GE Capital structured leveraged leases by optimization instead of trial and error

In the early 1990s, lease markets grew increasingly competitive, and GE Capital's business analysts structured leveraged leases — deals where the lessor's return depends on scheduling two intertwined payment streams: rentals collected from the lessee and debt payments owed to lenders. Structuring meant finding schedules that won the customer's business without breaking the profitability requirement.

Charles J. Litty developed an optimization model that allocates and schedules the rental and debt payments of a leveraged lease. Rather than an analyst iterating on a spreadsheet until a proposal priced competitively, the model searches the schedule space directly for payment streams that satisfy both sides of the tension at once.

The model shipped as an optional component of a large lease analysis package used to structure and analyze lease proposals — not a replacement for the analysts' workflow, but an engine they could invoke when a deal needed to be squeezed. A method was also developed for applying the same optimization approach to single-investor lease structuring, extending the idea beyond the leveraged case.

The instructive choice is architectural: the optimization was embedded inside the tool analysts already used for proposals. Adoption did not require a new process — just a better solver one menu item away in the same package.

Let math schedule rental and debt payments jointly

The dumb alternative is manual iteration: adjust the rental schedule, check profitability, adjust again. In a competitive market that search is slow, and slow quotes lose deals.

The two payment streams fight each other — terms that attract the lessee can starve the lessor's return — so scheduling them jointly is what makes both the price and the profit floor achievable.

Making the model optional inside the existing analysis package lowered the adoption barrier: analysts kept their workflow and reached for optimization only when a deal demanded it.

Extending the method to single-investor leases doubled the payoff of one modeling effort by reusing the same scheduling logic on a adjacent deal type.

When pricing is competitive, margin lives in the structure, not the headline rate. Optimized payment schedules are both competitive and profitable; manual iteration lands on one or the other.

The work appeared in Interfaces in 1994 as part of INFORMS' practice literature. The model ran as an optional component of GE Capital's large lease analysis package for structuring and analyzing lease proposals, with a companion method extending the approach to single-investor leases.