Energy and Climate Change CommitteeWritten evidence submitted by Alex Henney

“THE KING NEVER BUYS CHEAPLY”
ATTRIBUTED TO SAMUEL PEPYS

Executive Summary

The British roll-out of smart meters is one of the most incompetent, one of the most expensive, and definitely the most complex because of the supplier lead and the centralised comms system (whose licence runs to 170 pages). The project is likely to be a shambles which will have negligible consumer benefit. I titled an article published in New Power in October “Smart metering—a case study in Whitehall incompetence”.

The initial cost-benefit analysis in 2007 by a reputable firm of consulting engineers came up with a net-disbenefit of minus £4 billion. This “gave the wrong answer” so the civil servants cooked the numbers to come up with a net benefit of plus £4.9 billion in 2011. And if one believes that, one may also believe pigs can fly. The Energy Demand Research Project found (from a self selected sample) a saving of consumption of 3%, which is trivial compared with the Committee on Climate Change’s estimate of a potential 20% reduction from improved efficiency of lighting and appliances by 2020 which will be realized without any governmental song and dance. The following figures for the all-in installed cost for smart electricity meters (2009 prices) highlight the high costs:

	Britain	ENEL (Italy)	ACEA (Italy)	Iberdrola (Spain)
All in cost per meter (£)	135	65	75	70
Programme cost for 30 million electric meters (£bn)	4.05	1.95	2.25	2.1

Although part of the additional cost is due to the ill-judged proposal to provide in-home displays—many of which will not be used—for a cost of £600 milion, the rest is due to the structure of the arrangements. (I would not be surprised if there is not much more wasted on the gas smart meters, but they are not within my expertise).

The complexity is shown by comparing with the Italian roll-out. In 2000 ENEL designed and tested a smart meter. It ran a pilot study and then by 2008 had installed 32 million, an overall period of eight years. In May 2007 the British government envisaged a national roll-out of smart meters completed in 10 years—the start date is now 2014. A period of eight years has been filled with faffing and make-work by Ofgem and the civil service exercising their speciality of creating complexity out of irrelevancies and trivialities.

The wasteful roll-out should be stopped in its present form and replaced by a roll-out run by distribution network operators (DNOs) to a basic specification for new installations and replacement meters, with DNOs responsible for backhauling the data. The costs of the basic smart meters and comms would be socialised. DNOs and other companies would give customers the option of installing more sophisticated meters with in-home displays if they wish to have them—customers would pay the incremental costs.

Apparently the Major Projects Review Group of the Cabinet Office wanted to stop the centralised comms system, but DECC got its way. On 9 November I asked for a copy of the Group’s review under FOI I have been twice informed that DECC has “not yet reached a decision on the balance of the public interest” in providing me with the document. I recommend the Committee ask for a copy.

My Involvement in Smart Metering

In the late 1980s I knew Messrs. Bob Peddie, who had been the Chairman of the South Eastern Electricity Board (SEEBOARD), and John Fielden, an electronics engineer. They had designed the first smart meter in the world which was called the Credit and Load Management Unit (CALMU), which was (successfully) trialled by SEEBOARD in 300 homes commencing in December 1983 for approximately two years. We looked at trying to commercialise it, but the time was not ripe as the focus in the electric industry was on privatisation1.

Over the period 2008–09, I undertook a study of smart metering in the following jurisdictions –Britain, California, Denmark, Finland, France, Germany, Italy, Netherlands, New Zealand, Norway, Ontario, Spain, Sweden, Victoria—which produced reports on each ranging up to 115 pages. I visited all of the European countries, and drew on my extensive range of contacts for the American and ANZAC jurisdictions. In 2011, I updated the summary. The studies took about a man year of effort. I also studied the history of the roll-out in Britain for the chapter “Smart Metering Provided Unsmartly” in my book The British Electric Industry 1990–2010: the rise and demise of competition.

My expertise is limited to electricity and so I do not comment on the gas part of the roll-out except for common aspects of the economic assessment of the roll-out and the aggregate figures in the cost-benefit analyses. That said, since smart gas meters are more expensive than smart electric meters; most people have their central heating on timers and pay a simple tariff; the scope for conserving gas may be less than for electricity by managing it (as opposed to installing a modern condensing boiler), I suspect that it is also a waste of money on smart gas meters.

The Evolution of the Roll-Out

In 2004 the Carbon Trust ran a field trial of advanced metering for small and medium enterprises. Overall there was a societal benefit for larger companies, but a disbenefit for the smaller companies of £110/meter/year from which one might infer there was likely to be a disbenefit for domestic customers2.

