Many systems fail not because people work too little — they optimize the wrong metric.
Federal regulations meant to reduce fuel consumption rely on a simple metric: average miles per gallon (MPG). That sounds reasonable — until incentives start shaping behavior.
Corporate Average Fuel Economy (CAFE) standards were created in 1975 after the oil crisis. The objective was straightforward: reduce fuel consumption, reduce dependence on foreign oil, and improve energy security.
Instead of regulating fuel use directly, policymakers chose a proxy: fleet average fuel economy. Manufacturers would be required to meet minimum MPG targets across the vehicles they sold.
The regulation later evolved into a vehicle “footprint” rule, where MPG targets depend on vehicle size and classification. Larger vehicles and “light trucks” faced easier standards than smaller cars.
Manufacturers optimized the metric.
The rules unintentionally encouraged heavier vehicles. Vehicles became larger and heavier. The market shifted toward SUVs and trucks, nearly 80% of new vehicle sales.
Over the same period, the average weight of new vehicles increased by roughly 1,000 pounds compared to the early 1980s. Each vehicle became more efficient on paper, while the fleet itself grew heavier.
Electric vehicles affect the calculation too. They count as zero tailpipe emissions when regulators compute fleet averages.
Think of it like average sound limits in a movie theater. If the rule is based on the average volume of the film, the soundtrack can include very loud moments — explosions, action scenes — as long as those peaks are offset by quieter sections.
Vehicle regulations work similarly. Lower-efficiency vehicles — large SUVs and trucks — are the loud moments. Electric vehicles are the quiet ones. As long as the average across the fleet meets the target, the mix can include both.
The Equation Behind the System
Fuel consumption is determined by a simple relationship:
Total fuel use = miles driven ÷ average vehicle MPG
Improving MPG reduces fuel used per mile. If, however, total miles driven increases, total fuel consumption remains unchanged.
Since 2005, average new-vehicle fuel economy in the U.S. has improved from about 20 MPG to over 26 MPG. Americans still consume roughly 135 billion gallons of gasoline per year.
Over the past two decades, total gasoline consumption in the United States has remained remarkably stable — generally between about 130 and 145 billion gallons per year, even as vehicles became significantly more efficient.
As a result of the proxy (higher MPG), per-capita gasoline consumption has fallen. Total gasoline consumption has remained roughly flat, as population and miles driven increased. Americans also drive more miles than they did decades ago.
When vehicles become more efficient, the cost per mile of driving falls.
Lower costs encourages more driving — a phenomenon economists call the rebound effect.
The incentives improved the proxy (fuel efficiency per vehicle) - not the objective - reducing total fuel use.
Reducing fuel use requires affecting one of the variables in the equation — vehicle efficiency or total miles driven. Factors such as urban design, travel distance, transportation alternatives, and commuting patterns also influence the miles-driven side of the equation.
The system worked as designed — just not toward the original objective.
Intelligent People Solving the Wrong Objective Very Effectively
Automakers didn’t fail.
They optimized the system exactly as it was designed.
When incentives reward a proxy metric, intelligent actors will optimize that metric.
The lesson is simple:
Whenever a policy or organization measures success with a proxy metric, people will optimize that metric — even if doing so moves the system away from the real objective.
That’s why:
- companies optimize engagement instead of value
- schools optimize test scores instead of learning
- regulators optimize MPG instead of total fuel use
The proxy quietly becomes the goal.
The Pattern Appears Everywhere
- Revenue becomes a proxy for product value.
- Engagement becomes a proxy for usefulness.
- MPG becomes a proxy for fuel use.
- Weight becomes a proxy for health.
- Test scores become a proxy for learning.
- GDP becomes a proxy for societal well-being.
Proxies are useful because they are easier to measure.But once the proxy becomes the goal, intelligent effort can move the system away from the objective.
The easier the metric is to measure, the more tempting it becomes to optimize it.
The system doesn’t fail because people are careless. It fails because they are competent within the wrong frame.
Once you see this pattern, it becomes difficult to ignore.
Solve the right problem.
What metric in your organization might be standing in for the real objective?