[Posted by Prof Henry G. Overman]
In light of today's HS2 vote, a quick post on where things stand in terms of our understanding of costs and benefits.
1) Other ways of spending the money: Using standard methods HS2 has a benefit cost ratio of 1.8. Including wider economic impacts this rises to 2.3. This does not represent high value for money. For comparison see, for example, the BCRs for projects considered by Eddington. These numbers do not include wider economic impacts, so should be compared to the 1.8 figure for HS2. Evidence considered by Eddington doesn't suggest that the wider economic impacts are particularly big for inter-city schemes. If anything they are stronger for intra-city schemes, so if we could add in the wider economic impacts for the Eddington projects this would make the relative return on HS2 look worse.
2) Other ways of addressing the north-south rail capacity 'problem'. One option would be different pricing. This option has not been properly considered. A second set of options involve improvements to the existing networks. As outlined in section 6 of the strategic case the alternative to the full-Y has a BCR of 3.0 (including Wider Economic Impacts) considerably higher than HS2. The case for HS2 suggests that there would be lots of disruption associated with these alternatives. These should be costed, compared to the disruption costs for HS2 and the BCR's adjusted accordingly. Unfortunately, this hasn't happened because 'the alternatives are at an early stage of development'. Whatever you think of HS2 it is deeply depressing that parliament is being asked to vote on whether to initiate the project when we haven't fully considered the alternatives.
3) In this kind of appraisal, the standard BCR and the wider economic impacts represent the 'knowns' and the 'known unknowns', respectively. There could be other benefits out there that this modelling doesn't capture. That said, I don't think these other elements are likely to be large given the evidence to date in the academic literature. For example, the increased accessibility from HS2 might generate fast growth as a result of 'agglomeration benefits'. However, most of the academic literature suggests that these effects are more important at small to medium scales (i.e. within cities) rather than large scales associated with HS2. That suggests that these additional effects might improve the central BCR a little, but not a lot. This would also be true for other schemes so it doesn't particularly strengthen the case for HS2.
4) There is a large amount of uncertainty around the benefits and costs for HS2. In these circumstances it is most appropriate to take the central estimate (1.8-2.3) and discount it if we care about the risk. It is not appropriate to only look at the upside risk and argue that the gains could be massive or to pretend that the central estimate is uninformative and BCR pointless.
Overall, on the balance of the evidence that we have, HS2 is not an awful project (it delivers benefits that exceed the costs) but its not a great project either (there are other options for addressing the problem that we should consider and there are better ways of spending the money).
Thursday 31 October 2013
Tuesday 29 October 2013
The Strategic Case for HS2
[Posted by Prof Henry G. Overman]
A quick reaction to the government's new Strategic Case for HS2 and revised Economic Case.
1) The BCR hasn't changed much (it's been revised down as a result of a number of revisions, including to costs)
2) The BCR figures are presented differently. They now systematically include Wider Economic Impacts. As far as I can tell, this runs counter to DfT's guidance (p.5) which says that BCR should not include Wider Economic Impacts.
3) Excluding Wider Economic Impacts, consistent with DfT guidance, reduces BCRs. In terms of DfT's Value for Money guidance, phase 1 is low value for money and the full Y network is medium value for money (not high value, as stated in the Strategic Case).
4) The Economic Case provides the distribution of BCR around the central estimate. But other projects would also show variance around the central BCR figure. If the variance of the BCR for HS2 is higher this indicates that it is a more risky project so we should expect higher average BCR (to compensate us for taking on risk). The report often seems to imply the opposite of this ('look how high the BCR might be!')
5) The new Strategic Case shifts the focus to arguments around capacity and overcrowding on parts of the line. I find this confusing for a number of reasons. First, we have overcrowding on many parts of the network (some of it much more severe than the crowding experienced in to Euston). Second, transport improvements can't hope to meet all demand so there will always be capacity issues on the network. Third, focusing on crowding per se ignores questions about the value of journeys and the costs of providing that capacity. These should be captured in a good BCR.
6) In terms of alternatives to HS2, the Strategic Case rules out the possibility of raising prices because this wouldn't deliver the accessibility and growth benefits that you get from HS2. This is true, but it would also avoid the costs. This simply points us back to the BCR as an appropriate comparison to the alternative of pricing to deal with capacity constraints. If the BCR for HS2 doesn't look good then we should consider pricing as an alternative way of managing demand.