Until 2006 neither the government nor Ofgem was interested in smart metering beyond token words. In August 2005, Energywatch, the then statutory consumer body, initiated a debate by publishing “Get Smart: Bringing Meters into the 21st Century”, which reflected more a concern about inaccurate estimated bills than the potential sophistications of smart meters. In February 2006, Ofgem published “Domestic Metering Innovation” and discovered a fundamental economic barrier to a market roll-out of smart metering, namely that the benefits arising from it are split between the suppliers, the DNOs, and the customers3.

Some stimulus to smart metering came from Directive 2006/32/EC on Energy End-use Efficiency and Energy Services. While the Directive did not mandate smart meters, it pointed the way and provided a lobbying base for meter manufacturers and meter services providers. Subsequently Directive 2009/72/EC (electricity), part of the EU Third Package (in force and having direct effect in the UK since summer 2009), went much further requiring Member States to implement “intelligent metering systems that shall assist the active participation of consumers in the electricity supply market”. However, “The implementation of those metering systems may be subject to an economic assessment of all the long-term costs and benefits to the market and the individual consumer and of which form of intelligent metering is economically reasonable and cost-effective and which timeframe is feasible for their distribution.”

Another strand of the evolution was the development of the government’s commitment to greening the electric industry. In May 2007, the White Paper, “Meeting the Energy Challenge”, envisaged that “…within the next 10 years, all domestic energy customers will have smart meters with visual displays of real-time information…” (para 2.64).

Next the government allocated £9.75 million of matching finance for the Energy Demand Research Project (EDRP) which involved four suppliers and 47,000 households. The government took the decision to mandate a roll-out of meters in October 2009, a year and a half before the results of EDRP were published in June 2011. The study showed there were only savings—and modest savings at that, of about 3%—when customers were provided with a smart meter plus a real-time display (of which more below)4. (Subsequently the Committee on Climate Change has indicated that the more or less natural evolution of improving efficiency of lighting and electrical appliances could save 20% by 20205. But that natural evolution provides no work for DECC and Ofgem, and no scope for Ministerial pronouncements and futuristic stories about politically “sexy” synergist relationships between smart meters, (so called) smart grids, windmills, and electric vehicles (which few people yet want to afford).

Concurrently with that survey, the government led two other activities. First to commission or undertake a series of cost-benefit analyses (CBAs) or Impact Analyses (IAs), which are discussed in the next section; second to determine a market model. Initially there were three market models, including the one eventually chosen, the “Central Communications Model”, which in due course led to the idea for the Data Communications Company (DCC), for which DECC has assumed responsibility. Significantly, there was initially no consideration of a roll-out by the distribution network operators (DNOs), which is the approach used in nearly all—if not all—the mandated roll-outs in Europe, North America, and Australia. There appear to be two reasons for this omission:

It unthinkingly adopted the so called “supplier hub” model conceived by Ofgem 15 years ago for the introduction of mass market retail competition whereby notionally the supplier was responsible for all services that the customer needed including metering. This concept was linked with the opening to competition of mass market metering, which has been of no benefit whatever to customers. Thus DECC, taking its cue from Ofgem’s ideological view of metering as a quasi-commodity service rather than a fundamental infrastructure asset, assumed that suppliers should take the lead in rolling out the meters.

DECC did not understand (and still shows no sign of understanding) that power line carrier (PLC) using the distribution network as a data communications medium is (provided the topology of the network is suitable) generally cheaper than wireless comms. PLC is widely used in Europe and North America.

Yet the suppler led roll-out with a centralised comms system is significantly more complex organisationally than a DNO roll-out; is more difficult to implement and project manage coherently than a networks-based solution; is more expensive because it cannot use PLC; and requires an additional database to keep track of who owns which meter as customers switch from a supplier that installed a meter to another supplier—which will lead to more errors.

When, under pressure, DECC commissioned a cost-benefit analysis of a DNO roll-out, it made two mistakes. First, it did not use a lower regulated cost of capital (eg 6% real) for the meters which would be part of the regulated asset base rather than the 10% used for a supplier roll-out, which would have saved customers some £300m p.a. Second, it assumed that the data would be backhauled with a central communications system, thus ignoring the potential savings from PLC. Consequently, the analysis concluded that a DNO roll-out showed no benefit compared with DECC’s favoured Central Communication Model, and was put aside.

