MOD0001

Written evidence submitted by Professor David Kirkpatrick

I worked for the Ministry of Defence as Director of Project Time and Cost Analysis. Later I was Professor of Defence Analysis at University College London, and served as a Specialist Advisor to the House of Commons Defence Committee.

Summary              It is difficult to forecast accurately the cash cost for the procurement of a major defence equipment project, principally because of the three different elements of defence cost inflation. These three elements are –

• Inflation in the unit costs of the inputs to a particular defence equipment project,
• Growth in the total cost of the project between its approval and its entry to service, and
• Intergenerational escalation in unit procurement costs between successive projects in the same class of defence equipment.

An accurate forecast of a project’s procurement cost depends on rigorous analysis of data on related projects, a comprehensive understanding of the relevant technologies, and awareness of current development and manufacturing processes.

Defence input inflation

1              During the development and manufacture of any defence equipment, the unit costs of the required inputs (labour, materials, energy, etc.) tend to increase in response to economic developments in the UK and overseas. In former times the MoD combined the increases in such unit costs to create a defence inflation index for a particular industry. In many defence contracts, that index was used to increase the annual payments to a supplier as compensation for cost increases which were beyond the supplier’s control.

2              A defence inflation index is an input index, like the Producer Price Index (PPI) which measures the average increase in the unit cost of inputs to all national producers of goods and services. By contrast the Consumer Price Index (CPI), which measures the average increase in the unit cost of a specified array of retail goods and services, is an output index. An output index like the CPI increases more slowly than the PPI (or than a defence input inflation index) because producers of civilian goods and services can adapt their procurement and processes to minimise the adverse effect of inflation in the unit costs of their inputs. The difference between the annual increase of output and input indices is closely related to the concurrent increase in the productivity of the national economy. Defence equipment suppliers on the other hand are contractually committed to an agreed project plan, specifying not only the deliverable outputs but also the inputs and processes which cannot be changed because they are considered critical to the performance and safety of the equipment.

3              Defence input inflation attracts disproportionate attention from the Treasury because it increases the MoD’s cash payment to a supplier in the current financial year, and thus affects the government’s borrowing requirement.

Project cost growth

4              There is often considerable growth, in addition to the effect of defence input inflation, in the total cost of a defence equipment project between approval and entry into service. This is hardly surprising, since large and innovative civilian projects often suffer similar cost growth, but it is always inconveniently disruptive of the MoD’s budgetary planning, and it damages the reputations of both the MoD and of the supplier. Occasionally the cost growth after project approval is unavoidable because of geopolitical events which are beyond the control of customer or supplier. Sometimes post-approval changes to the project requirement and/or to the procurement plan (both generally incurring some extra cost) may be desirable to take advantage of opportunities to enhance the project’s cost effectiveness. But too often such changes are implemented for political, industrial or budgetary reasons, without a proper assessment of their effects on the project’s cost.

5              More frequently, cost growth in a defence equipment project occurs because the cost forecast submitted to secure funding approval did not take proper account of the scale and complexity of the work involved, particularly if there had been insufficient research and technology demonstration to identify technical risks. Cost growth can also occur if some of the assumptions underlying the cost forecast at approval prove to be incorrect. All forecasts should be based on a Master Data and Assumptions List (MDAL) which should be rigorously debated and ultimately accepted by all stakeholders. The MDAL should ensure that no cost component is omitted, and that proper account is taken of any interfaces with other projects. Forecasters should use all available data from other projects, employ appropriate forecasting methods, and give particular attention to the likely effects of any innovative technical or managerial features of the project considered.

6              Military officers and industrial entrepreneurs are inherently optimistic so their forecasts of project costs and timescales tend to be influenced by ‘optimism bias’, which is the pious hope that on this project everything will go according to plan. The officers responsible for a particular project may consciously or subconsciously be inclined to forecast a low cost which should increase the chance of their project securing funding from a limited defence budget. Industrialists are similarly inclined to forecast a low cost to improve their own company’s chance of winning the resulting contract. Both parties in this ‘conspiracy of optimism’ are well aware that, even if the project’s cost forecast rose significantly after approval, it would be politically embarrassing for any government to cancel a project defended by interested parties who might variously cite sunk costs, increased unemployment, loss of national security and  damage to the national technology base.

