When we read the question, the question is asking for an "assumption." We parse by looking for opinion. Here, the opinion is indicated by "therefore": the conclusion is the last sentence. The argument is that, of two ways to stimulate the economy, increasing spending is a better way of stimulating the economy. The middle sentence gives the sole piece of evidence for this conclusion. We fixate on the phrase "the long run," a critical detail. The argument depends on the idea that long-run benefits are important enough to be a decisive way of determining which option here is better.
Testing this prediction, Choices (B) and (E) both concern the timetable or long-run versus short-run, but our prediction isn't helping us evaluate them quickly. We can switch to the negation test.
We can look for logical confirmation of the correct answer: negating (A), say there were not substantial work projects. That is harmful, but there might be other ways to stimulate spending. So (A) is out. On to (B): what if increasing spending had to be done on a specific timetable? That would not necessarily be a problem. So (B) is out. In (C), whether or not the government has to stimulate the economy is a separate question from what the best way to do it is. Choice (E) is irrelevant. The argument concerns only the best way to stimulate, not whether to stimulate or how important it is to stimulate. So (E) is out. That leaves us with choice (D). We can apply the negation test to (D). What if both methods could be used together? The prompt does not specify that these two measures are mutually exclusive. That appears to be an assumption. And if they aren't mutually exclusive, the government might not be doing the so-called "most that it can do" by just stimulating.
The correct answer is (D).