Daniel sent us this one — he's asking about credit scores. What are they, actually? In the US, is there some central database that organizations query to check someone's rating? And is this whole thing a US-specific concept, or do other countries have their own versions under different names? It's a good question, because the credit score feels like a universal fact of adult life if you live here, and then you move abroad and discover it's basically meaningless.
The deeper question underneath all of it is — what does it mean for a society to quantify trust? Because that's what a credit score is trying to do. Take the messiness of someone's financial behavior and compress it into a three-digit number that a lender can glance at and make a decision. And different countries have answered that question in radically different ways.
Let's start with the thing itself. What is a credit score, not the marketing version, but the mathematical reality?
At its core, a credit score is a statistical summary of your past repayment behavior, designed to predict one specific thing — the likelihood that you'll default on a new loan within the next twenty-four months. That's it. It's not a measure of your financial health, it's not a report card on your responsibility as a human being, it's a default probability model.
It's a prediction, not an evaluation.
Well, not exactly, but yes. The two dominant models in the US are FICO and VantageScore. They use five weighted factors. Payment history is the heaviest, at thirty-five percent — did you pay on time? Amounts owed is thirty percent — how much of your available credit are you using right now? Length of credit history is fifteen percent, new credit applications are ten percent, and credit mix — the variety of account types you have — is the last ten percent.
These formulas are proprietary, right? You can't just look up the exact algorithm.
Proprietary and fiercely guarded. Fair Isaac Corporation, which created the FICO score, publishes the weightings but not the underlying model. It's like knowing the ingredients list on a food package but not the recipe. And the score itself ranges from three hundred to eight hundred fifty. The median US score in twenty twenty-five was seven hundred eighteen, according to Experian's data. Below about six seventy and you start running into higher interest rates or outright rejections.
That's the high-level picture. But how does the data actually move from your credit card payment to a lender's screen? Let's trace the pipeline.
This is where people get tripped up. There is no central government database. The system is a private oligopoly. Three bureaus — Equifax, Experian, and TransUnion — each maintain their own independent credit files on roughly two hundred million Americans. They're competitors, not branches of the same entity.
Which means your score can actually vary depending on which bureau the lender pulls from.
And the data flows in through something called data furnishers. Every month, your bank, your credit card issuer, your auto lender — they all send account status updates to one or more of the bureaus. Paid on time, thirty days late, balance is X, credit limit is Y. The bureaus aggregate this into a credit report.
The score itself — is that stored somewhere, sitting in a database waiting to be queried?
This is the key technical detail that most coverage misses. The score is computed at query time. It is not stored. When you apply for a car loan, the dealer doesn't ask TransUnion for your score. They ask for your credit report. Then their system — or a scoring engine they've licensed — runs the FICO or VantageScore algorithm against that raw report and generates the number on the spot.
The score is a derived product. It's ephemeral.
The underlying asset is the report. The score is just a computation performed on demand. And different lenders might use different versions of the scoring model. FICO has released multiple generations — FICO Score eight, FICO Score nine, FICO Score ten. Some lenders are still using FICO Score five for mortgage underwriting because Fannie Mae and Freddie Mac require it.
Of course they do. Nothing says cutting-edge risk assessment like a model that's been frozen in regulatory amber for two decades.
It's the mortgage equivalent of running Windows XP because your industrial equipment depends on it. But let me walk through a concrete timeline, because I think it makes the whole thing tangible. Say you pay your Chase credit card bill on May fifteenth. Chase reports your updated balance and payment status to the bureaus around May twentieth — there's always a few days of lag. On June first, you apply for a car loan. The dealer's financing department pulls your TransUnion report through an API — something like TransUnion's Credit Profile Number system. The raw report comes back. Their system runs it through FICO Score nine. Out pops seven hundred forty-two. That number determines your APR.
The entire chain is: you pay, Chase reports, bureau stores the raw data, lender requests the report, lender computes the score, lender makes the decision. And that all happens in seconds.
But the data behind it can be months old. That Chase payment from May fifteenth? If you apply on May twenty-second, it might not have hit the bureaus yet. The system is fast but it's not real-time.
