Guide

Allowance for doubtful accounts and bad debt expense explained

Harbor Telecom Equipment posted its best operating margin in five years in Q2 2025 — 14.2%, up 180 basis points year over year. The balance sheet footnote most analysts skipped told a different story: gross accounts receivable climbed 26% while the allowance for doubtful accounts fell from 2.1% to 0.4% of gross AR. Management attributed the reserve release to “improved collection experience.” Six months later, three regional wireless carriers filed for restructuring and Harbor took a $31M bad debt charge in a single quarter — wiping out two quarters of operating profit. The failure was not fraud in the classic sense; it was an allowance that no longer matched credit risk.

The allowance for doubtful accounts (also called the allowance for credit losses on trade receivables) is a contra-asset that sits against gross AR on the balance sheet. Bad debt expense is the income-statement charge that builds or releases that reserve. Under ASC 326 (CECL), public companies must estimate expected losses over the life of receivables — not only when default is probable. This guide covers gross vs net AR presentation, the allowance roll-forward, CECL methodology in plain language, ratio screens that pair with days sales outstanding (DSO), Harbor Telecom’s refactor, a technique decision table, earnings quality pitfalls, and an investor checklist alongside financial statements literacy.

Gross AR, the allowance, and net receivables

When a company sells on credit, it records the full invoice as a debit to accounts receivable (gross AR). Not every customer will pay in full. GAAP requires management to estimate uncollectible amounts and record an offsetting credit balance: the allowance for doubtful accounts.

Net accounts receivable = Gross AR − Allowance for doubtful accounts

The balance sheet shows net AR (the amount management expects to collect). The allowance footnote discloses gross AR, the allowance balance, and often an aging schedule. Investors who read only the net line miss whether reserves are shrinking because credit improved or because management is optimistic.

Bad debt expense (or provision for credit losses) hits the income statement in the period management revises its estimate. It is usually classified in selling, general and administrative (SG&A) expense or as a separate line near operating income. Actual write-offs — removing specific customer balances from gross AR — debit the allowance, not bad debt expense directly.

The allowance roll-forward: how reserves move each quarter

Every 10-Q and 10-K should include a roll-forward table. The mechanics are consistent across companies:

Roll-forward line	Effect on allowance	Investor read
Beginning balance	Starting reserve	Compare to prior year same quarter
+ Bad debt expense (provision)	Increases allowance	Income statement hit; builds cushion
− Write-offs	Decreases allowance	Actual defaults; should track aging
+ Recoveries	Increases allowance	Cash collected after prior write-off
± Other (FX, acquisitions)	Varies	Read footnote — can hide releases
= Ending balance	Reserve on balance sheet	Divide by gross AR for ratio

A healthy pattern: bad debt expense roughly tracks write-offs plus growth in gross AR at stable credit quality. A warning pattern: write-offs spike while prior-period provision was low — the reserve was understated and earnings were inflated in earlier quarters.

Harbor Telecom Q4 2025 roll-forward (simplified, $M):

Beginning allowance: $4.2M (0.4% of $1,050M gross AR)
Bad debt expense: $2.1M (still low vs peers at 1.8%)
Write-offs: −$31.4M (three carrier bankruptcies)
Ending allowance: $3.8M — now clearly inadequate at 0.35% of gross AR

The $31M write-off exceeded the entire allowance because management had released reserves in prior quarters while DSO stretched from 44 to 61 days.

CECL (ASC 326) in plain language for equity investors

Since 2020, U.S. public companies follow Current Expected Credit Losses (CECL) for most financial assets including trade receivables. The old “incurred loss” model waited until loss was probable; CECL requires reserving for expected losses over the contractual life of receivables, including losses not yet identifiable on a specific invoice.

Practical implications for investors:

Forward-looking estimates — management must incorporate macro conditions, customer concentration, and historical loss rates. A recession should normally increase the allowance even before write-offs appear.
Pooling and aging — companies group receivables by shared risk (geography, product, customer tier) and apply loss rates per aging bucket (current, 30, 60, 90+ days). Read which pools exist and whether rates changed.
Day-one CECL bump — adoption created a one-time retained earnings hit for many banks; trade AR adopters had smaller effects but check transition disclosures for lingering comparability issues.
Releases are earnings — reducing the allowance without offsetting write-offs flows through as lower bad debt expense, boosting net income. Treat material releases skeptically.

CECL does not eliminate judgment. Two auditors can disagree on reasonable forward loss rates. Your job is to test whether the allowance moves logically with DSO, customer concentration, and macro stress — not to rebuild the model.

Key ratios and screens

Pair allowance metrics with collection speed from DSO analysis:

Allowance ratio = Allowance ÷ Gross AR

Track quarter over quarter and vs peers. A falling ratio while DSO rises is the classic Harbor Telecom pattern.

Bad debt expense % of revenue = Bad debt expense ÷ Revenue

Smooths seasonality. Industrial distributors often run 0.3–1.0%; telecom equipment with concentrated carriers can run 1.5–3.0% in normal years.

