The National Infrastructure Commission commissioned a team of academics and researchers at the IFS and UCL to create a software tool that estimates how land values respond to changes in land purpose or infrastructure improvements. The tool is now available online here, and this page provides some background on, and documentation for, the project.
Background to the tool
The National Infrastructure Commission (NIC) is an independent body that provides central and local government with impartial advice on long-term infrastructure challenges.
A team of researchers at the IFS and UCL were commissioned by the NIC to produce a publicly-available software tool that estimates how land values respond to changes in land purpose or infrastructure improvements. Specifically, the commission was interested in:
- How does the value of existing properties would respond to investment in new or improved road and rail infrastructure
- What would the value of properties be in a new hypothetical settlement constructed where one does not currently exist
They specified that the tool should be detailed and flexible, providing estimates for a range of potential investments at the lowest geographical level possible. It was envisaged that the tool could include help inform government investment decisions by providing estimates of the size and distribution of specific project benefits, and help local authorities in their planning negotiations with developers, by giving them more information about the potential value of new properties built in their area.
Both of the commission’s questions are difficult to answer, for two key reasons:
The first is that it’s difficult to separate the value of local infrastructure from the other features of a house and its local area. This is especially problematic if proximity to good local infrastructure (like well maintained roads, and fast buses to the city centre) might be correlated with other desirable things like local parks or low flood risk.
The second is that current property prices reflect how much the person who lives there wants to pay for that house, not how much everyone would like to pay for it. This is a problem because when you change transport network, people are likely to move in response. For example, people who value good public transport more may move in when you improve the speed of a bus route, and those who value it less may move out. This makes it hard to predict the price changes that will result from the investment because we only have information about the valuation of people who currently live there. It also makes it conceptually and practically difficult to work out the value of a project – Do you use the valuation of the person who lived there before? Or the one who moved there afterwards? And how much would the new tenant have paid to live there before the new bus route? The economics of going between price changes and overall “benefits” is much more complex than it might at first seem.
As a result of these two challenges existing methods for estimating the benefits of infrastructure investments either rely on expert judgement, or relatively simple modelling.
The project team saw this work as the first step towards developing a new, implementable methodology for evaluating transport investments. The aim was to combine the best lessons from the forefront of the academic literature, with the large-scale administrative datasets that are increasingly available and accessible in recent years. In particular, we used a hedonic approach to address the first challenge, as it enables us to separately estimate the contribution of infrastructure to house price.
As a result, our approach estimates the value of new investments based on what people who currently live in an area value. Our findings do not take into account large scale changes in the macroeconomic environment, or long run effects resulting from changes in job locations, commuting or residential patterns. Nonetheless, the results produced by the tool can be considered the best short run estimates available given existing data and the timescale of the project, and an excellent starting point for future analysis of public investment and planning decisions.
The next steps of our work are tackling this second problem – that of mobility responses to infrastructure investments – using what’s called a sorting equilibrium model. This means that we will model people’s decisions about where to live and work, use that to predict how they would respond to changes in the transport network, and ultimately provide better estimates of the value of new investments.
The primary output of this project is the “Property Value Uplift” tool: a web application that can be used to evaluate how the price of existing property might change in response to a small scale investment in roads or public transport. Users can also input details of a prospective settlement and predict the value of properties at that site. The tool is very flexible, allowing the user to be very specific about the type and location of investment they’d like to evaluate. It is also very detailed, predicting the impact on the full distribution of house prices in the area (not just what happens to average house prices), and highlighting how localised the benefits of transport infrastructure can be. Of course the results are subject to the health warning discussed above.
Underpinning the tool are two essential components.
Firstly, a comprehensive property-level database for England that contains information about the price of a property, as well as its size, type and characteristics of its location (such as travel time to jobs, noise exposure and risk of flooding). The database combines a wide variety of publicly available data sources, and some restricted licence data. Available at the bottom of this page are a document containing details of the data used, a subset of the database (containing only publicly available data), guidance on obtaining access to the restricted access data sources, and computer code that can be used to clean and match in these data sources.
Secondly, using this data we have produced a set of econometric results that relate property characteristics to price. These results are the basis for the predictions made in the tool. More information about our econometric approach and results will be available soon in our final report.