7) Another alternative would be to improve existing lines. This possibility has been considered by DfT and has a BCR (including WEI) of 3 - considerably higher than for HS2. We are told that this would involve 14 years of disruption at the weekends. Again, I'm confused by the focus on number of years this would take, rather than some assessment of the disbenefits. If we had that, we could adjust the BCR for the alternative options and see how this compares to HS2. [I also think it's unrealistic that we won't need some improvements on these lines before HS2 is completed and I'm sure that HS2 causes disruption which is ignored in the case for HS2]
8) All of this is about option generation for fixing capacity problems on the North-South train lines. That ignores the opportunity costs of these funds. Eddington's 2006 report for HMT suggest that these are substantial because we can identify alternative projects that address other capacity/connectivity problems and that would have higher BCR.
9) The BCR is a useful input in to decision making. But it is true that BCR doesn't capture all of the economic benefits of HS2. The government's preferred estimate for these additional benefits is somewhere between £8bn and £15bn per year. My blog post from September explains why I think these estimates are too high by orders of magnitude. These additional growth impacts can't be directly compared to BCR (or added to them). Unfortunately, we don't have estimates of the additional growth benefits that would come from spending the £50bn on other transport projects, so we don't have a good point of comparison for these numbers. That said, the additional impacts of other schemes could be substantial (see, e.g., the low BCR for the Jubilee line which clearly under-estimated the overall benefits of the scheme).
There are other issues (e.g. around the demand cap assumptions) but that's all I have time for, for today.
A quick reaction to the government's new Strategic Case for HS2 and revised Economic Case.
1) The BCR hasn't changed much (it's been revised down as a result of a number of revisions, including to costs)
2) The BCR figures are presented differently. They now systematically include Wider Economic Impacts. As far as I can tell, this runs counter to DfT's guidance (p.5) which says that BCR should not include Wider Economic Impacts.
3) Excluding Wider Economic Impacts, consistent with DfT guidance, reduces BCRs. In terms of DfT's Value for Money guidance, phase 1 is low value for money and the full Y network is medium value for money (not high value, as stated in the Strategic Case).
4) The Economic Case provides the distribution of BCR around the central estimate. But other projects would also show variance around the central BCR figure. If the variance of the BCR for HS2 is higher this indicates that it is a more risky project so we should expect higher average BCR (to compensate us for taking on risk). The report often seems to imply the opposite of this ('look how high the BCR might be!')
5) The new Strategic Case shifts the focus to arguments around capacity and overcrowding on parts of the line. I find this confusing for a number of reasons. First, we have overcrowding on many parts of the network (some of it much more severe than the crowding experienced in to Euston). Second, transport improvements can't hope to meet all demand so there will always be capacity issues on the network. Third, focusing on crowding per se ignores questions about the value of journeys and the costs of providing that capacity. These should be captured in a good BCR.
6) In terms of alternatives to HS2, the Strategic Case rules out the possibility of raising prices because this wouldn't deliver the accessibility and growth benefits that you get from HS2. This is true, but it would also avoid the costs. This simply points us back to the BCR as an appropriate comparison to the alternative of pricing to deal with capacity constraints. If the BCR for HS2 doesn't look good then we should consider pricing as an alternative way of managing demand.
7) Another alternative would be to improve existing lines. This possibility has been considered by DfT and has a BCR (including WEI) of 3 - considerably higher than for HS2. We are told that this would involve 14 years of disruption at the weekends. Again, I'm confused by the focus on number of years this would take, rather than some assessment of the disbenefits. If we had that, we could adjust the BCR for the alternative options and see how this compares to HS2. [I also think it's unrealistic that we won't need some improvements on these lines before HS2 is completed and I'm sure that HS2 causes disruption which is ignored in the case for HS2]
8) All of this is about option generation for fixing capacity problems on the North-South train lines. That ignores the opportunity costs of these funds. Eddington's 2006 report for HMT suggest that these are substantial because we can identify alternative projects that address other capacity/connectivity problems and that would have higher BCR.
