Imagine a thermostat that, when the house gets too cold, doesn't just fail to heat — it also starts leaking gas. The leak makes the furnace less efficient, which makes the house colder, which makes the leak worse. That's the core mechanism here: when real wages fall behind productivity growth, opioid use disorder rises, which damages productivity, which changes the economic cycle itself. The thermostat doesn't just fail; it poisons the system it was supposed to regulate. The committed claim: embedding a discrete-choice OUD channel into Goodwin's predator-prey growth model produces a three-dimensional nonlinear system that admits a limit cycle via Andronov-Hopf bifurcation, and the resulting dynamics reproduce a set of paradoxical empirical patterns — including the counterintuitive finding that rising OUD temporarily raises the employment rate (because damaged productivity demands more labor per unit of output), even though in steady state higher OUD prevalence lowers labor force participation. Goodwin's 1967 model is a workhorse of heterodox macroeconomics: wages and employment chase each other in a Lotka-Volterra cycle. Chung and Davila Fernandez add a third state variable — OUD prevalence — connected to the wage share via a logistic discrete-choice function. When your wage share drops, the probability of OUD ticks up. OUD feeds back into the economy through a productivity damage function. The mathematical heavy lifting is showing this three-dimensional system still cycles rather than collapsing to a fixed point, which they establish analytically via the Andronov-Hopf theorem and confirm numerically. The empirical grounding uses U.S. state-level panel data from 1998 to 2019, documenting that deteriorations in the labor share are followed by rising drug-induced mortality. This is the motivating fact, not a full estimation of the structural model. The model is calibrated and simulated, not estimated econometrically against the full time series — an important distinction. The numerical experiments then explore comparative dynamics: higher OUD sensitivity to income widens fluctuations, while higher productivity sensitivity to OUD counterintuitively compresses them through a self-correcting feedback that speeds wage-share recovery. The ladder here is specific. The paper sits against a long line of Goodwin-type extensions — adding financial variables (Keen, Grasselli), open-economy dynamics, or institutional features. The novelty is the behavioral health channel. No prior Goodwin extension has incorporated substance use disorder as a productivity-damaging feedback loop with discrete-choice microfoundations. Within that niche, this is genuinely new. Against the broader macro-health literature (Case and Deaton's "deaths of despair," Krueger's opioid-labor-force paper), this is the first formalization as a limit-cycle model rather than a reduced-form regression or narrative. The policy punchline is clean: anything that weakens the link between income loss and OUD onset — stronger social safety nets, better addiction treatment access — stabilizes the cycle. This is intuitive but now has a formal dynamic structure showing why: it dampens the feedback loop rather than fighting its symptoms. The converse is also clear: policies that let real wages decouple from productivity growth don't just redistribute — they seed a health crisis that damages the productive capacity they were supposed to exploit. The main limitation is validation depth. The empirical section establishes the motivating correlation (wage share → drug mortality) but does not estimate the structural parameters from data or test the model's cycle predictions against observed U.S. business cycles. The limit cycle is proven to exist analytically and shown numerically, but whether the real U.S. economy sits in the parameter region that produces it is left as an open question. The obvious next step — structural estimation on state-level panels — is acknowledged but not attempted, likely because the data demands (clean OUD prevalence series, not just mortality) are severe.
