One possible reason is that many people eat back their exercise in the form of a reward for doing it, or because they have been taught by savvy marketers that they need to refuel or recover something or other.
Another possible reason is that upping intentional exercise may lead bodies to decrease unintentional calorie burn (decreased fidgeting/NEAT, decreased autonomic tone, etc.) and also by way of improved exercise efficiency.
Overarchingly the latter theory is called constrained energy expenditure, and the evidence on all of this is early, and somewhat mixed.
Well a few months ago another block was added to the pile (at least for older women without obesity) suggesting constrained energy expenditure is a real phenomenon. It was a study published in Physiological Reports and it detailed the impact that a 4 month long moderate-intensity walking program had on the total daily energy expenditures of older women without obesity.
The study's 87 included participants reported being physically inactive, and weren't found to have any significant medical issues. The group was randomized into either receiving one of 2 doses of moderate intensity exercise for 4 months. Importantly, the exercise itself was supervised. Before and after measurements included resting energy expenditure (via indirect calorimeter), total daily energy expenditure (via doubly labeled water), body composition (via DEXA), graded exercise test (via treadmill VO2 max), physical energy expenditure (TDEE*0.9-RMR (to account for thermic effect of food reduced by 10%), and NEAT (by subtracting exercise energy expenditure - physical activity energy expenditure).
When it was all said and done, the lesser group added an average of 105 minutes of walking to their weeks, while the higher group added 160 minutes to their weeks.
There was a teeny tiny bit of weight lost in both groups (1.7lbs), and a tiny change to body fat percentage (-0.7%), but there were no between group differences. Expectedly, V02max improved more in the larger amount of exercise group.
What didn't change?
Everything else.
Despite marked increases in intentional exercise, and marked differences in adhered to doses of exercise, there were no differences found for participants' total daily energy expenditures, resting metabolic rates, NEAT, non-exercise physical activity or even total physical activity.
These results changed some when they further analyzed the data as they determined that those patients with higher baseline levels of physical activity showed lower levels of NEAT (important to note, difference did not reach statistical significance) post 4 months of exercise, whereas those with lower levels of baseline activity experienced decreases to resting metabolism with exercise (authors suggested latter might be due to weight loss, but given how small weight loss was, I find this confusing as the RMR drops were not small).
All this to say when it comes to exercise and its impact on calories burned, it's clearly far from a simple math formula. It's also incredibly unfair to weight. But as always, when it comes to improving health, nothing beats it.
[Thanks to Matt Woodward for sharing this study with me]