Professional Documents
Culture Documents
Valuing Urban Realm Business Cases in Public Spaces PDF
Valuing Urban Realm Business Cases in Public Spaces PDF
Paul Buchanan
Colin Buchanan
1. INTRODUCTION
1.1. Background
Traditionally, the benefits assessed for urban realm business cases are time
savings and safety benefits. Time savings are particularly important for
schemes that relieve crowding or re-phase pedestrianisation and part-
pedestrianisation schemes. However, many schemes have impacts that are
not covered by quantitative measurement of time savings and safety benefits.
In particular, this approach underestimates the benefits of schemes that
improve ambience.
1.2. Objectives
The challenge of the project was to provide valuations to user benefits for
quality improvements to the pedestrian environment. In terms of a designing a
research methodology there were particular challenges in how complex
payment mechanisms and improvements could best be described to
respondents in relevant and easily understood way.
• Qualitative research – focus groups were used to test the visuals and
descriptions to be used in the stated preference work. This was a key
phase that ensured the most appropriate language and stimulus was put
to respondents whilst testing the levels for each attribute.
2. QUALITATIVE WORK
2.1. Methodology
A programme of six focus groups was held among pedestrian users of the
Edgware Road to help determine the most appropriate terminology and show
material to use in the state preference study.
Crossing wait time was explored with the aim of using this as a potential value
of time measure.
Younger respondents were more opportunistic and crossed where and when
they could. For the younger males, this would involve vaulting over the
barriers. The younger female respondents admitted to crossing first into the
middle of the lane, waiting in the middle and then going on to the middle of the
second lane before reaching the opposite pavement.
Where they did wait to cross the road, respondents talked of waiting “minutes”
before getting the green man.
A range of approaches were tested to see which was likely to be the most
successful in getting respondents to indicate any level, however small, of a
willingness to pay for improvements to the pedestrian environment along the
Edgware Road.
The payment vehicles used were Council Tax, public transport fares/journey
cost, rent, single donation and subscription.
Council Tax
It was important to emphasise that the focus was solely on the part of the
Council Tax that pays for street maintenance, so that respondents had a
reference point (we suggested £75) against which to pitch their suggested
payments. Amounts in the region of £5 and £10 were suggested.
Another key issue was that respondents needed reassurance that the money
would definitely go towards the Edgware Road. With that reassurance, they
Journey Cost
Older respondents, who do not pay bus fares, dismissed the idea of paying
anything at all. Other respondents felt that they would want any improvements
funded by fares to be made to public transport, not to the pedestrian
environment.
Donation/Subscription
For the last group, this approach was revised slightly to become a
subscription scheme, whereby respondents were asked how much they would
pay, annually, as a subscription to the Edgware Road improvement scheme.
3.1. Methodology
The method for the stated preference research was CAPI hall tests which
were held in two locations: Edgware Road and Holloway Road.
The sample comprised residents and non-residents and two trip purposes:
leisure/shopping (where it is assumed that the trip is less time constrained)
and commuting.
The sample size of 600 was based on eight cells of 75 interviews each (a
recommended minimum cell size for stated preference).
3.2. Attributes
There were fifteen attributes included in the research. Examples of some the
attribute levels and the photos (at the size they were shown to respondents)
that accompanied them are shown below.
There were four stated preference exercises. The first three traded the fifteen
variables.
An example layout for the first stated preference exercise is shown below.
The final exercise traded off the top and bottom levels of the preceding three
exercises. These were colour coded with blue as the “bottom” level and green
as the “top” level. An example of the screen layout for the fourth stated
preference exercise is shown below.
Stated Preference 1
All levels for both samples were significant and had the correct sign. The most
important attribute in the first exercise for both samples was ‘good, bright,
even lighting after dark’ with ‘vehicles do not park on pavement’ the second
most important attribute.