9) The BCR is a useful input in to decision making. But it is true that BCR doesn't capture all of the economic benefits of HS2. The government's preferred estimate for these additional benefits is somewhere between £8bn and £15bn per year. My blog post from September explains why I think these estimates are too high by orders of magnitude. These additional growth impacts can't be directly compared to BCR (or added to them). Unfortunately, we don't have estimates of the additional growth benefits that would come from spending the £50bn on other transport projects, so we don't have a good point of comparison for these numbers. That said, the additional impacts of other schemes could be substantial (see, e.g., the low BCR for the Jubilee line which clearly under-estimated the overall benefits of the scheme).
There are other issues (e.g. around the demand cap assumptions) but that's all I have time for, for today.
Tuesday 22 October 2013
Game of Zones
Posted by Sevrin Waights, LSE
Conservation areas seem like an uncontroversial policy. They protect the visual appearance of historic neighbourhoods – for example, by preventing your neighbours installing modern ‘improvements’ such as PVC windows.
Conservation areas seem like an uncontroversial policy. They protect the visual appearance of historic neighbourhoods – for example, by preventing your neighbours installing modern ‘improvements’ such as PVC windows.
But this
protection comes at a cost: you can no longer renovate your less-than
watertight roof, add that conservatory you always wanted, or chop down the tree
in your garden that is blocking your light. These costs and benefits are
considered by would-be buyers and will therefore capitalise into your house price.
So the obvious question is: if you were a homeowner, would you
want to be designated or not? This leads to a further question: can homeowners
game the designation process to their advantage? These are the questions I try
and answer in new research for SERC, working with my colleagues Gabriel Ahlfeldt, Kristoffer Moeller and Nicolai Wendland.
Let’s
imagine a scenario where conservation areas (CAs) are designated by an
all-powerful ‘social planner’, who acts according to homeowner interests. Heritage-rich
neighbourhoods would get CA status because the benefits of preservation will
outweigh the costs of the restrictions. In turn, CA designation should increase
house prices in those neighbourhoods.
If
designation status was solely determined
by heritage, the planner would quickly run out of neighbourhoods to designate.
But as we can see, new CAs have been designated in every year since the
policy's introduction in 1967:
Figure: Percentage of all CAs designated in
each year
So what’s
driving the newer designations? Neighbourhoods’ ‘heritage’ is fairly fixed, but
the value placed on heritage can change a lot with neighbourhood composition. Areas
that are gentrifying, for example, typically have rising shares of
higher-income homeowners with degrees . These incomers have substantial buying
power, and are both willing and able to pay more for 'heritage' - especially if
it is secured by conservation area status.
Since
gentrification increases the potential returns to CA status we should see
designation continue even as the stock of 'historic' neighbourhoods
is used up. In these newer CAs, 'heritage' status becomes more subjective.
In our paper
we test these predictions. First we examine the link between gentrification and
designation, using data on conservation areas provided by English Heritage and on
neighbourhood composition from the UK Census (1991 and 2011). We find that a 1%
increase in the degree share is associated with an 11% increase in the
designated land share – a strong effect.
Second, we
test our house price predictions by examining over 1 million property transactions.
As predicted, we find a zero price effect immediately on CA designation,
followed by subsequent house price growth.
This tells
us that designations appear to adhere to homeowner economic interests, implying
that they can influence the process. How might this happen in practice? Local
politicians like doing popular things, and CA designation is certainly popular
with homeowners, who tend to turn out in force in local elections. In the UK, public
engagement is also actively encouraged when neighbourhoods are nominated as
potential conservation areas. Since homeowners are typically better able to
form special interest pressure groups this may work to their advantage. All
this supports the ‘homevoter’ hypothesis that homeowners have a strong
influence over the political decisions that affect house prices. When you play
the game of zones, then, you
win if you own.
Tuesday 1 October 2013
Do student satisfaction ratings affect university choices? New evidence about the National Student Survey
Posted by Steve Gibbons, LSE and SERC
The season is here when next year’s school leavers start filling in their UCAS forms and applying to university. Yet, as any of us who know someone in this position will agree, picking a university is not always easy. For most subject areas, there are a large number of universities to choose from, and making a choice can involve a lot of research.
To help students with these difficult choices, The National Student Survey (NSS) was introduced in the mid-2000s to provide information about students’ satisfaction with their degree course. This survey has captured the attention of university lecturers and administrators, underpinned by concerns about the impact of scores on future recruitment. NSS scores are also one of several quality indicators used in the “league tables” published in newspapers and guidebooks. But do students really take any notice of satisfaction scores in making their university choices, or are other factors more important?
