2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
The Law of Many Prices
By Neville R. Norman (Melbourne and Cambridge)1
– in conjunction with Kenneth J. Coutts (Cambridge)
Prepared explicitly for consideration as a presentation to the Cambridge Business and Economics
Conference, Murray Edwards College, Cambridge, June 26-8, 2012.
N.B. The paper is incomplete as the statistical analysis is still underway with the emergence of recent new
and important data. The selections to follow will give the flavour of the new data and its significance for not
only economics but also business decisions and economic policy. N. R. Norman December, 2011
ABSTRACT
New industrial price data from the U.K. enable investigations not possible
before. The results are striking. Even at narrow product categories, prices
offered from local, European and other foreign markets diverge and
continue diverging for long periods of time, mixed with some signs of partial
convergence with more commodity-type products. The message for
management is that much scope for independent price setting exists; for
policy makers: many theories based on ‘one price’ will give faulty
predictions and policy advice on how prices are made, how they move and
react to tax and other price-based economic policies, including
environmental taxes. The ‘law of one price’ must be replaced by the ‘law of
one price’ must be replaced by the ‘law of many (and varied) price
experiences’.
There is a dominant strand of thinking is standard international trade analysis that (home and
foreign) prices tend to converge to equality, either in individual product markets, or move in unison,
overall. The approach is capture by t he slogan, ‘the law of ONE PRICE (LOOP). Companion theories
known as purchasing power parity and standard tariff theory embody the same ideas and
suppositions. Using data only recently available, we are able to demonstrate that reality is very
different: prices tend to diverge over long periods, both levels and ratios, with little apparent
convergences, except in almost pure commodity, exceptional, cases. We thus propose to contrast
the findings drawn directly from reality with the outmoded theories by introducing the evidencebased concept, The Law of Many Prices (LOMP).
The new data sources emerging from the British statistical authorities and have been assembled by
the authors , using approved confidential data access to make a complete domestic and import price
1
Corresponding principal author: n.norman@unimelb.edu.au
June 27-28, 2012
Cambridge, UK
1
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
index base for all manufacturing and the main divisions of it, covering quarterly spans since the
middle 1990s. For decades, only the UK authorities have, almost uniquely, provided output price
data for home-sales only to be compared with rival import price series. What is new is that the
import price series are now (a) exactly matched with the specific product classifications of the home
price series; and (b) import sources are shown separately for the European Union (EU) and also nonEU sources of supply, enabling price series for three supply sources to the UK market to be
assembled and compared.
The results are striking and have significant implications for the manner in which prices are formed,
policy works on the economy and competition is exercised from and upon the business management
perspective. We thus have messages for theory, economic policy and management in this important
research exercise.
A. Central Research Findings of the Cambridge ( Coutts-Norman) pricing project 19992011:




B.
1.
2.
3.
UK industry prices respond mainly to unit cost shifts, tempered to a degree by import prices,
and hardly in any discernible way in response to demand pressure;
The cost-price pass-through co-efficient is close to but less than unity, as required in markup pricing;
The import-to-domestic price pass-through co-efficient is positive but closer to zero than the
unity required in conventional trade and tariff analyses that dominate textbooks and
economist policy advice; and
Even through the global financial crisis, UK industry prices have been extremely
unresponsive to demand pressure, contrary to the dominating core of marginalist pricing
hypotheses.
Documentation for Research Findings:
Previous econometric studies without import prices: Neild (1964), Godley and Nordhaus
(1972), Coutts, Godley and Nordhaus (1978)
More recent econometric studies with import pricing influences tested explicitly: Martin
(1997), Coutts and Norman (2007, 2008, and 2010)
Survey evidence for UK (Oxford research group in Wilson and Andrews, 1951), Bank of
England (1999), also RBA for Australia (Park et al., 2010) and the Euro area (Fabniani et al.,
2006)
C. Implications for theory, teaching and economic policy advice:





Little or no support for economic models of pricing that suppose zero global pricing
influence, full cost pass-through and especially full import-price pass-through
Strong case for time-based (dynamic) specification to support statistical empirical work, and
advise public policy processes that are nearly always time-specific
No evidence that macro demand restraint will soften price increases in UK industry: macro
policy relevance
Standard micro models neglect global influences – or incorporate them in an erroneous and
simplistic manner
Standard trade and tariff models do not capture the forces actually at work in the price
process for UK industry in the period 1970-2010. Diversity associated with differential
June 27-28, 2012
Cambridge, UK
2
2012 Cambridge Business & Economics Conference

ISBN : 9780974211428
competition and product differential, dynamics and cost dominance in pricing are hardly
anywhere reflected in such models.