An Attempt to Stop the Roll-Out

Ross Anderson, who is Professor of Security Engineering at Cambridge University, was concerned that the DCC would be another government IT disaster waiting to happen. We combined forces to prepare “Smart Metering—a poisoned chalice”6 which we gave to Minister Charles Hendry last February. Professor Anderson argued that “Britain has a long history of public-sector IT disasters and the smart meter project displays all the classic signs of imminent failure. There is a quite unrealistic timescale; no stable specification; no clear technical leadership; an insufficiently experienced and accountable procurement team; an over-optimistic view of critical components, such as data communications and standards; the omission of other critical components, such as a means to communicate with the home area network; an inappropriate architecture; and a lack of a systems view.”

Under a section titled “Cooking the books”, I pointed out that “The first three economic assessments of residential smart metering found that it wouldn’t pay for itself…So the Government’s ministers kept on trying until they got a positive result.” The first CBA was published in April 2007 by consulting engineers Mott MacDonald, who concluded that “Provision of feedback through advanced metering solutions is heavily burdened by the high costs associated with legacy meters and developing the comms infrastructure.” Mott MacDonald arrived at a net negative disbenefit for the type of system now envisaged of minus £4bn. Then the civil service got to work on the numbers and started modestly by first reducing the net disbenefit to a range from £1.3–0bn in 20087, but by August 20118 had manipulated the figures to produce a net benefit of £4.9bn9. And if one believes that one may also believe pigs can fly.

To improve “profitability” the government stretched the assumptions:

The electricity price forecasts used to evaluate savings increase in real terms over 15 years by 41% (2.3% p.a.) (But that assumption was not set out in the Impact Assessment).

There were major changes in “Optimism Bias”, the allowance made for uncertainty in the Treasury methodology. Mott MacDonald assessed the factors as 30% for a meter, 30–40% for comms, and 135% for the Meter Data Management System. While the factor for the meters was not unreasonably reduced to 15%, the factor for IT had been reduced to 10% which is absurd for an unspecified system, and there did not appear to be a figure for comms. The Chief Executive of EDF Energy commented to the Public Accounts Committee on the magnitude of the “huge implementation risk” involved in physically installing the meters, which required six times the current number of installers. DECC’s factor was only 10%.

The discount rate used was the Treasury’s figure of 3.5% real, which compares with the 5% used in France and the 8% in New Zealand. Although a discount rate of 3.5% may be appropriate for a long-life public infrastructure project, it is not suitable for smart meters because:

1.consumers do not use such a low discount rate for their own purchases;

2.the Impact Assessment assumes suppliers will charge 10% real;

3.the meters will be replaced every 10–15 years while the rest of the equipment will be replaced even more often. IT systems cannot be given an accounting life of over 10 years; and

4.many of the in-house display units will have a very short life and many will not be used, see below.

The effect of the low discount rate is to magnify the difference between benefits and costs as compared with using a higher and more realistic rate10.

There was no attempt to discriminate between customers with small consumptions (say <2000kWh p.a.) whose cost of installation would be the same yet benefits would be lower than those with larger consumptions and larger potential benefits

We recommended “that Britain stop trying to invent the wheel and just use one that already works. The two models that immediately suggest themselves are the Dutch/Spanish11 and New Zealand12; any would be much cheaper than the current proposals and would largely remove the technological risk of a systems disaster that would become apparent just in time for the next election in 2015. DCC is not necessary; no other country is attempting to build such a system; and there are good reasons for expecting a very poor outcome if we try.” At our meeting with Mr. Hendry I suggested that officials should visit Iberdrola13 in Bilbao and ENEL and ACEA in Rome (and would provide contact details).

Following the meeting I put together an elaboration of the economic part of the paper as “A critique of the impact assessment of the smart meter roll-out for the domestic sector (GB) 18/08/2011”—Annex 1 consists of the Executive Summary (the whole report has been provided to the Clerk to the Committee14). The key finding was that “the roll-out for electric-only appears very expensive (2009 prices)”:

	Britain	ENEL (Italy)	ACEA (Italy)	Iberdrola (Spain)
All in cost per meter (£)	135	65	75	70
Programme cost for 30 million electric meters (£bn)	4.05	1.95	2.25	2.1

One factor in the higher cost is the proposal to provide “free” in-home displays which are likely to cost £20+ all-in, representing a cost of £600 million. This is a most wasteful idea—many will not be used, or will be thrown away because households are either elderly, illiterate/innumerate, or too busy. In a recent study for the Electricity Authority of New Zealand, the New Zealand Institute of Economic Research concluded that an in-house display was not economic and noted that “IHDs are subject to damage or loss by consumers. We assume they require replacement every five years.”15 It would be cheaper and more effective to let customers use their laptops or smartphones.

