TY - JOUR
AU - Bhaduri,, Saradindu
AB - Abstract Recent scholarship on technological path dependence has questioned the credibility of many previous studies for basing their analyses on cases settled in the past. Such cases often do not provide adequate insights into the historical contingencies, nature of incremental innovations, and the type of feedback mechanisms between consumers and producers of technologies. In addition, most of the cases dealing with path dependence focus on firms as users of technology, and miss out on the influence the consumers may have in shaping path dependence. We take the case of water purification technologies in India, where the technologies are still evolving, and analyze the path dependence implications of their evolution by taking the case of urban households with piped water supply. Using patent data, scientific studies, as well as interviews with firms and individuals, we analyze how through various incremental innovations and information contagion, reverse osmosis technologies demonstrate the features of path dependence, leading, possibly, to lock-in. 1. Introduction The theories of technological path dependence, and subsequent technological lock-ins, refer to the idea of selection and persistence of a “not-so-good” technology (David 1985, 2000; Arthur 1989, 1994; Cowan, 1990; Foray 1997; Puffert, 2002, 2008; Meyer, 2012; Dobusch and Schubler, 2012; Bergek and Onufrey, 2013). While historical contingencies play a significant role in selecting the “not-so-good” technologies, their retention is made possible through self-reinforcement due to factors like learning, incremental innovations, and information contagion. From the perspective of mainstream economic theories, these processes contradict the assumptions that efficient technologies are always chosen, and consumers are perfectly rational. Quite often, however, these studies rely on cases of technology adoptions which got settled in the distant past. Many of the competing technologies being already driven out of the market, data regarding their efficiency become scarce, leading to analyses of, the so called, “would have been” types (Foray, 1997; Liebowitz and Margolis, 1995; Page, 2006; Vergne and Durand, 2010; Meyer, 2012). This has seriously impaired the credibility of the notion. The present article analyzes the process of technology adoption and innovation in currently evolving water purification devices in India with the aim to understand what causes adoption of “inferior technologies” and how innovations, thereafter, shape the process of path dependence and lock-in. Accepting the criticisms offered by Liebowitz and Margolis (1995), we focus on the so called “third degree” path dependence, where the technology in focus is chosen even when other, purportedly superior, technologies were present in the market. Almost all existing studies conclude that incremental innovations have helped overcome the problems of inferiority of the selected technologies, and, over time, these “once inferior” technologies have become superior to the technologies which “should have been chosen” instead. In our view, it is premature to conclude that innovations in an already chosen, but inferior, technology necessarily make the technology superior, compared to its competing technologies, and help it escape “path dependence” or “lock-in,” post-adoption. To us, such a conclusion would depend on the nature of innovations, and carefully chosen parameters of quality and efficiency. Unfortunately, the existing studies do not engage with either of these issues. They do not analyze in detail the nature and pattern of innovations that the path-dependent technology goes through, nor do they always compare efficiency or superiority based on any clearly defined set of parameters.1 Problems of data are a major bottleneck here, especially when many of the “supposedly better” technologies may have already disappeared from the market. We examine path dependence and incremental innovations by incorporating the demand side into our analysis. Putting consumers into the analysis helps us to understand the role of preference heterogeneity in path dependence and lock-in.2 Brian Arthur in many of his writings points out that monopoly is difficult to obtain under preference heterogeneity. Understanding path dependence in such a situation in our view is more challenging, which not many studies have attempted.3 Moreover, by including the way consumers gather information, and articulate their preferences for appropriate purification technologies, we are able to examine the role of information contagion in shaping the process of path dependence and lock-in. In addition, how consumers articulate their preference for a “credence good” whose quality is often difficult to judge either by “search” or through “experience” (Nelson 1970) can have important policy implications, particularly, in a developing country context, where regulation making is in a nascent stage. Existing studies in this area do not engage with these questions. The market of water purifiers is expanding rapidly, across the world.4 Currently, its rate of growth is pegged at 10% per year, between 2013 and 2018.5 We take the case of water purification technologies in urban India. Most of the studies on path dependence are based on the United States (David, 1985; Cowan, 1990; Cusumano, et al., 1992; Cowan and Hulten, 1996; Cowan and Gunby, 1996 etc.). Indian market is important in this context for, at least, two reasons. First, India being a tropical country has a very high intake of raw water compared to Europe and the United States where much of the water consumption is in the form of liquids of various forms (Sharma and Bhaduri, 2013). Second, diversity in the quality of source water across the country may necessitate different technological solutions making path dependence or lock-in of a single technology less likely. While ultraviolet (UV) has been in the market for almost 30 years, reverse osmosis (RO), despite being a later entrant, has grown tremendously in the past 10–15 years and is expected to increase further.6 It enjoys a clear, nondecreasing, edge, despite being newer, resource in-efficient,7 having alleged adverse health impacts (Kozisek, 2004), and being costlier than other purifiers.8 Together, UV and RO share almost 90% of the market for inline purification technologies. One major type of innovation in this sector is the process of combination of UV, ultrafilter (UF), or nano-filter (NF) with RO. Therefore, we observe coexistence of multiple technologies separately, as well as in various combinations of them. RO is found to be a common technology in all these combinations.9 For this study, we have taken the case of households who are connected to piped10 water supplies from municipal bodies. Finally, a caveat, the more recent scholarship in this area, seems to be divided over the definition of path dependence. Scholars like Vergne and Durand (2010), and Kay (2013) suggest, much in line with the “first generation” scholars, a narrow definition of path dependence. Scholars like Dobusch and Schubler, (2012), Bergek and