1 Introduction

Public procurement in recent years has emerged as an essential innovation policy tool across countries. The strong demand pool exerted by such public procurement serves as an important incentive for firms to undertake risks of R&D and innovation (Edler and Georghiou 2007; Lember et al. 2014). Traditionally, governments around the world allocate substantial resources to procure new weapon systems and military equipment through either ‘off the shelf’ purchase or indigenous development. The defence procurement environment, however, is complex. On the one hand, there are uncertainties and severe resource constraints due to regularly changing threat perceptions, limited flow of information about new technologies, and the growing demand to reduce defence related expenses (Dertouzos 1994; Sceral et al. 2018). On the other hand, several stakeholders remain pre-occupied with the demand for ‘zero-error’ technologies (See, Davis 2015; Gholz and Sapolsky 2000).

The decisions concerning the indigenous development and procurement of complex military systems (CoPS), in particular, have generated much interest owing to considerable uncertainty over technological choices, innovation, and their end-user acceptance. By CoPS, we refer to the systems, networks, infrastructure, engineering constructs and services that are highly costly and technology-intensive (Hobday 1998). Due to their ability to provide comprehensive situational awareness, the demand for newer and cutting edge CoP systems has intensely grown over the years. The procurement of such systems, however, is frequently marred with problems of time and cost overruns, performance gaps, and, at times, even a complete failure, resulting in substantial financial losses (Flyvbjerg 2011; Schwartz 2014; Bogan et al. 2017). Allegedly, the ‘optimisation driven’ end-user requirements and the demand for ‘zero-error’, often, lie at the core of such problems of time and cost overruns (Park 2012; Smith and Tranfield 2005).Footnote 1

The paper examines the extent to which a heuristics-based frugal decision-making process can offer an efficient response to these problems. Under uncertainty, the ‘frugality’ approach emphasises on taking decisions based on heuristics and ‘simple rules of thumb’. A frugal approach prioritises “what works in the actual environment’ over “what ought to work” (Gigerenzer and Todd 1999). In CoPS, such strategies might enable system designers, and procurement practitioners to respond to uncertainty emerging from lack of data for making ‘reliable’ predictions about cost, schedule and performance levels vis-à-vis the ‘highest’ specifications, interpreted to mean a preference for ‘satisficing’ over ‘optimization’ (Klein 2002). Here, frugality helps to yield satisficing solutions where a payoff reasonably meets the “aspiration” level (Simon 1967; Gigerenzer 2010). Over time, however, such solutions can be further improved through adaptive learning and incremental changes.

Section 2 analyses the institutional superstructure of defence procurement mechanism with a view to underline ‘optimisation’ as the underpinning dynamic. The necessities for an alternative framework, and the challenges to achieving it, are briefly pointed out too in this section. Section 3 elaborates on the conceptual framework of frugality and its importance in innovation and technological capability building efforts in CoPS. Section 4 outlines the research method and data sources, and analyses case studies of CoPS acquisition of missile system, fighter aircraft and weapon locating radar by the Indian authorities using the framework of frugality. The last section presents an analytical ‘rounding up’ and lays out a few policy instruments to infuse a frugality narrative in defence procurement.

2 CoPS and the procurement environment

The procurement of military equipment constitute around 40–45% of annual defence expenditure in many developed and developing countries (Tian et al. 2018). The procurement of “high cost, customized” complex product systems (CoPS) poses novel developmental, and procurement challenges due to their resource intensity and complexity (Hobday 1998; Markowski and Hall 1998). The CoPS products have a large number of customised components, and their production requires a broad range of knowledge and expertise (Prencipe 2000). With increasing complexity, the procurement of these systems encounters significant challenges in all countries. One significant source of the complication arises from the demand for ‘zero-error’ technologies by the involved parties in the procurement, often, irrespective of their affordability (Johnson and Johnson 2002; De Spiegeleire 2011). Below we discuss the superstructure and the environment of defence procurement mechanisms to understand how the demand for ‘zero-error’ emerges and sustains itself in procurement practices, across countries.