The actual impact of trade policy changes is least affected by the extent of tariff policy
changes and mostly by other things, such as product and competition characteristics,
strategic industry considerations and business perceptions – including conscious uncertainty
that drive many to adopt pricing close to mark-up models found in heterodox economics.
D. Illustration of UK pricing research findings summarised above – before the fancy
econometrics, the story is very clear from these simple charts:
140.0
PPI and
Import prices:
the wandering
duette
120.0
100.0
80.0
PPI
60.0
Pimps
40.0
20.0
June 27-28, 2012
Cambridge, UK
Jun-10
Sep-08
Dec-06
Mar-05
Jun-03
Sep-01
Dec-99
Mar-98
Jun-96
Sep-94
Dec-92
Mar-91
Jun-89
Sep-87
Dec-85
Jun-82
Mar-84
Sep-80
Dec-78
Mar-77
Jun-75
Sep-73
Dec-71
Mar-70
0.0
3
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
140.0
PPI and
Unit costs
a closer fit
120.0
100.0
80.0
60.0
PPI
UC
40.0
20.0
Jun-10
Sep-08
Dec-06
Mar-05
Jun-03
Sep-01
Dec-99
Mar-98
Jun-96
Sep-94
Dec-92
Mar-91
Jun-89
Sep-87
Dec-85
Mar-84
Jun-82
Sep-80
Dec-78
Jun-75
Mar-77
Sep-73
Dec-71
Mar-70
0.0
Using CBI data:
30.00%
cbiCapUse
-vPPI 4QS
25.00%
20.00%
15.00%
10.00%
5.00%
0.00%
-5.00%
0
June 27-28, 2012
Cambridge, UK
10
20
30
40
50
60
70
80
4
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
30.00%
cbiNewOrders
v
PPI 4QS
25.00%
20.00%
15.00%
Serie…
10.00%
5.00%
0.00%
-80
-60
-40
-20
-5.00%
30.00%
25.00%
0
+20
+40
cbiExpNOrds
-vPPI4QS
20.00%
15.00%
Seri…
10.00%
5.00%
-60
-50
-40
-30
-20
0.00%
-10
0
-5.00%
+10
+20
+30
+40
However……………….
June 27-28, 2012
Cambridge, UK
5
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
30.00%
cbiExpIncrAVC
-vPPI 4QS
25.00%
20.00%
15.00%
Serie…
10.00%
5.00%
0.00%
-40
-20
0
+20
+40
+60
+80
+100
-5.00%
120
100
80
CbiOptInd
ex
60
40
CBI CapUt
20
Sep-10
Nov-08
Jan-07
Mar-05
May-03
Jul-01
Sep-99
Jan-96
Nov-97
Mar-94
May-92
Jul-90
Sep-88
Nov-86
Jan-85
Mar-83
May-81
Jul-79
Sep-77
Nov-75
Jan-74
-20
Mar-72
0
Whar*100
-40
-60
-80
-100
The chart above shows a good correlation between the CBI indicia of demand
pressure and the Wharton actual to computed peak-capacity method once
fashionable: Conclusion: as in Godley and Nordhaus (1972) we have a range
of demand-pressure indicators that are different but each credible, and none
of them links close to product price adjustments. Similarly with order
expectation data provided by the CBI for UK manufacturing
June 27-28, 2012
Cambridge, UK
6
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
120
100
80
VolNewOrd
60
40
20
ExpNewOrd
-40
Apr-10
Jun-08
Aug-06
Oct-04
Dec-02
Apr-99
Feb-01
Jun-97
Aug-95
Oct-93
Dec-91
Feb-90
Jun-86
Apr-88
Aug-84
Oct-82
Dec-80
Feb-79
Jun-75
-20
Apr-77
0
Whar*100
-60
-80
And with the ONS PPI data in four-quarter span percentage changes against
the same in the Wharton index:
4QS WharCU
0.2
0.15
0.1
0.05
0
0
50
100
150
200
-0.05
-0.1
-0.15
So what has happened to the relationship between demand pressure and price movements since
the 1930s when Hall and Hitch found the same?
