Framework tested across multiple Australian reporting cycles
The Signal Framework Built to Read What Earnings Reports Obscure
I designed this methodology because I kept encountering the same analytical gap — investors
trusting reported earnings figures without examining the accounting mechanics that produced
them, and missing the signals that were sitting in plain view inside the filings the whole
time.
-
Active framework +6 sectors Core Signal Types Tracked14+ASX Sectors 9 coveredFiling Cycles 6+ years
-
-
Signal Coverage Multi-sector -
The Five Signal Pillars That Form the Core of Our Earnings Quality Framework
Each pillar addresses a distinct dimension of earnings quality — not as an isolated ratio
but as a systematic lens through which a different layer of the relationship between
reported income and underlying financial reality becomes visible. Together they form an
integrated analytical framework rather than a checklist of disconnected metrics.
Pillar One: Accruals Quality and the Sloan Accruals Ratio
Accruals quality is the foundational signal pillar — the one from which almost everything else in the framework flows. It measures the proportion of reported earnings attributable to accruals rather than operating cash flow, using the balance sheet accruals construction method documented by Sloan and subsequently refined in the academic literature. A rising accruals ratio over successive reporting periods is the single most historically consistent signal of subsequent earnings disappointment available from publicly filed financial data. The construction logic is straightforward: changes in net operating assets, scaled by average total assets, produce the accruals ratio. The interpretation is more nuanced — it requires distinguishing between accruals accumulation driven by genuine business growth and accumulation driven by aggressive accounting elections, a distinction that requires reading the notes to the accounts rather than relying on the headline ratio alone.
Pillar Two: Earnings Persistence and the Cash Flow Coverage Signal
Earnings persistence measures the degree to which current-period reported earnings predict future-period earnings — a property that varies substantially depending on whether income is backed by operating cash flow or driven by accruals. The cash flow coverage signal, constructed as the ratio of operating cash flow to reported net income, provides the most direct measure of this relationship for a given reporting period. A ratio consistently below one across multiple consecutive periods is one of the clearest signals in the framework that reported earnings contain a significant accruals component unlikely to be sustained. This pillar also incorporates free cash flow adequacy assessment — examining whether the business generates sufficient cash after maintenance capital expenditure to support its reported earnings level without relying on working capital draws or financing activity that the income statement does not directly reflect.
Pillar Three: Earnings Smoothing Detection and Management Pattern Analysis
Earnings smoothing — the practice of managing reported income to reduce its apparent volatility across reporting periods — is one of the most analytically significant forms of earnings quality deterioration because it is specifically designed to be invisible to readers who are not looking for it. This pillar uses a combination of variance analysis, discretionary accruals estimation using the modified Jones model, and multi-period earnings pattern review to identify companies where reported income is unnaturally smooth relative to the cash flow volatility that the underlying business would be expected to produce. The signal is not that smooth earnings are automatically suspicious — some business models genuinely produce stable income — but that smoothness achieved through accounting elections rather than operational consistency carries very different implications for earnings sustainability and forward reliability.
Pillar Four: Cash Conversion Quality and Working Capital Signal Analysis
Cash conversion quality examines the relationship between reported revenue growth and the cash collection patterns that should accompany it — specifically, whether increases in accounts receivable and deferred revenue are tracking revenue growth proportionally or accumulating at rates that suggest revenue recognition is running ahead of genuine economic delivery. Working capital signal analysis extends this to the full operating cycle: inventory build relative to cost of sales trends, accounts payable behaviour relative to purchasing activity, and the aggregate working capital accruals ratio that captures the combined effect of these movements. This pillar is where revenue quality signals first become visible in the balance sheet — the accounting mechanics of premature revenue recognition or aggressive working capital management leave traces in the filing data that a systematic framework identifies before they surface in reported earnings revisions.
Pillar Five: Footnote and Disclosure Quality Analysis
The fifth pillar addresses what is perhaps the most underappreciated dimension of earnings quality analysis: the information content of what a company chooses to disclose, how it discloses it, and — most importantly — what it does not disclose that a comparably situated business in the same sector routinely would. Footnote quality analysis examines the specificity of accounting policy descriptions, the consistency of segment reporting across periods, the completeness of related party disclosures, the transparency of off-balance-sheet arrangements, and the degree to which management commentary is quantitatively supported rather than qualitatively asserted. Changes in disclosure practice between reporting periods — particularly reductions in specificity or the reclassification of previously disclosed items — are treated as significant signals in their own right, independent of the quantitative metrics produced by the other four pillars.