2.1 The demand for ‘zero error’

A ‘zero-error’ approach presupposes that the new technologies and systems must be delivered with precision based on the stipulation of broad operational requirements (Davis 2015). Scherer (1964: 29), for instance, points out how the “incentives in weapons acquisition…ran strongly in favour of quality maximization and lead time minimization”, ignoring, in turn, the aspects of costs. Rogerson (1995: 312) similarly argued that “the constant pursuit of improved performance and capabilities” constitutes the defining feature of defence procurement. In this pursuit, the military’s natural choice for various weapons systems, Rogerson (1995) argues, is often biased toward “too high” a level of quality. The so-called, Gray (2009)’s pronouncement “Bid High Spec, Bid Full Spec” captures this dynamic most suitably.Footnote 2 Such practices are interpreted as rational, by the end-users, in the ever-evolving threat environment of a nation. We now briefly discuss the factors, which lead to the demand for, and sustenance of, ‘optimisation’ and zero-error in defence procurement.

In the face of an evolving threat environment, existing defence capabilities are always perceived to be deficient. This gap pushes the end-users to demand ‘the maximum’ gains in the capability, perhaps, with the assumption that ‘some’ can only be obtained if one asks for ‘a lot’ (Gray 2009). The demand for the ‘maximum’ is often a typical response in a resource constraint setting, where existing resource scarcity makes one demand the ‘excessive’, whenever an opportunity arisesFootnote 3. The demand for highly “optimized” or ‘battle-ready’ systems is, therefore, legitimized, particularly when new acquisition opportunities come up infrequently. Moreover, the defence forces, in general, function within a ‘zero-error’ operational mindset, which drives them to seek the best achievable technological capabilities, ceteris paribus. The highly uncertain threat environment they work in, demands themselves to be battle-ready for any eventuality.

2.2 The critique of zero-error and the alternative approaches to procurement

Allegedly, the pursuit of ‘zero-error’ generates uncertainty about the system performance and entails delays and high risks of failures (Benett 2010). Brown (2013) terms such a practice as either unrealistic or cost-prohibitive. Many scholars and practitioners point out that it is hardly possible to determine the final system requirements (specifications) at a nascent phase of capability building, particularly for CoPS (Rich et al. 1986). Besides the quality parameters, the procurement process involves optimising the life-cycle cost (LCC) during the early stages of designing the programme (Lorell and Graser 2001). This exercise requires perfect knowledge of the costs of acquiring, owning, maintaining and the disposal of the system for its entire lifecycle (Pickup et al. 2010). In reality, these estimates often go way off the mark as they fail to take into account many uncertainties. The specifications once approved, however, become the statutory basis for assessing all the subsequent steps of a procurement process, irrespective of even the changing operational environment of the user (Hartley 1998). Audit agencies too, as mandated by legal and policy regime, evaluate the systems and contractual outcomes only by the approved requirements. The audit of the ‘financial transactions’ and ‘operational performance’ often contains findings, which are politically unpalatable and puts the pressure on congressional or parliamentary bodies to demand ‘accountability’ in the procurement decisions.Footnote 4

The developmental cycles of many complex systems last longer than a decade. During this period, incidentally, many, initially thought essential, requirements either change or become unviable. In such a setting, the need for preparing and announcing a precise and detailed system performance requirement, at the outset, proves to be deeply flawed (Wellman 2003; Kausal 1999)Footnote 5. As the estimates go awry, recriminations occur among the acquisition authorities, users, and contractor communities over who was at fault (ibid), undermining further the credibility of the whole process of cost calculations.Footnote 6 In a nutshell, the institutional structure of procurement acts as a critical driver of the ‘zero-error culture’ prevailing in the procurement environment. The unmet specifications, often, become a source of conflict between the system designers, procurement authorities, end-users and the oversight bodies (Lorell et al. 2006).Footnote 7 As a result, defence procurement  authorities, across countries, seem to bid for the highest possible capability, to be delivered within a given timeline, ‘at one go’, in an undifferentiated manner (Gray 2009). Incidentally, the quest for such ‘top of the line’ capability often leads to rejection of available operationally useful technologies, along the way (Wylie et al. 2006; Gray 2009).

