AMZN daily-return correlation: QQQ vs retail peers (WMT, COST) over ~3 years
Amazon’s day-to-day returns line up with Big Tech, not with Walmart or Costco. I compared close-to-close daily returns for AMZN against QQQ, WMT and COST over the last ~3 years (744 overlapping trading days), using resampled minute OHLCV clipped to the common trading calendar. The headline: AMZN’s Pearson r with QQQ is about 0.729 — roughly +0.54 higher than with WMT — and rank correlations show the same pattern, so this isn’t just a product of a few outlier days.
Below you’ll find the full statistics, 95% Fisher z confidence intervals, 60-day rolling correlations, and a four-ticker matrix that together show sustained tech-like co-movement rather than a one-off spike. The detailed charts and tests are next.
For AMZN over the past ~3 years, does it actually trade like Big Tech or like a retailer — is its daily-return correlation with QQQ far tighter than with retail peers WMT and COST? Thesis: AMZN's daily returns track QQQ far more closely than they track WMT or COST, so the market prices Amazon as a mega-cap tech/risk-appetite play rather than a consumer-retail stock despite half its revenue coming from the store.
How this was measured
Resampled minute OHLCV to daily closes for AMZN, QQQ, WMT, and COST, clipped to the last ~3 years by calendar date, aligned on the intersection of trading days, and computed close-to-close daily returns. Computed full-sample Pearson and Spearman correlations between AMZN and each comparator, along with 95% Fisher z confidence intervals for Pearson r. Also plotted 60-trading-day rolling Pearson correlations to show how the relationships vary over time, and a four-ticker correlation matrix for context.
The key numbers
Reading the numbers
Over 744 overlapping trading days, AMZN's daily returns correlate about 0.729 with QQQ but only about 0.191 with WMT and 0.266 with COST — a clear gap showing stronger day-to-day linkage to the market/tech bundle than to retail peers.
The charts
The 60-day rolling correlations show AMZN–QQQ persistently high (mean ~0.715, range 0.4427–0.9338, ending ~0.5325), while AMZN–WMT and AMZN–COST hover near zero on average (means 0.1497 and 0.2496) and even cross negative late in the sample (WMT last −0.1551, COST last −0.0935). Eye the top line versus the lower two: the QQQ linkage is both higher and more consistently positive. That pattern means AMZN's short-term co-movement looks like a tech/market play rather than a steady retailer peer.
The scatter of AMZN versus QQQ daily returns across 744 days shows a concentrated upward tilt with both series having means near 0.001 and AMZN ranging −0.1559 to 0.1408 while QQQ ranges −0.0618 to 0.1362. The cloud aligns along a positive slope consistent with the Pearson r of ~0.729, indicating fairly tight day-to-day co-movement. In other words, individual daily moves in QQQ tend to be echoed by AMZN far more often than by a random stock.
The full-sample correlation matrix puts numbers on the claim: AMZN–QQQ 0.729 versus AMZN–WMT 0.191 and AMZN–COST 0.266. Notice WMT–COST is 0.596, so the retail peers correlate with each other more than either does with AMZN. That matrix reinforces the takeaway that AMZN behaves like the QQQ bundle on daily returns, not like its retail peers.
AMZN correlation summary vs QQQ and retail peers
| pair | N_days | pearson_r | pearson_95%_CI_lo | pearson_95%_CI_hi | spearman_rho |
|---|---|---|---|---|---|
| AMZN–QQQ | 744 | 0.7289 | 0.6934 | 0.7609 | 0.7009 |
| AMZN–WMT | 744 | 0.1907 | 0.1205 | 0.259 | 0.1176 |
| AMZN–COST | 744 | 0.2658 | 0.1977 | 0.3314 | 0.2229 |
The takeaway
Short answer: over the last ~3 years Amazon behaves much more like Big Tech than like Walmart or Costco on a day-to-day basis. Over 744 overlapping trading days AMZN’s Pearson r with QQQ is about 0.729 (95% CI [0.693, 0.761]) versus 0.191 with WMT and 0.266 with COST — gaps of +0.538 and +0.463 respectively. Rank correlations tell the same story (AMZN–QQQ rho ≈ 0.701 vs 0.118 for WMT and 0.223 for COST), so this isn’t just a handful of outliers driving the link. With 744 observations and confidence intervals that exclude zero, this is a strong, not a tentative, signal — unlikely to be pure chance. The 60-day rolling correlations also show sustained tech-like co-movement across regimes rather than a one-off spike. Practical takeaway: the market is pricing AMZN closer to a mega-cap/tech/risk-appetite exposure than a pure retail/consumer stock, so portfolio tilts treating AMZN as a tech beta are consistent with recent price action.
The fine print
- QQQ includes AMZN as a constituent, so the AMZN–QQQ correlation contains a mechanical component from AMZN’s weight in the ETF.
- Results cover the last ~3 years and 744 trading days; different windows or crisis periods can shift these correlations materially.
- We did not control for the broad market (no partial correlation vs SPY); shared market beta can elevate pairwise correlations.
- Rolling correlations use a 60-trading-day window; shorter or longer windows change smoothness and peak/trough timing.