Professor Anderson and I provided our material to various people in the Cabinet Office, and the Major Projects Review Group decided to look at smart metering. One person commented to me “DECC’s evidence base was flimsy.” In due course we heard that the Review Group wanted to stop the DCC and modify the project, but in the end DECC got its way. On 9 November I asked for a copy of the Review Group’s report. I have now twice been informed that DECC is “considering the balance of public interest with respect to Section 35 (Formulation of Government Policy) of the Freedom of Information Act 2000”, but it has “not yet reached a decision on the balance of the public interest regarding the exemption and will not be able to respond to your request in full as originally intended by this date.” I note the Prime Minister’s assertion that “The government must set new standards for transparency”16—I hope this applies to a proposal to spend many £billions of customers’ money. Apparently one part of DECC’s story is that the reason for the much lower cost of the Italian and Spanish roll-outs is that the functionality is lower. While that claim may be correct for ENEL’s roll-out, it does not apply to Iberdrola’s roll-out which has a later and more developed PLC system with 10 times the bandwidth and could—if required by regulations—provide an extensive backhaul.

Furthermore, DECC’s claim begs the question as to what amount of detail and how quickly information that will be useful is wanted by customers. (Needless to say, in the customary manner of the electric industry projects, compromise is achieved by incorporating the highest common denominator of requirements at the expense of customers who are not party to the debates behind closed doors.) How many people want to be flooded with up-to-date detailed information about electricity consumption and prices? In any case, the current and long overdue proposals for simplifying tariffs run directly counter to the need for an elaborate backhaul.

Instead of visiting Iberdrola in Bilbao (to which there was an invitation), which would have provided an opportunity to see advanced PLC in operation and to learn how Iberdrola was controlling its costs, an official of DECC went to its subsidiary Central Maine Power whose network is not suitable for PLC. An e-mail from Iberdrola dated 29 November 2012 commented:

“1.[there was a] $96 million US dollar grant received from the President Obama economic stimulus bill, ie, by all US tax payers.

2.Total cost per supply point for the Maine deployment is approx. $ 250, all inclusive. This is twice as much as the cost of our deployment in Spain.”

What Should Be Done?

The roll-out should be stopped in its current form, including dropping the ongoing tender for the DCC. The DNOs’ licences should be amended so that they are made responsible for providing metering and metering services. They should furthermore be obliged when replacing or installing new meters to install smart meters with a basic specification17. Along with the meters, the DNOs would be obliged to install a communications system which would backhaul data. The costs of the basic smart meters and comms would be socialised. DNOs and other companies would give customers the option of installing more sophisticated meters with in-home displays if they wish them—customers would pay the incremental costs.

Answers to Some of the Questions in the Terms of Reference

Are the Government’s cost and timescale predictions for roll-out realistic and will it deliver value for money? The cost and timescale predictions are probably reasonable; the roll-out is so expensive that it provides no value for money. But this should not be surprising because rhetoric apart, DECC have never shown any concern with customers’ money. A fool and “his money are quickly parted|”—especially when it is the public’s money.

What are the potential benefits of smart meters for consumers, and what barriers need to be overcome in order for consumers to realise them? The benefits are that consumers will always receive bills based on actual rather than estimated consumption, and some customers will reduce consumption moderately. The reduction needs, however, to be put within the context of the current low average annual consumption of about 4000kWh.

Is there a possibility that suppliers will gain considerably more than consumers from smart meters? Is enough being done to ensure that any financial benefits accruing to suppliers will be passed on to consumers? The numbers suggest that suppliers will incur a net cost rather than a benefit18. They will pass this cost on to customers: it will not be absorbed under competitive pressures, but will simply become another non-transparent element in the general market overhead cost of energy.

What lessons can be learned from successful smart meter implementation and usage elsewhere in the world?

1.The DNOs should roll-out the meters.

2.PLC should be used where economic.

3.No other jurisdiction has assessed smart metering as favourably as Britain apart from perhaps Italy (2200) where the detail of the benefits for ENEL are not available, but it is widely thought the main benefit is reduced theft, see Annex 2. The range of figures suggest that one can get any answer one wants—and HMG wanted the most optimistic.

Will smart meters empower customers to take greater control of their energy consumption? Yes, in principle, and for those who wish it. But many without electric heating may have little scope to reduce consumption if they already have energy efficient lighting, a modern fridge/freezer and washing equipment, and have the habit of turning lights off.

What are the potential obstacles to rolling out smart meters in the UK and how should these be addressed? What pitfalls have hindered roll-out programmes elsewhere and are we doing all we can to avoid them?