2.3 Spiral development and evolutionary acquisition: some new directions in defence procurement

In recent years, the policymakers in leading military hardware producing countries, such as the U.S., attempted to resolve these problems in technology procurement by adopting the so-called “evolutionary acquisition” and “spiral development” (Lorell et al. 2006). The U.S. Department of Defence (DoD), for instance, enacted the new 5000 series acquisition directives and instructions in the year 2003, which mandated evolutionary acquisition (EA) strategies as the “preferred approach” to satisfy the operational needs. In contrast to the previous approach of waiting long for ‘total package procurement’, EA aims at procuring a ‘good enough’ (operationally useful) technology early. The procured technology is then adapted and improved upon, step by step (DoD 2000). In the UK, along the similar line, an independent report for the Secretary of State for Defence called for fielding ‘80%’ of the capability requirement as ‘good enough’ candidate for procurement, to minimise overruns in cost and schedule (Gray 2009). Indeed, a growing body of scholarship on system engineering suggests that systems are becoming increasingly complex, due to enhanced need for flexibility and incorporation of ICT. According to Gorod et al. (2017), decision-making approach in such a scenario needs to evolve significantly.Footnote 8 Amidst the growing complexity of systems, the ability to find the best outcome reduces considerably. The system designers in such situations take decisions based on ‘what works and what does not’ in the actual environment (Salmon 2015).

This so-called evolutionary approach nevertheless proved to be inadequate in resolving the problems of cost and time overruns mentioned above. At times, the approach led to serious ambiguities in implementation, and organisational conflicts over ‘requirements management’, ‘cost analysis’, ‘budgeting’, and ‘auditing’ (Sylvester and Ferrara 2003). A Rand Corporation study by Lorell et al. (2006) notes that the logic of ‘spiral procurement’ often remains under-appreciated within the superstructure of procurement. The key stakeholders, such as the politicians, military users and auditors involved in the procurement process continue to demand for disclosure of exact time and cost structures associated with the end-stage capability in advance. Besides, the so-called evolutionary structure does not leave out the use of various ‘optimisation logic’ (e.g. use of Life Cycle Cost approach, and the rigid rules of operation) in the setting, giving rise to inherent incompatibilities between the different components of the institutional architecture of procurement.Footnote 9

3 Optimisation, frugality, and complex systems

3.1 A framework for analysis

The discussion in the previous sections clearly shows that optimisation refers to the selection of the best, or the one with the highest expected utility (Klein 2002). The limitation of this approach has been well recognised in the literature on bounded rationality (Klein 2002; Simon 1972). Herbert Simon notes that “the computational requirements for optimizing are usually too high for decision-makers” and hence “it is not always possible to optimize” (Byron 1998; Klein 2002). According to Simon, human rationality is bounded due to constraints such as complexity of the environment, incomplete information, limited cognitive capacity of information processing, and the limitation in time and other resources. As a consequence, human decision-making is often not quite clear-cut or ‘logical’ (Simon 1956). Klein (2002: 103) describes the tendency among decision researchers and practitioners to cling, tenaciously, on to optimisation as “fiction”. He argues that “there is not a way to determine if a decision choice is optimal owing to time pressure, uncertainty, and ill-defined goals”. In certain types of complex situations, there is no way to identify all of the available options due to rapid changes in the given environment, calling for using heuristics over optimisation to deal with such eventualities (Mousavi 2018; Gigerenzer 2005).Footnote 10 The public policies premised on the logic of demanding or selecting ‘the best’ thus risk imposing unwarranted costs, and, at times, maybe unaffordable to implement (Gigerenzer and Kurzenhaeuser 2005). Nevertheless, the dream of optimization (i.e. “selection of the best among all available options for achieving best results”), continue to drive policy decisions in many public and private domains (Klein 2002).

