Paul Cheshire, Christian Hilber and Ioannis Kaplanis

1
Paul Cheshire, Christian Hilber and Ioannis
Kaplanis
Land Use Regulation and Retail Space Constraints
and Total Factor Productivity

• ERES Conference, Milan
• 24th June 2010

2
This paper hypotheses intended contribution

• Seems likely planning policy restricts land
  available for retail development so increases
  costs of space reduces retail TFP
• Try to quantify the impact by
• 1) estimating production function - including
  space
• 2) Investigating connection to differences in
  planning restrictiveness
• 3) Quantify impact on TFP and retail prices
• Problem Planning policy may negatively affect
  TFP via two distinct routes
• 1)Restriction of land supply for retail raises
  prices and cause profit maximising retailers to
  substitute land out of production
• 2) Town centre first policies may force to
  locate on smaller and less productive, higher
  cost ( for logistics, labour, customers) sites
• At this stage not distinguishing
• Using microdata and detailed planning performance
  data

3
The issues.

• Three factors of production land labour and
  capital
• Forget land (unless agricultural economist)
• But land an input into production in
  retailing think Ikea!!
• In 1980s land for retailing in prosperous SE of
  UK 250 X land for retailing in comparable US
  location (Cheshire Sheppard 1986)
• UK Planning system imposes (intentional)
  restrictions on supply of urban land via
  containment 60 brownfield
• And restricts for each (legally classified) use
• Not surprising increases cost of housing reduces
  supply elasticity
• gt so increases volatility
• Nearly all work so far on housing
• But Hilber Cheshire 2008 costs of office
  space
• Much higher in UK than continental Europe or New
  York
• tax on space in London West End equivalent of
  800 over 1999-2005
• Town centre first virtual prohibition on out
  of town large scale development gt even higher
  cost for retail?
• Another peculiarity of British planning
  reliance on development control
• gt more politicised, less planned

4
The issues.

• Increasing support for idea that planning
  policies reduce productivity in retail McKinsey
  Global Inst. 1998 Barker, 2006 Haskel Sadun,
  2009
• Haskel Sadun - first academic study by
  preventing emergence of large format out of town
  stores estimates lost 0.4 p.a. TFP growth
  1996-2006
• Also Competition Commission 2000 2008
• Well worth looking at access to store level
  micro data for 4 main supermarket groups
• Strong finding larger stores more productive and
  profitable
• More local competition reduces store prices (CC
  2008)
• And land for retail in UK x 5 to 10 in France
  (CC 2000)

5
Planning policy and its impact

• Prior to 1988 relatively relaxed approach to
  retail as such though clear evidence of overall
  space restriction via containment e.g. Reading
  1984
• 1988 PPG6 tried to steer out of town to
  regeneration sites e.g. Bluewater but still
  not restrict competition
• PPG6 1993 attempts steer to in-town sites
  because of belief free market might
  under-provide in town shopping
• Big change PPG6 1996
• -More or less prohibited out of town development
  for all town centre activities i.e. not just
  retail but offices, leisure, restaurants
• -Introduced Needs test Sequential test
• Fear - mainly a development control tool ODPM
  (2004)
• Confirmed even reinforced by PPS6 2010
• And implementation requires current local
  development plan estimated less than half LPAs
  have them

6
Figure 1 Number of Applications for Major
Retail Developments, 1979-2008

7
Figure 2 Applications for Extensions to
Foodstores, 1990 to 2001

8
Figure 3 Big 5 Supermarkets In- and Out of
Centre Openings, 1990-2000

In-centre opening rise relative to out- of
centre. But note in- or out- of centre
defined for planning purposes Merryhill
9
Figure 4 Age of Building Stock by Use Category

And an aging stock of retail buildings.
10
Data, approach and some problems

• Store level data for all stores for major
  retailer mainly food
• Detailed development control data for all LPAs
  (so far collected only England) applications,
  refusals, delays appeals
• Stores geocoded - so also data for store
  catchment areas population within given drive
  times, car ownership, competitor stores x
  distance, etc
• Some summary statistics

11
Table 2 Summary Statistics

12
Data, approach and some problems

• How measure planning restrictiveness?
• Use refusal or delay rate?
• Problem of endogeneity developers behaviour
  may be influenced by LPAs the discouraged
  developer effect
• So need instruments to identify
• Exploit change in targets for delays more than 13
  weeks 2002 separate for minor and major
• Expect more restrictive LPAs to both refuse more
  and delay more not possible post-2002
• gtSo use change in delay rate pre- post- 2002
• Or use political make-up of LPAs (Cheshire
  Hilber, 2008 explicitly Haskel Sadun, 2009,
  Hilber Vermeulen, 2010) rise of NIMBYism