(a) Real world data – tells us the same
(b) Main-core pricing postulates – impervious
June 27-28, 2012
Cambridge, UK
7
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
The impact of global forces (prices, tariffs, exchange rates) on domestically-produced
product prices is most significant, subject to controversy and needs evidence for
understanding and resolution. One extreme approach asserts a 100% global influence, or
pass-through; another extreme attributes little or no such influence. Based on the best data
and estimation methods available, Coutts & Norman (2007) found the relevant linking
parameter at just on 0.3. This means that the long-run (sustained) effect of a 10% maintained
rise in world prices, a tariffs rise or exchange rate depreciation, would be around 3%,
isolating this impact from other forces bearing on price-making, including domestic cost
movements, taxes and demand pressure. This coefficient is closer to zero (the extreme PostKeynesian position) than the 1.0 permeating nearly all conventional trade policy analysis. By
conventional trade analysis we include the law of one price dominating trade models, the
Marshallian, general equilibrium and effective-rate models of tariff policy analysis, and
purchasing-power-parity models of exchange rate determination. In addition, there is hardly a
neo-classical model of pricing that does not attribute some significant role for demand
pressure, unlike the cost-based (Post-Keynesian) approaches developed from the 1930s that
allow only a minor role, if any at all, to demand pressure.
E.
Main Results Presented in Summary Format
We have updated the data set we published previously for data running from 1971 to the end
of 2010, a decade beyond that reported in Coutts & Norman (2007). The coefficient of
greatest importance is reported below for a fixed-lag ARDL specification, covering the
quarterly data period from 1971 to progressive end-points running annually from December
quarter 1996 to the end of 2010. This approach to continuous re-estimation also illustrates the
value of advance partitioning of time-series data sets as proposed as a standard procedure in
Norman (2011).
The coefficient of global pricing influence moves upwards from the reported figure of 0.3
with end-points nearing the year 2000; it then stabilises near 0.37 for the main part of the
decade to 2010. Using this tracing procedure, we can say that there has been a form of slow,
mild warming to global influences in UK industrial price making, but the domestic cost
factors ignored in standard trade theory still remain predominant. Combining the price-cost
elasticities for both unit labour costs and materials prices, domestic costs exert an influence
that moves from just over to just under 60% through the data spans, confirming all previous
econometric and survey studies made for UK and many other countries. In every
specification and data span we investigated, five different measures of demand pressure are
never a significant influence of price movements.
June 27-28, 2012
Cambridge, UK
8
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
The data are presented in table A below. Arising from a log-log specification, they can be
interpreted as the long-run percentage effect on prices if any of the variables in the column
headings were to rise by one per cent, holding all other variables and influences in UK price
movements constant as the adjustment takes place. In the addendum, we present a full range
of statistical results, including standard errors and t-values. The fit, statistically is high, with
every coefficient being significant and autocorrelation tests clearly passed. The results to
2007 (q4) are highlighted as they are used in the forecasting analysis to predict the unseen
data from 2008 to 2010, as report herein.
