Researchers looking at the implications of a limit of 18 pupils per class in grades 1-3 (six to eight-year-olds) estimate that it will cost $5-6 billion (pound;3.2-pound;3.8bn) a year.
That would be the equivalent of $10,000 per teacher in grades 1-3. It is now argued that it might be better to invest this money in the professional development of existing teachers.
A nationwide pupil limit of 18 could only be introduced if another 100,000 teachers were recruited. It is surmised that this figure would not be reached unless teacher salaries were increased, costing an additional $1bn. Precise calculations are, however, impossible because the Federal government has not provided specific guidance on how the policy is to be adopted. It is, for example, not known whether the arrival of a 19th child would trigger the creation of a second class.
California's 20-pupils-per-infant-class limit means another class has to be formed when that figure is breached. The policy is costing the state $1.5bn a year. Experience there also suggests that new teachers are more likely to end up working in the most disadvantaged schools while more "marketable" staff migrate to better areas.
Tom Parrish of American Institute's Research told the conference that rising enrolment over the next 10 years would also be costly. Even without the additional accommodation it would cost about $30bn. But some savings might be made if more children got a better start in smaller classes. This could result in fewer pupils on remedial programmes and a reduction in the number of students forced to repeat a year.
"What remains in doubt is whether reducing the size of grades 1-3 classes is the most cost-effective approach to producing better high school graduates," he said.
"Three things should be considered - targeting reading instruction, flexibility in the implementation of the policy, and continuing evaluation of the programme."