# Budget Constraint in Duffie's book

On Page 5 of Duffie's Dynamic Asset Pricing Theory, the budget-feasible set is defined as: $$X(q,e) = {e+D^T\theta \in R_+^s:\theta \in R^N, q\theta \leq 0}$$

Compared to Kerry Back's presentation of investor budget constraint, at least for me, this is rather odd. Can anyone please explain the economics intuition of this set for me?

If $q\theta$ is the cost of the portfolio $\theta$, then why don't we set the budget constraint as $q\theta \leq e$, where e is the initial endowment?