Why Crypto Traders Act Irrationally: Behavioral Economics in Blockchain
Book: Cryptoeconomics
Authors: Jian Gong, Wei Xu
ISBN: 978-0-367-42993-5
Previous: Game Theory | Next: Blockchain Security
Chapter 5 is a gear shift. Less hash functions, more psychology. Gong asks: if cryptoeconomics assumes rational actors, why do crypto markets look so unhinged?
Short answer: people are not spreadsheets.
Rational vs real humans
Traditional economics says we weigh options and pick the best one. Markets correct mistakes.
Behavioral economics says we are predictably weird. We anchor on first prices. We chase free stuff. We mix up friendship and money.
Gong uses everyday crypto scenes to make the point.
Explaining blockchain to friends
Baidu-style definitions confuse people. Gong suggests analogies: blockchain is like the internet but for value; Bitcoin is like transferring money without a bank.
Relativity matters. You do not know an Antminer S9’s absolute price, but you know it costs more than an S7. Comparisons drive perceived value in mining, consensus debates, and token picks.
Anchoring and Bitcoin price
Person A thinks Bitcoin is worthless code. Person B prices it at mining cost. Person C compares it to gold’s market cap. Person D sees global financial reset.
First impression becomes the anchor. Unless something huge breaks that anchor, people buy or sell against that mental reference.
ICO boom example: famous advisor names in whitepapers anchored “legitimacy” for investors who never read the tech.
Gong’s advice: check your anchors. Blockchain’s immutable history helps. You can replay what you said and did when BTC was at 14K RMB.
Why airdrops work
Free tokens hit different. Zero is not just a low price. It is an emotional event.
MIT economist Dan Ariely’s point: free options hide downsides. You register sketchy exchanges. You chase candy you do not need.
Projects know this. Airdrops buy attention and social sharing. Tiered ICO discounts (40% week one, less each week) exploit the same FOMO.
Gong’s quiz: free 100 yuan token vs 200 yuan token for 70 yuan. Most pick free. Rational math says pay 70.
DAOs and social norms
Friend helps you get into a hot ICO. You offer 1000 RMB cash thanks. Awkward. Social norms (help freely) clash with market norms (pay for service).
DAOs promise algorithmic organizations where marginal cost of coordination drops toward zero. Xiao Feng’s quote: when p2p rules run on math, do you still need a company wrapper?
Gong sketches a “society without money” community: contributions logged on-chain, reciprocity without cash. Idealistic. The DAO hack (next chapter) shows where idealism cracks.
Smart contracts as discipline
We say we will only risk 10-20% of savings. Then we YOLO the house.
Behavioral fix: pre-commitment. Employers auto-deduct savings. Credit cards in ice water (old trick).
Smart contracts can cap spending by category, auto-save overages, or embarrass you on social media if you break rules.
Gong’s credit card example is playful but serious. Code as the “external voice” when your lizard brain takes the wheel.
Too many choices
2017 ICO mania: join every group, skim every whitepaper, miss the good ones anyway.
Yale’s Shen Jiying: more options can mean worse decisions, not better. FOMO is the tax.
Gong cites Erich Fromm’s Escape from Freedom. Modern life drowns you in opportunity. Crypto amplifies it 100x.
Intersection
Cryptoeconomics designs incentives for rational agents. Behavioral economics describes actual agents.
Smart contracts can bridge the gap: transparent records, enforced limits, fairer information access without platform gatekeepers.
My reaction
This is the most readable chapter in the book. No Byzantine faults. Just stuff you have lived through if you were in crypto during 2017-2018.
Some examples feel dated (imToken airdrops, RMB price talk). The biases are not. Anchoring still drives NFT floors. Free mints still clog wallets. Social norms still explode DAO treasuries.
Gong is not cynical. He thinks better tools can help. I am less sure humans outrun their own bugs. But putting guardrails in code is worth trying.