As opposed to optimisation, the frugal approaches emphasise upon ‘economizing’ the resources needed for decision-making. Its prescriptions, on the other hand, require an in-depth knowledge of institutions and relations operating in a given context, under resource constraints and uncertainty (Bhaduri et al. 2017). In several decision settings, information failures are widespread, and the ‘fast and frugal’ heuristic decisions, based on intuition, experience, learning, and gut feeling can be highly efficient (Gigerenzer 2008a, b, 1999). The works of Gigerenzer and his colleagues (Gigerenzer 2008a, b; Gigerenzer et al. 1999) are particularly significant, in this context, for bringing both the ‘person who takes the decision’ and the ‘(actual) environment in which the decision is taken’, at the center of the discourse on decision-making under uncertainty.Footnote 11

The ‘fast and frugal’ heuristics are, however, not “omniscience”. The effectiveness of these heuristics depends both on the person who takes the decision, and the environment in which the decisions are taken. Cognitive factors like ‘experience’, ‘adaptive aspirations’, ‘learning capacity’ importantly shape the effectiveness of these heuristics.Footnote 12 In this conceptual schematic, the idea of frugality mainly refers to the ‘simple search rules’ and ‘need satisfaction’, as opposed to ‘optimising under constraints’. Frugality, therefore, not only underscores ‘what’ is achieved but also ‘how’ it is achieved to obtain an ‘ecologically rational’ outcome (Gigerenzer 2008a, b). Such a process of decisionmaking would involve the following:

  1. 1.

    a search process using simple hierarchical steps and intuitive reasoning.

  2. 2.

    efforts to adapt to the environmental challenges through the demonstrated capacity for learning and imitation; and

  3. 3.

    an emphasis on actual performance, practicability, and effectiveness rather than logical/scientific validation (Bhaduri et al. 2017).

3.2 ‘Frugality’ in innovation

According to Nathan Rosenberg (1976), a seminal historian of technology, the innovators are seldom able to produce the systems or complex machines to the “special and exacting requirements and specifications of the (machine) users” (Rosenberg 1976:12). Due to large-scale uncertainty in realising the user requirements, the machine developers muddle through incrementally, often, at the shop-floor level, rather than being able to produce the system at “one fell swoop” (Rosenberg 1976).Footnote 13 Subsequent scholarship in the tradition of evolutionary economics has indeed identified uncertainty as a defining characteristic of the environment in which innovation is carried out. So far as the “rationality” of actors is bounded in such an environment, the agents try to pursue innovation through learning, adaptation, and trial and error (Nelson and Winter 1982). Heuristics come handy in such situations (Kheirandish and Mousavi 2018).Footnote 14

Incidentally, while the need for trial and error, learning and adaptation is, by now, well recognised in the context of major, frontier, innovation, their relevance for a technology catch-up process is often viewed with suspicion, at least in the mainstream innovation and policy scholarship. To be precise, we argue that an appreciation for learning, adaptation and trial and error in inherently incremental innovation-led technological catch-up has witnessed an inverse U-shaped enthusiasm among scholars and policymakers. While such an appreciation grew in the decade of 1980s and 1990s, the post-WTO policy environment has often belittled such activities by clubbing them as ‘imitation’ or ‘piracy’, ideally to be done away with. The homogenization and strengthening of intellectual property rights and related regulatory architectures policies in this period have undermined the risks and uncertainties in the catch-up environment. Policymakers in such an environment are often persuaded to ‘buy from the readily available technology shelf’ from the advanced industrial countries. This perception emboldens the defence technology users in these countries, for instance, to demand “get me what I saw at… a defence expo, brochure or magazine article” (Johnson and Johnson 2002, emphases added).

