Guidance

Forthcoming change: updates to wider economic impacts

Updated 17 October 2024

Description: updates to wider economic impact units and WITA software

Units: A2.1 (Wider Economic Impacts), A2.2 (Dependent Development; A2.3 (Employment Effects), A2.4 (Productivity Impacts); M5.3 (Supplementary Economic Modelling)

Change announced: October 2024

Expected release date: November 2024

Description

This forthcoming change sets out a compendium of updates to wider economic impact units to incorporate outputs of commissioned and in-house research, and to align with updates to the department’s value for money framework and supplementary guidance. Some updates will also require corresponding changes to the wider impacts and transport appraisal (WITA) software which will be issued after the updated guidance is published. 

Detail

Summary of the changes

In recent years the Transport Appraisal and Strategic Modelling (TASM) division have carried out a large volume of analysis and research projects – many externally commissioned – in line with the Appraisal and Modelling Strategy (AMS) on wider economic impacts (WEIs).

We are now at a stage where we can begin to implement some research outputs into TAG units A2.1-4 and M5.3, as well corresponding updates to the wider impacts transport appraisal (WITA) software, starting with the November 2024 TAG release.

This is a significant package of updates, which are summarised below by TAG Unit. There is a separate annex with further details of each change.

To note that changes with an asterisk* will also have corresponding changes made in WITA.

• All WEIs TAG units (Annex 1)
  • updates to align with the forthcoming update to DfT’s Value for Money framework
  • various corrections and clarifications reported by our research and by practitioners
  • use of new style ‘accessible’ TAG format
• A2.1 wider economic impacts (Annex 2)
  • adding guidance on mapping welfare appraisal outputs to GDP
  • textual updates to reflect changes to A2.2-4
• A2.2 Induced investment (Annex 3)
  • increasing the uplift factor for output change in imperfectly competitive markets (OCICM) from 10% to 13.4%*
• A2.4 productivity impacts (Annex 4)
  • include the agglomeration elasticity for public sector jobs based on the economy average elasticity*
  • assume a gradual ramp up of dynamic agglomeration benefits*
  • refine recommendations on spatial disaggregation (using lower level/granular economic data where possible) to alleviate issues arising from the modifiable areal unit problem (MAUP)
  • updating the map showing functional urban regions in Appendix A
• M5.3 supplementary economic modelling (Annex 2)
  • update to include guidance on mapping welfare to GDP impacts

Anticipated scale of change on appraisal outcomes

The vast majority of changes do not directly impact the methodology, and therefore have no impact on appraisal outcomes. However, some changes will have an impact on the estimated agglomeration benefits, but the net effect is likely to be ambiguous because:

• the public sector will increase welfare benefits, dependent on the share of public sector jobs in the area affected by the transport scheme
• introducing an agglomeration ramp up will reduce the benefit by up to 10% (holding all else equal), due to the fact the TAG appraisal has a long appraisal period and low discount rate^{[footnote 1]}
• adopting a lower spatial unit where possible is likely to have an ambiguous effect on estimated impacts. Given the nature of the MAUP it is not possible to assess the likely impact on business cases given approaches to spatial units will vary
• increasing the OCICM uplift will naturally increase the benefits holding all else equal, benefitting all business cases where this is estimated

Therefore, the net impact of the changes on estimated WEIs is likely to be ambiguous, but depending on the scheme there may be a clear increase or decrease.

WITA software

The updated WITA software will follow shortly after the TAG updates have been published. We aim to minimise the time to ensure practitioners can fully incorporate the updates in their appraisal – the expectation is that the software update will follow within two months of the TAG update.

Annex 1: update to all wider economic impact units

Updates to align with the forthcoming update to DfT’s value for money framework

The WEIs TAG units (A.2) and supplementary economic modelling (M.3) refer to the department’s value for money framework. All wider economic impacts – such as static and dynamic clustering – are categorised as either ‘evolving’ or ‘indicative’ (i.e. level 2 and level 3), and the value for money framework sets out how to handle these as part of the assessment. For example:

• evolving impacts are included in the ‘adjusted’ benefit cost ratio metric
• indictive monetised impacts are considered alongside non-monetised impacts in determining the final value for money category (e.g. low, medium, high)

The department is publishing an updated value for money framework to introduce an indicative BCR metric. Because of this, minor drafting change are needed to ensure these TAG units are aligned and consistent. All changes relate to the drafting only in relation to value for money – it does not change methodologies, assumptions, data or parameter values for estimating monetised impacts.

Various corrections and clarifications reported by our research and by practitioners

As described. Despite our best efforts and proofreading some mistakes do slip through, and we are keen these are addressed routinely. These do not change the substance of the methodology or estimated outputs.

