1
Sales prices, replacement prices and trade
margins- tracing price changes through the UK
retail sector -
Discussion of the paper by Matthew Powell (ONS)
Discussant Anita Wölfl, CEPII, EHESS

• 29th General Conference of The International
  Association for Research in Income and Wealth
• Joensuu, Finland, August 25 2006

2
Why looking at retail sale prices and margins

• Wholesale and retail trade is an important
  sector
• UK wholesale and retail trade accounts for about
  10 in total production,
• Retail prices are the source of most consumer
  prices.
• Main area of problems of price measurement
• Current practice assume a direct relationship
  between sales and retail price indices, ...
• an assumption which is not necessarily true,
  ...
• and disregards changes in the quantity/ quality
  of retail services themselves.

3
Implicit Value added deflators, wholesale and
retail trade - total economy is 100 -
Source OECD STAN Database 2002
4
A good start - the paper setup ...

• Step 1 Definition of margins and its discussion
• Step 2 Theoretical derivation of how to
  construct margin volume indices
• Step3 Formal discussion of how margin volume
  indices are estimated in practice and the biases
  that may result from it
• Step 4 Empirical analysis
• pattern of different price indices and deflated
  margins using these deflators
• regression analysis of the role of products sold
  versus retailers for margins
• Step 4 Conclusion and implications

5
but room for improvement in the detail

• The style of the paper
• the reader matters,
• precision, detail, clarity, ...
• The line of argumentation
• What are the points raised?
• Are these thoroughly discussed in the theoretical
  and empirical analysis?
• What about alternatives?
• gt Discussion by presenting the paper in more
  detail

6
What are trade margins?

• Definition in the SNA (1993)
• Trade margin the difference between the actual
  or imputed price realized on a good purchased for
  resale and the price that would have to be paid
  by the distributor to replace the good at the
  time it is sold or otherwise disposed of.
• Raises three issues
• not necessarily a proportional relationship
  between sales price index and constant price
  margins, instead difference between two prices
  gt double deflation preferable?
• replacement price matters and not the price
  that a retailer actually paid for purchasing the
  good gt existence of a price pipeline?
• margin price of the retail service gt
  measurable as such?

7
How to compute the volume index of trade margins?
I

• Based on volume index of sales
• volume index of margins index of sales volumes
  weighted by base period margin values.
• way in which margin volumes are computed in
  practice
• Based on double deflation
• volume index of margins difference between
  current sales deflated to the base period sales
  and current replacement costs deflated to base
  period replacement costs
• theoretically closest to the definition in SNA

8
How to compute the volume index of trade margins?
II

• Argumentation in the paper
• Both ways to estimate margin volumes may be
  subject to biases.
• Extend of biases depends on
• covariance between the current price margin and
  the resale quantity index
• covariance between resale/replacement price and
  quantity index.
• Double deflation better extra bias arising from
  the use of sales weighted indices as opposed to
  margin weighted indices.
• Which way would create smaller bias is then an
  empirical question.
• Questions
• Is it really the sales weighted versus the
  margin weighted that is relevant, or the lack
  of consideration of changes in the margins over
  time/ real margin index?
• Bias judged to which reference?

9
What do the data say ? I

• The data
• Retail Sales index (RSI) ONSs survey of retail
  activity
• monthly turnover inquiry for 27 NACE groups
• deflators for 44 different COICOP groups,
  calculated with data from the Retail Price Index
  (RPI)
• Annual Business Inquiry (ABI) 2000
• Questions
• RSI makes no attempt to measure double deflated
  trade margins or to measure replacement costs. -
  how are replacement costs estimated in this
  analysis?
• Data comparability ?

10
Diverging pattern of sales and replacement cost
deflators
Source Based on data in Powell (2006)
11
and their impact on deflated margins
Source Based on data in Powell (2006)
12
What do the data say? II

• Correlations and regression analyses
• current price margins are positively correlated
  with growth in quantities deflated by
  sales/replacement deflator
• sales/replacement price deflators are - to a
  similar extent - negatively correlated with
  growth in respective quantities.
• not the products sold matter most for determining
  current or constant margins, but the firms who
  sell them.
• Argumentation
• The results would speak against estimating margin
  volumes based on sales volume indices only and
  against the existence of a price pipeline
• gt double deflation would be preferable
• paradox current price margins grow despite the
  stronger growth in replacement prices than in
  sales prices.

13
What do the data say? II

• Discussion
• Not necessarily paradox depends on difference
  in relative level and development of resale and
  replacement price indices.
• Empirical results may speak against sales based
  margin volumes,
• but do not necessarily favour double deflation
  as the better alternative
• Theory versus practice once more (data needs,
  complex estimation of replacement costs)
• Also double deflation is based on price indices
  of the goods purchased and sold
• What about a real price index for the retail
  service?