Paradox of Thrift II

You can read the first post in this series here. I think I now see Fazzari's point.

Fazzari presents an example where a family has a choice of spending 5 dollars at a restaurant or saving 5 dollars. He proposes spending the 5 dollars is better than saving--I think--because you get 5 dollars of economic activity plus the economic activity created by the bank loaning out the money, whereas when the family simply saves the 5 dollars, they are 5 dollars of economic activity behind.

So "saving" then creates a multiplier, like spending, but the savings mutiplier is weaker than the spending multiplier because the entire amount is not spent initially. Instead the bank-savings multiplier is like the tax-cut multiplier. In keynsian economics, the tax cut multiplier is smaller than the government-spending multiplier because when the government cut taxes, the person recieving the tax cut has the option of saving part of the initial amount instead spending it, reducing the multiplier effect of tax cuts. Although under my new understanding, the saved money will also be spent by the bank-loan reciepient, some small portion (the reserve portion) will in fact be saved, making the tax-cut multiplier effect smaller than the spending multiplier.

Except it turns out that the actual evidence is that the tax multiplier is greater than the spending multiplier. Mankiw gives a possible explaination here. Maybe economics is intuitive and not a paradox at all.