This is the whole course in one sitting. Hold the through-line in your head: whenever your type is hidden behind a verification wall — good worker or bad, healthy peacock or sickly one, reliable seller or fly-by-night — talk is cheap because words cost every type the same, so they prove nothing. The only thing that separates the good type from the bad is an action whose cost differs by type: a costly signal. Honesty doesn’t live in the message; it lives entirely in the cost gap. That single move — do something too expensive for the wrong type to fake — is the peacock’s tail, the gazelle’s stott, the hard degree, the long warranty, the burned bridge, the Veblen watch, the initiation rite. And when the cost gap narrows, mimics flood in and the whole separating equilibrium erodes. Take a breath. There is no going back once you commit.
How this exam works
This is a real exam, not a practice quiz. Questions appear one at a time. Once you submit an answer it is locked for good — there is no going back, no retry, and no restart. Your score stays hidden until the very end. A few questions ask you to select all that apply (read those carefully — partial credit is not a thing here). You need 70% to pass. Ready when you are.
Why does CHEAP TALK — a costless message like "trust me, I'm a great worker" — carry no information in a setting where interests conflict?
Select an answer to continue.
Big picture
Signalling & costly signals, in one picture
- Signalling & costly signals
- Cheap talk vs costly signals
- Behind a verification wall your type is hidden - cheap talk costs every type the same so it proves nothing; a costly signal is an action whose cost DIFFERS by type. Spence single-crossing: a signal separates good from bad EXACTLY when it is cheaper for the good type to send. Honesty lives entirely in the cost gap
- The handicap principle
- Zahavi: a biological signal is honest BECAUSE it is costly - only high quality can afford the handicap (peacock's tail, gazelle stotting). Index (impossible to fake) vs handicap (too costly to fake); shaped by natural selection at the INDIVIDUAL level, not for the good of the species
- Markets
- Spence job-market signalling - a degree sorts by ability even teaching nothing, because studying is cheaper for high types (sheepskin effect). Warranties, brands, skin-in-the-game and costly ads resolve the market for lemons. Screening (uninformed builds a test/menu) vs signalling (informed pays to reveal); credential inflation is a signalling arms race
- Strategy & society
- Deterrence via sinking costs (spend up front to prove resolve) vs tying hands (remove your own options - burn the bridge so backing down is costly). Conspicuous consumption / Veblen goods, virtue signalling, and costly initiation rites that bind groups - all the same move: an action too costly for the wrong type
- Equilibria
- Pooling (all send the same / none, receiver falls back on the prior), separating (types split, needs the cost gap), semi-separating (probabilistic, partial reveal). Countersignalling: the very top skip the signal because they have nothing to prove - the billionaire in a hoodie
- Where the model lies
- When the cost gap narrows, dishonest mimicry floods in (diploma mills, counterfeits, Batesian mimics) and the separating equilibrium erodes. Positional arms races burn deadweight waste; revelation vs causation is a genuine confound (reveal or build ability?); don't over-read noisy one-shot signals; and cheap talk CAN work when interests are aligned - the theory is about adversarial settings
- Cheap talk vs costly signals
Key takeaways
You now hold the whole model. Behind a verification wall, your type is private and cheap talk proves nothing — words cost every type the same, so the bad type says them just as freely as the good. Information demands an action whose cost differs by type: a costly signal. The Spence single-crossing / separating condition is the engine of everything here — a signal separates the good type from the bad exactly when it is cheaper for the good type to send — so honesty lives entirely in the cost gap. In biology this is Zahavi’s handicap principle: the peacock’s tail and the gazelle’s stott are honest because they are costly in a currency only high quality can spare, an index (impossible to fake) or a handicap (too costly to fake), selected at the level of the individual, not “the good of the species.” In markets it is Spence’s job-market signalling (a degree sorts by ability even if it teaches nothing, hence the sheepskin effect), and the family of lemons-fixes — warranties, brands, skin-in-the-game, costly advertising — plus the screening (uninformed builds the test) vs signalling (informed pays to reveal) distinction and the credential-inflation arms race. In strategy and society it is deterrence by sinking costs (pay up front) or tying hands (burn the bridge so backing down hurts), conspicuous consumption / Veblen goods, virtue signalling, and costly initiation rites — all one move: an action too costly for the wrong type. The equilibria come in three flavours — pooling (all the same, receiver keeps the prior), separating (types split, needs the cost gap), semi-separating (probabilistic, partial reveal) — with countersignalling as the twist where the secure top skip the signal (the billionaire in a hoodie). And the model lies at its edges: when the cost gap narrows, dishonest mimicry (diploma mills, counterfeits, Batesian mimics) erodes the separating equilibrium toward pooling; positional races burn deadweight waste; revelation vs causation is a genuine confound (does the degree reveal or build ability?); over-reading a noisy one-shot signal over-fits to luck; and cheap talk works fine when interests are aligned — the whole apparatus is for adversarial, conflicting-interest settings. Never ask “what did they say?” Ask: what did it cost them to say it, and could the wrong type have afforded the same?