Which brings us to the "thin file" problem. What happens when there's just not enough data?
This is where the system's architecture creates structural exclusion. Roughly twenty-six million Americans are what the CFPB calls "credit invisible" — they have no credit history at all. No file exists. Another nineteen million have "unscorable" records — too thin, too sparse, too old to generate a reliable score. That's about forty-five million people total.
Almost one in seven Americans.
It disproportionately hits low-income communities, recent immigrants, and minority populations. If you've never had a credit card, never taken out a student loan, always paid rent in cash — you're invisible. The system has no way to assess you, so it defaults to rejection or predatory pricing.
Which is a self-reinforcing trap. You can't get credit because you have no history, and you can't build history because you can't get credit.
And this is where the alternative data experiments come in. Experian Boost lets you voluntarily add utility and telecom payment history to your file. UltraFICO — which was piloted a few years back — pulls in bank account transaction data, like your average balance and how long you've held the account. The CFPB has been pushing since twenty twenty-four to include rental payment data as a standard reporting category.
The idea is to expand the scoring population without fundamentally changing the architecture. You're still generating a three-digit number from a private bureau's data — you're just feeding it more inputs.
And there's a genuine tension here. More data means more people get scored, which is good for inclusion. But it also means more of your financial life gets fed into a surveillance system you didn't explicitly opt into.
"We've decided to include your rent payments — you're welcome.
That's basically the CFPB's position, yes. And look, the legal framework for all of this is the Fair Credit Reporting Act, signed into law in nineteen seventy. It created the rules for consumer reporting agencies — what they can collect, how long they can keep it, your right to dispute errors. The CFPB, created in twenty eleven by the Dodd-Frank Act, is the enforcement body.
The FCRA is over fifty years old at this point. It predates the internet, let alone the algorithmic scoring models we're talking about.
The law was written for an era when credit reports were literally paper files in filing cabinets. It doesn't really contemplate machine learning models that can infer your creditworthiness from your zip code or your shopping patterns.
That's the US system — a private, data-intensive, scoring-driven machine. Now let's look at a country that chose a completely different path.
This is where the contrast gets really sharp. Israel has no equivalent to the US credit score. There is no FICO, no VantageScore, no three-digit number that follows you around and determines your access to credit.
Which is wild when you think about it. Israel has a highly developed tech sector, a sophisticated financial system, and yet it doesn't do the thing that Americans assume every modern economy must do.
What Israel has instead is a centralized credit registry — the Credit Data Registry, or in Hebrew, Ma'agar Netunei Ashrai. It became operational in September twenty nineteen under the Credit Data Law. The Bank of Israel runs it. It collects both positive and negative data from banks, credit card companies, and non-bank lenders.
It's a government database — that's already a fundamental difference from the US model.
But here's the crucial distinction: it's a raw data repository, not a scoring system. Lenders query the registry and get back your credit history — accounts, balances, payment patterns, defaults. But the registry doesn't compute a score. Each lender has to do its own underwriting, its own risk assessment, using the raw data.
The government provides the data, but the judgment call stays with the lender.
And this evolved for several reasons. First, Israel's banking sector is incredibly concentrated — five banks control over ninety-five percent of credit. They already had deep relationships with their customers and their own internal risk models. They didn't need a third-party score.
They already knew who was reliable because they'd been lending to the same families for three generations.
That's part of it. Second, the mortgage market is heavily regulated. Mandatory down payments are twenty-five to fifty percent, depending on the property type and whether it's your first home. When someone's putting down that much cash, the need for granular risk scoring drops significantly — the down payment is the risk mitigation.
If I've already put down forty percent, the bank's exposure is much lower. They're not going to lose sleep over whether my FICO equivalent is six eighty or seven twenty.
Third, and this is harder to quantify but important — Israel is a small, networked society. Reputation and personal relationships substitute for formal scores in ways that don't scale to a country of three hundred thirty million people. Your bank manager probably knows someone who knows someone.
The "my cousin's neighbor vouched for me" underwriting model.
