When Accruals Actually Matter

For stock research and screening, we use Fiscal.ai (use code “STABLEBREAD” to save 15%). It’s the most complete all-in-one research terminal we’ve found that’s accessible to the average retail investor.

For custom analysis and flexible valuation models, we use our Automated Stock Analysis Spreadsheet, powered by Wisesheets (the best spreadsheet add-on for stock data).

These are the two tools we use to analyze and value stocks more effectively and reach quicker, more informed investment decisions. Both have free versions if you want to try them out.

🌾 Welcome to StableBread’s Newsletter!

Accruals are the gap between what a company reports as earnings and what it actually collects in cash. Under accrual accounting, companies record revenue when earned and expenses when incurred, regardless of when cash changes hands.

If a company reports $100M in net income but only $60M in operating cash flow, that $40M difference is accruals. Those accruals might represent legitimate timing differences that resolve themselves, or they might signal aggressive accounting that inflates current earnings at the expense of future periods.

This is why you'd prefer companies with low accruals, since high accruals raise questions about earnings quality. Richard Sloan formalized this intuition in his 1996 research with a simple ratio:

Sloan found that buying low-accrual stocks and shorting high-accrual stocks produced abnormal returns of ~10% annually (based on U.S. stocks from 1962 to 1991). Lafond (2005) and Pincus (2007) later confirmed the pattern across 17 developed countries including Australia, Canada, and the U.K.

But accruals on their own only tell part of the story—which many investors seem to miss.

Hirshleifer, Hou, Teoh, and Zhang (2004) introduced a balance sheet measure called net operating assets (NOA) that captures a company's cumulative history with accruals rather than just a single year.

Philip Gray, Iris Siyu Liao, and Maria Strydom build on this in their 2017 study "The Profitability of Trading NOA and Accruals." Examining 3,054 Australian companies from 1991 to 2016, they found that high accruals don't always predict poor returns.

Below, we'll explain how NOA measures a company's cumulative track record with accruals, show how combining NOA with accruals produces a more effective screen, and discuss why context determines when high accruals actually matter.

Net operating assets (NOA) capture the cumulative difference between accounting income and cash flow across a company's entire history. Rather than looking at a single year's accruals, NOA reflects what has accumulated over time:

where:

Hirshleifer et al. (2004) introduced NOA as an investing signal and described it as "balance sheet bloat." The authors showed it predicted poor stock returns for at least three years after the financial statements were released.

The logic follows directly from what accruals represent. Each year, accruals get raised with an expectation that they'll eventually convert to cash. When they consistently don't, NOA grows. A company might book revenue aggressively, capitalize expenses it shouldn't, or fail to write down obsolete inventory. Each practice inflates NOA over time.

Gray and his coauthors wanted to understand whether NOA and accruals capture the same information or provide distinct signals, and more importantly, how the two interact when used together.

The researchers first verified that both anomalies exist in their Australian sample:

Sorting stocks into deciles by NOA produced a monthly spread of 1.33% between the lowest and highest groups (bottom 10% vs. top 10%) when stocks were value-weighted.

On an annualized basis, low-NOA stocks returned 18.18% while high-NOA stocks returned just 0.87%.

The gap of ~17% annually persisted after adjusting for the Fama-French three-factor model, where the alpha remained significant at 0.93% monthly.

Sorting by accruals produced similar results:

The monthly spread between low-accrual and high-accrual deciles was 1.12% for value-weighted portfolios. Low-accrual stocks returned 0.90% monthly while high-accrual stocks returned -0.23%. Annualized, this translated to ~14%.

When both variables were included in regression analysis, each maintained statistical significance. The coefficient on NOA was -0.107, and the coefficient on accruals was -0.114.

The key takeaway is that higher NOA and higher accruals each independently predict lower future returns.

The central finding comes from examining how the two effects interact. The researchers double-sorted stocks into 25 portfolios based on quintiles of both NOA and accruals. This allowed them to measure each effect while holding the other constant:

When controlling for accruals and allowing NOA to differ across groups, the NOA effect appeared in all five accrual quintiles. Low-NOA stocks outperformed high-NOA stocks regardless of their current accrual level. Every spread was statistically significant.

But when controlling for NOA and letting accruals vary, a different picture emerged. The accrual effect only showed up in the highest NOA quintile.

For stocks in that top group, low-accrual stocks outperformed high-accrual stocks by 1.18% monthly. For stocks in the other four NOA quintiles, the accrual spread was not statistically significant.

The authors summarize:

When a low-NOA company reports high accruals in a given year, the market apparently views this as a one-time event. The company's track record suggests the accruals will eventually convert to cash. Future returns bear this out.

When a high-NOA company reports high accruals, it reinforces an existing problem. The new accruals add to a balance sheet already showing poor conversion. These are the stocks that subsequently underperform.

Therefore, a one-year spike in accruals means something different for a company with a clean track record than for a company that has consistently failed to convert accounting income into cash.

Gray and his coauthors found that buying stocks with low NOA and low accruals while shorting stocks with high NOA and high accruals produced monthly returns of 2.08%. This combined approach generated positive annual returns in 21 of the 24 years studied.

Stocks with both low NOA and low accruals returned 1.79% monthly, while those with both high NOA and high accruals returned -0.28%. The difference of ~25% annualized compounds significantly over time.

Overall, if you screen stocks using accruals, consider adding NOA as a filter. A company showing high accruals but maintaining low NOA isn't necessarily a red flag. The track record with converting accruals to cash matters more than any single year's figure.