Onufrey (2013), on the other hand, call for a rather broader understanding of path dependence, making it synonymous with “structural inertia,” “coevolution,” or “institutional persistence.” While these latter group of scholars promise many new insights into the process of path dependence, too broad a concept of path dependence, in our view, obscures our attention from the core economic concerns of efficiency, or rather inefficiency, inherent in the conceptualizations of path dependence and lock-in.11 We, therefore, proceed with a “narrow” definition of path dependence. The article has six sections. Section 2 gives a critical overview of the literature on path dependence and technological lock-in. Section 3 describes the various technologies in the field of purification of drinking water and their market shares in Indian context. Section 4 discusses our method of collecting data, sample features, and method of analysis. In Section 5, we analyze the case of RO purifier adoption by the urban households in India by first discussing the evolution of RO technology, followed by a comparative discussion of various purification devices, and juxtaposing these features with our survey findings. Section 6 makes the key concluding remarks. 2. Path dependence and lock-in: a critical overview David (1985, 1997, 2000) and Arthur (1989, 2004) argue that when technologies are accepted/diffused in the market, then their market shares or their fate in the market is not shaped by the fact that they are the “best” technologies but because of some “trivial” reasons in the history of their evolution. These “trivial” reasons or “historical small events” as described by Arthur (1994; 17) are “those events or conditions that are outside the knowledge of the observer, beyond the resolving power of his ‘model’ or abstraction of the situation.” The various reasons attributed to the process of path dependence could be (i) durability of capital equipment—obsolete, inferior equipment may remain in use because its fixed cost is already paid for, (ii) technical interrelatedness, and (iii) increasing returns—learning effects, positive network externalities, etc. (David, 1985; Arthur, 1989; Puffert, 2008).12 David’s early work on path dependence represents, in part, the culmination of an earlier economic literature on technical interrelatedness (Puffert, 2008). W. Brian Arthur, the other pioneering scholar on path dependence, on the other hand, based his ideas on an analogy between increasing returns in the economy, particularly when expressed in the form of positive externalities, and conditions that give rise to positive feedbacks in the natural sciences. Arthur, in a series of theoretical papers (1989, 1990, and 1994), emphasizes the role of “increasing returns to adoption” that develop over time. These increasing returns might arise on the supply side of a market, as a result of learning effects that lower the cost or improve the quality of a product as its cumulative production increases. Alternatively, increasing returns might arise on the demand side of a market, as a result of positive network externalities, which raise the value of a product or technique for each user as the total number of users increase (Katz and Shapiro, 1994). Arthur (1989), accordingly, remarks that “the usual policy of letting the superior technology reveal itself…is appropriate in the cases of constant and diminishing returns to adoption. But in increasing returns cases, there is no guarantee that the superior technology survives.” Technological lock-in is a way to describe the entry of a system into a trapping region that surrounds a locally (or globally) stable equilibrium (David, 2000). A technology that by chance gains an early lead in adoption may eventually “corner the market” of potential adopters, with the other technologies becoming locked out (Arthur, 1989). However, a mere accidental adoption does not necessarily lead to “lock-in” to inefficient technology, since this may be an outcome of an optimal choice being made at each stage with given information and constraints.13 When the buyers are risk averse, they tend to buy products that they know more about, bestowing advantages to certain products and technologies. This phenomenon is called as “information contagion” by Arthur (1994: 69; Arthur and Lane, 1993). Narduzzo and Warglien (1996) observed that while choosing a product, the individuals may make their choices based on increasing their benefits (mean rule14 or maximum rule15), reducing their risks (minimum rule16) or follow what others are doing (popularity rule17). Through their experiment, they observed that while minimum, maximum, and mean rules did not lead to path dependence and lock-in, popularity rules did (Lane and Vescovini 1996). In our view, this aspect needs further qualification by analyzing the quality characteristics of a product or technology. For goods whose quality can be ascertained through “search” and “experience,” efficiency is easy to judge (Nelson 1970). The quality of credence goods, on the other hand, is difficult to judge through simple search (of information) and experience (through consumption). Quite often, the quality implications of these goods are revealed after sufficient time gap (e.g., safety of certain pharmaceuticals) for which market does not reveal enough information. As a result, “popularity rule” may, at times, lead to selection of most efficient technologies when the quality in question can be ascertained through search or experience. For credence goods, however, popularity rule is more certain to lead to wrong choice, signifying market failure.18 Liebowitz and Margolis (1995) distinguish path dependence in three categories (degrees). First, degree path dependence is shown as a dynamic process, which is sensitive to initial decisions. This kind of path dependence assumes that either the inefficiency has been taken fully into account (assuming complete information) or it never had any inefficiency at all. Second, degree path dependence refers to cases with imperfect sight. Here the decisions are assumed to be taken to be best according to the information available, while it was found to be inefficient at a later date. Third, degree path dependence occurs “when there exists some feasible arrangement for recognising and achieving a preferred outcome, but that outcome is not obtained,” often for reasons which are noneconomic (buying, for instance, a house near one’s friend’s house, disregarding problems of transportation). In view of Liebowitz and Margolis (1995), only the “third-degree” path dependence would be considered inefficient, leading to “lock-in” through market failure. Therefore, we can say that path dependence does not necessarily cause market failure, but it creates the potential for market failure when it is of third degree.19Kay (2013) agreed with Liebowitz and Margolis (1995, 1999) in their assertion that none of the cases till date have come any close to being called third-degree path dependence. Kay shows that the QWERTY design was near optimal even at the time of its selection, and not merely chosen as a result of