Structural effects of mobility management

(1)

Panel analysis based on 10 clustered impact assessments

carried out for South Limburg Bereikbaar in 2013-2020.

Structural effects of mobility

management

December 2020 Dennis van Soest, Casper Stelling, Henk Meurs More information: [email protected]

(2)

Introduction

In 2020 MuConsult conducted the 10th Clustered Impact

Assessment for South Limburg Bereikbaar to monitor the

effects of the employer approach.

Within the framework of this jubilee edition of the Impact

Assessment, we carried out additional analyses into the long-

term effects of mobility management.

The analyses show that participants in bicycle and public

transport measures still maintain their new behaviour several

years after participation. Behavioural retention is highest

among participants who have switched to bicycles or e-bikes.

The data for 2020 has been adjusted for Corona effects.

However, it is possible that the crisis may have longer-term

structural effects on, for example, the attractiveness of

working from home and the use of public transport. This will

have to be demonstrated in future similar analyses.

Structural effects of mobility management - December 2020 2

Employer approach South Limburg

Bereikbaar

(https://www.zuidlimburgbereikbaar.nl/

nl/onze-producten)

(3)

Clustered Impact Measurement Commuters

The Impact Assessment is an annual survey among

commuters of affiliated covenant partners of Zuid-

Limburg Bereikbaar (formerly Maastricht-Bereikbaar).

Mapping commuters' mobility behaviour and car

ownership and changes therein;

Explain changes using measures of Zuid-Limburg

Bereikbaar, measures of the employers themselves,

other measures and autonomous developments.

Translation of the behavioural change attributable to

the programme into car avoidance, CO2 reduction and

intensity reduction on priority corridors.

Factsheets for companies for further measures for

smart and sustainable mobility.

Input for Monitoring & Evaluation of the total

programme for the purpose of monitoring, adjustment

and accountability.

Sample figure: origin, destination and

allocation of car avoidance on the

network in South Limburg

(4)

Data fusion 2013-2020

Total number of completed surveys included per

measurement (cumulative)

There are many commuters among covenant

partners who have participated in multiple

(consecutive) assessments

Linking their data with ID codes (panel) and e-

mail addresses (random samples) provides a

more complete picture of the development of

their travel behaviour.

The data for several years also enables us to

determine the effects of measures in the longer

term.

The analysis is based on 27,500 fully completed

surveys with at least two measurement points at

respondent level in the period 2013-2020.

0 5.000 10.000 15.000 20.000 25.000 30.000

Meting 1 Meting 5 Meting 10

(5)

Structure of the study

For all respondents, the first time a person participated was

the baseline measurement for short- and long-term

behavioural change.

Special panel regression models are constructed. The

explained variable (the change in car trips) is explained by

the explanatory variables (participation in ZLB, employer

products and past behaviour).

There is great continuity in the measures used to influence

the behaviour of commuters in South Limburg. Not all

measures/actions are questioned in every measurement. In

the appendix, tables are given with which measures/actions

are included in this research and in which measurements

they are questioned.

Data fusion 2013-2020

Year on year change in behaviour

Coding declared variable Coding of

explanatory variables

Performing and fine-tuning

regression

(6)

Results

Cycling actions lead to a 24% reduction in car

journeys per week among participants. After a few

years this drops to 19%.

PT actions lead to a reduction of 8% in car trips per

week among participants. After a few years, this

drops to 6%.

Car schemes and flexible travel allowances increase

participants' weekly car trips.

Results per coded measure are included in the

appendices.

Combination

Measures

Short

deadline

Long

deadline

Bicycle promotion -23,9% -19,2%

Public transport

stimulation -7,5% -5,9%

Car schemes +30,2% +24,5%

Mobility budget +4,8% +3,8%

Short and long-term effect of measures on the

number of home-work trips per week

(7)

Financial incentive to purchase a bicycle/e-

bike Measures that make it cheaper to buy a

bicycle or e-bike have the greatest effect

on reducing car trips, but at the same

time also the greatest drop in effect.

A possible explanation is that the newly

purchased bicycle or e-bike will show

defects after a number of years.

Other measures have a smaller but more

stable long-term effect

(see Annexes).