Use of new style ‘accessible’ TAG format

These have been recently introduced and are slowly being rolled as part of the TAG updates. Unit A2.4 was updated in May 2024, and Units A2.1-3 and M5.3 will follow in November.

Annex 2: updates to A2.1 wider economic impacts and M5.3 supplementary economic modelling

Adding guidance on mapping welfare appraisal outputs to GDP

Laird & Byett (2023), Relating transport appraisal to GDP impacts (link forthcoming) develops an approach for deriving GDP impacts from ‘standard’ TAG consistent welfare appraisal results.

Drawing from this, TASM have developed an updated methodology to include in TAG. We are planning to bring forward the following specific TAG A2.1 changes by updating figure 2 (the links between Welfare and Gross Domestic Product) and table 4 (relation of welfare to GDP).

To note Table 1 (Correspondence between national welfare and GDP impacts) in M5.3 will also be updated.

In addition, the guidance will also provide caveats for carrying out this kind of analysis, for example how to handle business user benefits and interpreting GDP metrics (e.g. GDP cost ratio).

Updated versions of table 4 and figure 2 are provided below.

Updated version of Table 4

	Welfare Impact	GDP
User benefits (A1.3)	User benefits from business, commuting and leisure trips	Business user benefits plus user benefits from price reductions for non-work travel
Induced Investment (A2.2) Dependent Development	Land Value Uplift (LVU)	LVU + 2 x development costs for residential development Development costs only for commercial development All estimates need adjusting for additionality
Induced Investment (A2.2) Output Change in Imperfectly Competitive Markets	13.4% of Business User benefits (including reliability benefits)	13.4% of Business User benefits (including reliability benefits)
Employment Effects (A2.3) Labour Supply Impacts	40% of change to GDP (tax revenue)	GDP (= welfare impact / 0.4)
Employment Effects (A2.3) Move to More/Less Productive Jobs	30% of change to GDP (tax revenue)	GDP (= welfare impact / 0.3)
Productivity Impacts (A2.4) Agglomeration Economies (including static and dynamic clustering)	Agglomeration Impacts	Agglomeration impacts
Accidents (A4.1)	Based on VPF and injury values	15% of road accident impacts Other modes: 30% of fatal injury impact, 10% of serious and 15% of slight
Physical Activity (A4.1)	Benefits calculated using the AMAT tool	Absenteeism benefits, plus 30% of the reduced mortality benefits
Air quality (A3)	Welfare impact taken from Defra AQ damage costs	20% of welfare impact

Updated version of figure 2.

Updates to reflect changes to A2.2-4

Given the significant package of changes, A2.1 will be amended accordingly to be consistent with the changes made in A2.2-4 and M5.3.

Annex 3: updates to A2.1 wider economic impacts and M5.3 supplementary economic modelling

Increasing the uplift factor for output change in imperfectly competitive markets from 10% to 13.4%

Output change in imperfectly competitive markets (OCICM) is appraised by applying an uplift factor to the benefits usually calculated for a perfect market. The current 10% factor were adopted in 2005 following initial work by SACTRA in the 1990s.^{[footnote 2]}

The department commissioned Geoffrey Hyman Consultancy and Oxera (link forthcoming) to undertake two separate pieces of work, to inform an update the uplift factors. The main outputs of the research is that the appropriate uplift factor has increased to 13.4%. This reflects (i) evidence of increasing price-cost margins over the past two decades and (ii) a technical improvement in how indirect taxation and the unit of account is handled within the OCICM uplift. This work also addresses some long standing issues with how the original estimate deals with the assumed market structure, number of competing firms, and cost-price pass through parameter.

Annex 4: changes to A2.4 productivity impacts

Include the agglomeration elasticity for public sector jobs based on the economy average elasticity

Following recommendations made in Laird and Tveter (2023) Agglomeration and transport appraisal: new developments and research directions (link forthcoming), Laird and Tveter (2024) Agglomeration and the public sector (link forthcoming) explored the case for the inclusion of an additional sectoral agglomeration elasticity for the public sector.

The paper concluded that it would be appropriate to use the economy average elasticity to quantify public sector impacts. On this basis, we are introducing this in the guidance and WITA software to enable the calculation of public sector impacts within transport appraisal, and will be included in the core indicative BCR as part of the value for money assessment. This follows the inclusion of public sector data in TAG A2.4 and the wider impacts dataset in May 2024.

Assume a gradual ramp up of dynamic agglomeration benefits

Laird and Tveter (2023) explains that there is likely to be a long ramp up before agglomeration economies appear. This is because there is a time lag for the dynamic effects to feed through (e.g. learning), whereas static agglomeration mechanisms (e.g. matching or sharing) occur instantaneously. Evidence suggests dynamic agglomeration mechanisms may be around 50% of the total agglomeration impact and take up to 10 years to fully appear following a worker’s relocation.