The first study to look directly at the effect of NSS scores on university applications was published yesterday as a SERC Discussion Paper. The research, carried out with my colleagues Professor Eric Neumayer and Dr Richard Perkins in the Department of Geography and Environment, links data on the NSS to applications data from UCAS for each subject and university from 2006 to 2011.
The study shows that NSS scores do matter for university applications. However, contrary to what many university managers might think, the effect of changes in NSS scores on demand for places is quite small. Moving from the bottom of the scale (around 65% satisfaction) to the top of the scale (about 95% satisfaction) will only result in a degree course gaining about seven more applicants for every 100 it already receives. And it isn’t the separately published NSS data themselves which matter. The timing of the effects of the NSS scores, relative to their publication, reveals that the NSS impact on student demand works indirectly by affecting university rankings in published league tables, such as The Times Good University Guide.
A likely explanation for this result is that leagues tables, which rank departments on a single scale, are more readily available and easily understood by applicants than more complex NSS data. In other words, the format in which information is presented to prospective students matters. But here too the effects are small: a 10 place move up a table of 100 universities only increases applications by around 2-3%. An improvement in position encourages a slightly more able pool of applicants (based on A- level tariff points). The effect of changes in league table position is also slightly bigger in subjects and places where there are more providers from which to choose.
Yet, across the board, student demand does not appear to be very responsive to improvements in NSS scores or league table positions. Instead, school leavers seem primarily to choose on a rather different set of criteria, related to more persistent factors (e.g. academic reputation, or distance from home - for which see my earlier work). None of this is very good news in the short-term for those universities who have invested significantly in measures to improve student satisfaction in the hope that they will profit from increased demand. The NSS may nevertheless be good news for students who benefit from these investments in terms of a superior educational experience.
The season is here when next year’s school leavers start filling in their UCAS forms and applying to university. Yet, as any of us who know someone in this position will agree, picking a university is not always easy. For most subject areas, there are a large number of universities to choose from, and making a choice can involve a lot of research.
To help students with these difficult choices, The National Student Survey (NSS) was introduced in the mid-2000s to provide information about students’ satisfaction with their degree course. This survey has captured the attention of university lecturers and administrators, underpinned by concerns about the impact of scores on future recruitment. NSS scores are also one of several quality indicators used in the “league tables” published in newspapers and guidebooks. But do students really take any notice of satisfaction scores in making their university choices, or are other factors more important?
The first study to look directly at the effect of NSS scores on university applications was published yesterday as a SERC Discussion Paper. The research, carried out with my colleagues Professor Eric Neumayer and Dr Richard Perkins in the Department of Geography and Environment, links data on the NSS to applications data from UCAS for each subject and university from 2006 to 2011.
The study shows that NSS scores do matter for university applications. However, contrary to what many university managers might think, the effect of changes in NSS scores on demand for places is quite small. Moving from the bottom of the scale (around 65% satisfaction) to the top of the scale (about 95% satisfaction) will only result in a degree course gaining about seven more applicants for every 100 it already receives. And it isn’t the separately published NSS data themselves which matter. The timing of the effects of the NSS scores, relative to their publication, reveals that the NSS impact on student demand works indirectly by affecting university rankings in published league tables, such as The Times Good University Guide.
A likely explanation for this result is that leagues tables, which rank departments on a single scale, are more readily available and easily understood by applicants than more complex NSS data. In other words, the format in which information is presented to prospective students matters. But here too the effects are small: a 10 place move up a table of 100 universities only increases applications by around 2-3%. An improvement in position encourages a slightly more able pool of applicants (based on A- level tariff points). The effect of changes in league table position is also slightly bigger in subjects and places where there are more providers from which to choose.
Yet, across the board, student demand does not appear to be very responsive to improvements in NSS scores or league table positions. Instead, school leavers seem primarily to choose on a rather different set of criteria, related to more persistent factors (e.g. academic reputation, or distance from home - for which see my earlier work). None of this is very good news in the short-term for those universities who have invested significantly in measures to improve student satisfaction in the hope that they will profit from increased demand. The NSS may nevertheless be good news for students who benefit from these investments in terms of a superior educational experience.
Subscribe to:
Posts (Atom)