Animal spirits
The waves of confidence and fear that drive economic activity beyond cold calculation.
What it means
Keynes's term for the spontaneous urge to action and the swings of optimism and pessimism that move investment and spending more than rational calculation of expected returns. Much economic decision-making, Keynes argued, rests on this instinct rather than on a weighted average of quantitative probabilities, which are often unknowable. Modern behavioral macroeconomics revived the phrase to bundle confidence, fairness, corruption, money illusion, and storytelling as non-rational drivers of booms and busts. It matters because it locates the source of business cycles partly in collective psychology, justifying a role for policy in stabilizing confidence.
Why the calculation runs out
Keynes's argument is narrower and stranger than the claim that people are irrational: mood is doing work that calculation cannot do, rather than work it does badly. The returns on a factory, a railway or a mine unfold over decades, and an agent who insisted on calculating them would never move at all. So the urge to act is not a defect bolted onto rational choice; it is what makes long-horizon investment happen in the first place. Keynes's point is that enterprise fades when spontaneous optimism falters, and no amount of arithmetic substitutes for it. That distinction is what most later uses of the phrase lose.
How mood becomes a cycle
One person's optimism is a mood; an economy's optimism is a coordination problem. Spending is someone else's income, so a belief about what others will do can make itself true: firms expecting weak demand cut investment, which weakens demand. That feedback is why economists formalize animal spirits as multiple equilibria, or as sunspots: extrinsic shocks that shift expectations without touching preferences or technology. Angeletos and La'O show sentiment shocks survive even where the equilibrium is unique, with no indeterminacy for a sunspot to select between, once random decentralized trading limits how well anyone can infer what everyone else believes, which matters because it means you do not need irrationality to get sentiment-driven cycles. Shiller supplies a transmission channel: contagious stories spread like an epidemic and hand scattered strangers a shared reason to act at the same moment.
What the evidence shows
Confidence indices track the cycle, and confidence innovations do predict output and consumption. Whether that predictive power is animal spirits is contested, and the answer turns entirely on identification. Barsky and Sims decompose confidence innovations into news, spirits and noise, and find the link between confidence and later activity runs almost wholly through the news component: consumers are picking up real information about future productivity rather than creating the future with their mood. Choi and coauthors repeat the exercise using Great Recession and zero-lower-bound data, precisely where spirits should have the most room to operate, and the conclusion holds. Note what this does and does not settle: it is a verdict on consumer surveys, not on the long-horizon investment decisions Keynes was actually describing, which no comparable decomposition has been run on. Angeletos, Collard and Dellas reach a friendlier verdict on a different object — not the survey index, but a structural shock to higher-order beliefs inside an estimated model, which they find carries a sizable share of cyclical volatility. The two results can both hold.
Limits and what to do with it
The concept's weakness is that it absorbs anything. Whatever a model fails to explain gets labelled animal spirits, which makes the phrase a residual rather than a hypothesis, and the broad modern revival, which annexes everything from fairness to storytelling, widens the net further. Nothing that explains every boom explains any particular one. Treat it as a question rather than an answer: when confidence moves, ask what the crowd might know before concluding it is only mood, because on the one measure that has been carefully identified — innovations in consumer confidence surveys — the informative reading has twice won. The practical implication runs the other way too. A confidence index is not an independent forecast to stack on top of your own, since most of what it carries is news you can find elsewhere.
Examples
A surge of business optimism fuels an investment boom that outruns any sober forecast of profits, then reverses when the mood sours.
Buyers queue for flats they have barely seen while prices climb; when the mood turns, the same flats sit unsold at the same asking prices, with nothing else changed.
Households whose incomes have not moved cancel holidays and delay a new car because the headlines feel grim, and enough of them doing it makes the downturn they feared.
Loan officers tighten credit standards on borrowers whose balance sheets have not changed, spooked by the collapse of a lender in an unrelated market that carried no exposure to these loan books and no information about them; the resulting credit drought produces the defaults they were bracing for.
A mining company sanctions a decade-long project on a wave of price optimism, when the price that will clear in year twelve is not something anyone can put a distribution over; the mood, not the arithmetic, is what moves the first shovel.
First described in Keynes (1936); Akerlof & Shiller (2009).
Key references
- Choi, S., Jeong, J., Park, D., & Yoo, D. (2024). News or animal spirits? Consumer confidence and economic activity: Redux. Journal of Applied Econometrics, 39(5), 960-966. doi.org/10.1002/jae.3070
- Angeletos, G.-M., Collard, F., & Dellas, H. (2018). Quantifying confidence. Econometrica, 86(5), 1689-1726. doi.org/10.3982/ECTA13079
- Shiller, R. J. (2017). Narrative economics. American Economic Review, 107(4), 967-1004. doi.org/10.1257/aer.107.4.967
- Angeletos, G.-M., & La'O, J. (2013). Sentiments. Econometrica, 81(2), 739-779. doi.org/10.3982/ECTA10008
- Barsky, R. B., & Sims, E. R. (2012). Information, animal spirits, and the meaning of innovations in consumer confidence. American Economic Review, 102(4), 1343-1377. doi.org/10.1257/aer.102.4.1343