Short and long-term effect of financial benefit

when buying a bicycle or e-bike

58%

69%

45%

60%

87% 89%

0%

25%

50%

75%

100%

Percentage of car trips before, during and after measure

Bicycle purchase allowance Bicycle tax break Discount purchase bicycle

After one year Long Term

(8)

Conclusions and recommendations

Commuters who switch to bicycle, e-bike or

public transport as a result of actions taken

by Zuid-Limburg Bereikbaar or the

employer's policy, will continue to do so for

a longer period of time.

The long-term effect of mobility

management is a reduction of 6% to 19% in

the number of commutes per week among

the group of participants of the action(s).

It is precisely during the crisis that many

commuters and employers reconsider their

mobility options. An extra effort on desired

changes is likely to be effective right now.

(9)

Annex

(10)

Annexes

1. Further explanation of regression model

2. Result tables for univariate analyses

3. Data tables: which variables are included from which measurement

4. Regression tables: detailed results from the univariate regression

analyses

3 februari 2021

<footer>. 10

(11)

Regression model

A dynamic panel data model was used, in which the variable to be explained, 𝑦

_𝑖𝑡

(the number of car trips made by individual i during measurement t), also

depends on the value of this variable during the previous measurement, 𝑦 _{𝑖,𝑡−1} .

Moving behaviour does not change suddenly and is based on past behaviour

Allows estimation of long-term effects

Participation in an action or scheme is the explanatory variable (X)

β is the short-term effect of this action or regulation

For each individual, an individual time-independent component was estimated

( 𝛼

_𝑖

)

We have also checked for any annual effects ( 𝑚

_𝑡

)

𝑢

_𝑖𝑡

is an error component for unobserved variation in the data.

𝑦

_𝑖𝑡

= 𝛾𝑦

_{𝑖,𝑡−1}

+ 𝛽𝑋

_𝑖𝑡

+ 𝛼

_𝑖

+ 𝑚

_𝑡

+ 𝑢

_𝑖𝑡

(12)

Regression model

The long-term effect can be determined as ^𝛽

1−𝛾 with 𝛾 < 1

If γ is between 0 and 1, the long-term effect is greater than the short-term effect.

People then find it difficult to adapt to the measure in the short term. Only in

the long term will the full effect be achieved.

If γ is between -1 and 0 (is negative), then the long-term effect is somewhat

smaller than the short-term effect. When people buy an e-bike, they may use it

very well at first, with the effect fading away a little over time.

(13)

Bicycle results

Bicycle Short term Long term

Bike promotions -13,1% -10,4%

Reimbursement of bicycle expenses -10,8% -8,0%

Purchase allowance bicycle/e-bike -42,5% -31,0%

Tax benefit for purchasing a

bicycle/e-bike -54,9% -39,9%

Reduced price bike/e-bike purchase -13,1% -11,2%

Cycling facilities at the workplace -8,9% -7,8%

Discover the e-bike -6,4% -5,1%

Percentage change in car trips compared to the average number of car trips during the first

measurement that respondents took part in

(14)

Results OV

OV Short term Long term

Public transport actions -8,1% -6,4%

Reimbursement of public transport

card -30,7% -22,4%

(15)

Car results

Car Short term Long term

Car travel allowance 35,4% 26,7%

Free parking at employer's premises 10,5% 9,2%

Lease car 38,9% 31,0%

(16)

Results combinations

In addition to the individual measures, models combining bicycle, public

transport, car and general actions have been estimated.

Total variable bicycle: someone has participated in or used a bicycle campaign,

travel allowance bicycle, purchase allowance bicycle, tax benefit bicycle, discount

on bicycle purchase, interest-free loan for bicycle purchase, lease bicycle, bicycle

facilities at work or company bicycle.

OV: OV promotions, OV card discounts, OV card compensation, business OV card,

OV travel allowance, welcome OV offer, OV commuter allowance

Car: Car reimbursement, lease car, car sharing scheme, PR subscription

General: fixed amount per month, fixed amount per km, mobility budget,

cafeteria scheme, smart work, smart travel, flexible work scheme

(17)

