapped to my chest with all my important papers inside.

For things to get serious, (finding a place to live and a job), I need to start a few administrative steps locally. Luckily, it is pretty simple and fast here. I set up my phone plan from the airport, then applied for a Social Insurance Number (SIN) from the federal government, and the next day I’ll open a bank account.

Once at the branch, they offer to do everything standing at the counter, but I prefer to take a moment with an advisor, as I expect there will be a few details to understand in the North American system. I ask for my first credit card, and since here this is a real line of credit being opened, it is also the beginning of what is called my credit history, and my credit score.

What is a credit score?

All Canadians (but also Americans, Australians, and Britons) who have ever applied for credit, whether a credit card, a mortgage, or a margin account, have a credit score. It is a points system that lets the lender estimate whether the borrower will be able to repay.

It also works for car dealerships who check whether someone is likely to repay the car loan they are asking for; and also for landlords who want to make sure the tenant has a good payment history.

So, concretely, how does it work?

In short, when you get your first credit card, your history begins, and it is tied to your Social Insurance Number.

If you forget to pay back your credit card at the end of the month¹, then your credit score will go down by a certain number of points. On top of that, your file will include a negative note, and it will stay there for 6 years.

It is the same if you miss a loan payment, or if you default on a bank account. These are the things that have the strongest negative effect on a credit score.

To get a good credit score, and therefore access the best borrowing conditions or rent an apartment more easily, it is a little more subtle. Beyond having a long history with no incidents, you increase your score by accumulating several credit cards. In fact, it is not uncommon in North America to have two, three - or even 12 credit cards. It is almost a hobby for people who chase points or cashback rewards, which are more attractive than in Europe².

Also, using only 10% or less of your credit limit at the end of the month helps your score rise faster. So the strategy is to ask for high limits ($10,000 or more, for each card) in order to stay below that threshold.

Standardizing risk, making the market flow

Credit score versus… what exactly?

I have looked for rental apartments in France, in several tense housing markets, and then in Montreal, where demand is also high compared with supply. And yet the experience is radically different.

In both cases, you have to move fast to find and secure an apartment with good value for money, but the paperwork and supporting documents are almost completely absent on this side of the Atlantic. A credit score check is almost always requested, sometimes with references from previous landlords. Full stop. In fact, it is discouraged or even forbidden to ask for a copy of an ID document or a pay slip.

Not only are people asked for more and more documents, often intrusive and even humiliating to provide, but tenants are also exposed to a risk of personal data leaks. The best part? Because of this bureaucratic paranoia and the growing use of generative AI, up to 70% of rental applications now include at least one forged document.

So if I simply compare the two systems as a whole³, whether you are financially fragile or comfortable, I find the credit score more efficient and less intrusive.

Every risk has a price

The same advantage exists when buying real estate. In Europe, banks usually try to assess solvency through a full review of supporting documents. They are also more conservative in how they define debt.

Getting a mortgage is usually simpler and faster in North America, where the credit score also plays a central role. If you are above 800 points, you have every chance of getting the best rate, for the highest amount. All the lights are green for the bank, so it makes you pay less for the risk than others.⁴

So in the end, nothing surprising: the market puts a price on risk. Any economic actor that wants to survive tries to assess risk as precisely as possible, whether it is investing or lending money. The parallel can be made with rating agencies, which assess whether states are able to repay their debt.

An algorithm this important must be explainable, right?

This is where we get to the heart of the matter. So I will keep telling you about my life (quickly, I promise), to explain how I ended up thinking about this. Spoiler: I’m not the only one.

After living in Calgary, my partner and I eventually move to Montreal. Since we had already got used to a city where a car was almost essential, we go to a dealership to buy a vehicle (a project we quickly give up on, because it is so easy to get around here by bike and public transit).

And when it’s time to pre-authorize the car financing on part of the amount, the sales agent runs a credit check. Then she tells me just after, because she had forgotten:

It will take a few points off, but it should be negligible.

A few days later, I check my credit score on my banking app, and I realize that what she called “negligible” was actually a 47-point drop in my score. So I look into it, and I learn that if someone makes several credit applications too close together, the system decides that they are trying to borrow left and right. As a result, it penalizes them by temporarily removing points.

So even though I did not end up taking that loan, the inquiry (or hard inquiry) was too close to my first credit card application two months earlier.

And there are many rules like this. They have a logic, but the real problem is that the algorithm is so opaque that nobody really knows how many points you lose for which action. Or how many you gain for another one… nor when exactly you will get them back if you fix the situation.

”What asshole designed credit score algorithms?”

I am not the one asking the question. It comes from an internet user whose score dropped by 50 points just because he may have paid back his credit card a little too quickly. The comments below pile on with personal stories, one after another expressing their disbelief at how absurd the calculation feels.

The first credit scoring systems appeared in the United States in the late 1950s (or even the 1930s, if we include non-computerized systems). They gradually became standardized, eventually becoming the FICO score in 1989. From there, the three major credit bureaus adopted it: Equifax, Transunion, and Experian.

These companies were not new. From the second half of the 20th century, as household banking spread, credit bureaus quickly started racing to collect data on people’s credit histories… and far beyond that.

Data on consumer habits, then online transactions, all the way to the use of capital letters in online forms, was aggregated by data brokers to make it easier to profile American consumers⁵.