13
Figure 5 Plotting the coefficients from
regressing refusal rate on delay rate
Residential (major) 1979-2008

14
Figure 6 Plotting the coefficients from
regressing refusal rate on delay rate Retail
(major) 1979-2008
Nos of major retail low relative to major
resid. - so more noise

15
But are Town centres actually town centres?

• The case of Merryhill the comparative lack of
  current local development plans
• Town centre versus out of town may be planning
  definitions more than geographical, functional or
  economic!
• Test
• does size of store vary with planning location?
• does price of space vary with official locational
  classification?
• are planning locations strongly related to
  distance from town centre e.g. major rail
  stations?
• Or do PPG6 1996 PPS6 2010 really just more or
  less prevent all retail development and
  particularly large format retail development?
• Done 1) 2)

16
Table 3a Number of stores and average floorspace
by location type
Only Destination stores clearly larger on
average
17
Table 3b Floorspace costs by location type
But unit price of Destination stores highest
town centre cheapest - contrast with distance
decay of price in Reading 1984
18
Simple Cobb-Douglas production function
No detailed info on margins but assured they are
constant by item across stores. So using sales
as measure of output
19
Figure 7 Relationship of productivity
(sales/employment) to net floorspace
20
Table 4 Basic results from a TFP approach with
Total Sales as output
21
Findings.

• Clear evidence productivity rises with store size
• Elasticity 0.1 to 0.13
• Productivity also rises with number of hours open
  and employment
• Falls with non-food format and if mezzanine
• Non-food format stores have different production
  functions
• Add controls
• Competition
• Characteristics of catchment area
• Age of store (date of opening)
• Test model only on English sample (availability
  of planning data)

22
Table 5 Add further store area controls
UKEngland
23
Figure 8 Productivity by year of opening
Impact of store age is interesting/suggestive
using estimates from model (9) gtOldest stores
least productive (no surprise) but productivity
falls cet. par. in stores founded from late
1980s And falls strongly thereafter. Looks like
PPG6 .
24
Role of planning?.

• Is store size influenced by restrictiveness of
  local LPA?
• Test against
• Refusal rate both major residential and major
  retail
• (note major retail numbers can be small and seem
  noisy)
• Instrument 1 change in delay rate following new
  targets in 2002 - measured as change in mean
  delay rate 1994-98 2004-08
• Instrument 2 share of labour seats at the
  local elections (average over 2000-2007)

25
Table 6 Regressing floorspace on planning
restrictiveness(major residential projects
refusal ratio) IV share of Labour seats
Notes The dependent variable is log(net
floorspace). The sample excludes non-food
formats. The sample is restricted to the stores
that are located in England only regulation
data collected The refusal rate is calculated as
the ratio of declined major residential projects
applications to the total number of applications
and averaged over 1979-2008 t-statistics in
parentheses
26
Table 7 Regressing floorspace on planning
restrictiveness- alternative measures
Notes The dependent variable is log(net
floorspace). The sample excludes non-food
formats. t-statistics in parentheses. The sample
is restricted to the stores that are located in
England only planning data collected. refusal
rate ratio of declined major retail project
applications to the total number of applications
and averaged over 1980-2008 (the period for which
regulation data exist). delay rate change in
the average delay ratio of applications pending
for more than 13 weeks between the period 1994-98
and the period 2004-2008.
27
Conclusions

• 1. Strong confirmation that productivity rises
  with store size
• So - restricting stores sizes by either direct
  constraints on sites/formats, or restricting
  supply of land so raising prices
• gtIncreases resource use in retail and raises
  retail prices
• Clear welfare cost but not yet quantified
  (possible)
• 2. Clear evidence that more restrictive local
  planning policy causes stores to be smaller
• By implication planning policy responsible for
  lower retail productivity
• See impact of restrictiveness from late 1980s and
  esp. 1990s
• Since poorer spend proportionately more of
  disposable income in stores (esp. food) this is
  distributionally regressive
• Net costs? What are the benefits esp. of Town
  centre first?

28
Concluding Discussion

• Benefits? Claimed
• Town centre sites most sustainable because
  most accessible by alternative transport modes
  allow linked trips so reducing need to travel
• But need to distinguish between what people
  should do and what they actually do
• Continue to decentralise use cars for shopping
  car use continues to rise at about same rate
  just more congestion
• So town centre locations likely
• Separate households from shops lead to longer
  more congested trips
• Reduce shop sizes more trips plus more
  restocking
• Increase logistics costs
• To test - but seem likely benefits additional
  costs (carbon)