Whose Money Pays the Person Advising You | HL Hunt
Whose Money Pays the Person Advising You
Most consumer credit is not bought directly. It arrives through a broker, a finance desk, a comparison site, a lead generator, and almost none of them are paid by the borrower. The compensation structure predicts the recommendation better than anything about the products themselves, and it does so without requiring anyone to make a dishonest decision. The stylized ranking below is the clearest case: ordering lenders by what the site earns produces the exact reverse of ordering them by what the borrower pays, and that inversion is the equilibrium rather than an accident.
What you'll learn
The core thesis
Most consumer credit is not bought directly. It reaches the borrower through someone in between: a broker, a dealer's finance desk, a comparison site, a lead generator, a marketplace, an advisor. Each of them is paid, and almost none of them are paid by the borrower.
Our position is that the compensation structure of the intermediary predicts the recommendation better than any characteristic of the products being recommended. This is not a claim that intermediaries are dishonest. It is a claim about what happens to a distribution channel over time when one option pays more than another and nothing forces the difference to be reconciled with the borrower's interest.
The mechanism does not require anyone to make a corrupt decision. It requires only that, across thousands of marginal calls, the option that pays better gets presented slightly more often, ranked slightly higher, described slightly more favourably, and pursued slightly harder when the borrower hesitates. Per our distribution analysis, the channel shapes the outcome more than the product does, and compensation is the channel's operating instruction.
What the difference is worth
Two examples, one simple and one less obvious.
The size incentive
A broker compensated as a percentage of the amount arranged faces this on every file:
- 2.5% of a $30,000 loan: $750
- 1.5% of an $18,000 loan: $270
- Difference: $480 for recommending the larger obligation
Nothing in that structure is hidden or unusual. It simply means that when a borrower could reasonably take either, the intermediary is paid $480 to see it one way. Per our exposure analysis, the larger loan is also the one that leaves less room for everything else the borrower will need to pay.
The ranking incentive
A comparison site earns a payout when a click converts. Its expected revenue per click is the payout multiplied by the conversion rate, and the ordering that maximizes it has nothing to do with rate.
| Lender | Payout | Conversion | Revenue per click | Rate for the borrower |
|---|---|---|---|---|
| C | $180 | 3.5% | $6.30 | Worst |
| D | $95 | 6.0% | $5.70 | Mid |
| B | $70 | 7.0% | $4.90 | Mid |
| E | $55 | 8.0% | $4.40 | Good |
| A | $40 | 9.0% | $3.60 | Best |
The inversion is not a coincidence, it is the equilibrium. A lender with a poor offer must pay more per conversion to buy distribution, and a lender with a strong offer does not have to. So payout and borrower value are negatively related by construction, and any ranking that follows payout will tend to follow it downward.
The forms compensation takes
The question to ask of any intermediary is not whether they are paid but what changes the amount.
| Structure | What it rewards | Alignment with the borrower |
|---|---|---|
| Flat fee from the borrower | Completing the work | Strongest available |
| Flat fee per funded loan, same for all lenders | Funding something | Neutral on which, biased toward yes |
| Percentage of amount | Larger obligations | Diverges as the amount rises |
| Percentage varying by lender | Specific lenders | Diverges on the choice itself |
| Rate participation | Higher rates | Directly opposed |
| Volume bonuses or tiers | Hitting a threshold | Distorts sharply near period end |
| Paid on performance of the loan | Loans that get repaid | Strong, and rare |
The second row is the one most people misjudge. A flat fee per funded loan is often described as neutral, and it is neutral between lenders. It is not neutral between funding and not funding, which is frequently the decision that matters most. An intermediary paid only when something closes has no way to be paid for advising against it.
The sixth row produces the sharpest distortions and the least attention. Per our incentive analysis, a threshold converts a smooth incentive into a cliff, and behaviour near a cliff is not the behaviour the scheme was designed to produce. The same file gets a different answer on the 3rd of the month and the 29th.
The last row is the alignment everyone claims to want and almost nobody adopts, because it defers the intermediary's income by months and transfers credit risk to a party that does not want to hold it. That it is rare is informative about how much the industry values the alignment relative to its cost.
Why disclosure does not fix it
The standard remedy is to disclose the compensation. The evidence that this works is weak, and the reasons are structural rather than about consumer inattention.
- The disclosure arrives after trust is established. A document presented at signing cannot undo a relationship formed over weeks.
- It says the arrangement exists, not what it changed. Knowing a broker is paid by lenders does not tell a borrower which option was not presented.
- Per our consent analysis, acknowledgement is not comprehension. A signature confirms receipt and nothing else.
- It can increase trust rather than reduce it. Candour about a conflict reads as honesty, and a disclosed conflict can be more persuasive than a hidden one.
- The borrower has no comparison. Without knowing what was excluded, the disclosure describes a structure whose effects are unobservable.
The fourth point is the uncomfortable one. Disclosure is not merely insufficient; in some settings it makes the problem worse, because it licenses the intermediary to feel candid while doing nothing about the underlying incentive.