Table A: Estimated coefficients linking pricing determinants to UK
producer price movements
1971
to
end of Pimp Coeff LULC
June 27-28, 2012
Cambridge, UK
LPMat Dom Cost Cost+Pimp
1996
0.308 0.428 0.182
0.610
0.918
1997
0.320 0.428 0.177
0.605
0.925
1998
0.332 0.417 0.178
0.595
0.927
1999
0.334 0.422 0.173
0.595
0.929
2000
0.356 0.437 0.153
0.590
0.946
2001
0.390 0.456 0.123
0.579
0.969
2002
0.385 0.455 0.126
0.581
0.966
2003
0.379 0.450 0.132
0.582
0.961
2004
0.358 0.421 0.164
0.585
0.943
9
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
2005
0.365 0.429 0.154
0.583
0.948
2006
0.374 0.444 0.138
0.582
0.956
2007
0.360 0.414 0.168
0.582
0.942
2008
0.379 0.420 0.154
0.574
0.953
2009
0.377 0.419 0.155
0.574
0.951
2010
0.377 0.422 0.153
0.575
0.952
The trend in estimated coefficients is show in Chart 1 below
1.200
1.000
0.800
Pimp Coeff
LULC
0.600
LPMat
Dom Cost
0.400
Cost+Pimp
0.200
0.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
F. Forecasting Analysis using Different Explana for Price Movements
It is a standard technique in data analysis to save some of the actual data, conventionally but
not necessarily at the end of the data span, in order to use it for forecasting and predictionerror analysis. (Henri Theil was the most ardent proponent of this method. A useful guide and
evaluation is given in Pindyck and Rubinfeld (1998).) The purpose of this exercise is not
simply to demonstrate how well alternative approaches can forecast; it is mainly as a further
test of the veracity of alternative hypotheses. To do this we need to select (i) a projection
period, starting from the dates which become unknown to the projection method (data
June 27-28, 2012
Cambridge, UK
10
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
saving); (ii) some projection approaches, reflecting specific theories or hypotheses on how
industrial prices move; and (iii) some criteria by which to evaluate the alternative predictions:
prediction error analysis.
Our method is to truncate/save some of the data at the end of pour data span, in order the
enable the alternative approach to predict the price movements that actually took place in the
projection span. We have chosen predictions for the period 2008, March quarter, to 2010,
September quarter, the latest data available at the time the exercise was carried out. The cutoff point is justified by both economic and statistical considerations. Statistically, the eleven
data points for the producer price series is a suitable number to test prediction power;
economically, the end of 2007 comes at the height of the sterling’s long strength, after which
it was subject to considerable deprecation against other currencies; the span also captures the
onset of the global financial crisis that intensified during 2008 with effects enduring for the
British economy through the entire projection span (2008/1 to 2010/3).
We test three specific approaches to predicting the course of British industrial prices in the
period 2008-2010:
A. Conventional international trade and tariff theory, based on the dominating textbook
proposition that domestic prices are determined by the duty-corrected prices of
foreign (imported) products. If this hypothesis were completely correct, an index of
UK producer prices, PPI, would match exactly the index of prices of imported
industrial products (Pimp). Almost the entire body of trade and tariff theory clings to
this hypothesis. It is supported by the age-old assumptions of perfect product
substitutability, perfect competition among home producers and the small-country
assertion (that each ‘home’ country lacks any power to price independently of world
prices).
B. An extreme post-Keynesian mark-up pricing approach, developed in Norman (1996)
in which domestic producers defy trends and movements in rival import prices,
preferring for strategic reasons to gear their prices quite rigidly to domestic unit cost
movements. If this approach were completely correct, then the PPI would mirror an
index of unit costs, or a weighted average of unit labour costs and materials costs.
While many conventional economists might have little regard for this approach,
because it assigns no role whatever to either import prices or demand pressure, it is in
principle no more extreme that the dominating trade theory model (A. above) that
assigns no role whatever to either domestic costs or any demand influences.
C. A hybrid approach derived from our preferred regression method adopting the
predictor of PPI from a log-log specification involving import prices, unit labour costs
and material prices.
June 27-28, 2012
Cambridge, UK
11
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
We have generated PPI predictions based on each of these three approaches for the estimation
period 2008(q1) to 2010(q3). We show in each case the actual and predicted PPI, and the
working to derive root-mean-squared prediction errors, which are comparable when applied
to the same data sets. We commence with prediction-basis A, import prices the sole predictor.
