In many countries, loan guarantee programs are important elements of government policy with respect to small- and medium-sized enterprises (SMEs). If loan guarantee schemes are to be effective, a majority of firms obtaining assistance through such a scheme ought not to be able to obtain financing from existing sources: a property known as incrementality or additionality. This paper describes a new approach to measuring incrementality. This work uses a two-stage process to estimate the incrementality of loans made under the terms of the Canada Small Business Financing (CSBF) program. First, a logistic regression-based model of loan outcomes (essentially a credit-scoring model) is estimated based on a large representative sample of SMEs. The resulting model was consistent with prior expectations and exhibited high levels of goodness-of-fit. The model was then employed to classify a sample of firms that had received loans under the terms of the loan guarantee scheme. Incremental loans ought to be classified as “turndowns” by the model; hence the proportion of loan guarantee recipients that the model classified as turndowns is a direct measure of incrementality. For the CSBF loan guarantee program incrementality was estimated (with 95% confidence) as 74.8±9.0%. Copyright Springer 2007additionality, incrementality, small business, loan guarantees, G18, G28, M13, O17,