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.

<strong>Producer</strong> <strong>Price</strong> <strong>Index</strong> <strong>Manual</strong><br />

on a fixed basket of goods <strong>and</strong> services. That the<br />

Walsh index is both a superlative <strong>and</strong> a pure index<br />

throws light on the interrelationships between the<br />

theoretical output price index <strong>and</strong> pure price indices.<br />

The distinctive feature of a Walsh index is not<br />

just that the basket of goods <strong>and</strong> services is a simple<br />

(geometric) average of the quantities in each of the<br />

two periods; by being a geometric average, it also<br />

assigns equal importance to the relative, as distinct<br />

from the absolute, quantities. Such an index clearly<br />

treats both periods symmetrically. 12 Pure price indices<br />

do not have to diverge from the theoretical output<br />

price index <strong>and</strong> are not inherently biased as estimators<br />

of the theoretical index. Bias is only likely<br />

to arise when the relative quantities used in a pure<br />

price index favor one of the periods at the expense<br />

of the other, as in a Laspeyres or Paasche index.<br />

E.3.3 Representativity bias<br />

1.110 That the Walsh index is a Lowe index that<br />

is also superlative suggests that the bias in other<br />

Lowe indices depends on the extent to which their<br />

quantities deviate from those in the Walsh basket.<br />

This can be viewed from another angle.<br />

1.111 Because the quantities in the Walsh basket<br />

are geometric averages of the quantities in the two<br />

periods, equal importance is assigned to the relative,<br />

as distinct from the absolute, quantities in both<br />

periods. The Walsh basket may therefore be regarded<br />

as being the basket that is most representative<br />

of both periods. 13 If equal importance is attached<br />

to the production patterns in the two periods,<br />

the optimal basket for a Lowe index ought to be the<br />

most representative basket. The Walsh index then<br />

becomes the conceptually preferred target index for<br />

a Lowe index.<br />

1.112 Suppose that period b, whose quantities are<br />

actually used in the Lowe index, lies midway between<br />

0 <strong>and</strong> t. In this case, assuming fairly smooth<br />

trends in the relative quantities, the actual basket in<br />

period b is likely to approximate the most represen-<br />

12 The Marshall-Edgeworth index (see Chapter 15) uses a<br />

simple arithmetic average of the quantities, but the resulting<br />

basket will be dominated by the quantities for one or other<br />

of the periods if the quantities are larger, on average, in one<br />

period than the other. The Marshall-Edgeworth is not a superlative<br />

index.<br />

13 The Walsh basket is the one that minimizes the sum of<br />

the squares of the logarithmic deviations between the quantities<br />

in the two actual baskets <strong>and</strong> those in the index basket.<br />

tative basket. Conversely, the farther away that period<br />

b is from the midpoint between 0 <strong>and</strong> t, the<br />

more the relative quantities of b are likely to diverge<br />

from those in the most representative basket.<br />

In this case, the Lowe index between periods 0 <strong>and</strong><br />

t that uses period b quantities is likely to exceed the<br />

Lowe index that uses the most representative quantities<br />

by an amount that becomes progressively larger<br />

the farther back in time period b is positioned.<br />

The excess constitutes “bias” if the latter index is<br />

target the index. The bias can be attributed to the<br />

fact that the period b quantities tend to become increasingly<br />

unrepresentative of a comparison between<br />

0 <strong>and</strong> t the farther back period b is positioned.<br />

The underlying economic factors responsible<br />

are, of course, exactly the same as those that<br />

give rise to bias when the target index is the economic<br />

index. Thus, certain kinds of indices can be<br />

regarded as biased without invoking the concept of<br />

an economic index. Conversely, the same kinds of<br />

indices that tend to emerge as being preferred,<br />

whether or not the objective is to estimate an economic<br />

index.<br />

1.113 If interest is focused on short-term price<br />

movements, the target index is an index between<br />

consecutive time periods t <strong>and</strong> t + 1. In this case,<br />

the most representative basket has to move forward<br />

one period as the index moves forward. Choosing<br />

the most representative basket implies chaining.<br />

Similarly, chaining is also implied for the target<br />

economic index t <strong>and</strong> t + 1. In practice, the universe<br />

of products is continually changing as well. As the<br />

most representative basket moves forward, it is possible<br />

to update the set of products covered as well<br />

as take account of changes in the relative quantities<br />

of products that were covered previously.<br />

E.3.4 Data requirements <strong>and</strong> calculation<br />

issues<br />

1.114 Because superlative indices require price<br />

<strong>and</strong> revenue data for both periods <strong>and</strong> revenue data<br />

are usually not available for the current period, it is<br />

not feasible to calculate a superlative PPI, at least at<br />

the time that a PPI is first published. In practice, it<br />

may be necessary for the official index to be a<br />

Laspeyres type index. However, in the course of<br />

time more revenue data may become available,<br />

enabling a superlative PPI to be calculated subsequently.<br />

Some statistical offices may find it useful<br />

to do so, without necessarily revising the original<br />

official index. Comparing movements in the official<br />

22

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

Saved successfully!

Ooh no, something went wrong!