Files
Abstract
The US Department of Agriculture (USDA), through its ARMS, collects detailed information from farm operators on debt used in the farm business. Specific loan characteristics such as interest
rate, loan term, origination date, type of loan, loan purpose, and type of financing are collected for up to the five largest loans. This information is used to construct portions of the farm sector balance sheet in addition to supporting research on credit use, farm solvency, and debt repayment capacity. Valid estimation and inferences are critical to the generation of this data,
however, because of sensitivity, is subject to nonresponse or "do not know." Ignoring item nonresponse completely, by setting all missing values to zero or by taking into account only the existing answers will result in a bias. Imputation, the practice of filling in missing data with plausible values, can mitigate this bias. This analysis examines the use of multivariate
techniques for debt component imputation. This would be an improvement over the
generalized mean imputation approach used in ARMS and for many of the debt components the first attempt at imputation.