some accident in history. Vergne (2013), in reply to this article, argues that the theory of path dependence should not be invoked everywhere, but it should be talked of when other theories are not able to explain the phenomenon in question. Lock-in to a monopoly situation with an inefficient technology is also difficult in the face of preference heterogeneity. In such cases the two competing technologies may be present together, without any one of them getting sufficiently ahead to cause lock-in (Arthur, 1989). However, studies have empirically examined neither the reasons nor the implications of preference heterogeneity for path dependence and lock-in. Vergne and Durand (2010) talk of the possible difficulties arising out of empirically testing path dependence and advice for a narrow definition of path dependence, as done by their predecessors (David, 1985; Cowan, 1990; Puffert 2002; Cusumano et al., 1992; Liebowitz and Margolis, 1995; Cowan and Gunby 1996; Foreman-Peck, 1996) and recently by Kay (2013). Some recent studies, on the other hand, have called for broadening the concept of path dependence, to encompass the phenomena of structural inertia, coevolution, or institutional persistence (Dobusch and Schubler, 2012; Bergek and Onufrey, 2013). Bergek and Onufrey (2013), using such a broader notion of path dependence, analyze patent data to show how persistence is observed not only within a path but also across paths. This study promises new insights into the process of incremental innovation in path dependence—a point widely mentioned but only rarely examined in detail. To be more precise, Bergek and Onufrey (2013) highlight how persistence of a technological path can be observed even in the presence of new knowledge combinations (patent class). However, the study confesses to a serious limitation of not effectively addressing the self-reinforcement mechanism, which is a crucial building block of path dependence. Without a clear self-reinforcement mechanism, it is difficult to ascertain how new paths are generated. In our view, too broad a definition of path dependence, which include structural inertia and institutional persistence, obscures our attention from analyzing the impact of factors like preference heterogeneity of consumers and the self-reinforcement process, which lie at the core of the idea of path dependence and lock-in. We, therefore, opt for a narrow definition of path dependence, as a persistent process attributed to self-reinforcement, for this study. As suggested by Vergne and Durand (2010), we conduct an empirical study by stipulating the relevant properties of technologies that are to be taken to evaluate alternative paths (like cost, quality of source water, etc.) and the potential contingencies that occurred at the time of selection, and thereafter, specifying the “components of self-reinforcement.” 2.1 Incremental innovations and path dependence The major role that minor or incremental innovations play in shaping the diffusion of a technology is well-known (Abernathy and Utterback, 1978; Rosenberg, 1976, 1982). Rosenberg (1982) points out that such stream of improvements crucially shapes the performance of invention, its modification and adaptation to suit submarkets, and the availability and introduction of other complimentary inputs. The literature on path dependence argues that the technology once selected is improved through various incremental innovations, and these innovations make the technology “superior” in the contemporary sense, even when it might not have been the case when they were selected (David 1985, 2000; Arthur 1989, 1994; Cowan, 1990; Foray 1997; Puffert, 2002, 2008; Meyer, 2012). Cowan (1990) in his study of the nuclear reactors finds that the light water technology might be dominating the market now (owing to various improvements over the years), but it did not have a clear advantage over other technologies (heavy water or gas graphite) at the time of its selection. Similarly, in the cases of VHS vs. Betamax video recorders (Cusumano et al., 1992), pest control strategies (Cowan and Gunby, 1996), and railway track (Puffert, 2002), the authors reflect that the prevailing technology is well-accepted in the current scenario due to their increased “efficiency” or workability owing to various incremental innovations, in the aftermath of their initial “selection” in the market. However, without a detailed account of the nature of these incremental innovations and clear evaluative criteria of how these incremental innovations may have raised efficiency of the already selected, yet inefficient, technologies, these accounts remain, at best, conjectural hypotheses difficult to defend. Kay (2013), for instance, challenging these postulates argues that QWERTY was efficient even at the time of its introduction. One important way to analyze the nature and dynamics of incremental innovation is by incorporating the demand side of a technology. Incorporation of preference heterogeneity in the empirical analysis of path dependence is rarely done. In the absence of clear regulatory guidelines, preferences can indeed be a key driving force behind adoptions and innovations of a technology. 3. Water purification technologies: the unfolding scenario For common people, purity of drinking water has two, often nonoverlapping, dimensions: aesthetic characteristics (e.g., taste, turbidity, and smell) and disease-causing properties (WHO, 2008). While aesthetic characteristics can be ascertained through search (turbidity or smell), or experience (taste), disease-causing properties are difficult to ascertain at the level of individual consumers. An ideal water purification device, therefore, possesses credence good characteristics. In recent years, chemical and microbial contamination of drinking water has emerged as a major public health concern. These concerns have led to innovation of several purification techniques, including hyperfiltration, and radiation treatment (US-EPA, 1999). The users, currently, have a vast diversity of water purification technologies to choose from depending, ideally, on the quality of source water. No one method can be said to be capable of removing all the contaminants or solving all the water quality problems (Table 1). Table 1. Technologies/methods for some common drinking water pollutants Contaminant . Causes of its presence in water . Possible purification methods . Pollutants/contaminants causing health hazards Chemical contaminants Arsenic Naturally occurring in water in some areas RO; ion exchange Lead Corrosive water, lead pipes or lead solder RO; distillation Nitrate Well not sealed; faulty septic system; animal waste; fertilizers Distillation; RO; anion exchange (water softener) Pesticides and organic chemicals Use of pesticides and chemicals near water source Activated carbon filter; RO; distillation Biological contaminants Bacteria/viruses Well not sealed; sewage, manure or surface runoff Chlorination; ozonation; UV disinfection Contaminants that do not cause health hazards may be treated for their aesthetic value Bad odor and color, taste Variety of sources Ion exchange; activated