It sounds funny, but it genuinely reduces information asymmetry in a small market. And fourth — the twenty nineteen law wasn't purely organic. It was partly a response to EU regulatory pressure and a twenty sixteen Bank of Israel report that found forty percent of small business loan applications were being rejected due to lack of credit history.
Israel recognized it had its own version of the thin file problem, and built a centralized registry as the solution — but deliberately stopped short of creating a scoring system.
And by twenty twenty-three, the registry covered one hundred percent of bank credit and about eighty percent of non-bank credit. It's comprehensive. But it's a data layer, not a judgment layer.
Which makes it fundamentally different from the US approach. The US says: we'll compute the judgment for you. Israel says: here's the data, you figure it out.
This pattern repeats across countries in different configurations. Let's look at Germany. Germany has SCHUFA — Schutzgemeinschaft für allgemeine Kreditsicherung, which is a private credit bureau much like the US bureaus. SCHUFA holds data on sixty-eight million individuals and six million businesses, and it produces a score — the SCHUFA-Score.
Germany looks more like the US.
On the surface, yes. But the factors are different. SCHUFA doesn't use credit utilization ratio the way FICO does. Instead, it places heavy emphasis on address stability — how often you move — and on public records like bankruptcy filings. The German system basically assumes that someone who's lived at the same address for ten years is more reliable than someone who moves every year.
Which makes a certain kind of cultural sense in a country where renting is normalized and long-term residence is common.
And the SCHUFA score is used for everything — loans, mobile phone contracts, even rental applications. You can be denied an apartment because your SCHUFA score is too low, which creates exactly the kind of exclusion trap we talked about with the US system.
What about the UK?
The UK has three bureaus — Experian, Equifax, and a company called Callcredit, which is now called TransUnion UK after an acquisition. But the scores are lender-specific and not standardized. Each lender develops its own scorecard using bureau data. There's no single national number.
France is fascinating. They have a centralized negative registry called the FICP — Fichier des Incidents de remboursement des Crédits aux Particuliers. It only records defaults. If you've never defaulted, you have no entry. There is no positive scoring at all. No record of on-time payments, no credit utilization tracking, no length-of-history calculation.
In France, you're innocent until proven guilty. In the US, you have to prove you're creditworthy before you're trusted.
That's exactly the philosophical split. The French system is built on a presumption of reliability — you're assumed trustworthy unless you've demonstrated otherwise. The American system is built on a presumption of opacity — you need to generate data that proves you're trustworthy.
Then there's China, which is the one everyone asks about.
China's Social Credit System is a separate beast entirely. It's state-run, and it blends financial behavior with social behavior — everything from loan repayment to traffic violations to, in some pilot programs, social media activity. But here's the thing most Western coverage misses: the pilot systems in twenty twenty-four through twenty twenty-six have been scaled back significantly. The original vision of a unified national score that determines your access to everything from flights to school enrollment has been walked back to a patchwork of local and sectoral systems.
The dystopian version that makes for good headlines isn't quite the reality.
Not yet, and maybe not ever. The implementation has been much messier and more fragmented than the policy documents suggested. But the concept is still fundamentally different from anything in the West — it's a state mechanism for social control, not a private mechanism for lending decisions.
Let's talk about the EU dimension, because there's a regulatory framework that shapes all of this.
The General Data Protection Regulation, effective May twenty eighteen. Article twenty-two is the relevant bit — it restricts fully automated individual decision-making, including credit scoring, without meaningful human review. If an algorithm denies you a loan, you have the right to have a human look at it.
Which is why many European countries have weaker scoring systems — not because they can't build them, but because the law limits how they can be used.
And this creates a genuine tradeoff. The US system enables instant credit decisions. You can apply for a credit card online and get approved in thirty seconds. Risk-based pricing means people with good scores get lower rates. But the system creates exclusion and opacity — most Americans don't know what's in their credit report, and errors are notoriously hard to fix.
Israel's system reduces exclusion — the centralized registry means lenders at least have the raw data — but it requires manual underwriting. Credit decisions are slower. You're not getting instant approval on a loan application at two in the morning.