2(i) The standard trade theory predictor
Pimp
Predict Actual Error
Predictor
106.7
106.7
0.0
0.04
111
109.1
1.9
3.48942
Jun-08 0.0192
113.1
113.5
-0.4
0.1584
Sep-08 0.0189
115.2
114.8
0.4
0.1901
Dec-08 0.0463
120.6
112.4
8.2
66.7652
Mar-09 0.0442
125.9
112.2
13.7
187.854
Mar-08
June 27-28, 2012
Cambridge, UK
Sqd
Error
Jun-09
-0.025
122.7
113.5
9.2
84.732
Sep-09
-0.009
121.6
114.3
7.3
53.8462
Dec-09 0.0175
123.8
115.5
8.3
68.426
Mar-10 0.0172
125.9
117.1
8.8
77.5456
Jun-10 0.0153
127.8
119.7
8.1
66.076
Sep-10
127.1
119.7
7.4
54.6004
-0.006
sum
663.684
Mean
sqd e
60.3349
rmspe
7.7676
12
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
The import-price (conventional trade theory) approach begins with closely concordant
predictions in early 2008, but as the global financial crisis unfolds and sterling depreciates
against most other countries, import prices in the UK rise significantly, especially from the
March quarter 2009. British producer prices remain closely geared to costs from early 2009,
apparently ignoring rising foreign-product prices. The prediction errors remain significant in
the final years of the test period, as reflected in the high root-mean-squared prediction error.
2.(ii) Price predictions from Rigid Mark-up pricing geared to domestic costs
In this approach the predictor is a weighted combination of materials cost movements and
those in domestic unit labour costs. The weights (0.43, 0.57) are those used in Coutts and
Norman (2007). The workings of the prediction analysis for method B are now shown.
CBP
CBP
Actual
Predr
PPI
Error
SqdError
Mar-08 0.0348 110.4
109.1
1.3
1.72384
Jun-08 0.0564 116.6
113.5
3.1
9.8854
Sep-08 0.0089 117.7
114.8
2.9
8.28714
Dec-08
-0.017 115.7
112.4
3.3
10.9371
Mar-09
0.01 116.9
112.2
4.7
21.7488
Jun-09 0.0017 117.1
113.5
3.6
12.6932
Sep-09 0.0068 117.9
114.3
3.6
12.6239
Dec-09 0.0177 119.9
115.5
4.4
19.693
Mar-10 0.0082 120.9
117.1
3.8
14.5593
-0.004 120.4
119.7
0.7
0.47761
Sep-10 0.0045 120.9
119.7
1.2
1.52456
Jun-10
June 27-28, 2012
Cambridge, UK
Crude
Sum
114.154
Mean
10.3776
rmspe
3.2214
13
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
It is clear that prices track domestic costs much more closely than import prices, despite
standard trade theory claiming that they would be irrelevant. Accordingly, the cost-based
pricing prediction error is less than half that of the standard trade theory method. If we were
to judge between the two extreme approached (methods A or B) the evidence is
overwhelmingly in favour of the (post-Keynesian) extreme. The question remaining is
whether the hybrid method selected by the econometric approach can improve on prediction
based only on extreme methods.
2(iii) Regression-method price predictions using both costs and import prices
This method uses the information available to the end of 2007, as highlighted in table A. The
underlying theory is that UK producer prices are geared mainly to labour and materials costs,
tempered by some consideration for rival import prices. There is no scope for any influence
from cost or demand pressures in standard trade and tariff theory; however, there are
approaches within the Post-Keynesian tradition that do permit rival product prices to be a
formal part of the price-formation explanation, notably approaches following Kalecki. (See
Coutts and Norman (2011)).