carbon filter; chlorination Cloudy or dirty water Fine sand, clay, or other particles Mechanical filter Hardness Naturally occurring minerals in water Ion exchange (water softener) Rotten-egg odor Hydrogen sulfide gas Chlorination and activated carbon filter Staining of sink and/or laundry (iron or manganese) Naturally occurring in water, especially deep wells Ion exchange, chlorination and filtration Contaminant . Causes of its presence in water . Possible purification methods . Pollutants/contaminants causing health hazards Chemical contaminants Arsenic Naturally occurring in water in some areas RO; ion exchange Lead Corrosive water, lead pipes or lead solder RO; distillation Nitrate Well not sealed; faulty septic system; animal waste; fertilizers Distillation; RO; anion exchange (water softener) Pesticides and organic chemicals Use of pesticides and chemicals near water source Activated carbon filter; RO; distillation Biological contaminants Bacteria/viruses Well not sealed; sewage, manure or surface runoff Chlorination; ozonation; UV disinfection Contaminants that do not cause health hazards may be treated for their aesthetic value Bad odor and color, taste Variety of sources Ion exchange; activated carbon filter; chlorination Cloudy or dirty water Fine sand, clay, or other particles Mechanical filter Hardness Naturally occurring minerals in water Ion exchange (water softener) Rotten-egg odor Hydrogen sulfide gas Chlorination and activated carbon filter Staining of sink and/or laundry (iron or manganese) Naturally occurring in water, especially deep wells Ion exchange, chlorination and filtration Source: Compiled from Daniels and Mesner (2010). Table 1. Technologies/methods for some common drinking water pollutants Contaminant . Causes of its presence in water . Possible purification methods . Pollutants/contaminants causing health hazards Chemical contaminants Arsenic Naturally occurring in water in some areas RO; ion exchange Lead Corrosive water, lead pipes or lead solder RO; distillation Nitrate Well not sealed; faulty septic system; animal waste; fertilizers Distillation; RO; anion exchange (water softener) Pesticides and organic chemicals Use of pesticides and chemicals near water source Activated carbon filter; RO; distillation Biological contaminants Bacteria/viruses Well not sealed; sewage, manure or surface runoff Chlorination; ozonation; UV disinfection Contaminants that do not cause health hazards may be treated for their aesthetic value Bad odor and color, taste Variety of sources Ion exchange; activated carbon filter; chlorination Cloudy or dirty water Fine sand, clay, or other particles Mechanical filter Hardness Naturally occurring minerals in water Ion exchange (water softener) Rotten-egg odor Hydrogen sulfide gas Chlorination and activated carbon filter Staining of sink and/or laundry (iron or manganese) Naturally occurring in water, especially deep wells Ion exchange, chlorination and filtration Contaminant . Causes of its presence in water . Possible purification methods . Pollutants/contaminants causing health hazards Chemical contaminants Arsenic Naturally occurring in water in some areas RO; ion exchange Lead Corrosive water, lead pipes or lead solder RO; distillation Nitrate Well not sealed; faulty septic system; animal waste; fertilizers Distillation; RO; anion exchange (water softener) Pesticides and organic chemicals Use of pesticides and chemicals near water source Activated carbon filter; RO; distillation Biological contaminants Bacteria/viruses Well not sealed; sewage, manure or surface runoff Chlorination; ozonation; UV disinfection Contaminants that do not cause health hazards may be treated for their aesthetic value Bad odor and color, taste Variety of sources Ion exchange; activated carbon filter; chlorination Cloudy or dirty water Fine sand, clay, or other particles Mechanical filter Hardness Naturally occurring minerals in water Ion exchange (water softener) Rotten-egg odor Hydrogen sulfide gas Chlorination and activated carbon filter Staining of sink and/or laundry (iron or manganese) Naturally occurring in water, especially deep wells Ion exchange, chlorination and filtration Source: Compiled from Daniels and Mesner (2010). From the table, we can deduce that RO is suitable to remove chemical contaminants from water, whereas UV technologies, chlorination, and ozonation are more suitable for the removal of living microorganisms. Besides, chlorination, ozonation, ion exchange are found to be effective to improve sensory properties of water, which, at times, is given more importance by people compared to chemical and microbiological pollutants (WHO, 2004). It may be noted that UV does not alter taste, and RO, in fact, makes water sour by removing minerals (total dissolved solid, TDS) in an indiscriminate manner. It may be worth noting that the impact of hardness on disease is not clearly established, and the WHO regulations do not make any specific suggestion on removing these minerals from water. European law, in fact, bans such indiscriminate acts of demineralization. In India, people rely on multiple sources for their drinking water. Emphasis on these sources depends on location, where urban people rely comparatively more on municipal supply of water, and groundwater through boring and tube well/well is the most common source in rural areas. Nowadays, very few people rely on surface water for drinking purposes. It is important to note here that disease-causing microorganisms are found in all kinds of water sources. Inorganic contaminants including pesticides are specific to groundwater in some areas, and the municipal supplied water is assumed to have been treated for these contaminants.20 The municipal water is also treated for microorganisms. The only way these pollutants may find their way to water at the point of consumption is through leakages in distribution pipes. Awareness about water quality in India rose only in the late 1990s (Bhaduri and Sharma, 2012). At the policy level too, the quality of drinking water was a neglected phenomenon until the late 1990s. The standards for Natural Mineral Water and bottled water came as late as 1992 and 1998, respectively, but they were made mandatory even later, only in 2001. However, till date, there is no regulatory mechanism in place for ensuring quality of water purification devices. The public outcry by the Centre for Science and Environment in the year 2003, protesting against the presence of pesticides in bottled water, was a landmark event. This event led to the formation of only fourth, and first in public health, Joint Parliamentary Committee in the history of independent India. Subsequently, we have observed a phenomenal growth in water purification technologies in India. A TV Veopar ADI Media (2012) publications research considered five most prominent brands in India and found out that off-line water purification technologies account for around 75% of market share in this industry. Given the skewed income distribution in India, this choice is not a surprise. Due to their low price (ranging between INR 2000 and 3000), however, off-line21 