Stated Preference 2
For the Edgware Road sample all levels were significant but both the levels of
the number of people in daylight were negative: ‘busy but not crowded.’ and
‘very crowded so that walking pace is badly affected’. In other words both of
these were seen as worse than the base level: ‘street (nearly) deserted.’
For the Holloway Road sample one level was insignificant: ‘busy but not
crowded.’ As for the Edgware Road sample both the levels of the number of
people in daylight were negative.
Stated Preference 3
All levels for both samples were significant and had the correct sign.
The most important attributes for both samples were in the crossing the street
variable: ‘direct green man crossing’ followed by ‘green man crossing in
stages’.
Stated Preference 4
The fourth stated preference exercise showed the attributes from first three
exercises each at two levels, the base level and the ‘top’ level.
For both samples all levels were significant and had the correct sign. Package
1 was the most important attribute and Package 3 was the second most
important for both samples.
The coefficients for the first three exercises can be ranked in order of the size
of their coefficients – the larger the coefficient the more important it is.
To scale the coefficients for the three ‘lower’ level exercises to the final
exercise, first of all the coefficients for the top levels for each exercise were
summed.
So, for example, for Stated Preference exercise 3, the sum of the top level
coefficients for the Holloway Road sample was 3.7703 when compared with
the lowest levels:
Finally, all the values were scaled to 100. Figure 1 shows the scaled values
for the whole sample. The attributes are ordered by the ‘top’ level attribute. If
there were values for two attributes within a topic, for example ‘direct green
man crossing’ and ‘green man crossing in stages’ within ‘crossing the road’,
then the ‘lower’ level is shaded light green in the chart.
Plants alongside street and public art in w ell chosen locations 3.23
Plants alongside street 2.38
Traffic flow is safe to cross and there is little or no noise 3.15
A lot of people around after dark *
A few people around after dark -1.94
-2 -1 0 1 2 3 4 5 6 7 8 9 10
Base: 600
• Council Tax
• Rent
• Public transport fares
For each a transfer pricing approach was used. This asks respondents
whether they would pay a set amount. If they say no, the amount is reduced. If
they say yes the amount is increased. If they say yes again they are asked
what is the maximum they would pay.
Council Tax
The Council Tax payers were asked questions about their willingness to pay
more Council Tax for the improvements they has just seen in the stated
preference.
There were three price points: £5, £2 and £10 per year.
This data was used to calculate the maximum preparedness to pay. For
example a respondent who said ‘no’ to £2 increase was assumed to be
prepared to pay 0p, a respondent who said yes to £5 but no to £10 was
assumed to be prepared to £5 and a respondent who said yes to £10 was
asked the maximum he/she would be prepared to pay.
On this basis the average annual increase in Council Tax to pay for all the
improvements was £17.35 at Edgware Road and £14.78 at Holloway Road.
Journey Cost
The public transport users who paid fares were asked questions about their
willingness to pay increased public transport fares for the improvements they
had just seen in the stated preference.
There were three price points: 5p, 2p and 10p per public transport journey.
The mean additional fare was 17p at Edgware Road and 18p at Holloway
Road.
The rent payers were asked questions about their willingness to pay
increased rent for the improvements they had just seen in the stated
preference.
There were three price points: 10p, 2p and 20p per week.
This data was used to calculate the maximum preparedness to pay. The
average weekly increase in rent to pay for all the improvements was £2.02 at
Edgware Road and £1.90 at Holloway Road.
For public transport per trip figures were annualised by multiplying them by
the average number of public transport trips made per week by public
transport fare payers and then multiplying by 48 weeks.
The preparedness to pay increased rent or public transport fares for the
improvements are much higher than for Council Tax. This is probably largely
because the values were asked for in weekly amounts for rent (although the
annual amount was mentioned) and per trip for public transport users whereas
for the Council tax they were always asked in annual amounts.
4.1. Transferability
Clearly the council tax value is much lower than the results obtained for the
other two. To look more closely at an explanation for this, the distribution of
responses is shown in Figure 2.