The results are as below:
RegPredict PPI
0.02915
SqdErr
106.7
109.1
-2.4
0.03242 110.15947
113.5
-3.3 11.15911
0.01233 111.51765
114.8
-3.3 10.77385
0.0172 113.43566
112.4
1.0 1.072592
0.0245 116.21533
112.2
4.0 16.12284
115.2698
113.5
1.8 3.132208
0.00056 115.33403
114.3
1.0 1.069228
0.01357 116.89968
115.5
1.4 1.959111
0.00696 117.71332
117.1
0.6 0.376162
-0.00814
June 27-28, 2012
Cambridge, UK
Actual Error
5.76
14
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
-0.00183 117.49801
119.7
-2.2 4.848774
0.00258 117.80149
119.7
-1.9 3.604349
sum
59.87822
mse
5.443475
rmspe
2.33313
Clearly, the hybrid method predicts best of all these approaches; fuller information from both
foreign and domestic influences on British industrial prices offers the best explanation of
price movement. It is notable that neither extreme approach performs as well.
2(iv) Prediction Analysis Overview, with comparisons with Australian results
We can thus summarise the prediction errors for each of the approaches tested in the present
exercise. They are compared with the same extreme prediction approaches (A. and B. as
above) based on Australian data performed by the current authors in Melbourne in 2008.
Summary
of Prediction
Errors
Prediction period
2008/1 to 2010/3
UK
Predictor
UK
2003 to 2008
Australia Australia
% to
RMSPE bestFC
Pimp: Conventional Trade Theory
7.768
333%
47.11
638%
Crude CBP: Post-Keynesian
3.221
138%
7.38
100%
CN Regression basis: Multiple
Explana
2.333
100%
12.6
171%
For the recent UK investigation we can conclude that the crude (post-Keynesian) pricing
approach disregarded (and sometimes derided) by many conventional economists performs
almost as well as a price predictor as the preferred multi-variable regression method;
however, the standard trade theory model adopted by conventional economists exhibits
June 27-28, 2012
Cambridge, UK
15
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
prediction errors more than three times as large as the regression approach and more than
double the errors found in the mark-up pricing model.
While the prediction errors are not exactly comparable as between UK and Australia, because
the time span is longer for Australia and the data variance is greater, the regression fit is
closer overall for the UK. However, in each case there is very considerable and persistent
exchange rate movement in the forecast period and the mark-up pricing model significantly
outperforms the standard trade-theory approach in both countries.
Addendum: A Fuller presentation of the data results
1971
to
end of
Pimp
Coeff
St Error
t-ratio
LULC
St Error
t-ratio
LPMat
St Error
t-ratio
1996
0.308
0.102
3.020
0.428
0.093
4.602
0.182
0.087
2.092
1997
0.320
0.096
3.333
0.428
0.094
4.553
0.177
0.087
2.034
1998
0.332
0.093
3.570
0.417
0.093
4.484
0.178
0.087
2.046
1999
0.334
0.086
3.884
0.422
0.089
4.742
0.173
0.084
2.060
2000
0.356
0.079
4.506
0.437
0.085
5.141
0.153
0.078
1.962
2001
0.390
0.076
5.132
0.456
0.085
5.365
0.123
0.074
1.662
2002
0.385
0.069
5.580
0.455
0.089
5.101
0.126
0.069
1.826
2003
0.379
0.064
5.922
0.450
0.080
5.625
0.132
0.065
2.031
2004
0.358
0.063
5.683
0.421
0.080
5.263
0.164
0.063
2.603
2005
0.365
0.061
5.984
0.429
0.072
5.958
0.154
0.053
2.906
2006
0.374
0.058
6.448
0.444
0.067
6.627
0.138
0.045
3.067
2007
0.360
0.056
6.429
0.414
0.065
6.369
0.168
0.042
4.000
2008
0.379
0.063
6.016
0.420
0.070
6.000
0.154
0.042
3.667
2009
0.377
0.061
6.180
0.419
0.067
6.254
0.155
0.039
3.974
2010
0.377
0.061
6.180
0.422
0.066
6.394
0.153
0.039
3.923
June 27-28, 2012
Cambridge, UK
16
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
G. Some very recent UK survey evidence: another way of testing pricing Hypotheses
Bunn and Ellis (2010)2 study price adjustments in UK industry, finding considerable heterogeneity
and extreme skewness in the distribution of price changes. A few items change prices very rapidly,
biasing the implications drawn from aggregated price adjustment data. Bun and Ellis are very open
about the inability of ‘conventional economic theory ‘to ‘match the results’ they find. (p.1) They say,
‘if we really want to understand and model prices with any degree of accuracy, we need to find a
way of capturing the richness of the heterogeneity that is present in the data…’ (p.27) The data
periods concern both the 1960s when official (ONS) price data were much more specific, and some
more recent surveys based on 2003-7.
