HS2: A rocking-chair with three legs

The HS2 has two problems according to Sir Terry Morgan, who resigned as the Chairman of HS2 last month. The first problem according to Sir Terry Morgan is that the name is misleading. The emphasis should not have been on high speed but on adding capacity to the existing network. The second problem is the cost. He suggested that in order to keep the cost down, the line should stop at Old Oak Common in Northwest London. The question remains what is the point of adding capacity if the line does not reach Central London.

It also lays bare the limitations of the cost-benefit paradigm. The cost-benefit paradigm presumes that costs and benefits can be explicitly estimated and compared to pin down the net benefit to the society. If estimated benefits exceed estimated costs, then the projects is weakly a Pareto-improvement, i.e., it can potentially make people better off after the people who are worse off are suitably compensated.

From a practical perspective, costs can all be converted to pecuniary units and easily added up. Conversely, benefits are multi-dimensional and largely non-pecuniary. It is very difficult to estimate each dimension, let alone find some mechanism to add these various non-pecuniary dimensions. A bit like adding apples and oranges. The math just does not add up. Costs are more precisely estimated and benefits are effectively assumed. That is, there are usually a number of hidden assumptions in the methodology that allow benefits to be added up and given a pecuniary value that can then be compared to the benefits. If you are sceptical ask yourself how would you compare the benefits to HS2 line being built to Euston or Old Oak Common.

Sir Terry Morgan’s frank testimony at the Economic Affairs Committee on 22 January 2019 reveals how cost-benefit analysis works in practice. For any large troubled project, the focus often moves to minimising the cost, without mapping the resultant loss in benefits. This is because benefits are vague and cannot be easily estimated. For people who claim that cost-benefit analysis is flawed yet best available methodology, the very least you can say is that putting numbers on unknown unknowns is not a methodology. It is called stabbing in the dark.

HS2 may have been an unwise project to start without fully understanding the pattern of benefits to society. Yet, having sunk the cost into HS2, it needs to be brought to Central London. No one buys a chair with three legs however cheap its. So, why should we as a society pay for a high capacity line from Manchester to Old Oak Common.


Hart, O., Shleifer, A., & Vishny, R. W. (1997). The proper scope of government: theory and an application to prisons. The Quarterly Journal of Economics, 112(4), 1127-1161. Link

Kanbur, R. (2005). Pareto’s revenge. Cornell University. Link

Spatial Peer Effects Among Children

First day of the Econometric society World Congress meeting. Interesting paper by Helmers and Patnam this morning. The paper tries to disentangle peer effect of and common (covariate) effects on children’s reading and writing achievements. The paper was interesting because it was using networks to disentangle these effects and not through the usual experimental framework. It is a neat idea. Using the Young Lives survey from Andhra Pradesh, the paper maps spatial (strictly geographic networks). Then it uses the networks to identify the the clusters. Certain individuals have multiple links within the clusters whereas certain other individuals are linked to the cluster through just one link. This variation allows the authors to identify the peer effect. They also look at the insurance component of the peer group. For this, in a really neat trick, they use idiosyncratic shocks are IVs.

The paper is very useful and shows us that we can use observation data and social network maps to disentangle various effects. Of course, it takes the spatial network as given. It may (may not be) be a reasonable assumption. After all, there may be some strategic relocation by the household. The other caveat is that it does not map the network completely. The network mapped (or the information available) is only for the children’s spatial peer network. The paper is not able to distinguish the effects that run through the children’s peer network from the effects that run through the adult’s network. In spite of these caveats, I really like the paper, not for what it is right now, but more for what it maybe lead to.

Looking forward to John Moore’s presidential address this afternoon. The title of the talk is contagious iliquidity. Very topical and may have a flavour of his credit cycles paper. I am sure it is vintage John Moore stuff. Entertaining with lots of stories involving red and blue stuff. More on Moore later on in the evening.

Christian Helmers and Manasa Patnam. Does the Rotten Child Spoil His Companion? Spatial Peer Effects Among Children in Rural India

Progress in Rwanda

Just read an interesting account of Rwanda 16 years on from the ethnic conflict that brutalised and fractured the civil society. The visible progress that Rwanda has made is note worthy and impressive. The economic growth, albeit from a low base, has led to the incomes tripling in the last 16 years. There is considerable interest from the global business community in the country. The question is whether economic progress can reduce the probability of future violence in the country or may actually increase the probability of violence. We need a framework to think about this.

Lets take the tunnel effect described by Hirschman and Rothschild (1973). If you are sitting in the a traffic jam in a tunnel and you suddenly find that that the other side starts moving, you get excited thinking that the movement indicates that things are getting resolved and you would move soon as well. If the other side keeps moving and you are stuck, frustration increases and may reach a boiling point where you decide to crash you car and stop the other side from moving. This describes the aspect of political violence seen across the world when the economic gains are concentrated and not shared widely.

