Historical Emergence — The Payday Effect

The 401(k) timeline

If the payday effect is truly driven by retirement-fund flows, its history should mirror the history of 401(k) plans themselves. Here are the key dates:

Year	Milestone	Significance
1978	Revenue Act creates Section 401(k)	The legal mechanism is born — but nobody uses it yet
1980s	Early corporate adoption	Large employers begin offering plans; assets are still small
1990s	Rapid growth	Participation surges; mutual fund companies aggressively market 401(k)s
2000s	Assets surpass $2 trillion; target-date funds proliferate	Automated equity allocation means more dollars flow into stocks mechanically
2006	Pension Protection Act (PPA) enables auto-enrollment	Workers are enrolled by default — participation rates jump from ~60% to ~90%
2010s	Assets surpass $5 trillion; passive indexing dominates	Massive, predictable flow into S&P 500 index funds
2020s	Assets surpass $7 trillion	Largest pool of recurring equity-buying in history

If the payday effect is real and caused by 401(k) flows, we should see nothing before 1980, a gradual emergence in the 1990s, a peak in the 2000s–2010s, and possibly some decay if sophisticated traders discover and arbitrage the pattern.

Decade-by-decade results

We ran the same regression model on each decade separately, using semi-monthly paydays with a 7-day clearing lag. This is the test: does the signal track 401(k) history?

Pre-run beta (%) by decade — semi-monthly paydays, lag = 7

Bars colored by statistical significance: green = p < 0.05, yellow = p < 0.10, gray = not significant. Vertical lines mark the creation of 401(k) plans (1978) and auto-enrollment (2006 PPA).

Pre-1978: β_run = +0.008% (p = 0.787) — NOTHING. Peak decade (2000s): β_run = +0.137% (p = 0.067).

Why the decade you started investing in shapes YOUR cost:

If you started contributing to a 401(k) in the 1980s (Early era, β = 0.115%/day), the settlement timing drag was roughly double the long-run average. Compounded over a 30-40 year career with reinvested dividends, that's the most expensive era to have been a 401(k) investor. If you started in the 2000s, the pattern had partially reversed — you may have benefited slightly.

Started in 1980s–1990s? You likely experienced the highest per-dollar timing drag
Started in 2000s? The pattern had weakened — quant funds may have partially arbitraged it away
Starting now (2020s)? The pattern appears to be returning as 401(k) assets surpass $7 trillion

Project your era's impact over your career in the simulator →

Rolling 5-year window

Decade boundaries are arbitrary. To see the pattern evolve continuously, we ran the regression on every rolling 5-year window from 1965 to present, plotting the coefficient over time.

Rolling 5-year β_run (%) — semi-monthly paydays

Two traces: lag = 0 (raw payday) and lag = 7 (clearing-adjusted). Vertical lines mark 1980 (401(k) created) and 2006 (PPA auto-enrollment).

The anomaly lifecycle

Financial researchers have documented a well-known pattern: once a market anomaly is published, it tends to shrink. McLean & Pontiff (2016) studied 97 anomalies and found they decay by roughly 32% after publication, on average. Sophisticated traders — quant funds, high-frequency firms — learn about the pattern and trade against it, reducing the profit opportunity.

The payday effect's history is consistent with this lifecycle:

1960s–1970s: No signal. 401(k) plans don't exist yet.
1980s–1990s: Signal emerges as 401(k) assets grow into hundreds of billions.
2000s: Peak signal. 401(k) assets hit $2 trillion, target-date funds automate equity allocation, and auto-enrollment hasn't yet saturated.
2010s: Signal fades. Quant funds and academic research likely identify the pattern. This is the "arbitrage decay" phase McLean & Pontiff describe.
2020s: Partial return. Assets now exceed $7 trillion, and the sheer volume of flow may be overwhelming the arbitrageurs.

The pattern appeared right when 401(k) plans became popular, peaked when 401(k) assets hit $2 trillion, and faded when sophisticated traders likely discovered it. It's now partially returning as assets grow past $7 trillion.

What exactly emerged post-1978?

Our deeper investigation revealed something important: the pattern that emerged isn't a standalone phenomenon — it's a Friday-concentrated addition to a pre-existing calendar pattern.

Pre-existing: the turn-of-month effect (lag+0)

The turn-of-month effect — elevated returns on the last day of the month and first few days of the next — exists in our 1960s data, well before 401(k) plans. This is driven by multiple payment-standardization mechanisms (Ogden 1990): salary payments, bond coupon dates, dividend reinvestment, government transfers.

New post-1978: the clearing-lag component (lag+7-8, Fridays only)

What emerged with 401(k) adoption is a SECOND component — elevated returns in a 5-day window ending ~7-8 trading days after semi-monthly paydays. This component:

Is absent before 1978 (confirmed in our data)
Tracks 401(k) asset growth precisely
Is 100% concentrated on Fridays (vanishes when Fridays removed)
Is NOT visible when testing the 15th or EOM anchor alone — only when combined

Two layers of the same phenomenon: Think of it like geological layers. The turn-of-month effect is the bedrock — it's been there since the 1960s. The clearing-lag component is a newer layer deposited on top starting in the 1980s, driven by the massive growth in predictable retirement-plan flow. Both are real; the newer one is what our study contributes to the literature.

Important caveat: The clearing-lag component is entirely Friday-dependent. This means "what emerged post-1978" is more precisely described as "a Friday-concentrated settlement effect tied to the payroll clearing pipeline." It's structural — driven by how money moves through the financial system — not evidence of manipulation. See the Friday Deep Dive for the full analysis.