We, therefore thought it proper to present a snapshot of the literature on technological capability-developed since the 1908s-to proceed further. Lall (1987) defines technological capability as the, sequentially developed, capacity to select, absorb, assimilate, adapt, imitate, and improve upon, given (imported) technologies, which could eventually culminate into capabilities to create new technologies. This process underlines the importance of “uncertain environment”, in which minor innovations at the shop floor level contributes to the accumulation of technological capability, through creative adaptation or technological effort (Lall 1992).Footnote 15 Nelson (2008) locates such arguments within the broad premise of the evolutionary theory of innovation.

Note that the uncertainty in the environment, emphasised by this scholarship, would be even more relevant for catch-up in complex product systems for obvious reasons. Defence CoPS innovation environment, in particular, remains marred with uncertainties due to its strategic importance, which leads to an abysmally low level of information sharing about these technologies. Any innovation in this sector would have to rely significantly on adaptation, learning by doing and trial and error.

4 Data, research methods and the cases

India has followed a policy of promoting technological self-reliance in the defence sector through import substitution and indigenous technological capability building.

4.1 Data and research methods

We analyse three case projects undertaken by India’s Defence Research and Development Organisation (DRDO) in collaboration with the various public and private sector defence companies. The study draws upon both primary and secondary sources of information. The primary sources include interviews conducted with in-service and retired senior officials from DRDO, Ministry of defence and the defence companies during 2015–19. We also conducted discussions with officials from the defence ‘service headquarters’ to obtain users perspectives. From each organization, we chose 2–3 officials with procurement, R&D, and development related portfolios. These interview schedules mostly had open-ended questions, and each interview lasted for about 45 min to 1 h. In some instances, the interviews were followed up through emails, telephone calls or personal visits for validation and obtain additional information. The data from such interviews were supplemented with the data available in the public domain in the form of questions raised in Indian Parliament, and the various reports of India’s main audit body, the Comptroller and Auditor General (CAG).Footnote 16 For understanding the broad contours of satisficing parameters and behaviour in indigenous technology projects, the real-life decisions were examined critically (Glockner and Betsch 2008).

4.2 The cases

4.2.1 The Akash Missile project

In the 1980s, India faced a severe shortage in its air defence capability as the existing systems imported from the Soviet Union in 1960s became increasingly obsolete, and needed to be replaced by new-generation systems. In 1983, the Ministry of Defence (MoD) commissioned an indigenous technology procurement project to the DRDO for the development of two Medium-Range Surface-to-Air Missile (MRSAM) systems named “Akash”, one each for the army and the air force.Footnote 17

Being the first of its kind to be developed in the country, the project involved many uncertainties. The DRDO followed an incremental approach to develop the system and improve it further through a series of user-trials and feedback. After its commencement in 1983, the developmental tests were carried out during 1989 and 1997. During the initial user trials of the system in 2004, however, only Air Force participated while the Army chose to stay away, as the version was to be tested, apparently, yet to meet their specification requirements. In repeated trials up to 2004, the air force rejected the Akash even though the missile repeatedly struck its targets, the lead project developer revealed to us. During the period 2005–2006, the missile was tested as many as 16 times to demonstrate its interception capability against two live aerial targets and to establish its high accuracy and precision. Another official revealed to us that the strike probability of the missile, over time, has increased from 80 to 88 to 98.5%. Nevertheless, the stakeholders in the procurement ecosystem insisted upon the acquisition of complete capability, at one go, significantly delayed the induction of the complete system within a pre-defined timeframe.

The Air Force’s demand for meeting optimum performance and complete development multi-mode capability significantly delayed the first induction. It is only in 2008–09, the air force agreed to procure the two squadrons of Akash after a set of trials in December 2007 (Shukla 2009). Similarly, the army too, being skeptical about the specifications, agreed to induct the system in the year 2012 only after conducting its own trials in 2011.

Explaining the zero-error dynamics in the indigenous technology procurement, the MoD official with several years of experience in defence acquisitions remarked that in short-range missile systems, the DRDO has come to face stiff competition from foreign vendors.Footnote 18 This reinforced the demand for zero-error, as any failure to meet the zero-error specification by the DRDO would have strengthened the call  for importing the technology from abroad. However, the users, for long, did not share the DRDO’s assessment that Akash system fully met their ‘operational’ requirements. In this case, while the procurement superstructure conforms to the logic of frugality in allowing the scientists to indigenously develop the technology, it remained reluctant to procure the same, unless the zero error condition is met.