In a Federal Trade Commission report⁶, it says that in the 1990s, Transunion had created a behavioral analysis model called SOLO. It categorized 140 million Americans into 6 clusters. For example: “Urban Ethnics”, “Urban Upscale”, “Empty Nesters”, “Single Strugglers”, “Kids and Cars”…

And even though, since 1970, the Fair Credit Reporting Act and its more recent amendments have brought some protections to American citizens, the credit score system has not become transparent. Far from it. The same goes for Canadian law⁷.

And now?

Today, FICO directly licenses mortgage scores, bypassing traditional credit bureaus. This allows the company to regain control over pricing, while strengthening its position against competitors.

But that was not enough, because very recently, on April 22, 2026, the Federal Housing Finance Agency and the Department of Housing and Urban Development announced the adoption of VantageScore 4.0 (in addition to FICO 10T) for mortgages. It somewhat marks the end of FICO’s monopoly and the beginning of a new era of competition in the American mortgage credit sector.

FICO shares reacted violently too, with a 14% drop that evening.

Business first, we can discuss fairness in court

The rules behind credit score calculations are not clear, and this leads to many strange stories every year. A few examples:

A man from Victoria, BC, pays off all his debts and his credit score is reset to zero, after 50 years of good credit. The algorithm decided that he was no longer a credit product customer, so in a way it stopped assessing his solvency. Consequence: trouble finding a rental apartment.

Equifax eventually revealed that this happened after two years of inactivity, without saying whether it applied to all Canadians. The man spent more than a year fighting for justice… without winning. According to the Financial Consumer Agency of Canada:

Even worse, a fraudulent debt taken out in your name can destroy your credit score. A woman spent 18 months fighting Equifax before the error was finally corrected. Too late to avoid losing $11,000, after signing a less attractive loan offer than the one she could have obtained.

It looks as if even the credit bureaus cannot explain the algorithm they use. It is a bit like a black box whose lid we occasionally try to lift, just to check the state of the thousand cables inside.

Of course, the guides explaining the credit score calculation, especially FICO’s, offer a first basic level of transparency. But even if we roughly understand what counts in the calculation, we do not know at what level, nor when: how long we are penalized, and for what exact reason.

And on top of that, we have to pay for it?

The cherry on top? To access your full credit file online, you have to pay.

Even though access to the score itself and to some data is free on a monthly basis, you have to pay, for example, $24.95 per month to get your file from Transunion, which also sells you monitoring. Because yes, your score can collapse after identity theft⁸.

Companies have even been created to meet the demand for explainability around one’s own score, like Credit Karma, which gives you a view of your score and ways to improve it.

The citizen method: reverse engineering and data sharing

Citizens with expertise in finance and data analysis teamed up to carry out and publish a colossal investigation. To gradually strip the algorithm bare, they relied among other things on years of testimony posted on Reddit and FICO forums, adding up to 4 million posts and 300,000 users.

We managed to understand GENERALLY how the FICO score works, We managed to understand A LOT about how some parts of the score work, BUT we managed to understand that we do not know EXACTLY how the whole FICO score works.

FICO’s approach to credit scoring is proprietary, meaning it is private. People at FICO know how their scoring systems work. The rest of us make increasingly educated guesses to learn the details of these models, and we come to forums like this to learn for ourselves how to understand, improve, and manage our scores, and to help others learn too.

We learn, for example, that the score does not work the same way for everyone, with 12 profiles called Scorecards, judged either clean or dirty. We also learn that these scorecards are calculated using more or less advanced mathematical formulas: maximum likelihood, multicollinearity tests…

For example, having a loan from an institution with “Finance” or “Financial” in its name (often high-interest loans for riskier profiles) can cost between 20 and 40 points, because these accounts are seen negatively by the algorithm.

In short, the algorithm is proprietary and has never been published by FICO or by the credit bureaus. We can infer several reasons (more or less legitimate):

• It is a trade secret for FICO. And credit bureaus boast about their unmatched ability to compile data through an exhaustive calculation algorithm.
• It is too technically complex. With 500 parameters and only 20 selected depending on the individual (scorecards), themselves determined by relatively complex formulas, we can imagine that the algorithm does not fit into one simple formula.
• And probably the most important one: to prevent smart consumers from optimizing their score too effectively with a few tricks. The predictive value of the score, its ability to distinguish good payers from bad ones, could be compromised by excessive user optimization.

A thousand and one arguments for more transparency

I want to believe it’s because I am European that I’m outraged by the lack of accountability of such an algorithm. As if the GDPR, and its endless cookie banners and alerts about our access and correction rights, had ended up entering my DNA. Let us sing for the new peace, Of our united Europe ⁹ In reality, it is probably professional deformation, after 11 years of open data. And above all, criticism of this system is not new.

This opacity is even more problematic when we know that the credit score amplifies social inequalities. First, wealthier people are favored: the earlier you can borrow, the more your score rises, and the more you can borrow. Second, two researchers from Stanford and Chicago showed that less advantaged individuals were more likely to have incorrect data in their credit files.

Many studies have linked credit scores and ethnic minorities, with the consequence of higher average interest rates for Black Americans than for white Americans. People also talk about Credit Invisibility to describe what particularly affects Hispanics in the United States. Since they pay more often in cash or by mobile payments, they use credit cards less, and so you know what comes next: no history, no score, no borrowing.

The right to access and correct information is also central to avoid a double punishment for people facing medical debt. After a missed hospital bill payment, some people see their credit score deteriorate, without being able to act and correct the situation.

There are many criticisms of the score’s effectiveness, and others that question the ethics of the system itself. But to stay pragmatic, if we cannot hope for a radically different system, could we at least impose stronger transparency on credit bureaus?

Note: While the original version of this article was written by a human, AI was used to assist with a quick translation