The European approach, with GDPR's human-review requirement, adds a layer of protection but also friction. It's slower, more bureaucratic, but harder for an algorithm to quietly discriminate against you.
Let me ask you about a concrete case. A German freelancer with no credit history versus an American freelancer with no credit history. Same person, economically speaking — self-employed, irregular income, always paid their bills. What happens when each of them tries to get a loan?
The German goes to their local Sparkasse — the savings bank — with three years of bank statements, tax returns, and maybe a letter from their accountant. The loan officer reviews the documents, looks at the cash flow, and makes a judgment. They might get the loan. It might take a week. But there's a path.
The American applies online. The system pulls their credit report. There's no file, or a file too thin to score. The algorithm doesn't see bank statements, doesn't see tax returns, doesn't see that they've paid rent on time for five years. It sees nothing. The application is rejected — or they're offered a thirty percent APR from a subprime lender.
The American freelancer is punished for being invisible, while the German freelancer can at least make their case to a human being.
This is the core insight. Every credit system embeds assumptions about what counts as reliable behavior. The US system assumes that reliability is demonstrated through the formal credit market — if you've never used credit, you can't be assessed. The German system assumes that stability and cash flow are visible through banking relationships. The Israeli system assumes that the lender already knows enough about you to make a judgment, or can supplement the registry data with personal knowledge.
None of these are neutral. They're all choices.
They have distributional consequences. The US system benefits people who are already inside the credit system — people with mortgages, car loans, credit cards they've held for years. It systematically disadvantages people who operate outside that system, even if they're financially responsible.
We've seen three models. The US scoring machine — private, algorithmic, instant. The Israeli data registry — centralized, raw, relationship-supplemented. And the European regulatory approach — constrained by GDPR, varied by country, with a bias toward human review. What does this mean for someone navigating these systems, depending on where they live or where they're moving?
Let's start with the American moving abroad. Your FICO score does not travel. It's meaningless in Israel, in Germany, in France, in the UK. You will be starting from scratch in the new country.
Which is a shock if you've spent twenty years building an eight hundred ten score and then discover it's worth exactly nothing the moment you cross a border.
The practical implication is that you need to build credit from scratch, or rely on alternative proofs of reliability. Bank statements showing consistent income. Rental payment history. In some countries, a letter from your previous bank can help. But there is no global credit score, and there probably never will be, because creditworthiness is so deeply embedded in local legal systems and cultural assumptions.
What about the reverse? Someone moving to the US who's never had a credit score?
There are services designed for exactly this. Nova Credit is the most prominent — it translates foreign credit data into a US-equivalent score that American lenders can use. It works with data from countries including the UK, Canada, Australia, India, and Mexico. But it's not universal, and it's not free.
It's a bridge, but a toll bridge.
The other strategy is the secured credit card. You deposit, say, five hundred dollars with a bank. They issue you a credit card with a five-hundred-dollar limit. You use it and pay it off every month. After six to twelve months, you've generated enough data to get a FICO score, and you can graduate to an unsecured card.
The "thin file" problem is solvable, but it requires deliberate strategy. You have to know the rules of the game before you can play it.
This brings us to the third actionable insight, which is maybe the most important. A credit score is not a measure of financial health. It's a measure of creditworthiness for a specific type of debt product. You can have an eight hundred credit score and be one missed paycheck away from disaster because you have no savings. You can have a six hundred credit score and a fully funded emergency fund, no debt, and a paid-off house.
The score doesn't know about your savings account. It doesn't know about your income. It doesn't know whether you inherited money or live paycheck to paycheck.
It knows none of that. It only knows your history with debt products. And understanding this distinction is the first step to what you might call ethically gaming the system — using the rules as they exist, rather than treating the score as a moral judgment.
"Gaming the system ethically" is a great phrase. It's basically: learn what the algorithm rewards, and do those things, while understanding that the algorithm is not measuring your worth as a human being.