purifiers only have around 32% share of revenues earned by the industry. The in-line22 purifier industry is dominated by RO, UV, and the various combinations involving them. Mostly, these combinations have RO common in them. Together, RO-related water purifiers have a higher share of revenues than UV technologies (about 508 crores as against 265 crores of in-line/UV purifiers), though the latter have more sales compared to the former (about 520,000 units and 395 units, respectively). This reflects higher prices of RO (ranging between INR 15,000 and 45,000) compared to, for instance, UV purifiers (priced around INR 10,000).23 Industry reports predict further growth of this industry. A study by the Associated Chambers of Commerce of India (ASSOCHAM)24 predicted that the water purifier market would grow 100% between 2012 and 2015. Of the total installed base of purifiers, UV systems account for a market share of 51%. Products based on RO too have a significant share of 42%. Their share is rising because of widely held belief that RO is the “best method” (Marwas, 2010). It is important to note here that UV purifiers came as early as 1984 in India, while RO purifiers came in 1999. RO-based purifiers, therefore, seem to be closing the gap very fast. The water purifier market in India is based only on assembly of the parts, with membranes and pumps imported from China or Korea. Reportedly, Delhi has around 5000 assembling units around the city. Firms across size class import the membranes and pumps from these countries and assemble their devices within the country. There are no regulations or guidelines regarding manufacturing of these devices (Bhaduri et al 2015). The brand name is important, which is gained by durability, and post-marketing services. We now lay out the various methods of data collection and analysis we pursue to understand its evolving trajectory while, at the same time, exploring the nuances of path dependence in this industry. 4. Data and methods of analysis We have multiple sources of data, which we discuss below. 4.1 Patent analysis and review of scientific literature We reviewed scientific papers and patents to explore the emergence of the various technological options for drinking water purification, the pattern of their change. We use ion exchange, chlorination/bromination, ozonation, UV radiation, and RO for comparative understanding of their usability and dominance. These technologies were shortlisted on the basis of their prevalence in Indian markets and their usefulness to remove different types of contaminants. Patent search gives us an understanding of the nature of innovations in these technologies and their relative importance in the market. Since India imports most of the platform technologies for urban household water purification, we conducted the patent search on global patent databases. Our main source for patents is the United States Patent and Trademark Office (USPTO).25 In addition, we have obtained necessary patent data for India-specific innovations in the field of RO technologies from the online database of Indian Patent Office (IPO). Here, a “keyword search” of granted patents was adequate to obtain the patent on “total dissolved solid” (number 199716). For USPTO, patent classification (class/subclass, coded as CCL) has been identified for each of these technologies, to obtain the total number of patents in each category.26 For instance, to obtain the number of patents in the field of RO for liquid separation, we use “CCL/210/652” as our search query. Furthermore, yearly, and decadal numbers have been obtained by using “range of issue dates (ISD)” in the search. For country-wise data, the query included the class name and the assignee country code. For example, to find the number of patents assigned to Japan in the field of RO, the query used was “CCL/210/652” and “ACN/JP.” ACN stands for assignee country, and JP is code for Japan. For United States, the data are available through the assignee state within the country, which then was added to get the total number of patents assigned to the United States. Some of the documents were studied in detail on the basis of their chronological order, to gain insights into the nature of innovations, for instance, in the areas of energy, and performance efficiency. 4.2 Firm interviews People from purifier manufacturing firms were interviewed (telephonic/personal) to obtain information about the nature of the industry and its innovation dynamics. We obtained major insights into these activities through our visits to the Aquatech India, the key trade fair for purification firms, in 2012 and 2013. Around 130 firms, mostly large domestic and multinational, belonging to manufacturing and regulation of water purification participated in these events. In addition, we conducted an interview with a small-scale manufacturer of RO technology in Delhi in July 2014. Altogether six personal interviews were taken. Note that RO is not only used for household water purification but also for bottled water processing. We surveyed one firm, one owner of a now closed down firm, and a technology consultant in Kolkata to gain perspectives from the bottled water manufacturers particularly on emerging incremental innovations in this industry. 4.3 Household survey We conducted an off-line and an online survey of users of water purifiers. Both surveys followed a snowball-cum-random method of data collection. 4.3.1 Field survey It was conducted during February–April 2013. Households were interviewed with the help of a semi-structured questionnaire to gather information on the pattern of consumer choices and reasons behind these choices. The questionnaire had both closed- and open-ended questions. Through this survey, we collected data from 25 households in Delhi. The localities targeted for this survey had mostly municipality supplied water, which is treated for chemical contaminants at the source.27 4.3.2 Online survey An online survey was conducted using surveymonkey.com to supplement the sample. This survey was kept active for 3 months (May 14, 2014–July 14, 2014). The questionnaire had all the closed-ended questions of the other questionnaire, along with a few additional questions. The link to the site was sent to various online e-mail groups of academic institutions, and posted on our social networking sites like LinkedIn and Facebook. We urged our social peers to circulate it among their social groups as well. We got 108 responses online. We had to discard four incomplete responses for the final analysis. These 104 online responses were merged with the other 25 responses that we obtained through the field survey in Delhi (Table 2). Merging the two data sets enabled us to capture data from tier 128 cities other than Delhi. Through online survey, we also could obtain data from 28 tier 2 cities, and 6 tier 3 cities. This combined data set, therefore, captures the geographic variation and social diversity better than any of the two individual data sets. Table 2. Sample description Variable . Type of variable . Number of responses . Gender Binary 128 Age (in years) Category 128 Occupation