The case of AKASH project suggests that the preoccupation with optimisation limited the opportunities for feed-back based learning and incremental technology development. The procurement superstructure did not accord much flexibility for meeting the desired specifications, frugally. By refusing the early induction of the initial version of the missile, the procurement policy, arguably, delayed sequential development of the system through learning and technological upgradations.

4.2.2 Light Combat Aircraft (LCA)

The Light Combat Aircraft (LCA) is among the few success stories among latecomer industrializing countries for undertaking the developmental of fourth-generation fighter aircraft. Although the project was conceived in the year 1983, it received the official sanction only in the year 1993. Interesting to note that only the 60 per cent of the requested budget was sanctioned to develop two non-weaponised technology demonstrators.Footnote 19 Only after the successful demonstration of first technology, the MoD sanctioned the remaining budget in the year 2001 for developing the concurrent fighting systems. According to many analysts and practitioners, the two-part developmental approach unduly delayed the developmental cycle of the LCA project (Shukla 2016). Being a CoPS project, the LCA development process has mainly relied upon learning by ‘trial and error’. Besides in-house learning-by-doing efforts, the DRDO forged collaborations with other S&T institutions, local industry, and academia for solving critical challenges of indigenous development.

In its current form, the LCA represents a ‘fourth-generation plus’ fighter aircraft and incorporates full networking capabilities developed by the Indian defence establishment. Most importantly, the LCA is considered to be most suitable for network-centric warfare, as it can easily be integrated into the network of indigenously developed radars and sensors, which the imported systems cannot meet. According to many credible estimates, the LCA at the cost of USD 1.2 billion offers itself as highly cost-competitive compared to rival systems in the global market (Thiagarajan 2017). Similarly, performance-wise, the LCA has proved itself to be superior in many critical parameters compared to other fighter jets like Mirage 2000 and Mig 29 (BIS) (Karnad 2017a, b). According to a lead scientist, the LCA aircraft has already achieved over 3000 h of flight-testing remarkably well.

The ongoing developments to the aircraft structure are primarily aimed at reducing its weight and improving the engine and aerodynamic performance in order to make it a more robust system as demanded by the air force. He also emphasized that only an indigenously developed system such as the LCA, can offer a “secure” and “interoperable network-centric capability”. The LCA, this way, enjoys a distinct advantage over its foreign competitors. Notwithstanding such advantages, the procurement superstructure has withheld its induction in their quest for optimum capability. This delay, according to the experts, means a failure to appreciate the incremental, adaptable nature of technological change. The emphasis on acquiring the latest configurations (in terms of maneuverability and firepower) ‘at one go’ means that the procurement superstructure continues to prefer optimization over inter-operability. The developer strongly emphasized that a spiral procurement would not only have fulfilled the immediate operational requirement but also open up export possibilities through the demonstrated capacity of its efficiency. The policy approach has, once again, overlooked the importance of incrementality in innovation, thereby, causing cost and time overruns-all in the quest for ‘zero-error’ and optimality.

4.2.3 Weapon locating radars

The in-house development and procurement of Weapon Locating Radars (WLRs), shows how frugality driven approaches can fructify into an ‘operationally useful’ military systems with minimal costs and time overruns. In the 1980s, the Indian army had projected a requirement for fielding fire-finding radars and, in late 1980s, it reportedly decided to purchase the American AN/TPQ-36/37 radar. After a series of failed attempts to import the American radar, the MoD decided to commission the DRDO and BEL for its indigenous development.Footnote 20 According to a former official, the indigenous project, however, did not receive the sanction immediately, with users being sceptical about the ability of the DRDO to deliver the radar in time. The ensuing ‘buy’ vs ‘make’ dilemma had, reportedly, cost the country during the ‘Kargil conflict’, in which many Indian soldiers lost their lives due to their inability to detect enemy positions in high mountainous terrain.Footnote 21