Dates - Bicycle

Measure / regulation 1 2 panel 3 panel 4 panel 5 6 7 8 9 10

Bike promotions x x x x x x x x x x x x x

Reimbursement of bicycle expenses x x x x x x x x x x x

Purchase allowance bicycle/e-bike x x x x x x

Tax benefit for purchasing a bicycle/e- bike

x x x x x

Reduced price bike/e-bike purchase x x x x x x x x x x x x

Interest-free loan for the purchase of a bicycle/e-bike

x x x

Leasing a bicycle/e-bike x x x x x

Cycling facilities at the workplace x x x x x x x x x x x

Discover the e-bike (Come on! Take the bike)

x x x x x x x x x x x x x

Burn Fat Not Fuel x x x x x x

Company bicycle x x x x x x x x

Change regulation RK bicycle x x x x

Overview of which measures/actions are included in each measurement

(18)

Dates - OV

Public transport actions x x x x x x x x x x x x x

Public transport commuting card x x x x x x x x

Discount card for public transport between home and work

x x x x

Business public transport card x x x x x x x x x x x

Mileage allowance commuting x x x x x x x x x x

Subscription OV-fiets x x x x

Nextbike subscription x x

Shuttle bus station-employer x x x

Discover 't OV (Come on! Take the public transport)

x x x x x x x x x

Welcome offer OV x x x x x x

Amendment of RK OV regulations x x x x

(19)

Data - Auto

Mileage allowance commuting x x x x x x x x x x x x

Paid parking at employer's premises x x x x x x x x x x x

Free parking at employer's premises x x x x x x x x x x x

Lease car / company car x x x x x x x x x x x x x

Shared car x x

Carpool scheme x x x x x x x

P+R action x x x x x

Modification of RK car scheme x x x x

Change of parking fees employer x x x x

Change of parking availability x x x x

(20)

Dates - General

Fixed amount per month regardless of means of transport

x x x x

Fixed amount per km regardless of means of transport

x x x x

Mobility budget x x x x x x x x x x

Cafeteria scheme x x x x x x x x x x

Work smart - Travel smart x x x x x x x x x x

Flexible working x x x x x x x x x x

Modification of flexible work arrangement

x x x

Change of home working arrangement x x x x

Modification of working at a different location

x x x x

(21)

Regression table - Bicycle

Variable Coefficient Value Standard error t-value p-value

Bicycle_actions Beta -0.5428 0.099027 -5.4813 4.38E-08

Bicycle_actions Gamma -0.26323 0.014673 -17.94 2.72E-70

Bicycle_RK Beta -0.44493 0.113501 -3.9201 8.98E-05

Bicycle_RK Gamma -0.3558 0.016494 -21.572 1.74E-98

Bicycle purchase Beta -1.4798 0.286323 -5.1683 2.76E-07

Bicycle purchase Gamma -0.37134 0.02889 -12.854 1.50E-35

Bicycle_fiscal Beta -1.89384 0.237248 -7.9825 3.50E-15

Bicycle_fiscal Gamma -0.37718 0.029453 -12.806 3.73E-35

Bicycle_discount Beta -0.54663 0.135622 -4.0305 5.63E-05

Bicycle_discount Gamma -0.16271 0.019277 -8.4408 3.90E-17

Bicycle loan Beta -0.38703 0.875722 -0.442 0.65994547

Bicycle loan Gamma -0.46234 0.123881 -3.7321 0.00039361

Bicycle_lease Beta 2.27799 1.183305 1.9251 0.05446736

Bicycle_lease Gamma -0.3936 0.030149 -13.055 2.21E-36

Bicycle Beta -0.38358 0.088247 -4.3467 1.41E-05

Bicycle Gamma -0.13939 0.017717 -7.8677 4.32E-15

Bicycle_discover Beta -0.2643 0.126571 -2.0882 0.0368203

Bicycle_discover Gamma -0.26149 0.014698 -17.79 3.51E-69

Bicycle_bfnf Beta -0.76613 0.927961 -0.8256 0.40942782

Bicycle_bfnf Gamma -0.42249 0.045752 -9.2344 7.66E-19

Bicycle shop Beta -0.14463 0.140766 -1.0274 0.30427987

Bicycle shop Gamma -0.30701 0.02181 -14.076 6.88E-44

(22)

Regression table - OV

3 februari 2021 22

OV_actions Beta -0.33583 0.103834 -3.2343 0.00122533

OV_actions Gamma -0.26198 0.014692 -17.831 1.74E-69

OV_card Beta -1.08544 0.317444 -3.4193 0.00064852

OV_card Gamma -0.37437 0.029095 -12.867 1.28E-35

OV_discount Beta -0.25663 0.699646 -0.3668 0.71426266 OV_discount Gamma -0.06252 0.074885 -0.8349 0.40506269

