What the Credit Cycle Does to Earnings Quality
Easy credit does not just raise earnings — it degrades the information inside them. At the top of a credit cycle, a reported dollar of profit tells you the least, priced at its most.
Serious investors already know that earnings are cyclical, and most correct for it: normalise the margin, average the cycle, apply a haircut at the peak. The subtler problem is that the credit cycle does not just move the level of earnings — it changes what a reported dollar of earnings is made of. Two identical income statements, one produced in a credit expansion and one in neutral conditions, are not equally informative documents. The first contains revenue that is really someone else's borrowing, per-share growth that is really the company's own borrowing, and an audit process quietly weakened by the fact that nobody demands answers of a borrower when everyone is eager to lend. The mistake is not failing to normalise earnings. It is normalising the level while trusting the composition.
Borrowed earnings: the three channels
We call the phenomenon borrowed earnings: profits that the credit cycle lends to the income statement and that the balance sheet — somebody's balance sheet — must eventually repay. The borrowing happens through three distinct channels, and they compound because they peak together.
The first channel is credit-financed demand. A meaningful share of what appears as revenue in a credit expansion is, on inspection, the customer's borrowing wearing the company's top line as a costume. Consumer durables bought on financing, housing-adjacent volumes driven by mortgage availability, capital equipment sold into debt-funded projects, enterprise contracts signed by venture- and leverage-funded customers — in each case the company reports revenue, but the economic event is a credit decision made somewhere upstream. When credit tightens, this revenue does not merely slow; it reveals itself to have been partly an advance on future demand, because credit-financed purchases pull consumption and investment forward in time. The company's own operations can be flawless and its revenue still contracts, because the contraction happens on its customers' liability side.
The second channel is financing-cost engineering. Cheap debt flatters per-share earnings through two mechanical routes: interest expense held below its through-cycle level while the coupon gap sits unrealised on the maturity schedule, and share counts shrunk by debt-funded repurchases. Neither route creates a dollar of operating profit, yet both raise the number the market capitalises. The tell is decomposition: when earnings-per-share growth persistently outruns net income growth, and net income growth outruns operating income growth, the gap between each pair is the credit cycle's contribution — and it reverses with the cycle, because the same leverage that bought the share count amplifies the decline when operating income turns.
The third channel is the least discussed and the most corrosive: the erosion of lender discipline. Credit markets are, in normal times, an external audit function — lenders demand covenants, information rights, and conservative working-capital behaviour as the price of capital. In an aggressive expansion, that price collapses. Documentation loosens, covenant protections thin out, and management teams face fewer counterparties entitled to ask hard questions. The consequence for earnings quality is indirect but real: receivables stretch further before anyone objects, aggressive revenue recognition draws less challenge, acquisitions clear a lower bar because the debt to fund them is available on demand. The credit cycle does not merely finance lower-quality earnings — it removes the mechanism that would have flagged them.
Why quality degrades exactly when confidence peaks
The illustration: the mid-2000s, where all three channels ran at once
The 2004–2007 expansion remains the cleanest teaching case because the three channels are all visible in the historical record, at scale, in different sectors. The demand channel ran through housing and everything attached to it: homebuilders, lenders, building products, and consumer discretionary businesses reported earnings that were, in retrospect, a levered bet on mortgage availability — revenue booked by companies, borrowed by households. The engineering channel ran through the corporate sector broadly, as inexpensive leverage funded buybacks and acquisition sprees that converted balance-sheet capacity into per-share growth. And the discipline channel ran through the credit markets themselves, as documentation standards and underwriting scrutiny famously loosened while spreads compressed.
The instructive detail is not that 2008 punished this — it is what the punishment revealed. Financial-sector earnings from the expansion years turned out to contain a large component that was the credit cycle recognising its own expansion as profit: origination fees, securitisation gains, and reserve releases that existed because credit was growing, and reversed because it stopped. Businesses downstream of household credit discovered that their addressable demand had been partly an advance from future years, repayable through the lean period that followed. None of this required fraud, and that is the point of the illustration: borrowed earnings are overwhelmingly legal, audited, and honestly reported. The distortion lives in composition and context, which is exactly why it does not show up in the documents investors are trained to trust.
The strongest case against the framework
The disciplined objection comes in two parts. First, unfalsifiability: any revenue can be labelled credit-financed by a sufficiently motivated bear, and the framework risks becoming a permanent reason to distrust good results — a thesis that is never wrong because it is never testable. Second, the futility of cycle awareness: credit cycles run long, turning points resist prediction, and an investor who discounted earnings quality throughout an expansion would have surrendered years of compounding waiting for a reckoning that arrived late. Cash is cash, audits are real, and the burden of proof sits with whoever claims the reported numbers overstate reality.
Both points deserve straight answers. On unfalsifiability: the framework is testable, just not at the aggregate level — it cashes out in specific, checkable properties of individual companies. What share of revenue depends on customers' access to financing, and what happened to that revenue in past tightenings? Is cash conversion tracking reported earnings, or diverging as receivables and accruals absorb the difference? Is per-share growth explained by operations or by the gap between EPS and operating income growth? These are questions with answers, and companies differ sharply on them. On futility: the framework is deliberately not a timing tool, and using it as one misreads it. It is a cross-sectional discrimination tool — at every point in the cycle, it separates companies whose earnings the credit cycle can repossess from companies whose earnings it cannot. You harvest that distinction without ever calling the turn, because when the turn arrives, on whatever schedule, it grades the portfolio you already hold.
Holding earnings to a cycle-proof standard
The operational shift is to treat earnings quality as a credit-cycle variable rather than a company-only variable. The same diligence questions carry different weight at different points in the cycle: a widening gap between earnings and operating cash flow is always worth noting, but in a late-stage credit expansion it is the single most important line of inquiry, because that is when the incentive and the opportunity to borrow earnings are both at their maximum. Sterling's Embedded Intelligence operationalises this by reading income statements against their credit context — asking not only whether earnings grew, but which channel grew them, and whether that channel survives a tightening. The related frameworks in this library extend the picture: the private credit piece examines what happens to lender discipline when the negotiation moves behind closed doors, and the refinancing wall piece covers the moment the financing-cost channel reverses on schedule.
How to apply this framework
- Trace the customer's balance sheet, not just the company's. Estimate what share of revenue depends on someone else's access to credit — customer financing, project debt, mortgage availability. That share is the revenue the credit cycle can repossess, whatever the company's own leverage looks like.
- Watch the cash-conversion gap as a cycle instrument. Reported earnings persistently outrunning operating cash flow — with receivables and accruals absorbing the difference — is the standard signature of borrowed earnings, and it deserves the most weight precisely when credit conditions are loosest.
- Decompose per-share growth into its layers. Compare EPS growth to net income growth to operating income growth. The gaps between them measure the financing-cost channel's contribution — the part of growth that reverses with the cycle rather than compounding through it.
- Read documentation looseness as a market-wide quality gauge. When covenant protections thin and credit is extended with few questions, the external audit function is weakening for the whole cohort — so tighten your own standard exactly when the market is relaxing its, because the cross-sectional benchmark is most contaminated then.