Category 129 Education Category 129 Family income (in rupees) Category 129 Place of stay Category 129 Period of stay (in years) Category 125 Source of water Category (multiple choices) 129 Water quality (perception) Category (ranking) 120 Water purification method (current) Category 121 Period of use (in years) Category 90 Reasons for choosing this method (perception) Category (multiple choices) 84 Water purification method (previous) Category 92 Reasons for switch (perception) Category (multiple choices) 67 Purifier (brand name) Category 80 Reasons for choosing brand (perception) Category (multiple choices) 81 Source of information (brand) Category (multiple choices) 81 Quality of RO water (perception) Category (ranking) 52 Benefits of RO (perception) Category (multiple choices) 54 Drawbacks of RO (perception) Category (multiple choices) 57 Use of rejected water Category (multiple choices) 47 Frequency of servicing/cleaning Category 46 Reduction in disease because of RO (perception) Category 48 Reasons (reduction disease) (perception) Category 13 RO vs. previous method (perception) Category (ranking) 34 Reasons (no method required) (perception) Category (multiple choices) 40 Reasons (boiling/chemical) (perception) Category (multiple choices) 28 Variable . Type of variable . Number of responses . Gender Binary 128 Age (in years) Category 128 Occupation Category 129 Education Category 129 Family income (in rupees) Category 129 Place of stay Category 129 Period of stay (in years) Category 125 Source of water Category (multiple choices) 129 Water quality (perception) Category (ranking) 120 Water purification method (current) Category 121 Period of use (in years) Category 90 Reasons for choosing this method (perception) Category (multiple choices) 84 Water purification method (previous) Category 92 Reasons for switch (perception) Category (multiple choices) 67 Purifier (brand name) Category 80 Reasons for choosing brand (perception) Category (multiple choices) 81 Source of information (brand) Category (multiple choices) 81 Quality of RO water (perception) Category (ranking) 52 Benefits of RO (perception) Category (multiple choices) 54 Drawbacks of RO (perception) Category (multiple choices) 57 Use of rejected water Category (multiple choices) 47 Frequency of servicing/cleaning Category 46 Reduction in disease because of RO (perception) Category 48 Reasons (reduction disease) (perception) Category 13 RO vs. previous method (perception) Category (ranking) 34 Reasons (no method required) (perception) Category (multiple choices) 40 Reasons (boiling/chemical) (perception) Category (multiple choices) 28 Source: Field work. Table 2. Sample description Variable . Type of variable . Number of responses . Gender Binary 128 Age (in years) Category 128 Occupation Category 129 Education Category 129 Family income (in rupees) Category 129 Place of stay Category 129 Period of stay (in years) Category 125 Source of water Category (multiple choices) 129 Water quality (perception) Category (ranking) 120 Water purification method (current) Category 121 Period of use (in years) Category 90 Reasons for choosing this method (perception) Category (multiple choices) 84 Water purification method (previous) Category 92 Reasons for switch (perception) Category (multiple choices) 67 Purifier (brand name) Category 80 Reasons for choosing brand (perception) Category (multiple choices) 81 Source of information (brand) Category (multiple choices) 81 Quality of RO water (perception) Category (ranking) 52 Benefits of RO (perception) Category (multiple choices) 54 Drawbacks of RO (perception) Category (multiple choices) 57 Use of rejected water Category (multiple choices) 47 Frequency of servicing/cleaning Category 46 Reduction in disease because of RO (perception) Category 48 Reasons (reduction disease) (perception) Category 13 RO vs. previous method (perception) Category (ranking) 34 Reasons (no method required) (perception) Category (multiple choices) 40 Reasons (boiling/chemical) (perception) Category (multiple choices) 28 Variable . Type of variable . Number of responses . Gender Binary 128 Age (in years) Category 128 Occupation Category 129 Education Category 129 Family income (in rupees) Category 129 Place of stay Category 129 Period of stay (in years) Category 125 Source of water Category (multiple choices) 129 Water quality (perception) Category (ranking) 120 Water purification method (current) Category 121 Period of use (in years) Category 90 Reasons for choosing this method (perception) Category (multiple choices) 84 Water purification method (previous) Category 92 Reasons for switch (perception) Category (multiple choices) 67 Purifier (brand name) Category 80 Reasons for choosing brand (perception) Category (multiple choices) 81 Source of information (brand) Category (multiple choices) 81 Quality of RO water (perception) Category (ranking) 52 Benefits of RO (perception) Category (multiple choices) 54 Drawbacks of RO (perception) Category (multiple choices) 57 Use of rejected water Category (multiple choices) 47 Frequency of servicing/cleaning Category 46 Reduction in disease because of RO (perception) Category 48 Reasons (reduction disease) (perception) Category 13 RO vs. previous method (perception) Category (ranking) 34 Reasons (no method required) (perception) Category (multiple choices) 40 Reasons (boiling/chemical) (perception) Category (multiple choices) 28 Source: Field work. The field survey could be conducted only in Delhi, and among the lower- and middle-income group households (84% of the respondents). It was difficult to access the households belonging to higher-income groups (income above INR 1 million per year), as these households often live in gated communities, and do not entertain unknown visitors. In addition, through household survey, we could only meet women and people belonging to the middle-aged group. We could not obtain the perspective of younger section of people, who in our view play an important role in deciding about new gadgets in urban households. To ameliorate this problem of underrepresentation of social groups and geographic locations, we conducted the online survey. Through the online survey, we could obtain around 37% of responses from the income group having annual income of INR 1 million and more.29 The online survey also allowed us to go beyond Delhi and capture tier 2 and tier 3 cities, which finally constitutes around 27% of our sample. Many of our online respondents also came from younger age group. In the final sample, the younger age group (aged between 18 and 35 years) constitute 62% of the sample. We use these data for quantitative analysis of household behavior through descriptive statistics and regression. In addition, the qualitative information obtained through our field surveys is used for substantiating our findings from patent databases and survey of firms. 