7              All cost forecasting is difficult, but in the period after the Second World War it was particularly difficult to forecast accurately the development costs of military aircraft projects. In that period advances in aircraft technologies yielded dramatic improvements in the performance of military aircraft, and unprecedented challenges in their development programmes. In the 1950s, in the UK and elsewhere, there were some spectacular overruns in the development costs of such aircraft. In response the Ministry of Defence’s Procurement Executive, MoD(PE), set up in 1964 the Directorate of Project Time and Cost Analysis (DPTCAn) of expert and experienced engineers to collect and analyse data on previous aircraft projects, and to develop more-realistic forecasting methods. The Downey report[i] (presented in 1966 and printed two years later) insisted on a staged approach in which each phase of airframe development had to be satisfactorily completed before the next phased was funded. The combined effect of DPTCAn and Downey procedures reduced the average overrun in the development cost of UK aircraft projects from 400% for those starting in the 1950s to 4% above the DPTCAn forecast for those projects starting in the 1970s[ii]. Despite this evident success, MoD(PE) formed no comparable groups to analyse the costs of other classes of defence equipment, and even in the Air Systems Controllerate (responsible for aircraft and missiles) many MoD project managers with delegated authority chose to believe their own cost forecasts, or even the forecasts of their contractors’.

8              DPTCAn remained unpopular, like any bringer of realistic but unwelcome news. In the 1980s the MoD was persuaded that it should rely on taut contracts with its prime contractors, and should not second-guess their forecasts of project costs. This policy allowed DPTCAn to be under-resourced and to be ignored when its advice was inconvenient. In the 1990s the funding for UK defence research was drastically reduced, and the government’s non-nuclear defence research establishments were detached into an Agency (and subsequently most were privatised), depriving the MoD(PE) of easy access to impartial advice and expert recruits. Eventually in 1995 DPTCAn was incorporated into a larger organisation called Specialist Procurement Services (SPS) which was principally concerned with accountability rather than analysis.

9              In 1988 a review[iii] concluded that the MoD had not yet adopted Downey principles, and in particular was not allocating sufficient resources to project planning and risk reduction. In 1999 the Smart Procurement reform[iv] judged that the MoD’s procurement system was not working well; referring to the problem of defence cost growth, it noted that the MoD was still allocating insufficient resources to the early stages of a project so that key funding decisions were made with inadequate information, stressed again the importance of projects meeting their targets for performance, cost and timescale, and aspired to keep the final procurement cost almost always within 2% of its approved budget. Another proposed reform[v] in 2006 stressed the need for MoD to produce realistic cost forecasts, but did not suggest how the staff producing such forecasts could be shielded from hostile and career-limiting criticism. In 2007 a comparison of UK and US performance in defence procurement[vi] showed that the unit cost a sample of UK projects had increased after approval by 34% (even after creative accounting had transferred some costs to other budgets). In 2009 Gray’s independent report to the Secretary of State for Defence[vii] derived a similar assessment of cost growth post-approval. Yet another strategy for acquisition reform in 2010[viii] promised to improve the MoD’s ability to produce realistic cost forecasts by recruiting more staff and by closer partnering with private-sector suppliers. The MoD’s recent Evidence[ix] to the PAC on defence cost inflation did not indicate whether the 2010 promises were fulfilled; that Evidence failed to present the current levels of cost growth for large and small MoD projects, but in 2021 the NAO[x] yet again deplored significant cost growth in a sample of defence equipment projects

10              This chronology demonstrates that the MoD has for many years been concerned about project cost growth between approval and entry into service, and that it has repeatedly announced a plethora of good intentions and plans to introduce better cost forecasting. But it has never seriously assaulted the enduring incentives which have consistently driven Service customers and industrial suppliers into ‘conspiracies of optimism’. Perhaps this is because the adverse effects of an over-optimistic cost forecast at approval are long delayed (thus becoming someone else’s problem) and ca be obscured by -

• Downgrading the project’s specification and/or
• Reducing the number of units to be procured and/or
• Delaying the project so that its annual expenditure fits below a budgetary profile.

11              The outcome, when the project finally enters service (perhaps 20 years after approval), could be a smaller contribution to the UK’s military capabilities than might have been obtained from alternatives available in the same timescale. However this outcome is never publically acknowledged, and the guilty Ministers, officers, officials and executives have long since moved to pastures new. Another adverse effect of over-optimistic forecasts throughout the procurement programme is a funding gap which may require the MoD to economise on some less-glamourous projects, such as body armour and robust patrol vehicles.

12              It should be recognised that over recent decades it has become progressively more difficult to forecast with confidence the procurement cost of defence projects and even more difficult to forecast their acquisition cost (= procurement + operation + support + disposal). During this period the gaps between projects in the same class have lengthened (allowing the expertise in the MoD and its suppliers to atrophy), projects may now involve international collaboration and innovative management arrangements (like Private Finance Initiatives and Public Private Partnerships), projects rely more on bespoke software, and many projects’ service lives may be extended by upgrades requiring technology insertion.