Greenslade and Parker (2010)3also update earlier Bank surveys cited in Coutts and Norman (2007),
again finding that prices were very sticky and price movements were dominated by cost movements
rather than demand pressure, despite the financial crisis being covered in the investigation. The
data period was December 2007 to February 2008, and nearly 700 firms were interviewed. The
authors cite ‘economic theory …based on the actions of a profit-maximising firm’ (p.7) but do not
see the in consistency of many of their findings with this model, or comment on the discrepancy or
its implications. There was a mixture of rules of thumb and both forward and backward-looking
approached to price setting. Firms frequently cited the costs of price changes as a major reason for
the infrequency of price adjustments. Contractual conditions explained more than pure menu-cost
arguments, which seems to have surprised the investigations (p.33). There was no clear connection
between price strategies or revision frequencies and firm size or concentration ratios.
A possible criticism of the method is that the questions asked do not enable respondents to
apportion the relative importance of cost and demand or other factors in the pricing decision in the
way that the Australian RBA study and certainly the econometrics approaches (as in C&N (2007)) can
do.
The study asked specific questions about exchange rate influences and found that considerable
changes in FX rates took place and were ignored for a combination of convenience and strategic
reasons, all consistent with Coutts- Norman (20070 and earlier findings. There was some evidence
of increasing frequency of price adjustments in line with greater apparent competition.
The 2007-8 survey confirms all previous such surveys in findings that ‘the use of the mark-up over
costs form of pricing’ was dominant. (p.33)
Some further research published earlier by the Bank of England shows that allowing imported
(materials) input inflation through intermediary imports greatly assists the prediction and
understand of inflationary processes and strongly supports the use of a New Keynesian Phillips Curve
in place of conventional economic theory. Peacock and Baumann (2008) examine the period 1965 to
early 2007 and find that (i) price predictions are significantly improved by including the influences of
2
Paper finalised 29 March, 2010.
Paper finalised 7 April 2010.
3
June 27-28, 2012
Cambridge, UK
17
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
import prices, as compared with closed economy models; and (ii) the relative importance of import
price as ain inclusion in marginal costs has remained relatively constant over the data period, except
for the UK where in the last decade (1998-2007) the influence has slightly risen. Both findings accord
closely with both the findings reported in detail in Coutts and Norman (20070 and the revisions
made by the same authors to the end of 2010. While the specification is supported by functional
forms purporting to be Post-Keynesian and subject to the Lucas critique (section 2 of Peacock and
Baumann (2008)), the approach is basically conventionally neo-classical in most aspects. Despite this
apparent restriction, the chosen form permits the data to reveal dominating Post-Keynesian features
consistent with PK pricing theory as presented din Coutts and Norman (2011)
H. Significance of the findings for economic analysis and policy advice
There are considerable implications of these further findings for economic analysis, the
understanding of how economies work, and for economic (including trade) policy.
For analysis, the pricing postulates embodied in standard trade and tariff theories appear to be
neither valid descriptions nor accurate predictors of forces shaping industrial price
movements. Surveys and statistical evidence over long periods in many industrial countries
affirm both points. In relation to international economics especially trade and tariff theory, it
is difficult to see why the economic profession adheres so tenaciously to the conventional
models that were set up for a former age of high competition and highly substitutable, basic
industrial products. Using it and teaching it does considerable potential damage. The central
point of Post-Keynesian pricing analysis is that demand functions and demand factors should
be avoided or limited in any credible explanation of industrial price movements.