4.4 Method of analysis We studied scientific papers, technical documents, and patent descriptions to understand the history of RO and other competing technologies for a comparative evaluation. Firms were surveyed to explore the nature and drivers of innovation in the Indian household water purifier industry. We use the primary household survey data to analyze the nature and impact of information contagion. These data have been used in multiple ways. Mostly, we have used these data for simple descriptive statistics and correlations. To clearly bring out the dynamics of information contagion, we also use these data for a probit analysis. We use STATA (version 12) for this analysis. 5. Analyses and discussion 5.1 RO and other purification technologies: a comparative overview To analyze RO’s technical superiority and cost-efficiency, we refer to Table 3. Technical superiority is measured through the capabilities to remove microbe, and chemicals. In addition, we take into consideration the vulnerability to contamination during the storage of water, purified through the various available techniques. To understand the costs-effectiveness of RO, we compare it with other available technologies in terms of actual costs, health costs, environmental costs, infrastructure requirement costs, and other costs. Table 3. Comparing different methods of water purification Parameters . Methods of water purification . . . Ceramic filters (candle filter) . Boiling . Chlorination . Activated carbon . UV . RO . Technological efficacy Removal of microbes √ √√√√√ √√√ √√√ √√√√ √√√√ (membrane may be degraded by microbes— fouling) Removal of chemicals √ √ √√ √√√ – √√√√ Storage (in context of the activity of water toward absorbing chemicals after purification) √√√ √√√ √√√ √√√ √√√ √ (highly active water; vulnerable to absorption of ions) Costs Actual costs Installation costs √ – – √√ √√√ √√√√√ Maintenance costs – – – √ – √√ Operation costs – √ √ – √ √ Health costs Mineral content of purified water √√√√√ √√√ √√√√ √√√√ √√√√√ √ Reduction in pH – – – – – √√ Sweetness (taste) of purified water √√√ √ √√ √√√√√ √√√√ √ Environmental costs Wastage of water in the process – – – – – √√√ Electricity consumption – – – – √√ √√√ Infrastructural costs Dependence on water pressure – – – – – √√√ Dependence on electricity – – – – √√√ √√√ Other costs Aesthetic qualities of purified water √√ √ √√√ √√√√√ √ √ Requirement of user expertise/knowledge √ √ √ √ √ √√ Ease of handling √√√ √ √ √√√ √√√√√ √√√ Parameters . Methods of water purification . . . Ceramic filters (candle filter) . Boiling . Chlorination . Activated carbon . UV . RO . Technological efficacy Removal of microbes √ √√√√√ √√√ √√√ √√√√ √√√√ (membrane may be degraded by microbes— fouling) Removal of chemicals √ √ √√ √√√ – √√√√ Storage (in context of the activity of water toward absorbing chemicals after purification) √√√ √√√ √√√ √√√ √√√ √ (highly active water; vulnerable to absorption of ions) Costs Actual costs Installation costs √ – – √√ √√√ √√√√√ Maintenance costs – – – √ – √√ Operation costs – √ √ – √ √ Health costs Mineral content of purified water √√√√√ √√√ √√√√ √√√√ √√√√√ √ Reduction in pH – – – – – √√ Sweetness (taste) of purified water √√√ √ √√ √√√√√ √√√√ √ Environmental costs Wastage of water in the process – – – – – √√√ Electricity consumption – – – – √√ √√√ Infrastructural costs Dependence on water pressure – – – – – √√√ Dependence on electricity – – – – √√√ √√√ Other costs Aesthetic qualities of purified water √√ √ √√√ √√√√√ √ √ Requirement of user expertise/knowledge √ √ √ √ √ √√ Ease of handling √√√ √ √ √√√ √√√√√ √√√ Note: The strengths are given by us on the basis of our understanding of the source documents. The √signs in the above table show the strength of the parameter in question. The higher the number of the √sign, stronger is the effect of the relevant parameter. Sources: World Health Organization (1996, 2003, 2008, 2011a,b), PBI (2013), Kozisek (2004), field survey. Table 3. Comparing different methods of water purification Parameters . Methods of water purification . . . Ceramic filters (candle filter) . Boiling . Chlorination . Activated carbon . UV . RO . Technological efficacy Removal of microbes √ √√√√√ √√√ √√√ √√√√ √√√√ (membrane may be degraded by microbes— fouling) Removal of chemicals √ √ √√ √√√ – √√√√ Storage (in context of the activity of water toward absorbing chemicals after purification) √√√ √√√ √√√ √√√ √√√ √ (highly active water; vulnerable to absorption of ions) Costs Actual costs Installation costs √ – – √√ √√√ √√√√√ Maintenance costs – – – √ – √√ Operation costs – √ √ – √ √ Health costs Mineral content of purified water √√√√√ √√√ √√√√ √√√√ √√√√√ √ Reduction in pH – – – – – √√ Sweetness (taste) of purified water √√√ √ √√ √√√√√ √√√√ √ Environmental costs Wastage of water in the process – – – – – √√√ Electricity consumption – – – – √√ √√√ Infrastructural costs Dependence on water pressure – – – – – √√√ Dependence on electricity – – – – √√√ √√√ Other costs Aesthetic qualities of purified water √√ √ √√√ √√√√√ √ √ Requirement of user expertise/knowledge √ √ √ √ √ √√ Ease of handling √√√ √ √ √√√ √√√√√ √√√ Parameters . Methods of water purification . . . Ceramic filters (candle filter) . Boiling . Chlorination . Activated carbon . UV . RO . Technological efficacy Removal of microbes √ √√√√√ √√√ √√√ √√√√ √√√√ (membrane may be degraded by microbes— fouling) Removal of chemicals √ √ √√ √√√ – √√√√ Storage (in context of the activity of water toward absorbing chemicals after purification) √√√ √√√ √√√ √√√ √√√ √ (highly active water; vulnerable to absorption of ions) Costs Actual costs Installation costs √ – – √√ √√√ √√√√√ Maintenance costs – – – √ – √√ Operation costs – √ √ – √ √ Health costs Mineral content of purified water √√√√√ √√√ √√√√ √√√√ √√√√√ √ Reduction in pH – – – – – √√ Sweetness (taste) of purified water √√√ √ √√ √√√√√ √√√√ √ Environmental costs Wastage of water in the process – – – – – √√√ Electricity consumption – – – – √√ √√√ Infrastructural costs Dependence on water pressure – – – – – √√√ Dependence on electricity – – – – √√√ √√√ Other costs Aesthetic qualities of purified water √√ √ √√√ √√√√√ √ √ Requirement of user expertise/knowledge √ √ √ √ √ √√ Ease of handling √√√ √ √ √√√ √√√√√ √√√ Note: The strengths are given by us on the basis of our understanding of the source documents. The √signs in the above table show the strength of the parameter in question. The higher the number of the √sign, stronger is the effect of the relevant parameter. Sources: World Health Organization (1996, 2003, 2008, 2011a,b), PBI (2013), Kozisek (2004), field survey. Among all the parameters, RO’s superiority is unquestionable in removing chemicals at the level of household water purification devices.30 We have already noted that chemical contamination is more relevant in groundwater, and the nature of such contamination is location-specific. Such forms of contamination are mostly absent in municipality-treated water. Even there, activated carbon, which is much cheaper than RO, works almost as good as RO. In addition, activated carbon filters do not have to incur the recurrent costs of membrane damage. In removing microbes, UV and boiling seem to score better than RO. Although RO can remove bacteria, which are of larger size than the pores of membrane, it has two shortcomings when compared with UV. First, it does not kill the bacteria, which UV does. Second, RO cannot deal with the issue of fouling of membrane by the same set of bacteria, which clings on to the membrane, post-removal. RO water, being acidic, also becomes chemically active when comes in contact with the storage materials. In terms of costs, its actual cost is higher than all competing technologies. The picture is very similar when we add health costs and environmental costs. Its beneficial impact on health is lowest, given that it removes essential minerals and reduces pH, and makes water sour in taste. From environmental costs perspectives, the wastage of water rejected by RO, and the energy costs of electricity, once again, puts it behind other competing technologies. In addition, it demands more infrastructure and calls for greater user’s knowledge. Both these factors add to the already high actual costs of RO. In the light of this discussion, we now provide a brief history of RO to understand the various contingencies, and the nature of incremental innovations, which have shaped its adoption as household water purification system. 