OV_business Beta 0.00906 0.112623 0.0805 0.93585743

OV_business Gamma -0.13769 0.017744 -7.7602 1.00E-14

OV_RK Beta -0.14256 0.173981 -0.8194 0.4125916

OV_RK Gamma -0.36581 0.016312 -22.426 6.83E-106

OV_bike Beta -0.89744 1.488133 -0.6031 0.54734232

OV_bike Gamma -0.06183 0.074834 -0.8262 0.40994615

OV_pendel Beta -0.0446 0.133276 -0.3346 0.73894072

OV_pendel Gamma -0.43911 0.111507 -3.938 0.00019605

OV_Discover Beta -0.25014 0.285601 -0.8758 0.38123968

OV_Discover Gamma -0.37979 0.025169 -15.089 2.59E-48

OV_welcome Beta 2.71874 1.805022 1.5062 0.1335598

OV_welcome Gamma -0.31455 0.052406 -6.0022 8.77E-09

DRK_OV Beta 0.40322 0.611886 0.659 0.51005606

DRK_OV Gamma -0.40306 0.030946 -13.025 4.43E-36

(23)

Regression table - Auto

Auto_RK Beta 1.45934 0.095053 15.353 5.50E-52

Auto_RK Gamma -0.32551 0.015972 -20.379 1.16E-88

Car_paid Beta -0.01956 0.12007 -0.1629 0.87062978

Car_paid Gamma -0.13768 0.017743 -7.7597 1.01E-14

Car_free Beta 0.45138 0.114878 3.9292 8.63E-05

Car_free Gamma -0.13947 0.017724 -7.8687 4.28E-15

Auto_lease Beta 1.60733 0.363711 4.4193 1.01E-05

Auto_lease Gamma -0.2557 0.014625 -17.483 6.15E-67

Car_carpool Beta -0.11076 0.396272 -0.2795 0.7798715

Car_carpool Gamma -0.30736 0.021842 -14.072 7.33E-44

Auto_PR Beta 0.24825 0.594246 0.4178 0.67624705

Auto_PR Gamma -0.29318 0.091847 -3.1921 0.00147533

DParking rates Beta 0.12688 0.399089 0.3179 0.75059936 DParking rates Gamma -0.40271 0.030945 -13.014 5.01E-36 DBavailableParking Beta 0.28921 0.421747 0.6857 0.49302368 DBavailableParking Gamma -0.40178 0.030959 -12.978 7.51E-36

DRK_car Beta 0.96199 0.572839 1.6793 0.0933806

DRK_car Gamma -0.40679 0.031005 -13.12 1.50E-36

(24)

Regression table - General

3 februari 2021 24

Amount_month Beta 0.62455 0.363222 1.7195 0.08751018

Amount_month Gamma -0.06703 0.074266 -0.9026 0.36814312

Amount_km Beta -0.23597 0.329339 -0.7165 0.47475816

Amount_km Gamma -0.05787 0.075055 -0.771 0.44187799

Mob_budget Beta -0.31756 0.245316 -1.2945 0.19555358

Mob_budget Gamma -0.13875 0.01838 -7.5494 5.13E-14

Caf_reg Beta 0.10631 0.138763 0.7662 0.44361711

Caf_reg Gamma -0.13866 0.018381 -7.5435 5.36E-14

SWSR Beta 0.04341 0.116558 0.3724 0.7096164

SWSR Gamma -0.15623 0.018033 -8.6637 5.95E-18

Flex_work Beta 0.00909 0.103228 0.0881 0.92982604

Flex_work Gamma -0.17672 0.018199 -9.7104 4.14E-22

DReg_flexwork Beta -0.25214 0.399698 -0.6308 0.52844028 DReg_flexwork Gamma -0.74447 0.021251 -35.032 6.32E-137 DReg_homework Beta 0.04471 0.293672 0.1522 0.87903306 DReg_homework Gamma -0.40259 0.030944 -13.01 5.20E-36 DReg_otherLoc Beta -0.62139 0.664145 -0.9356 0.34967845 DReg_otherLoc Gamma -0.40126 0.030963 -12.959 9.27E-36