35%
30%
25%
20%
15%
10%
5%
Rent
0%
Fares
£0-£5 £5 - £15 £15 - £25 Council Tax
£25-£50 £50-£100 £100 -£200 £200-£500 £500+
In order to test the most appropriate figure for annual willingness to pay from
the three mechanisms, sensitivity tests were carried out by:
• removing the top and bottom 1%, 2.5%, 5% and 10% of answers
• removing the additional zero valuations associated with the fares
mechanism
• capping values at £100 p.a
• re-coding selected rent values to an annualised measure.
This suggests that £13 and £59 can provide low and high values for appraisal.
If a single number is required, a simple average of these (£35) is a reasonable
approach. However, given the need for comparability with other TfL
appraisals, public transport fares provide, perhaps, the most useful indicator.
As a result, we take £45 as the value used for willingness to pay valuation. By
applying both a £100 cap and a 17% proportion valuing improvements at zero,
this is a figure determined from the most reliable data obtained for fares.
5.1. Outline
The following describes how the values obtained for willingness to pay for
improvements to the urban realm have been developed into an evaluation
mechanism. It is split as follows:
• Improvements for high streets in general (rather than just the high
street they were on).
If the value placed by respondents was related to the starting condition on the
road then one would anticipate a positive correlation between start condition
and willingness to pay. In fact, this is not the case. Figure 3 shows that the
relationship appears to be either nil or normally distributed. As a result, it
750
'
Willingness to pay
500
250
0
0 5 10 15 20 25 30
Given the lack of correlation between base conditions and willingness to pay
as well as a pricing mechanism that involves two places, it appears that
respondents were valuing improvements over a larger area than that defined
by the study limits. We suggest that respondents answered with values that
cover all of their walks on high streets for an entire year. Since this means that
the values specified cover a wider area, it is a conservative assumption that
significantly reduces the values applied.
In order to place values on elements, it was important to ensure that the both
the start and end conditions were taken into account. In this way, the value
applied was for the change in condition. With up to three levels within each
category, this results in levels defined as low, medium or high are shown in
Table 13.
The coefficients for each category in the table above provided the basis for
splitting the values by element. Where mid-values exist, these were converted
to money values by linear interpolation. Based on this, the valuation by
element is shown in Table 14.
Since it appeared that respondents placed values for annual journeys in high
streets in general, it was important to determine how much time was spent in
high streets. The responses provide data on:
Through breaking usage down into types, and capping answers that included
time in shops as walking time, this information was used to estimate total time
spent along high streets in a year. This came to about 80 hours per year (1
hour 40 minutes per week).
To obtain values for improvements, this figure wasused to convert the annual
values into pence per minute spent on improved walkway. These values are
shown in Table 15.
PERS provides a framework that can be used to achieve all of these. PERS is
an audit tool developed by TRL and which is already in use at TfL to assess
the walking environment. PERS can assess infrastructure provision of links
and public spaces (as well as crossings, routes, public transport waiting areas
and interchange spaces). For links and public spaces, there are 14 and 6
characteristics respectively that are scored. For example, the characteristics
of links include ‘effective width’, ‘dropped kerbs’ and ‘permeability’. In a PERS
audit, a scorecard, detailed instruction manual and a series of prompts are
used to score these characteristics on a 7 point scale between -3 and +3. This
helps auditors to provide a consistent measure of the walking environment
that can be used to compare different locations. The first of two scorecards
used for assessing links is shown in Figure 4.
With these splits and the monetary values developed for each of the stated
preference elements, it is possible to ascertain a value for each of the PERS
characteristics. That value can then be related to PERS scores by using
PERS to assess the show material graphics. The scores for the show material
graphics can then be interpolated to provide a value for the full scale of scores
for each characteristic. The results from this are shown in Table 19 and Table
20.
With a PERS audit carried out prior to bidding for funds and an estimate of
post-implementation scores conducted for the scheme designs, these tables
provide valuation of the improvement (measured in pence per person per
minute walked in that environment). Measuring improvements in this way
ensures: