01.02.2015 Views

Producer Price Index Manual: Theory and Practice ... - METAC

Producer Price Index Manual: Theory and Practice ... - METAC

Producer Price Index Manual: Theory and Practice ... - METAC

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

1. An Introduction to PPI Methodology<br />

1.37 This has different implication for consumers<br />

<strong>and</strong> producers. The theory of consumer behavior<br />

indicates that consumers typically react to price<br />

changes by substituting goods or services that have<br />

become relatively cheaper for those that have become<br />

relatively dearer. Thus they purchase smaller<br />

quantities of the higher-priced products <strong>and</strong> more of<br />

lower-priced ones. This is known as the substitution<br />

effect, <strong>and</strong> it implies a negative correlation between<br />

the price <strong>and</strong> quantity relatives. In this case the<br />

Laspeyres CPI would be greater than the Paasche<br />

CPI with the gap between them tending to widen<br />

over time. 7 That the Laspeyres tends to rise faster<br />

than the Paasche is a matter of concern to many<br />

analysts <strong>and</strong> CPI users because it suggests that the<br />

widely used Laspeyres index may have an upward<br />

bias.<br />

1.38 The theory of the firm indicates the opposite<br />

behavior on the part of producers. As prices for<br />

particular products begin to rise, producers will<br />

shift production away from lower-priced, less profitable<br />

products toward the higher-priced more profitable<br />

ones. This type of substitution by producers<br />

implies a positive correlation between price <strong>and</strong><br />

quantity relatives. In this case the Paasche PPI<br />

would be greater than the Laspeyres PPI with the<br />

gap between them widening over time. That the<br />

Paasche tends to rise faster than the Laspeyres is a<br />

matter of concern to many analysts <strong>and</strong> PPI users<br />

because it suggests that the widely used Laspeyres<br />

index may have a downward bias, a point taken up<br />

later.<br />

1.39 In practice, however, statistical offices often<br />

do not calculate Laspeyres or Paasche indices<br />

but instead calculate Lowe indices as defined in<br />

equation (1.1). The question then arises of how the<br />

Lowe index relates to the Laspeyres <strong>and</strong> Paasche<br />

indices. It is shown in Section D.1 of Chapter 15<br />

that if the there are persistent long-term trends in<br />

relative prices <strong>and</strong> if the substitution effect for purchasers<br />

is dominant, the Lowe index will tend to exceed<br />

the Laspeyres, <strong>and</strong> therefore also the Fisher<br />

<strong>and</strong> the Paasche. Assuming that the time period b is<br />

7 If the revenue shares—that is, the weights associated with<br />

the price relatives—happen to be the same in both periods,<br />

the Laspeyres must be greater than the Paasche because a<br />

weighted arithmetic average is always greater than a harmonic<br />

average with the same weights. In order to maintain<br />

the revenue shares intact, the substitution of the quantities in<br />

response to changes in relative prices must be perfect.<br />

prior to the time period 0, the ranking under these<br />

conditions will be:<br />

Lowe ≥ Laspeyres ≥ Fisher ≥ Paasche.<br />

Moreover, the amount by which the Lowe exceeds<br />

the other three indices will tend to increase, the further<br />

back in time period b is in relation to period 0.<br />

1.40 The positioning of period b is crucial.<br />

Given the assumptions about long-term price trends<br />

<strong>and</strong> substitution, a Lowe index will tend to increase<br />

(decrease) as period b is moved backward (forward)<br />

in time. While b may have to precede 0 when<br />

the index is first published, there is no such restriction<br />

on the positioning of b as price <strong>and</strong> quantity<br />

data become available for later periods with the<br />

passage of time. Period b can then be moved forwards.<br />

If b is positioned midway between 0 <strong>and</strong> t,<br />

the quantities are likely to be equirepresentative of<br />

both periods, assuming that there is a fairly smooth<br />

transition from the relative quantities of 0 to those<br />

of t. In these circumstances, the Lowe index is<br />

likely to be close to the Fisher <strong>and</strong> other superlative<br />

indices <strong>and</strong> cannot be presumed to have either an<br />

upward or a downward bias. These points are elaborated<br />

further below <strong>and</strong> also in Section D.2 of<br />

Chapter 15.<br />

1.41 It is important that statistical offices take<br />

these relationships into consideration in deciding<br />

upon their policies. There are obviously practical<br />

advantages <strong>and</strong> financial savings from continuing to<br />

make repeated use over many years of the same<br />

fixed set of quantities to calculate a PPI. However,<br />

the amount by which such a PPI exceeds some conceptually<br />

preferred target index, such as the economic<br />

index discussed in Section E below, is likely<br />

to get steadily larger the further back in time the period<br />

b to which the quantities refer. Most users are<br />

likely to interpret the difference as an upward bias. 8<br />

A large bias may undermine the credibility <strong>and</strong> acceptability<br />

of the index.<br />

8 Of course, if producers are price takers from the market<br />

<strong>and</strong> the dem<strong>and</strong> shifts dominate, then producers will respond<br />

by increasing the quantities produced of goods with higher<br />

relative prices. The correlation between prices <strong>and</strong> quantities<br />

in this instance will be positive, <strong>and</strong> the relationship among<br />

the indices will be<br />

Paasche ≥ Fisher ≥ Laspeyres ≥ Lowe,<br />

<strong>and</strong> the bias interpreted as downward.<br />

9

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!