DATA STORY · WIDGET 39

The Replacement Paradox

Drag the slider — see what the U.S. honey bee colony count would have looked like without beekeeper replacement.

0% (no replacement)100% (one-for-one)150% (net growth)
At 100% replacement, the count holds flat

At full one-for-one replacement, the simulated line stays at 2.39 million from 2006 through 2024 — within ~12% of the actual NASS census. The fact that NASS shows ~2.7 million in 2023 (slightly higher than 2006) means real-world replacement averaged just above 100%. The "stable colony count" headline is a labor artifact, not a health signal.

U.S. Managed Honey Bee Colonies — Actual vs Simulated
Y-axis: thousands of colonies · X-axis: 2006 → 2024 · Hover years for detail
00.5M1.0M1.5M2.0M2.5M3.0M2006200820102012201420162018202020222024Colonies (millions)
Actual (USDA NASS)Simulated at 100% replacement0% reference (collapse)
Cumulative deaths
14.72M
colonies that died, 2007–24
Cumulative replacements
14.72M
splits / packages produced
2024 simulated count
2.39M
vs actual 2.71M
Replacement / standing ratio
5.5×
colonies produced per current colony
Why this matters

Headlines about "stable colony counts" read as good news for bee health — they are not. They are evidence that beekeepers are doing extraordinary annual labour to mask attrition that would otherwise be visible at the census level.

The simulated trajectory at any rate <100% shows what the U.S. honey bee population would have done if beekeepers had only matched the historical-norm replacement effort rather than today's near-total replacement push. Even at 80% the count crashes to under 1 million in 18 years.

The 0% reference line is the closest thing to the "true" annual attrition the U.S. honey bee population has actually experienced — a 99.9% collapse over 18 years, hidden entirely by the labour and cost of replacement. The unit (colony) is durable in name only.

Related data stories
Colony Loss Monitor (Widget 37) →World Bee Day guide →Subspecies refugia →
Method

Recurrence: N(t+1) = N(t) × (1 − L(t) × (1 − r)) where N is colony count in thousands, L(t) is the BeeInformed-reported winter loss for the season ending in calendar year t (decimal), and r is the slider's replacement rate (decimal). Initial value N(2006) = 2,393K from USDA NASS Honey Bee Colonies report.

We do not model summer losses, queen-failure rates, splitter productivity, package-buying rates, or migratory-pollination dynamics. The single replacement-rate parameter rolls all sources of net new colonies (splits, swarms, packages, queens) into one fraction of overwinter losses. This is intentionally simple — the point is to make visible that the same NASS count is consistent with very different bee-health realities depending on how much beekeeper labour is going into replacement.

The 2024 NASS value is preliminary. The 2024 BeeInformed loss (34.0%) is the published preliminary survey result.

Embed on Your Site — Free

Beekeeping extension services, honey producer associations, and pollinator-advocacy sites can embed this simulator free of charge. No account, API key, or tracking required. Add ?theme=dark for dark mode.

<iframe
  src="https://rawhoneyguide.com/tools/honey-bee-replacement-paradox"
  width="100%" height="1300"
  frameborder="0" loading="lazy"
  title="The Replacement Paradox — Honey Bee Colony Simulator">
</iframe>
FAQ
What is the Replacement Paradox?
The U.S. managed colony count has held near 2.7 million for 18 years per USDA NASS even though BeeInformed documents 33% average winter losses over the same period. Both numbers are true because beekeepers replace virtually every dead colony each spring via splits, packages, and swarm captures. The replaced colony is more Varroa-naive than the one it replaced, so next winter's losses are at least as high. The "stable count" is a labour artifact.
What does the cumulative-replacements counter measure?
At 100% replacement, summing each year's deaths over 18 years gives ~14M cumulative replacement colonies — roughly six times the standing count. Each one represents queen-rearing labour, brood-frame transfers, package purchases ($150-200 each), or sugar feed. None of this is reflected in the published "U.S. has 2.7M colonies" headline.
Why do real-world beekeepers replace at slightly over 100%?
Almond pollination demand has grown ~30%+ from 2007-2023 (USDA NASS) and the U.S. needs ~2.0M colonies trucked to California each February. That created sustained economic incentive for migratory operators to push net-positive replacement. Hobbyist colony counts also grew. The actual NASS rise from 2.4M to 2.7M corresponds to roughly 101% replacement averaged over the period.
Why does the FAO global count show ~2× growth from 1961 to 2019?
FAO doubling is driven by South Asia (China, India), Sub-Saharan Africa (Ethiopia), and Eastern Europe — regions where colony counts genuinely grew via expanded smallholder beekeeping. North America and Western Europe are flat over the same window, masked exactly by replacement-via-splitting. Global doubling is not evidence of bee health improvement.
Sources
  • USDA National Agricultural Statistics Service (2024). Honey Bee Colonies — annual reports 2006–2023. nass.usda.gov.
  • BeeInformed Partnership (2024). Annual National Colony Loss and Management Survey, 2006–07 to 2023–24. University of Maryland / USDA-ARS. beeinformed.org.
  • vanEngelsdorp D. & Meixner M.D. (2010). A historical review of managed honey bee populations in Europe and the United States and the factors that may affect them. Journal of Invertebrate Pathology 103: S80–S95.
  • FAO (2021). FAOSTAT Livestock Primary dataset — Hives of bees. Food and Agriculture Organization of the United Nations. fao.org/faostat.