Working Paper: NBER ID: w21550
Authors: Andrew B. Abel
Abstract: The crowding-out coefficient is the ratio of the reduction in privately-issued bonds to the increase in government bonds that are issued to finance a tax cut. If (1) Ricardian equivalence holds, and (2) households do not simultaneously borrow risklessly and have positive gross positions in other riskless assets, the crowding-out coefficient equals the fraction of the aggregate tax cut that accrues to households that borrow. In the conventional case in which all households receive equal tax cuts, the crowding-out coefficient equals the fraction of households that borrow. In the United States, about 75% of households borrow, so the crowding-out coefficient is predicted to be 0.75, which differs from econometric estimates that are around 0.5. I explore extensions of the model, such as a departure from Ricardian Equivalence or the introduction of cross-sectional variation in taxes, that might account for this difference.
Keywords: Crowding out; Ricardian equivalence; Tax cuts; Government bonds; Household borrowing
JEL Codes: E62; G11; H6
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| cross-sectional variation in tax changes (H31) | crowding-out coefficient approximately 0.85 (E62) |
| Ricardian equivalence holds (E62) | crowding-out coefficient equals fraction of tax cut accruing to borrowing households (H31) |
| 75% of households in U.S. borrow (G51) | predicted crowding-out coefficient of 0.75 (E62) |
| closed economy with identical households (D10) | crowding-out coefficient equals 0 (E62) |
| heterogeneity among households (G59) | crowding-out coefficient varies between 0 and 1 (E62) |
| size of Ricardian tax cut (H29) | crowding-out coefficient weakly decreasing (E19) |
| size of Ricardian tax increase (H29) | crowding-out coefficient weakly increasing (E62) |