Financial Accelerator Model
Adds leverage, collateral, and spread dynamics to a structural core so financial amplification becomes explicit.
Graph workspace
Impulse response function panel
IRFs are the canonical visualization for the financial accelerator because they show the mechanism directly: each panel draws the same shock twice, once with the external finance premium free to move and once with it held at its steady state, so the gap between the two lines is exactly what the balance-sheet channel adds - or, for some shocks and parameters, takes away.
Current results
Amplification ratio
Peak output deviation with the accelerator divided by peak output deviation in the benchmark. Above 1 the accelerator amplifies the output peak, below 1 it dampens it. Under the default technology shock it is only about 1.03: in models of this type the accelerator adds little to technology shocks (Kocherlakota 2000) and more to monetary tightenings. Shown as - for the net-worth shock, whose benchmark does not move at all.
1.03
ratio
Peak credit spread
Largest deviation of the external finance premium from steady state, with its sign, in annualized basis points. Comparable to movements in corporate credit spreads; it is countercyclical when it rises while output falls.
9.76
basis points (annualized)
Net worth trough
Deepest percentage fall in entrepreneurial net worth. On impact net worth falls by leverage times the surprise in the return on capital, so a fall in Q lands on it about twice over at leverage 2.
-1.07%
percent
Investment half-life
Quarters until the investment deviation first falls below half its impact value. Compare it with the dashed benchmark's: the accelerator lengthens it after a monetary tightening (2 quarters against 1) but shortens it after a technology shock (10 against 13), when net worth rebuilds faster than the capital stock.
10.00
quarters
Controls
Share of a failed borrower's revenue lost to checking the books; higher values make lending dearer, so firms borrow less.
Small, mostly through leverage. Higher raises the steady-state spread but lowers steady-state leverage, so the same shock dents net worth a little less and the solid lines move slightly toward the dashed benchmark. The slope of the premium is , not .
Premium elasticity (leverage): 0.05
How much extra borrowers pay when their debt rises relative to their own funds.
Higher makes the premium more sensitive to net worth, so the solid lines pull away from the dashed benchmark: the amplification ratio rises steadily (monetary tightening 1.07 at , 1.12 at 0.05, 1.21 at 0.1). At the solid and dashed lines coincide.
Shock persistence: 0.95
How slowly the technology or policy shock fades; it does not apply to the one-off net-worth loss.
Higher persistence stretches the technology responses rightward and lengthens the investment half-life. For the monetary shock it can flip the policy-rate panel: above about 0.6 the rule's reaction to the deeper fall in inflation and output outweighs the innovation and the rate falls on impact, as in the textbook New Keynesian model.
How costly it is to change the pace of investment; higher values make asset prices swing more.
Higher makes Q swing more for a given change in investment, so net worth falls further and the spread rises more - but investment itself becomes costlier to change, so both investment lines shrink. Near zero, Q is almost fixed and the accelerator can dampen instead of amplify.
Shock traced: 0
Pick the disturbance: a productivity fall, a write-down of borrower wealth, or a rate rise.
The technology shock works through productivity and is amplified only modestly; the net-worth shock moves nothing at all in the dashed benchmark, so everything it does is the accelerator; the monetary tightening is amplified most visibly in how long investment and output stay down.