Alternative Explanations — The Payday Effect

Why alternative explanations matter

Finding a statistical pattern is not the same as finding a cause. Correlation does not imply causation — a mantra worth repeating. Before we can attribute the pre-payday run-up to 401(k) flows, we need to rule out other known market events that happen on a similar calendar. If the signal disappears once we control for these confounds, then the "payday effect" was likely just a proxy for something else entirely.

We tested the three most plausible alternatives: FOMC meetings, options expiration, and CPI releases.

FOMC pre-announcement drift

The Federal Open Market Committee meets roughly eight times per year, and a well-documented anomaly (Lucca & Moench, 2015) shows that equities rise +25 to +50 basis points in the 24 hours before the FOMC announcement. If FOMC meetings happen to overlap with our payday run-up window, the entire signal could be FOMC drift in disguise.

We checked the overlap. In the lag-7/lag-8 window, FOMC meeting days are actually under-represented — only about 40% of FOMC dates fall in the run-up window, compared to the ~57% you'd expect by chance. The pattern is not being inflated by FOMC overlap; if anything, it's being diluted.

More importantly, when we add an FOMC dummy variable to the regression, the coefficient on the payday run-up (β_run) barely moves. The signal is independent of FOMC meetings.

Options expiration (OpEx)

Options expiration occurs on the third Friday of every month — a date when massive derivatives positions are unwound, delta-hedging adjustments spike, and trading volume surges. This is the closest mechanical confound to the payday effect because the third Friday frequently lands near the 15th, overlapping with semi-monthly pay dates.

We added an OpEx dummy to the regression. Result: β_run survives. The payday run-up is not explained by options-expiration effects.

CPI release

The Consumer Price Index is typically released between the 10th and 13th of each month, landing close to the semi-monthly payday window. CPI releases can move markets sharply in either direction — if they systematically fall in our run-up window, they could create spurious results.

CPI release dates are under-represented in the run-up window. When we add a CPI dummy to the regression, β_run survives with essentially no change.

Combined robustness check

The definitive test: add all three controls simultaneously. If the payday signal is really just a combination of FOMC drift, options-expiration effects, and CPI reactions, it should collapse when we throw everything in at once.

β_run (%) across control specifications — semi-monthly lag +8

Grouped bar chart comparing the run-up coefficient under baseline and four control specifications. Asterisks indicate statistical significance: * p < 0.05.

After controlling for ALL THREE alternatives simultaneously, β_run = +0.120% (p = 0.014) in 2000–2019 — actually slightly stronger than baseline.

Remaining untested alternatives

We controlled for the three most plausible confounds, but intellectual honesty requires acknowledging what we didn't test:

Alternative	Why untested
Dividend reinvestment (DRIP)	Would need ex-dividend date data for all S&P 500 constituents across 65 years
Window dressing	Primarily a quarter-end phenomenon — only 4 of 24 semi-monthly paydays per year coincide with quarter-end
Treasury auctions	Would need the complete Treasury auction schedule; auction timing has changed over the decades
IRA contributions	Too diffuse — IRA contributions can happen any time during the year and lack a fixed calendar

We threw everything we could at the pattern — FOMC meetings, options expiration, inflation reports — and it didn't budge. Whatever is driving this isn't explained by these known market events.