1.The obstacles are the complexity of the arrangements based on a supplier led roll-out and centralized comms.

2.Stop the current roll-out and reorganise it as a DNO roll-out when new meters have to be installed with DNOs backhauling the data.

3.But whether it is DNO or supplier led, the logistics of the project are daunting, given that it will be the largest UK home-visiting programme for 40 years and is directed at a consumer population that is alienated from the energy industry’s aspirations and in many cases hostile to its activities and indifferent to the roll-out’s objectives.

Will the commercial benefits of smart meter roll-out be captured within the UK? No—many of the meters will be manufactured in China and other countries (eg Eastern Europe, India).

Will DECC’s current approach to roll-out, including on procurement and establishment of the central Data and Communications Company, deliver an optimal data and communications strategy? Most definitely not. It is an expensive, cumbersome, and organisationally fragmented approach that is unwise. It is a consequence in part of DECC’s ignorance of PLC and in part the uncritical acceptance of the “supplier hub” concept.

Annex 1

EXECUTIVE SUMMARY OF THE CRITIQUE

The IAs for 2007 and 2008 showed a net disbenefit of -£4½bn and -£1.3bn respectively, which was transmuted into a significant net benefit in 2009 and increased further in 2011 to a benefit of £4.9bn, a change of >£9bn which stretches credulity. There are significant shortcomings with the August 2011 Impact Assessment (IA):

1.It is assessing a very large project where there are many unknowns eg regarding the effectiveness of the in-home-display (IHD) in saving electricity; the costs of the comms, DCC, and modifications to suppliers systems; and the benefits to network companies are perforce speculative. But more significantly, the correct and low risk approach to the smart meter project would have been to commence with a 3 year pilot project of rolling out meters to (say) two cities each of 100–200,000 households and to have evaluated both the system costs and customer responses in a non-experimental context.

2.There is basic information missing from the IA—no mention of the future prices for electricity; no mention of the savings from not replacing dumb meters; no undiscounted costs, in particular the costs of the initial roll-out. And there are many examples of unclear drafting which make the document very difficult to understand.

3.The risks have not been realistically appraised, particular for IT; for comms, which is not mentioned; and for installation.

4.We consider some of the methodology is either flawed or inappropriate. For a start the IA should not have adopted a one size fits all approach, but should have analysed alternates of less sophisticated data; discriminated between the benefit of higher/lower consumption groups of customers; separated gas and electric roll-outs; and scoped a minimum cost solution. The evaluation of the DNO option was incorrect. Furthermore, the discount rate was too low, a flaw exacerbated by assuming (without disclosing the details) an increasing electricity price in real terms.

5.The roll-out for electric only appears very expensive (2009 prices):

	Britain	ENEL (Italy)	ACEA (Italy)	Iberdrola (Spain)
All in cost per meter (£)	135	65	75	70
Programme cost for 30 million electric meters (£bn)	4.05	1.95	2.25	2.1

6. The result of these shortcomings is that the British assessment of the benefits is by far the most optimistic of the eight others that we have analysed. We have difficulty in believing it.

Annex 2

A SYNOPSIS OF SOME OST-BENEFIT ANALYSES

Victoria (5,70019)—after more than a quarter of the roll-out Deloitte were commissioned to undertake a CBA, which concluded:

“Over 2008–28, the Victorian AMI Program will result in net costs to customers of $319 million (NPV at 2008). This reflects a significant change from previous cost benefit analyses undertaken…This change is driven by the fact that costs have significantly increased since the previous forecasts, benefits have been reduced and barriers to the early provision of automated metering information services have further slowed benefit realisation.

Half of the Program costs will be sunk “or incurred by distributors by the end of 2011. Most benefits are yet to be realised.”

Spain (3,200)—although the government was not interested, an unofficial CBA undertaken within the regulator’s office found net disbenefit.

Sweden (average 8,000 but 12,000 for single family homes; 6,000 for apartments)—full smart metering only justified for dwellings taking >8,000kWH pa.

Norway (16,000)—no benefit without including “uncertainties”.

Denmark (3,800)—some benefit if include customers reducing consumption.

France (4,750)—benefits just cover costs from the perspective of the DNO.

Netherlands (3,350)—a positive net benefit, but “if 20% of consumers opt for the “switch off” situation the NPV will be [marginally] negative”.

New Zealand (7,400)—benefits likely to exceed costs for the majority of residential customers.

No CBAs were undertaken for Germany (3,500) and Finland (7,600).

January 2013