Keep old credit cards open even if you don't use them, because length of credit history matters. Don't use more than thirty percent of your available credit, because utilization ratio matters. Pay your bills on time, obviously. These are mechanical optimizations, not spiritual improvements.
Like adopting a feral cat. You're not a better person for having done it, but the system now recognizes you.
an unusual analogy, but I'll allow it.
Let me push on something. You mentioned the CFPB's Section ten thirty-three rule, finalized in late twenty twenty-five, which mandates consumer data access. And we're seeing experiments with bank account scoring. Is the monopoly of the FICO score eroding?
I think it is, slowly. The idea behind open banking is that if consumers can authorize lenders to access their bank account transaction data directly, you don't need a credit score to assess risk — you can look at actual cash flow. Does this person have consistent income? Do they regularly overdraft? What's their average balance? That's a much richer picture than "did they pay their credit card on time.
The shift would be from credit-score-as-prediction to cash-flow-underwriting.
And the twenty twenty-six experiments are significant. Several fintechs are now offering "bank account scoring" products that bypass the traditional credit bureaus entirely. The CFPB's rule gives consumers the legal right to access and share their own financial data, which means they can route around the Equifax-Experian-TransUnion oligopoly if they choose to.
Which would be a fundamental restructuring of how credit decisions are made in this country.
But there's a parallel development that complicates the picture. Buy Now, Pay Later — services like Klarna, Affirm, Afterpay. These are exploding in popularity. And most of them don't report to the credit bureaus at all.
Someone could have ten different BNPL loans outstanding and a completely blank credit report.
It's creating a parallel credit system that's invisible to the traditional scoring infrastructure. Regulators are grappling with this right now — should BNPL providers be forced to report to the bureaus? If they are, it would reshape the scoring landscape significantly. If they aren't, we're going to have an increasingly fragmented system where traditional scores capture less and less of people's actual credit behavior.
The musical equivalent of beige wallpaper. It's there, it's doing something, but nobody's quite sure what it's covering anymore.
That's the forward-looking tension. We built this incredibly sophisticated apparatus for quantifying trust — private bureaus, proprietary algorithms, API-level data pipelines — and now it might be slowly rendered obsolete by two forces pulling in opposite directions. Open banking, which offers a richer picture of financial health. And BNPL, which operates entirely outside the old framework.
The credit score might not be permanent. It might just be a fifty-year interlude between "the bank manager knew your father" and "the algorithm reads your bank statements directly.
That's a provocative way to frame it, and I think there's truth to it. The credit score was a solution to a specific problem — how do you assess a stranger's creditworthiness in a mass market where personal relationships don't scale? If open banking solves that problem better, the FICO score becomes a legacy technology.
Now: Hilbert's daily fun fact.
Now: Hilbert's daily fun fact.
Hilbert: Seamounts — underwater mountains that rise from the ocean floor without breaking the surface — are among the most isolated habitats on Earth. Some seamounts in the Pacific have endemism rates above thirty percent, meaning nearly a third of the species found there exist nowhere else. In the high medieval period — around the twelfth century — the coastal waters off what is now French Guiana sat above a submerged volcanic ridge that scientists believe functioned as a biological time capsule, preserving species that had vanished from surrounding regions thousands of years earlier. The seamount acted as a refugium, a surviving artifact of an older ocean.
...right.
Here's the open question. If cash-flow underwriting replaces credit scoring, what counts as reliable behavior gets redefined again. And someone will build a new algorithm, and that algorithm will have its own blind spots, its own exclusions, its own assumptions about what a trustworthy person looks like.
The technology changes, but the fundamental problem doesn't. Every society has to decide how to measure trust in strangers. The credit score was one answer. Israel's registry is another. Whatever comes next will be a third — and it'll embed its own politics, whether we notice them or not.
If this episode changed how you think about that three-digit number, leave us a review and tell us your credit score origin story. Or, if you're outside the US, tell us what your country uses instead. We'd love to hear it.
This has been My Weird Prompts, with me, Herman Poppleberry.
Thanks to our producer Hilbert Flumingtop. Find us at myweirdprompts dot com.
We'll be back next week.
