# The End of Accounting **Baruch Lev and Feng Gu**  --- _When you see high ROE, ask whether the denominator is artificially low before you admire the numerator._ In the 1950s, earnings explained around 90% of stock price movements. Today that figure is roughly 50%. This isn't gradual drift. It's a structural collapse in the relevance of financial reporting, and it happened because the economy changed while the accounting rules didn't. Balance sheets were designed when factories and inventory were the assets that mattered. Those assets are commodities, available to anyone with capital. The things that actually create competitive advantage, patents, brands, customer relationships, proprietary knowledge, don't appear on the balance sheet at all. Accounting measures what's easy to count and misses what counts. --- **The distortion runs deeper than omission.** If you build a brand internally, the way Coca-Cola did, GAAP treats the investment as an expense. It hits the income statement and disappears. If you buy that same brand through an acquisition, it appears on your balance sheet as an asset. The accounting rules therefore incentivise buying capabilities over building them, which is backwards. But the real damage is to the [[Ratios]] you rely on for analysis. When internally developed intangibles are expensed rather than capitalised, the asset base is understated. ROE and ROA look better than they should, because the denominator is artificially small. A company that has spent decades investing in R&D, brand, and talent shows a lean balance sheet and flattering profitability ratios, while a company that acquired identical capabilities shows a bloated one. Same economic reality, different accounting picture. The next time you admire a business with exceptional returns on equity, it's worth asking how much of that return reflects genuine operating performance and how much reflects an understated denominator. --- **Three forces explain why earnings have become unreliable as a signal.** One-time items, restructuring charges, impairments, write-offs, have proliferated to the point where "recurring" earnings are largely fictional. Accrual manipulation through depreciation schedules, bad debt provisions, and option expensing lets management shape reported results. And genuinely important events, clinical trial outcomes, the loss of key talent, strategic partnerships signed or cancelled, don't show up in financial statements at all because they aren't transactions with third parties. Lev and Gu keep returning to one empirical finding: forecasting cash flows is simpler, more accurate, and yields higher investment returns than forecasting earnings. Cash avoids accrual distortions and can't be manipulated the same way. It either arrived or it didn't. The divergence between reported earnings and actual cash generation is precisely the gap that [[Cash and profit]] maps, and it explains why serious investors read the cash flow statement before the income statement. --- Since the 1980s, corporate investment in intangible capital has surpassed investment in physical assets, and the gap keeps widening. Entire industries, software, biotech, internet services, are essentially intangible. Yet accounting hasn't changed in over a century. The result is a reporting system that meticulously tracks office buildings and production machinery while remaining blind to the resources that actually drive value. True competitive advantage comes from assets that are valuable, rare, and difficult to imitate. GAAP can't see any of them. Regulators respond by adding disclosure complexity, more footnotes, more rules, without restoring relevance. The quarterly earnings call has become theatre. The signal is elsewhere. ---