Integration Layer: Composite Signal Scoring and Peer Benchmarking
The five pillars produce individual signal readings that are analytically meaningful in isolation but most powerful when integrated into a composite signal score that reflects the overall earnings quality picture across all dimensions simultaneously. The integration layer constructs a composite score for each company in the coverage universe, weights each pillar's contribution according to its historical signal strength in the relevant sector and reporting environment, and contextualises the result against the cross-sectional distribution of scores for comparable businesses. A company that scores poorly on multiple pillars simultaneously — elevated accruals, weak cash flow coverage, smoothing indicators, deteriorating cash conversion, and reduced disclosure specificity — presents a substantially different analytical picture than one where a single signal is elevated while all others remain within normal historical ranges for its sector peers.
Signal Framework Versus Conventional EPS Analysis
Most investors stop at earnings per share and call the analytical work done. Here is what that approach systematically misses and what a signal framework finds instead.
Pheokereaqaes
EPS-Only Analysis
Detecting Accruals Accumulation Before Earnings Disappoint
Reporting period
Pheokereaqaes
Systematic accruals ratio construction identifies accumulation patterns in the balance sheet data well before they surface in consensus earnings revisions or analyst commentary.
Early detection
•
Current filing
Прогресс
88%
EPS-Only Analysis
Headline EPS accepted as given — accruals component not examined, accumulation pattern invisible until earnings disappointment is already reported.
Missed signal
•
After revision
Прогресс
22%
Identifying Cash Flow Divergence From Reported Income
Multi-period assessment
Pheokereaqaes
Cash flow coverage signal tracks the relationship between operating cash generation and reported net income across successive cycles, flagging persistent divergence as a quality concern.
Systematic tracking
•
Ongoing monitoring
Прогресс
91%
EPS-Only Analysis
Operating cash flow reviewed in isolation if at all — multi-period divergence pattern from reported earnings not systematically constructed or interpreted as a signal.
Not assessed
•
Rarely checked
Прогресс
18%
Footnote Disclosure Change Detection
Period-over-period review
Pheokereaqaes
Structured disclosure quality analysis identifies reductions in footnote specificity, reclassifications of previously disclosed items, and changes in segment reporting that carry independent signal value.
Actively monitored
•
Each filing cycle
Прогресс
85%
EPS-Only Analysis
Footnote changes not systematically tracked — disclosure quality treated as a compliance matter rather than an independent source of analytical signal about management intent and financial health.
Overlooked
•
Not tracked
Прогресс
12%
Pheokereaqaes
88%
EPS-Only Analysis
17%
The analytical gap between a systematic earnings quality signal framework and conventional EPS-led analysis is widest precisely when it matters most — when a company's reported earnings are diverging from underlying financial reality in ways the headline number is specifically designed to obscure.
Signal Construction Process
From Raw Filing Data to Actionable Earnings Quality Signal Score
Step 1
Primary Source Data Ingestion and Filing Verification
Every signal calculation begins with the primary source: the actual regulatory filing, read in full. Income statement, cash flow statement, balance sheet, notes to accounts, auditor report, and related party disclosures are all reviewed before any metric construction begins. Data is sourced from the filed document rather than pre-calculated vendor databases, because the construction assumptions that matter for signal quality are embedded in the accounting detail that aggregated data sources routinely strip out.
The filing review process follows a consistent analytical sequence developed over multiple years of signal research practice. We begin with the auditor report and related party disclosures — the sections most analysts read last, if at all — because changes in audit language, scope limitations, or related party transaction volume are independent signal sources that inform the interpretation of every quantitative metric constructed subsequently. The income statement is then reviewed with specific attention to revenue recognition note disclosures, non-recurring item classifications, and segment reporting consistency relative to the prior period. The cash flow statement is reviewed before the balance sheet to establish the cash generation baseline against which accruals will be measured. All data points used in signal construction are recorded with their specific filing location — note reference, page number, and line item — so that every metric can be independently traced and verified by any analyst who wishes to check our work.
Step 2
Signal Metric Construction from Accounting First Principles
Each earnings quality metric is constructed directly from the filing data using the formulas documented in our published methodology — never sourced from pre-calculated vendor databases where construction assumptions are opaque. The accruals ratio, cash flow coverage signal, discretionary accruals estimate, cash conversion quality measures, and disclosure quality assessment are all built from the primary accounting data in the sequence described in the methodology documentation.
