Financial Accelerator Model
Adds leverage, collateral, and spread dynamics to a structural core so financial amplification becomes explicit.
Understand the mathematics
Follow the same equations and assumptions one section at a time or in a continuous reading.
Hover or focus a symbol defined in this model to find its exact matches. Select it to keep the highlight; Escape clears the selection.
Entrepreneurial Sector and the Optimal Debt Contract
The financial accelerator is built on a micro-founded debt contract between entrepreneurs and lenders. Entrepreneurs purchase capital at the end of each period, operate it next period subject to an idiosyncratic productivity shock , and then decide whether to repay their debt or default. The lender cannot freely observe the entrepreneur's realized - verification is costly. This asymmetric information problem is the Townsend (1979) costly state verification (CSV) framework.
The optimal contract under CSV takes a standard debt form (Gale and Hellwig 1985): the entrepreneur repays a fixed amount if exceeds a threshold, and defaults otherwise. In default, the lender pays monitoring cost and seizes the entrepreneur's gross revenue. The contract is fully characterized by two objects: the non-default repayment rate and the default threshold. The key economic insight is that the contract terms depend on the entrepreneur's leverage - more levered entrepreneurs face worse terms because the probability of costly default is higher.
Entrepreneur 's realized return on capital: the aggregate return times an idiosyncratic shock . is drawn i.i.d. from a log-normal distribution with .
Balance sheet identity: the entrepreneur's capital purchase is financed by net worth and borrowing .
The default threshold: the entrepreneur defaults when falls below , defined as the value where gross revenue exactly equals the contracted debt repayment .
Lender participation constraint: the lender's expected revenue (payments from non-defaulters plus seized assets from defaulters, net of monitoring costs ) must equal the opportunity cost .
is the lender's gross share of expected revenue (non-default payments plus default seizures). is the expected revenue from defaulting entrepreneurs. The monitoring cost represents the deadweight loss from CSV.
The entrepreneur keeps the residual share of expected revenue. The optimal contract maximizes this subject to the lender participation constraint.
Set up the entrepreneur's capital purchase problem
The entrepreneur enters the period with net worth N and borrows B from the lender to fund capital purchases at price Q. Leverage is the ratio of total assets to equity.Characterize the idiosyncratic return
The log-normal distribution with mean-preserving parameterization ensures . The CDF and pdf are standard, which makes the contract tractable. The variance governs the dispersion of idiosyncratic returns.Define the optimal debt contract
The entrepreneur maximizes expected equity value subject to the lender earning at least the risk-free return on the loan. The control variables are the default threshold and the scale of the project .First-order condition with respect to
Differentiating the Lagrangian with respect to gives the shadow value of the participation constraint. Note and . The deadweight monitoring cost creates a wedge.First-order condition with respect to K
Differentiating with respect to and using the balance sheet identity. This can be rearranged to express the expected return on capital as the risk-free rate plus a premium that depends on leverage through .Derive the external finance premium
Combining the two FOCs and the participation constraint yields the spread function . When leverage is 1 (fully self-financed), the premium is zero. As leverage rises, the default probability increases, expected monitoring costs rise, and the lender charges a higher spread. This is the core BGG result.Log-linearize the premium function
First-order Taylor expansion around the steady-state leverage ratio. In BGG the elasticity is not a free parameter: it follows from the contract once the dispersion of , and leverage are set. At this page's calibration (, leverage 2) the contract implies about 0.04, close to the 0.05 commonly used for BGG; estimated models usually treat as a free parameter.Townsend (1979) and the CSV tradition
Costly state verification was introduced by Townsend (1979) as a model of lending when only the borrower observes the project's return and the lender must pay to check it. Gale and Hellwig (1985) showed that the optimal one-period contract is then a standard debt contract: the borrower pays a fixed amount in good states and surrenders everything, after verification, in bad states. This micro foundation is what separates BGG from ad hoc financial friction specifications - the premium function is derived, not assumed.
Calibrating the monitoring cost
BGG set and calibrate the rest of the contract to an annual default rate of about 3%, a steady-state spread of about 200 basis points a year and capital-to-net-worth of 2. On this page reproduces those targets at (199 bp, 2.9% of firms defaulting a year). Holding and the start-up transfer fixed, a higher raises the spread but lowers the leverage entrepreneurs choose, because borrowing is dearer; inside BGG's contract it also steepens the premium-leverage relationship.
Transition checks
Impact jump
In a linear state-space representation, B maps the innovation into the contemporaneous state. A delayed variable has a zero row in B and moves only after A propagates the shock. A jump is therefore a timing restriction, not a charting artifact.
A nonzero point at h=0 means the shock changes the plotted variable on impact. It is not the end of a smooth pre-shock segment.
Do not smooth through h=0 when the model says the shock enters contemporaneously.
Decay or hump
Stable eigenvalues shrink repeated applications of A. If a variable is reached through another state first, the response rises before the stable decay dominates, producing a hump-shaped path.
Post-impact movement follows the model recursion. A hump means delayed propagation, not interpolation by the chart.
Do not treat a hump as evidence of a linear ramp unless the model includes that ramp.
Variable timing
Innovation
The shock is dated at horizon zero. A response at h=0 is an impact effect, not the end of a hidden ramp.
State vector
A state can move on impact only through the relevant row of B; otherwise it moves after A propagates the shock.
Reported response
The plotted variable can jump directly through D or indirectly through the state vector.
Borrower net worth
Net worth carries balance-sheet history, so it can amplify shocks after impact.
Proof lineage
Typed relationships across models