This paper reviews the literature on the consumer price index (CPI) and the main causes of the upward biases that are usually present in traditional measurements. The debate on this subject is old, but it has recently been renewed by a report to the US Senate by an Advisory Commission made by economic experts. An upward bias of one percentage point per year in the CPI can create huge effects on the economy in general and the government budget deficit in particular within a time span of a decade. There are many common elements in the methodologies that are used by national statistical agencies. All the five main sources of bias that are considered by the US commission are likely to be present in the CPI statistics of almost every country. Some changes in the methodology, as for example the updating of the weighting system that has been accelerated in some countries for the last stage of aggregation, cannot be very effective, since the biased formula that aggregates basic elementary information has remained unchanged. This paper takes into exam each of the main causes of distortion in the CPI and recommends some methodology improvements. The formula bias can be reduced by simply adopting index numbers that are more flexible than the simple arithmetic mean of the elementary indices. The so-called substitution bias can be reduced by adopting a flexible-weight aggregation formula in the last stage of aggregation, where a system of weights are available. The outlet-substitution bias can be substantially reduced by including in the aggregating formula the price-induced shifts of consumers to economically more convenient locations where they make their purchases. The quality and new-product biases can be reduced by evaluating the price components due to changes in quality and introducing the new goods in the market at least in some critical sectors, such as those of computers, automobiles, and some electronic and biochemical products.