Intergenerational unit cost escalation

13              Throughout military history it has been observed that the unit costs of many classes of defence equipment were higher than the unit costs of contemporary consumer goods. The main reason for the higher unit cost of many classes of defence equipment is that they are ‘tournament goods’ and it is vitally important that their performance matches or exceeds that of an opponent’s equipment. Thus defence equipment resembles Formula 1 racing cars and America’s Cup yachts, or the golf clubs and tennis racquets used by professionals, rather than washing machines and television sets.

14              During some historical periods, it was impractical to improve the performance of particular classes of defence equipment, due to technological or operational constraints, and consequently the same design remained in production at much the same unit cost for many decades (such as the British ‘Brown Bess’ musket). But since the middle of the 19th century technological advances have unleashed an arms race in which any nation relying on second-rate or obsolete equipment is virtually guaranteed to suffer bloody defeat. It is this technological competition to match or exceed the performance (as perceived by military intelligence) of equipment being developed by unfriendly nations which has driven intergenerational unit cost escalation. For some classes of equipment, the real unit cost (adjusted for input inflation) has increased tenfold between successive generations. The logical rationale for such unit cost escalation was presented in 1995[xi].

15              Intergenerational unit cost escalation has perhaps been slightly increased by other current trends – less competition between a reducing number of suppliers and more rigorous environmental and safety regulations – but their effects have been far less significant than those of the technological arms race. The MoD has suggested[xii] that some cost escalation arises from the tendency of governments to favour local suppliers, but this tendency has endured for many decades (and was disturbed in the UK, only temporarily and partially, by the Levene reforms during the 1980s insisting on international competition).

16              The observed rates of intergenerational unit cost escalation over recent decades have been smallest for mature systems (like machine guns) and largest for complex systems (like anti-submarine helicopters) whose sensors and communications have become ever more sophisticated. These rates have not been perceptibly reduced by the succession of reforms in the UK’s defence procurement policies and procedures. Future escalation rates will depend on the interaction of new technology and operational realities, but it is prudent to expect that present trends will continue unless new obstacles to innovation emerge.

MoD’s procurement staff shortage

17              The 2009 Gray report, successive NAO reports and the recent MoD Evidence to the PAC all bemoan the lack of MoD staff (military and civilian) with the knowledge and skills needed to manage its programme of defence equipment procurement (costing about £25B per year). This problem has become more acute since the end of the Cold War, because of cuts in the MoD’s procurement staff and because of the delusion, widespread in the higher levels of the MoD, that its procurement staff required no great understanding of technology, analysis or business, but needed only the ‘leadership’ skills required to follow the latest set of MoD procurement procedures. Such project leaders cannot be expected to recognise and avoid the multiple pitfalls and risks which beset large complex defence projects involving new technology, or to generate accurate procurement cost forecasts.

18              Sir Donald Spiers, who in 1991 was Controller of the MoD’s aircraft and missile projects and was also Head of Profession of the MoD’s scientists and engineers, was not seduced by the delusion that defence procurement needed only taut contracts and diligent bureaucracy. In that year he established at University College London the Defence Engineering Group (DEG) as the MoD’s designated centre of excellence for analysis and education in defence equipment procurement[xiii]. The DEG provided an intensive one-year post-graduate course, blending parts of a civilian MBA (such as project management and risk analysis) with defence-specific modules covering disciplines such as cost effectiveness and operational logistics. This course yielded a cadre of educated MoD project managers who understood their project’s future contribution to a UK military capability and who could communicate effectively with other stakeholders. The DEG’s students became truly competent defence acquisition managers, each with a deep knowledge of their own specialist area but also with a broad understanding of all of the other disciplines involved in a major defence equipment project. In parallel the DEG did research work on various aspects of defence procurement, and made many contributions to professional journals and conferences. However one of the MoD’s chronic challenges is its resistance to reform, and the DEG was closed in 2005. Its ethos is embodied in the book ‘Conquering Complexity – lessons for defence systems acquisition’ published by TSO in the same year[xiv].

19              It is the MoD’s duty to recruit, educate, motivate and retain sufficient staff to manage the procurement of defence equipment. It should adopt policies which fulfil that duty.

Concluding remarks

20              The procurement costs of future large defence projects are affected by input inflation, project cost growth and intergenerational cost escalation, and such costs are as difficult to forecast as were the costs of projects in the recent past. Better forecasting needs expert staff and rigorous analysis, but many organisation appear to prefer sublime optimism. This conclusion is based on my own personal memory and opinions, which do not necessarily correspond to those of any of my former employers.

David Kirkpatrick                                                                                                                   June 2022.