Standby generator installation: the case is a safety case

CPSC’s incident record locates the danger precisely, and it is not the machine

Generator safety advice is usually delivered as a list of warnings with no sense of which ones carry the risk. CPSC’s incident file allows something better, because it records where the generator actually was.

According to CPSC’s report "Fatal Incidents Associated with Non-Fire Carbon Monoxide Poisoning from Engine-Driven Generators and Other Engine-Driven Tools, 2011–2021" (Matthew Hnatov, Directorate for Epidemiology, June 2022), retrieved 17 September 2026: CPSC recorded 900 fatalities from 710 incidents involving engine-driven tools over the period, of which 770 deaths — 86% — were associated with generators only. Table 13 breaks down deaths in fixed-structure homes by where the generator was: living space other than a basement, 226; basement or crawlspace, 118; garage, enclosed carport or attached barn, 103; inside the house with no further detail, 61; a transit area such as a closet, laundry room, utility room or hallway, 24; outside the home, 28; and unknown location but at home, 20.

HyreElectrical calculation from those published counts: the five inside-the-envelope categories total 532 deaths, which against the 581 recorded in the table is approximately 92%. The "outside the home" category accounts for 28, or about 5%. That is arithmetic on CPSC’s counts, not a separate CPSC finding.

The garage line is the one worth sitting with. At 103 deaths — roughly 18% of the table — a garage, enclosed carport or attached barn is the third most common location. It is also the location people choose precisely because they believe they are being careful: out of the rain, out of the living room, not indoors. An attached garage is inside the home envelope, and the data does not treat it as a compromise position. It treats it as inside.

CPSC’s own caveats travel with all of this, and they matter. The report states that the information on these fatalities is anecdotal, does not represent a complete set of all incidents that may have occurred, and represents a minimum count. Reporting delays mean 2020 and 2021 are incomplete and the figures for those years will most likely rise. None of that weakens the pattern — a minimum count that is 92% concentrated in one direction is still 92% concentrated — but it does mean these are recorded incidents, not a national rate.

Non-fire CO poisoning deaths in fixed-structure homes by location of the generator, 2011–2021. CPSC, June 2022, Table 13. Recorded incidents, a minimum count, not a national rate.

The seasonal pattern contradicts how the category is sold

Standby generators are marketed on hurricane season. CPSC’s seasonal breakdown points somewhere else. According to the same report: 43% of fatalities from engine-driven tools occur during the four colder months of November through February, 32% during the transition months of March, April, September and October, and 24% in the warmer months. The report also records that 70% of all engine-driven-tool fatalities occurred in fixed-structure homes.

HyreElectrical analysis: the most plausible reading is that winter outages produce a combination that summer ones do not — a house that is getting cold, a longer outage, sealed windows, and an incentive to bring the machine somewhere sheltered. That is a hypothesis about mechanism, not something CPSC states. What CPSC’s figures do support directly is the narrower claim: the risk is not concentrated in the season the product is advertised in.

The practical consequence for a purchase decision is that "do I live somewhere with hurricanes" is the wrong filter. The better question is how often your supply actually fails, for how long, and what in your house cannot tolerate that — heat in a cold climate, a well pump, a sump pump, refrigerated medication, or someone at home who depends on powered equipment. If the honest answer is a few short outages a year and nothing critical, a standby generator is a large purchase solving a small problem, and we would rather say so. The outage cost tool puts your own hours and your own rate against that question instead of a national average, because HyreElectrical holds no national outage figure.

Sizing starts from the circuits you need, not from square footage

A generator quote that begins with the square footage of your house has started from the wrong number. What determines the size is the set of loads you intend to keep running and their behaviour when they start.

The method is to list what must stay alive — heating or cooling, refrigeration, a well pump, a sump pump, some lighting and outlets, any medical equipment — and to distinguish running load from starting load. Motor loads such as pumps, compressors and well pumps draw substantially more current at the instant they start than they do while running, and a generator sized only to the running total will stall when one of them kicks in. That is the single most common sizing error, and it is why a serious quote asks what appliances you have rather than how big your house is.

HyreElectrical’s generator sizing tool adds running watts to a starting adder you type in yourself. It is deliberately not a product selector: it does not name a machine, it does not return a SKU, and it is not a load calculation an inspector will accept. It exists to let you arrive at the conversation with a list rather than a square-footage number. The load calculator handles the different question of what your service can carry, and if the answer there is uncomfortable, the conversation is a panel and service one before it is a generator one.

Where HyreElectrical stops and the electrician starts

HyreElectrical is an independent information site. We read the federal safety report, the model code, the municipal fee sheet and the utility tariff, and we publish what they say with the document and the retrieval date attached. We do not perform, supervise, inspect or warrant electrical work; we do not size, sell, recommend or install generators or transfer equipment; and we take no payment for placement, ranking or a favourable mention.

We also publish no national price for a standby generator installation, because we hold no dataset of bids and the installed cost is driven by site conditions — gas availability, the run from the pad to the service, whether the panel is adequate, and local permit and inspection requirements — that no average survives.

Our contractor store holds 16,369 electrical companies across 5 verified states — Florida 12,114, Utah 1,489, Nevada 1,066, California 903, Texas 839 — counted on 5 September 2026, 74% of it Florida. That is a research sample for five states, not a national census, not a ranking and not a recommendation of any firm in it. Verifying a licence on the issuing board’s own register, using the licence lookup, is the check that counts.