The WLR developmental project, nevertheless, witnessed a unique developmental pathway. During the tests of the Akash missile, the DRDO engineers ‘accidentally’ observed that the ground radar (named Rajendra radar), a key component of Akash system, was able to detect and track artillery shells being test-fired at a nearby range. Using a ‘similarity heuristic’, based on intuitive observation, and domain knowledge, the LRDE scientists, proposed to adapt the Array antenna of the Rajendra Radar for the WLR, which eventually got sanctioned (Subramanian 2012).Footnote 22

The procurement superstructure relied upon the experiential knowledge and expertise of DRDO scientists. It commissioned a mission-mode project for developing the radar system, including the antenna and key components indigenously. Bypassing the demand for acquiring a fourth (latest) generation high-accuracy radar (the US-made AN/TPQ-37 radar), the procurement superstructure went ahead with a locally made, third generation radar with ‘operationally useful’ efficiency in target detection. Incidentally, neither the users nor other stakeholders insisted on ‘zero-error’ for its procurement, even though the demand for the imported radar has not put off the list, till date.

5 Frugality-based heuristics in Defence procurement policy: a summing up

This paper has attempted to provide an understanding of the application and relevance of fast and frugal heuristics in the context of development and procurement of complex military systems. We find that while a certain degree of frugality was allowed concerning the decisions of allowing indigenous development of CoPS technologies, the ‘optimization logic’ looms large about the final procurement and use.Footnote 23 The research reveals that procurement policies, which are heavily biased towards achieving “zero-error” “at one go”, discourage sequential learning through user-producer feedback, eventually leading to severe cost and time overruns, and failures. The optimizing tendency on the part of the military users, policymakers, audit agencies and, perhaps, the political establishments limits the scope for heuristic-based decision-making and thereby impose serious learning and opportunity costs on the indigenous defence innovation ecosystem. We argue that such tendencies to rely on optimization would possibly be higher in a technology catch-up country for which, the ‘technology shelf’ is presumed to exist and available to draw upon without much hassle. It is, therefore, not surprising that an appreciation for innovation-related uncertainties remains absent among the procurement stakeholders in a country like India. As Johnson and Johnson (2012) points out “get me what I saw at… a defence expo, brochure or magazine article” becomes a common syndrome in such a scenario.

Bringing back the discourse on technological capability, we reiterate the limitation of such an assumption and emphasise that incremental innovations too are prone to significant risks and uncertainties. This is particularly true for defence technologies for which information flow is inherently inadequate. It is, therefore, crucial for procurement policymakers to adapt to the realities of innovation uncertainties and adopt sequential (incremental) approaches and learning in procurement decision-making based on feedback from the ‘actual’ environment of use. In some countries, this has been initiated in the form of ‘evolutionary acquisition’ and ‘spiral development’. In our view, however, these approaches reveal significant methodological incongruence. While adopting the term ‘evolutionary’, they continue to repose faith in the various optimization-based methods of cost calculations. A fast and frugal approach, on the other hand, has the potential to embrace the logic of evolutionary policymaking, as illustrated by Witt (2003), more fully, since both streams of scholarship share the common theoretical premise of bounded rationality, learning and adaptation to achieve a satisficing outcome.

Indeed, combining the thoughts of frugality and evolutionary policy discourse, one could ponder upon the following takeaways. First, there is a need to promote the scope of ‘learning along the way’, especially for policies in the areas of innovation and capability building. The emerging policy instrument of ‘spiral induction’ could achieve this if configured around the nuances of frugality and non-optimization. Secondly, and finally, encouraging interactions between the users, producers and regulators of technology from the very beginning could help to develop a shared understanding of the need for adaptation and incremental learning in the course of innovation, eventually, reducing the reliance on optimization based ‘zero-error’, and paving way for a judicious use of economic resources.