For the purpose of understanding how economies actually work, a Post-Keynesian framework
offers richness and accuracy that is missing from conventional neoclassical theories of price,
either general or as specifically embodied in models used in international economics.
For economic policy, demand factors either directly or via economic policy initiatives have
relatively small price impacts, and tariff and exchange rate movements have also relative
small price impacts. In each case, the messages are quite contrary to conventional macro and
micro economic policy hypotheses and advice e to policy makers.
It remains a mystery why findings like this have had such little impact on the manner in
which economics is taught and policy advice is offered, especially from economists seeking
to explain actual economies and guide policy makers in the quest of helping them to perform
June 27-28, 2012
Cambridge, UK
18
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
better. The intransigence of most economists and the limited marketing success of realistic
economists seem to explain much. In the spirit of Keynes, each of these inferences and
conclusions are demonstrably provisional.
References
Bank of England (1999) “What makes prices sticky? Some evidence for the United Kingdom.” Bank of England
Quarterly Bulletin, 39 (3), August 262-271
Bunn, P and Ellis, C., (2010), “How do individual UK producer prices behave?” ,Bank of England Working Paper
No. 394 available at www.bankogengland.co.uk/publications/workingpapers/index.htm
Also [philip.bunn@bankofengland.co.uk; cellis@bham.ac.uk]
Coutts, K., Godley, W. and Nordhaus, W. (1978), Pricing in the Trade Cycle, CUP, Cambridge.
Coutts, K. and Norman, N. (2007) “Global Influences on UK Manufacturing Prices: 1970-2000”, European
Economic Review”, 51, Issue 5, July 2007, 1205-1221.
Coutts, K. and Norman, N. (2012) “Post Keynesian Approaches to Industrial Pricing: A Survey and Critique”,
forthcoming in G. Harcourt and P. Kriesler (eds.), Oxford University Press Post-Keynesian Economics Handbook,
OUP, Oxford.
Fabiani S, Druant M., Hernando I., Kwapil C., Landau B., Loupias C., Martins F., Mathä T., Sabbatini R., Stahl H.
and Stokman A. (2006), ‘What firms’ surveys tell us about price-setting behaviour in the euro area’,
International Journal of Central Banking, 2(3), 03–47.
Godley, W.A.H. and Nordhaus, W.D. (1972) “Pricing in the Trade Cycle”, Economic Journal.
Greenslade, J. and Parker, M. (2010), “New Insights into price-setting behaviour in the United Kingdom”, Bank
of England Working Paper No. 395 available at
www.bankogengland.co.uk/publications/workingpapers/index.htm
[jennifer.greenslade@bankofengland.co.uk]
Neild, R.R. (1964), Pricing and Employment in the Trade Cycle, CUP, Cambridge
Norman, N. (2011), Data Partitioning in Time-series Econometrics (mimeo)
Office for National Statistics (ONS) (2011), Producer Price Indices, MM22, July 2011
Park A, Rayner V., D’Arcy P. (2010) “Price-setting behaviour – insights from Australian firms”, Reserve Bank of
Australia Bulletin, June Quarter 07-14.
Peacock, C and Baumann, U., (2008), “Globalisation, Import prices and inflation dynamics”, Bank of England
Working Paper No. 394 available at www.bankogengland.co.uk/publications/workingpapers/index.htm
[chris.peacock@bankofengland.co.uk; ursel.baumann@ecb.int]
Pesaran, M.H. and Shin, Y. (1999) “An ARDL approach to time-series econometrics”, in S. Strom (ed.)
Econometrics and Economic Theory in the 21st Century: Ragnar Frisch Centennial Symposium, CUP, Cambridge.
June 27-28, 2012
Cambridge, UK
19
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Pindyck, R.S. and Rubinfeld, D.L. (1998) Econometric Models and Economic Forecasts, McGraw-Hill, Boston
Mass.
Wilson, T. and Andrews, P.W.S. (1951) Oxford Studies in the Price Mechanism, OUP, Oxford.
June 27-28, 2012
Cambridge, UK
20