The accruals ratio is constructed using the balance sheet method: the change in net
operating assets between the current and prior period balance sheets, scaled by
average total assets across both periods. Net operating assets are defined as total
assets less cash and short-term investments, less total liabilities excluding
interest-bearing debt — a construction that captures the operating accruals
component without contaminating it with financing activity. The cash flow coverage
ratio divides operating cash flow from the cash flow statement by net income from
the income statement, with adjustments for non-cash items that are explicitly
disclosed in the notes and that represent genuine economic events rather than
accounting elections. Discretionary accruals are estimated using the modified Jones
model, with sector-specific parameter estimation using peer group data to improve
the accuracy of the discretionary versus non-discretionary accruals decomposition.
All construction steps are documented in the methodology with the specific formula,
the accounting standard reference where relevant, and the known limitations of each
approach.
Step 3
Multi-Period Trend Assessment and Pattern Identification
Individual period signal readings are interpreted in the context of the prior
reporting cycles for the same company — typically a minimum of three consecutive
periods and ideally five or more. This multi-period view transforms a
single-period observation into an analytically meaningful trend assessment,
distinguishing between temporary fluctuations consistent with normal business
cycle variation and systematic deterioration patterns that carry genuine
predictive weight about future earnings sustainability.
The trend assessment examines the direction, magnitude, and consistency of signal
movement across successive reporting periods rather than the absolute level of any
single metric. A company whose accruals ratio has moved from a historically normal
level to an elevated level over three consecutive reporting periods presents a very
different analytical picture from one whose ratio has been elevated for a single
period following an acquisition or capital expenditure cycle that the notes to the
accounts explain clearly. The cash flow coverage trend analysis specifically tracks
whether the gap between reported earnings and operating cash flow is widening,
stable, or narrowing — because a widening gap sustained over multiple periods is one
of the most historically reliable early indicators of future earnings disappointment
in the academic literature and in our own research across Australian company
filings. Multi-period trend data is presented in the signal summary as a directional
indicator alongside the point-in-time metric reading, so that the analytical
conclusion reflects both the current level and the trajectory.
Step 4
Peer Benchmarking and Cross-Sectional Signal Contextualisation
Every individual company signal reading is contextualised against the distribution of the same metric across a defined peer group of comparable businesses operating in the same ASX sector and under similar accounting environments. This cross-sectional comparison is what transforms a signal observation from potentially interesting to analytically meaningful — because the significance of any given accruals ratio or cash flow coverage reading depends fundamentally on whether it is unusual relative to comparable businesses.
Peer groups are constructed using a combination of GICS sector classification,
business model comparability assessment, and size-adjusted grouping to ensure that
the cross-sectional comparison reflects genuinely comparable accounting environments
rather than superficially similar sector labels. For each signal metric, we
calculate the sector distribution — median, interquartile range, and tail thresholds
— using the current reporting cycle data for all companies in the peer group with
sufficient filing history to produce reliable metric readings. Individual company
signal readings are then expressed relative to this distribution: a company in the
top quartile of accruals accumulation relative to its sector peers receives a
materially different signal interpretation than one at the sector median, even if
the absolute ratio level is similar in numerical terms. The peer benchmarking output
is documented in the signal summary alongside the absolute metric reading and the
multi-period trend assessment, giving the composite signal score the
three-dimensional analytical context that makes it more reliable than any single
metric examined in isolation.
Integrating Earnings Quality Signals Into Your Existing Analytical Workflow
You do not need to rebuild your research process from scratch to benefit from systematic signal analysis. Here is how I have seen analysts add signal rigour to what they already do, without creating parallel workloads that are unsustainable in practice.
Start With the Cash Flow Statement Before the Income Statement
The single most effective habit change I recommend to analysts encountering the signal framework for the first time is to read the cash flow statement before the income statement when opening a new filing. This simple sequencing shift changes the analytical frame entirely — instead of reading the cash flows as a check on the income statement you have already accepted, you are using the cash flows as the baseline against which you then interrogate the income statement. The difference in what you notice is immediate and often striking, particularly for companies where the cash flow story is quietly diverging from the earnings narrative.
Build a Three-Period Accruals Trend Before Forming Any Quality Judgment
A single-period accruals ratio is a data point. Three consecutive periods of the same metric is a pattern — and patterns carry analytical meaning that isolated readings cannot. Before forming any conclusion about a company's earnings quality, I calculate the accruals ratio for the current period and the two prior periods, note the direction of movement, and ask whether the trend is consistent with the business model changes described in the notes or whether it is accumulating without an obvious operational explanation. This adds perhaps thirty minutes to a standard filing review and changes the quality of the analytical conclusion substantially.