It is important to distinguish between static and dynamic agglomeration mechanisms, and static and dynamic clustering. Static clustering is where effective density changes due to changes in generalised travel cost (GTC). Dynamic clustering is where there are additionally changes in land-use and employment location, which cause further changes in effective density.

In contrast, the terms static and dynamic agglomeration refer to different underlying agglomeration mechanisms (e.g. sharing, matching and learning). Static clustering leads to agglomeration impacts via both static and dynamic mechanisms. Dynamic clustering, where estimated, will subsume and include static clustering impacts.

In addition, dynamic clustering impacts will operate via both static and dynamic mechanisms. TAG A2.4 §2.6.6 recommends dynamic clustering impacts are broken down into 2 components:

• the change in productivity due to GTC driven changes in effective density
• the change due to land-use and employment location effects

An approximate method is given in TAG for estimating this breakdown. The figure below illustrates these points.

Gradual ramp up in agglomeration benefits set out in Laird and Tveter (2023).

We will implement a proportionate approach to the agglomeration ramp up in TAG and WITA, which will apply to 50% of all agglomeration benefits over 10 years (static and dynamic), with the other 50% accruing in full at scheme opening, in line with Laird and Tveter (2023). The ramp up factors increment linearly from 0.5 in year 1 to 1.0 in year 10, where year 1 is on scheme opening (although there may be cases where before scheme opening is appropriate). The same method will apply when static and dynamic clustering is estimated, and this is how it will be implemented in WITA.

For static clustering, the ramp up captures the static agglomeration mechanisms which are instantaneous, while dynamic agglomeration mechanisms (e.g. learning, one of the three micro-mechanisms of agglomeration) will take years to fully materialise.

However, we acknowledge that this risks overestimating the dynamic clustering component of the benefits. This is because there are frictions between the land-use effect (e.g. the worker moving from one city to another) and agglomeration economies forming as set out in Laird and Tveter. Nevertheless, it makes sense to assume at least some of the effect occurs on scheme opening. This is because there may be pre- scheme opening effects (mostly relevant to the largest schemes), and because dynamic clustering includes static clustering.

Where possible, these land-use change lags should be represented in the SEM being used, and then fed into the appraisal of wider impacts directly. However, in the absence of specific modelled lags, we will recommend an additional default 5 year ramp each time there is a change effective density from land use change (i.e. any productivity impact attributable to employment relocation) to represent this lag. This would mean there could be multiple ramps if there is land use change in multiple years. However, this would need to be implemented outside of WITA (i.e. off model). This lag is additional to, and needs to be applied in conjunction with, the 10 year ramp up discussed above.

We also plan to clarify in M5.3 that in some cases there may be pre-opening effects in anticipation of investment, as part of a future suite of updates to M5.3.

Refine recommendations on spatial disaggregation

Laird and Tveter (2023) makes recommendations on spatial level of agglomeration calculations.

The current TAG method for estimating agglomeration is susceptible to the Modifiable Areal Unit Problem (MAUP), a phenomenon which occurs when aggregating spatial data which can be minimised but not eliminated.  As a result, this may lead to upward biases in the estimated agglomeration benefits, with errors of up to 300% possible.

The paper recommends that agglomeration calculations should be undertaken at as small a level as practical, and the proposal is that TAG recommends this approach. However, we acknowledge there are barriers to practically implement this (e.g. WITA cannot be run at sub local authority district level).

The department will work to provide resources to practically implement this over time – for example, we are planning to carry out major research to update agglomeration parameters and datasets at a finer spatial level (e.g. MSOA).

Updating the map showing functional urban regions (FURs) in Appendix A

The updated map is still based on the same middle layer super output area (MSOA) boundaries as the previous map, but urban centres are now labelled and the colour scheme is changed to greyscale for increased clarity. This means there are no changes required to the current FURs dataset (XLSX, 496 KB).

Updated FURs map.

Contact

For further information on this guidance update, please contact:

Transport Appraisal and Strategic Modelling (TASM) Division
Department for Transport
Zone 2/25 Great Minster House
33 Horseferry Road
London
SW1P 4DR

tasm@dft.gov.uk

---

1. Laird and Tveter (2023) Agglomeration and transport appraisal: new developments and research directions [link forthcoming] states “For countries with low discount rates and long appraisal periods, such as the UK, then the error is more like a 10% overestimation.” This is interpreted as “up to 10% reduction” since 1/1.1 -1 ≈ 9.1% . ↩

2. See Department for Transport (2005), Transport, Wider Economic Benefits, and Impacts on GDP (PDF, 446KB. ↩