As mentioned in the book, when adding up the utility functions, we can assume that each agent´s utility function is cardinal, meaning it captures not only the agents ranking of alternatives but also the intensity of his satisfactions with each alternative.
Given cardinal utilities, economists have been able to measure this. Yew-Kwang Ng mentioned in The Economic Journal, 107, that being able to measure happiness, which is fully cardinal and interpersonally comparable is possible, this based on the fact that no one is infinitely sensitive.
Further from promoting the principle of treating a dollar as a dollar and then using this to solve the paradox of interpersonal cardinal utility, Ng (1976) shows the impossibility of reasonable social decisions without interpersonal comparison of cardinal utility. Hence the need of interpersonal cardinal utilities.
Having said this, is it possible to aggregate the different interpersonal comparisons of utilities made by different individuals into a social or aggregate interpersonal comparison?
Roberts (1997) remarks that a way out may be the use of some “objective” comparisons which have an independent existence.
The paradox of interpersonal cardinal utility may be solved using some monetary measure of net benefits, though, as mentioned in “A case for happiness” some marginal efficiency costs of distorting choice may be created.