Statistical1 min read
◆ The PatternTreating money differently based on its source or intended use (Richard Thaler).
Mental accounting (Richard Thaler, 1985) — treating money differently based on arbitrary categories.
Examples:
• House money effect: Profits from winning trades are risked more freely — "it's house money"
• Found money: A tax refund or bonus spent more frivolously than salary
• Separate buckets: "My retirement account" vs. "my trading account" — but it's all your wealth
Money is fungible. A dollar of profit = a dollar of salary = a dollar of savings.
• House money effect: Profits from winning trades are risked more freely — "it's house money"
• Found money: A tax refund or bonus spent more frivolously than salary
• Separate buckets: "My retirement account" vs. "my trading account" — but it's all your wealth
Money is fungible. A dollar of profit = a dollar of salary = a dollar of savings.
Worked example
14,000 in one account, thought of as 10,000 of “my money” and 4,000 of profit. Risking 1% of the first and 25% of the “house money” puts 1,100 at risk, 7.9% of the account. The same 1% rule on all of it would risk 140. Where the money came from changed nothing about what losing it would cost.
Where it misleads
- Separate buckets can be useful self-control: an emergency fund you do not trade with is a sensible account, even if money is fungible.
- Thaler & Johnson (1990) found the house-money effect alongside a “break-even effect”: after losses, people take risks that offer a chance to get back to even.
- Measuring each trade on its own also hides correlation: five “separate” bets on the same theme are one bet.
In code
capital, profit = 10_000, 4_000 # the same 14,000, in two mental accounts risk = 0.01 * capital + 0.25 * profit # 1% of "my money", 25% of "house money" share_at_risk = risk / (capital + profit) one_account = 0.01 * (capital + profit) # the same 1% rule on all of it
Assumes import numpy as np and import pandas as pd. Each snippet carries its own example numbers; the comments say which are made up.
Sources
- R. Thaler, “Mental Accounting and Consumer Choice”, Marketing Science 4(3), 1985
- R. H. Thaler, “Mental Accounting Matters”, Journal of Behavioral Decision Making 12(3), 1999
- R. H. Thaler & E. J. Johnson, “Gambling with the House Money and Trying to Break Even: The Effects of Prior Outcomes on Risky Choice”, Management Science 36(6), 1990
Last reviewed October 2026 · How this site checks its content · Report a mistake
Pattern bridge: Treating money differently based on its source or bucket. In ML, batch normalization per-group treats each layer’s activations as separate accounts. In statistics, Simpson’s paradox shows how grouping changes conclusions.
Linked from 1