Estimation of Laspeyres Price Indexes Using the Prediction Approach for Finite Population Sampling
用有限总体抽样的预测方法，研究链式和固定基期拉斯佩耶斯价格指数的估计量，包括美国政府指数项目使用的估计量，并分析其偏差和方差，对编制消费者价格指数有参考价值。
Estimators of chain and fixed-base Laspeyres price indexes are studied using the prediction approach to finite population sampling. The estimators include some that are based on those used in several U.S. government index programs and others derived from prediction models. Biases and variances of the estimators are studied for a case in which the reference period index weights are unknown for nonsample items. Under a model for a one-period price change in which items have common within-stratum means, unbiased estimators can be constructed, but under a more general regression model, special sample balance conditions are needed for unbiasedness of those estimators. The theory for the estimators of fixed-base indexes is illustrated in an empirical study using a population of items priced for the U.S. Consumer Price Index.