A falsifiable claim
Where an intermediary's compensation varies by lender, the distribution of placements should correlate with payout after controlling for borrower characteristics. Any lender that buys through intermediaries can test this on its own data: for borrowers with comparable files who were eligible at more than one lender, compare where they actually landed against relative payout. If placement is uncorrelated with payout once eligibility is held constant, the argument here is overstated and the channel is functioning on merit. We expect a clear correlation, and its absence would be the strongest available evidence that disclosure and duty regimes are working better than this article assumes.
The case for intermediaries
The argument against intermediation entirely is worse than the problem, and the reasons are worth stating at full strength.
- They reduce search costs enormously. Per our verification analysis, a borrower approaching lenders individually pays the cost of being assessed repeatedly, and per our inquiry analysis that search is itself recorded against them.
- They aggregate access. Many lenders do not sell directly at all, so the intermediary is not a markup on a direct option, it is the only option.
- They carry expertise the borrower cannot acquire for one transaction. Knowing which lender accepts which circumstance is genuinely valuable and genuinely hard to learn.
- Borrowers will not pay directly. This is the binding constraint, and it explains nearly all of the structure. A fee-for-advice model has repeatedly failed in consumer credit because willingness to pay for advice before a loan exists is close to zero.
The fourth point is the honest centre of the whole problem. The conflicted model exists because the unconflicted model does not sell. Per our cross-subsidy analysis, this is the general shape of free distribution: someone pays, and the identity of the payer determines what the service optimizes.
What a borrower can actually do
- Ask what changes the amount. Not "are you paid" but "does it differ by lender, and does it differ by loan size." The answer is more informative than any disclosure document.
- Ask what was excluded and why. The presented set is the output of the incentive; the excluded set is where the information is.
- Get one direct quote. One lender approached directly calibrates everything the intermediary presented, per our shopping analysis, and the cost is one inquiry.
- Treat an unrequested larger offer as information. An offer above what you asked for is the size incentive operating.
- Notice the timing. Pressure concentrated at month end is a volume threshold, not an expiring rate.
- Compare total cost, not payment. Per our pricing analysis, a longer term lowers the payment, raises the total, and raises the intermediary's percentage compensation at the same time.
Item two is the highest-value question and almost nobody asks it. "Which lenders did you not put forward, and why" converts an opaque process into one where the intermediary must either give a reason about the borrower's file or reveal that the reason was something else.
Scenarios and what we are watching
| Scenario | What we would expect |
|---|---|
| Duty-of-care standards extend to more intermediaries | Placement distributions flatten against payout, and some intermediaries exit because the conflicted margin was the business |
| Compensation is levelled across lenders | The lender-choice distortion goes, the fund-or-not-fund bias stays, and the second is rarely addressed |
| Automated matching expands | The incentive moves from a person to a ranking algorithm, becomes harder to observe, and is applied with perfect consistency |
| Nothing changes | Disclosure continues to be the remedy, and continues to describe a structure whose effects the borrower cannot see |
The third row is the trajectory we would flag. A human intermediary applies a conflicted incentive unevenly, with judgment and occasional conscience. An algorithm optimizing revenue per click applies it identically to every borrower, at scale, with no variance and no record of the alternative. That is a better business and a worse outcome, and it will be described as removing human bias from the process.
Frequently asked questions
A broker at 2.5% on $30,000 earns $750 against $270 at 1.5% on $18,000, so $480 rides on recommending the larger obligation.
Because revenue per click is payout times conversion, and lenders with worse offers must pay more for distribution. The ranking inverts borrower value by construction.
Weakly, and sometimes it backfires: candour about a conflict reads as honesty and can make an intermediary more persuasive while changing nothing underneath.
Which lenders did you not put forward, and why. The presented set is the output of the incentive; the excluded set is where the information is.
Key takeaways
- Compensation structure predicts the recommendation better than any characteristic of the products recommended.
- A broker at 2.5% on $30,000 earns $480 more than at 1.5% on $18,000, on every marginal file.
- Ranking by revenue per click inverts ranking by borrower rate, because weak offers must pay more for distribution.
- A flat fee per funded loan is neutral between lenders and not neutral between funding and not funding.
- Volume thresholds produce the sharpest distortions: the same file gets a different answer on the 3rd and the 29th.
- Paying intermediaries on loan performance is the alignment everyone claims to want and almost nobody adopts.
- Disclosure is a weak remedy and can backfire, because candour about a conflict makes an intermediary more persuasive.
- The conflicted model exists because the unconflicted one does not sell: willingness to pay for advice before a loan exists is near zero.
This analysis is educational and does not constitute financial or legal advice. All figures are stylized illustrations using assumed compensation rates, payouts, and conversion rates, and do not describe any particular intermediary, lender, or comparison service. Compensation disclosure requirements, duty-of-care standards, licensing obligations, and restrictions on lender-paid compensation vary by product, by intermediary type, and by jurisdiction, and change over time.