Lets look at it from Krugman’s “History versus Expectations” perspective. Krugman shows that equilibrium could either be determined by either historical factors or by the expectations of the future. The shared history in the tunnel example is the fact that you got stuck with your fellow passengers in the tunnel. The shared expectations is the based on the information you have about when things would improve. Whether the shared history or the shared expectations determine the outcome depends on which one dominates. The information you have about what the future has in store comes from the movement of the vehicles on the other side. At first when the vehicles start moving, the shared expectations of the problem getting resolved very soon dominates. As the other side keeps moving and you remain stuck, your and your fellow traveller’s expectations about the future changes, leading to either a small subset crashing their car to stop the other side from moving.

There is an obvious behavioural explanation for why stopping the other side from moving increases you perceived utility. There is also an non-behavioural explanations, which comes from the membership of social networks or communities. Social networks or community ties emerge as a way to facilitate the societies’ activities in the absence of a wider effective government and functioning markets. It solves the information and enforcement problems associated with public goods. An effective government can make the need for these social ties obsolete by providing the public goods and access to market. In the absence of the public goods and markets, need for social ties increases. There is clear relationship between the strength of social ties and acess to markets and public goods. The members of a community choose the investments they make in social ties given the state of effectiveness of government. Given a weak or absent state, crashing the car may be the investment that community members are required to make in order strengthen their own social ties and in turn strengthen their social identity. There are interesting papers from across the world that show that local public goods provision is easier in the areas where the ethnic diversity is lower. (Alesina, Baqir, and Easterly, 1999) I am not aware of papers that take social ties as endogenous and show that the strength of these ties depends on other exogenous factors, which in turn leads to government’s effectiveness in providing public goods and market access.

The solution to ethnic violence lies in an government’s ability to co-ordinate the respective communities’ expectation about the future. This needs to be done in such a way that gains from the future economic progress outweighs the problems of the past. The incentives should be such that there is no gains for even a small sub-group to under-take the violent route. People speculate that it was economic progress both in Ireland and United Kingdom that led to the Northern Island conflict getting resolved. Of course, the expected gains from future economic progress needs to be sufficiently wide spread for this to happen.

In Rwanda’s case, it would be interesting to see whether the progress hitherto has resulted in the benefits from the economic growth of the last 16 years being concentrated in the urban areas where the emerging middle class live or it has had a wider impact in the country. It would be important to get a sense of what different part of the Rwandan society expect from the future. Kagame certainly seems to have improved the public goods in the urban areas but the relevant question is whether his government has allocated sufficient public funds to improve the public goods and facilitate access to the markets in the rural areas?

One obvious way to eliminate the source of the past conflict is for the government to act as a co-ordination device and to dramatically increase the future economic gains from peace so that the groups collectively choose the peace strategy over the violent strategy. So, the future equilibrium can be driven by expectations and overwhelm the difficult history in the country. The government needs to ensure that expectations starts dominating the history in Rwanda.

With this in mind, it would be very interesting to see how the incomes, pubic goods and market access have fared in the rural countryside across the country. Actually, looking at the micro data it may be possible to predict the future trouble spots, if any. Gujarat was racked by politically motivated violence along religious lines in 2001. Pande, Field, Levinson and Visara (2008) is a interesting retrospective econometric study that carefully looks at the pattern of violence in Gujarat’s capital city, Ahmedabad. It explains that pattern of violence in the terms of historical economic factors, i.e., neighbourhood’s proximity to derelict textile mills. The subtext is that historical economic shocks determined the pattern of violence in the city.

Recent Articles in the Press:

Rwanda’s other genocide, Foreign Policy, 2 September 2010.

Revisiting the killing fields, Economist, 2 September 2010.


A. Alesina, R. Baqir, and W. Easterly (1999). Public goods and ethnic divisions *. Quarterly Journal of Economics, 114(4):1243–1284. Link.

Albert O. Hirschman and Michael Rothschild (1973). The Changing Tolerance for Income Inequality in the Course of Economic Development. The Quarterly Journal of Economics, Vol. 87, No. 4 (Nov., 1973), pp. 544-566. Link.

Paul Krugman (2001). History Versus Expectations. The Quarterly Journal of Economics, Vol. 106, No. 2 (May, 1991), pp. 651-667. Link.

Rohini Pande, Eric Field, Matthew Levinson and Sujata Visara (2008). Segregation, Rent Control, and Riots: the Economics of Religious Conflict in an Indian City. American Economic Review Papers and Proceedings, May 2008, Vol. 98 (2): pp. 505-510. Link.