The average new light-duty vehicle produced for the U.S. market in model year 2024 combined two figures that would have looked unusual together in 1975: 27.2 mpg of estimated real-world fuel economy and 258 horsepower.
The same EPA series puts the model year 1975 fleet at 13.1 mpg and 137 hp. Average inertia weight, meanwhile, ended the 50-model-year span at 4,354 pounds, 294 pounds above the 1975 value of 4,060 pounds.
Those numbers describe a fleet, not a typical nameplate and not the choices available to one buyer. They also do not explain why the changes happened. The value of putting them together is narrower: the 2025 EPA Automotive Trends Report makes it possible to see how several production-weighted vehicle attributes moved over the same long record, while preserving the limits of each measure.
The final record stops at model year 2024
EPA labels every model year from 1975 through 2024 as final. Those observations use final production volumes submitted after the model year. Model year 2025 is different: its values use projected production volumes supplied before vehicles go on sale, and EPA marks them preliminary.
That is why this analysis treats 1975–2024 as the historical series and puts 2025 in a separate column. The preliminary values are useful context, but they are not another final point on the line. EPA’s own appendix shows why the distinction matters: preliminary fuel-economy estimates are routinely revised when final volumes replace projections.
Reproducibility check
Final history and the preliminary next year
| Measure | Final observations | Preliminary MY2025 |
|---|---|---|
| Estimated real-world fuel economy | MY1975: 13.1 mpg; MY2024: 27.2 mpg | 28.1 mpg, based on projected production volumes |
| Inertia weight | MY1975: 4,060 lb; MY2024: 4,354 lb | 4,441 lb, based on projected production volumes |
| Horsepower | MY1975: 137 hp; MY2024: 258 hp | 264 hp; EPA notes higher uncertainty for some electrified vehicles |
| Footprint | MY2008: 48.9 ft²; MY2024: 51.5 ft² | 51.8 ft²; EPA projects this field using prior-year and public design data |
Four measures, two starting lines
Fuel economy, inertia weight, and horsepower share a final 1975 baseline. Footprint does not. EPA’s footprint record begins in model year 2008, so a 1975-to-2024 footprint comparison would manufacture a missing starting value.
Using the published endpoints:
- estimated real-world fuel economy rose 14.1 mpg, from 13.1 to 27.2 mpg;
- horsepower rose 121 hp, from 137 to 258 hp;
- inertia weight rose 294 pounds, from 4,060 to 4,354 pounds; and
- footprint rose 2.6 square feet between its model year 2008 baseline of 48.9 square feet and model year 2024’s 51.5 square feet.
The weight line did not move steadily upward. EPA reports that average weight fell sharply from 1975 to 1981, then climbed for decades. Fuel economy also did not rise continuously: EPA describes a rapid increase through 1987, a slow decline through 2004, and an increase after 2004. Horsepower fell early and then increased in almost every model year after 1981.
That shape matters because an endpoint comparison compresses the route between the endpoints. It can accurately report net change while still hiding reversals and inflection points.
What each fleet average actually means
The EPA data notes say the database covers new light-duty vehicles produced for sale in the United States. The averages reflect production volume by model year; they are not calendar-year sales averages.
Fuel economy
The report uses estimated real-world fuel economy rather than the separate CAFE compliance value. EPA currently combines its real-world city and highway results using a 43% city and 57% highway weighting. Fleet fuel economy is harmonically production-weighted, because miles per gallon cannot be averaged correctly with a simple arithmetic mean.
Battery-electric and plug-in hybrid values enter the fleetwide figure in miles per gallon equivalent where applicable. EPA reports that the model year 2024 fleet average would have been 1.7 mpg lower without those vehicles. That is a description of the calculation, not a claim that one technology alone caused the full long-term change.
Weight
The report’s weight field is inertia weight. EPA defines an inertia-weight class as a range based on loaded vehicle weight: curb weight plus 300 pounds. The 4,354-pound model year 2024 fleet value therefore should not be relabeled as average curb weight or gross vehicle weight rating.
Horsepower
EPA reports 258 hp for the final model year 2024 fleet, 88% above model year 1975 after rounding to a whole percent. The agency also warns that combined horsepower can be harder to determine for some battery-electric and plug-in hybrid vehicles with multiple motors or mixed power sources. The field remains usable, but the uncertainty is part of its definition.
Footprint
Footprint is wheelbase multiplied by average track width: the area outlined by the centers of the tires where they meet the ground. EPA reports 48.9 square feet in model year 2008 and 51.5 square feet in model year 2024.
The early footprint series also has a source limitation. Automakers did not submit the field until model year 2011, so EPA supplemented model years 2008–2010 with independent research. The preliminary model year 2025 value is more provisional still: EPA projects existing designs from the prior year and uses public information for new designs.
Reproducibility check
Reproducible endpoint calculations
| Measure | Calculated change | Reconciliation boundary |
|---|---|---|
| Estimated real-world fuel economy | +14.1 mpg, or +107.6%, from MY1975 to MY2024 | Uses EPA’s production-weighted harmonic fleet average, not a simple average of vehicle mpg values |
| Inertia weight | +294 lb, or +7.2%, from MY1975 to MY2024 | This is inertia-weight class data, not curb weight or gross vehicle weight rating |
| Horsepower | +121 hp, or +88.3%, from MY1975 to MY2024 | Reconciles to EPA’s narrative value of 88% after whole-percent rounding |
| Footprint | +2.6 ft², or +5.3%, from MY2008 to MY2024 | EPA reports 5.4% from unrounded data; the table displays one-decimal endpoints |
What the data supports and what it does not
The final endpoints support four descriptive statements: average fuel economy was higher in model year 2024 than in 1975; average horsepower was higher; average inertia weight was modestly higher after an early decline and later rise; and average footprint was higher than at the beginning of its shorter 2008 series.
The table does not isolate causes. EPA says the fleetwide figures reflect both changes within vehicle types and changes in the mix of vehicle types produced. It also notes that battery-electric and plug-in hybrid vehicles affect recent fleet averages. Those facts identify components of the measured fleet; they do not provide a controlled estimate of how much any one regulation, technology, fuel price, or buyer preference caused the 50-year result.
This analysis therefore makes no ranking, recommendation, or policy judgment. It does not compare manufacturers, predict model year 2025, or convert model-year production into calendar-year retail sales.
Method and reproducibility
Spec & Signal used the February 2026 edition of the 2025 EPA Automotive Trends Report, publication number EPA-420-R-26-001. The exact PDF downloaded for this analysis was 2,392,150 bytes with SHA-256 09a1da141acd07678166faee97be20c508fc00a53556306ecd650d2257f2c66b.
The endpoint values were transcribed from Table 3.1, except the model year 2008 footprint baseline, which comes from EPA’s Figure 3.15 discussion. Calculations use end − start for absolute change and (end ÷ start − 1) × 100 for percentage change. Percentages are rounded to one decimal place only after calculation. No missing value was imputed, and the model year 2025 values were not used to calculate the final-series changes.
The full private method record retains the source URLs and access time, byte digest, extracted inputs, definitions, exclusions, formulas, versioned calculation code, and reconciliation results. The official records were accessed and the calculations reproduced on July 30, 2026. Quinten Sinclair independently reviewed the source data, calculations, presentation, and editorial boundaries on August 2, 2026.