Cross-Reference Signal Readings Against Sector Peers Systematically
The most common mistake I see in informal earnings quality assessment is treating a company's metrics in absolute isolation — concluding that an accruals ratio of a given level is high or low without reference to what comparable businesses in the same sector are producing. Before interpreting any individual signal reading, build a simple peer comparison using the same metric for three to five sector comparables from the most recent reporting period. The cross-sectional context often changes the interpretation entirely, and the companies that stand out as outliers in a peer comparison are frequently the ones where closer forensic scrutiny of the filing reveals the most analytically significant findings.
Read the Notes to the Accounts Before Calculating Any Metric
Every quantitative earnings quality metric is only as reliable as the analyst's understanding of the accounting policies that produced the underlying numbers. Before constructing an accruals ratio or a cash flow coverage signal, I read the revenue recognition note, the provisions and estimates note, and the segment reporting note in full. Changes in accounting policy elections, new estimates, or reclassifications between periods affect the comparability of the metrics and can produce signal readings that look alarming but reflect policy changes rather than genuine earnings quality deterioration — a distinction that is only visible if you have read the notes carefully before you started calculating.
Treat Disclosure Reductions as Independent Warning Signals
When a company provides less specific disclosure in a current filing than it did in the prior period — fewer line items in a segment breakdown, a shortened related party transactions note, a more generic accounting policy description — treat that reduction as an independent signal worth noting before you have examined any quantitative metric. In my experience reviewing Australian company filings systematically, reductions in disclosure specificity are not random. They tend to cluster around periods where the quantitative signals are also deteriorating, and they occasionally precede quantitative deterioration by a reporting cycle — making them one of the earliest warnings a systematic reader can identify.
Update Your Signal Readings Every Reporting Cycle Without Exception
The most analytically dangerous version of earnings quality assessment is one that was done thoroughly once and then not revisited. Earnings quality is not a static property of a business — it changes with management decisions, business cycle position, accounting standard updates, and operational pressures that may not be visible from the outside until they show up in the filing data. Every reporting cycle, re-run the full signal framework for every company in your coverage universe, compare the current readings to the prior period trend, and note any metrics that have moved materially. The companies that warrant closest attention are almost always the ones where multiple signals have moved in the same direction simultaneously.
Methodology Questions
Honest answers to the questions analysts ask most often about our signal framework.
Each signal in our framework is validated against peer-reviewed academic finance research demonstrating a historical relationship between the metric and subsequent earnings outcomes or financial distress events. We additionally test each signal against our own database of Australian company filings before including it in the published methodology, documenting both the conditions under which the signal has shown predictive strength and the conditions under which it has not. We publish the limitations alongside the findings.
The methodology undergoes a structured annual review cycle, with interim updates triggered by material changes to Australian accounting standards, significant new peer-reviewed research that is directly relevant to our signal construction, or our own research findings that reveal a meaningful limitation in an existing metric's construction. Every update is documented with the date, the specific change made, and the analytical rationale — the methodology changelog is published on this page.
Conflicting signals are common and analytically informative rather than a problem to be resolved by discarding the inconvenient reading. When accruals quality is deteriorating while cash flow coverage remains strong, for example, the conflict itself suggests a specific analytical hypothesis worth examining in the filing notes. Our methodology documentation includes a section on common signal conflict patterns, their most frequent accounting explanations, and the specific disclosures that typically resolve the ambiguity.
The framework was developed and validated specifically against Australian company filings under Australian GAAP and AASB reporting standards. The underlying academic foundations — the Sloan accruals model, the Beneish M-Score structure, the modified Jones model for discretionary accruals — are internationally derived and have been tested across multiple jurisdictions. However, the peer benchmarking thresholds and the disclosure quality assessment criteria are calibrated for Australian reporting conventions and should be recalibrated before applying the framework to companies filing under IFRS or US GAAP in different jurisdictions.
Accounting standard changes are the primary source of methodology recalibration requirements, and we treat them seriously. When a new standard changes the accounting mechanics underlying a signal — as the lease capitalisation changes under AASB 16 did for balance sheet accruals measures — we document the effect on signal construction, adjust the metric formulas where necessary, and note any periods where historical comparability is impaired. We do not simply continue using pre-change formulas and pretend the standard change did not affect the signal's meaning.
Financial sector companies — banks, insurers, and financial holding entities — have accounting structures that differ sufficiently from non-financial businesses that several of our standard signal constructions require material modification before they produce analytically meaningful readings. The accruals ratio construction, in particular, is not directly applicable to financial institutions without adjustments to the definition of operating assets that reflect the different balance sheet structure of a lending or insurance business. We document these sector-specific modifications in the methodology, and we recommend caution in applying standard framework outputs to financial sector companies without first reviewing those modifications carefully.