Fast-growing firms tend to earn lower future returns — the asset-growth anomaly — a fact that has always sat awkwardly beside the observation that good, profitable firms are the ones that invest and expand. We ask not how much a firm grew, but how it paid for that growth. An exact balance-sheet identity partitions the standard asset-growth rate of Cooper, Gulen and Schill (2008) into two additive components: internal growth, the change in retained earnings scaled by beginning assets, and external growth, the residual raised from new debt and new equity. Because the partition is additive, the canonical one-slope specification becomes a testable restriction nested inside a two-slope model.

The sample is large: 495,842 firm-year observations for 42,057 firms across twenty-four markets, 1986–2024, drawn from S&P Compustat Global and North America, with delistings retained and a 120-day information delay. Three layers of results follow. In the pooled contemporaneous regression, both slopes are positive but internal growth carries roughly five times the reward (0.534 versus 0.106). In the predictive regression the signs diverge: the internal slope is positive (0.095, t = 3.77), the external slope negative (−0.055, t = −7.41), with a contrast of 0.150 (t = 5.88).

An aggregation argument explains why the literature missed this. Because the two components are negatively correlated (−0.47) and external growth dominates the variance of the total, the projection weights of the total-growth regression are approximately −0.01 on the internal part and 1.01 on the external — the scalar asset-growth measure is, in this panel, nearly an exact proxy for the external component and almost blind to the internal one. Opening the external residual, debt, contributed capital and other equity carry the negative prediction, and net share issuance from the cash-flow statement confirms the equity channel. The contemporaneous contrast is positive in every one of the twenty-three markets where the full model can be estimated and significant at 5% in twenty. The anomaly was never about growing too much: it is about who pays for the growth. Self-funded growth is a mark of quality; growth bought with outside capital is the part the market marks down.