Net AR growth vs revenue growth

If gross AR grows 20% and the allowance is flat, net AR grows ~20% too — working capital absorbs cash. If allowance falls while gross AR rises, net AR growth understates the cash trap and overstates earnings quality.

Coverage of 90+ day aging

When companies disclose aging buckets, compare the allowance to the sum of 60-day-plus balances. Under-reserved past-due AR is a leading indicator of future write-offs.

Sector	Typical allowance % of gross AR	Notes
Consumer staples distribution	0.5–1.2%	Short terms; high volume
Industrial B2B	1.0–2.5%	Project billing adds tail risk
Enterprise software	0.8–2.0%	Contract assets may be separate pool
Telecom / carrier equipment	1.5–4.0%	Customer concentration matters
Healthcare services	2.0–5.0%	Payer denial complexity

Red flags linking allowances to earnings quality

Allowance ratio falls while DSO rises — reserves should normally move with collection risk, not against it.
Material reserve releases without disclosed credit improvement — check MD&A for customer-specific evidence, not boilerplate.
Write-offs consistently exceed prior provisions — lagging recognition of bad debt inflates past earnings.
Customer concentration hidden in AR footnote — one buyer past 10% of AR needs a higher pool-specific reserve.
Related-party receivables with zero allowance — slow-paying affiliates are a common quality issue.
Revenue acceleration into quarter-end with flat allowance — possible channel stuffing; see earnings quality screens.
Non-GAAP add-backs of bad debt expense — recurring credit losses are operating costs, not one-time adjustments.
Factored or sold receivables without recourse disclosure — can remove risky AR from the balance sheet entirely; read factoring footnotes.

Harbor Telecom refactor: rebuilding the reserve

After the Q4 2025 write-off, Harbor’s board mandated an independent credit review. Week 1–3: segmented gross AR into carrier, enterprise, and government pools; rebuilt historical loss curves by aging bucket from 2018 forward. Month 2: applied CECL-compliant forward overlays for rising carrier leverage ratios (net debt/EBITDA > 5× on three accounts). Month 3: restated Q1–Q3 2025 bad debt expense upward by $18M cumulative; allowance ratio restored to 2.3% of gross AR.

Outcomes: allowance ratio 0.4% → 2.3%, bad debt expense normalized at 1.9% of revenue, operating margin fell to a sustainable 11.8%, and auditors added a material weakness on credit estimation that was remediated in 2026. Stock fell 22% on the restatement — but subsequent quarters showed no further surprise write-offs. Investors who had tracked allowance-to-DSO divergence reduced exposure six months earlier.

Technique decision table

Metric / approach	Best for	Weak when
Allowance ÷ gross AR	Reserve adequacy vs peers; spotting releases	Business mix shifts (e.g., more government AR)
Bad debt expense ÷ revenue	Trending normalized credit cost	Lumpy large write-offs — use 4-quarter average
Allowance roll-forward	Separating provision from write-offs	Footnote omitted (rare for public filers)
DSO + allowance combined screen	Quick quarterly quality check	Terms genuinely renegotiated (verify in 10-K)
90+ day aging vs allowance	Near-term default risk	Aging not disclosed (many companies omit)
Cash conversion cycle	Cash impact of net AR changes	Service businesses with immaterial AR
Beneish M-score DSRI	Forensic screen when AR/revenue diverges	Not a standalone fraud proof

Common pitfalls

Using net AR in the allowance ratio — divide by gross AR or the ratio looks artificially high.
Ignoring contract assets (unbilled receivables) — ASC 606 SaaS companies often carry large contract asset balances with separate CECL pools.
Treating one-time write-offs as non-recurring forever — if underwriting was weak, future losses may match.
Comparing allowance ratios across industries — healthcare and telecom carry structurally higher reserves than grocery distribution.
Missing allowance on other receivables lines — employee advances, vendor rebates, and notes receivable may have separate reserves.
Forgetting currency translation — FX moves gross AR and allowance together but can distort quarter-over-quarter ratio changes.
Equating lower bad debt expense with better operations — it may be a reserve release, not improved collections.

Investor checklist

Locate gross AR, allowance, and net AR on the balance sheet each quarter.
Calculate allowance ÷ gross AR; plot vs trailing eight quarters.
Read the allowance roll-forward in credit-loss footnotes (ASC 326).
Compare bad debt expense to write-offs — flag provision shortfalls.
Pair allowance trend with DSO and revenue growth deltas.
Check customer concentration (>10% of AR) and related-party balances.
Review aging buckets if disclosed; test coverage of 60+ day balances.
Scan MD&A for reserve releases and CECL methodology changes.
Reconcile net AR change to operating cash flow working-capital line.
Benchmark allowance ratio against two direct peers in the same credit model.

Key takeaways

The allowance is a contra-asset estimate of uncollectible AR — net AR is what the balance sheet carries.
Bad debt expense builds the reserve; write-offs consume it — the roll-forward connects income statement to balance sheet.
CECL requires forward-looking loss estimates — falling reserves during rising DSO deserve skepticism.
Harbor Telecom’s $31M charge followed years of reserve releases — the warning signs were in the footnotes.
Always pair allowance ratios with DSO and cash flow — no single line item tells the whole credit story.