5.2 Emergence and adoption of RO for household water purification Patent data (USPTO) suggest that there has been a notable increase in the number of patents for all the technologies considered for this study (ion exchange, chlorination/bromination, UV, ozonation, and RO), over the past 10–15 years (Table 4). The number of patents in the field of use of RO for liquid purification is highest when compared with other technologies, even though it came much later (in 1961) than its competing technologies. The decadal data for 2000–2009 show highest number of patents for ozonation, which declines thereafter. The number of patents for RO continues to increase. Table 4. Patent data . Ion exchange (CCL/210/652) . Utilizing ozone (CCL/210/760) . Chlorination/ bromination (CCL/210/754) . RO (CCL/210/652) . Utilizing UV (CCL/210/748.1) . First patent (year) 1870 1884 1889 1961 1970 Total number of patents till April 2014 706 756 623 809 103 The decade having highest number of patents (number) 2000–2009 (253) 2000–2009 (293) 2000–2009 (165) 2000–2009 (282) 2010–2014 (63) Countries with highest numbers of patents The United States (341), Japan (75), Germany (46), and France (23) The United States (317), Japan (70), Canada (33), France (33), and Germany (21) The United States (305), Japan (23), Germany (17), France (12), and Israel (10) The United States (405), Japan (82), Germany (35), France (25) and Canada (22) The United States (43), Canada (10), Japan (8), and Israel (5) . Ion exchange (CCL/210/652) . Utilizing ozone (CCL/210/760) . Chlorination/ bromination (CCL/210/754) . RO (CCL/210/652) . Utilizing UV (CCL/210/748.1) . First patent (year) 1870 1884 1889 1961 1970 Total number of patents till April 2014 706 756 623 809 103 The decade having highest number of patents (number) 2000–2009 (253) 2000–2009 (293) 2000–2009 (165) 2000–2009 (282) 2010–2014 (63) Countries with highest numbers of patents The United States (341), Japan (75), Germany (46), and France (23) The United States (317), Japan (70), Canada (33), France (33), and Germany (21) The United States (305), Japan (23), Germany (17), France (12), and Israel (10) The United States (405), Japan (82), Germany (35), France (25) and Canada (22) The United States (43), Canada (10), Japan (8), and Israel (5) Source: USPTO (own calculations based on data from USPTO). Table 4. Patent data . Ion exchange (CCL/210/652) . Utilizing ozone (CCL/210/760) . Chlorination/ bromination (CCL/210/754) . RO (CCL/210/652) . Utilizing UV (CCL/210/748.1) . First patent (year) 1870 1884 1889 1961 1970 Total number of patents till April 2014 706 756 623 809 103 The decade having highest number of patents (number) 2000–2009 (253) 2000–2009 (293) 2000–2009 (165) 2000–2009 (282) 2010–2014 (63) Countries with highest numbers of patents The United States (341), Japan (75), Germany (46), and France (23) The United States (317), Japan (70), Canada (33), France (33), and Germany (21) The United States (305), Japan (23), Germany (17), France (12), and Israel (10) The United States (405), Japan (82), Germany (35), France (25) and Canada (22) The United States (43), Canada (10), Japan (8), and Israel (5) . Ion exchange (CCL/210/652) . Utilizing ozone (CCL/210/760) . Chlorination/ bromination (CCL/210/754) . RO (CCL/210/652) . Utilizing UV (CCL/210/748.1) . First patent (year) 1870 1884 1889 1961 1970 Total number of patents till April 2014 706 756 623 809 103 The decade having highest number of patents (number) 2000–2009 (253) 2000–2009 (293) 2000–2009 (165) 2000–2009 (282) 2010–2014 (63) Countries with highest numbers of patents The United States (341), Japan (75), Germany (46), and France (23) The United States (317), Japan (70), Canada (33), France (33), and Germany (21) The United States (305), Japan (23), Germany (17), France (12), and Israel (10) The United States (405), Japan (82), Germany (35), France (25) and Canada (22) The United States (43), Canada (10), Japan (8), and Israel (5) Source: USPTO (own calculations based on data from USPTO). Almost 200 years after the phenomenon of osmosis discovered by Abbe Nollet in 1748, S. Sourirajan developed the principle of RO in the year 1950. Three years later, Reid showed that cellular acetate membrane of RO can filter salt from water. This invention paved the way for the development of RO-based water purification technology in the year 1961 (Table 5). This development held immense promise to ameliorate water crisis in coastal areas. The first-generation RO membranes were the cellulose acetate membranes, which had lower salt rejection efficiency, lower water flux, and lower produced water rate. These membranes were environment-friendly, as they were made from natural fibers and were biodegradable, but had shorter life. The second generation of RO membranes were the thin-film composite (TFC) membranes, developed by the Filmtech Company in 1970. These membranes have higher salt rejection and water production potential. They are made from synthetic polymers, and hence, are nonbiodegradable and non-eco-friendly. They became popular from 1990s onward. A third type of membrane, which is an advanced form of TFC membrane, called the polymer nanocomposite membrane was developed in the year 2005 by Samsung SDI Co. Ltd. (WIPO Patent number 1511112). The first use of this technology for water purification, however, took place in 2011 (US Patent number 7,931,838). Table 5. Evolution of RO membrane Serial number . Invention/innovation . Inventor . Year . 1. Studies on osmosis (discovered the phenomenon that pig bladder can permeate water) Abbe Nollet 1748 2. Discovered osmosis and developed it Graham 1861 3. First inorganic membrane Traube 1876 4. Introduced electrodes in dialysis and furthered dialysis research Morse and Pierce 1903 5. Developed “artificial simulation bio selection osmosis membrane,” the principle of RO S. Sourirajan 1950 6. Demonstrated that cellulose acetate RO membranes were capable of separating salt from water Reid 1953 7. Asymmetric cellulose acetate membranes with relatively high water fluxes and separations Loeb and Suriranjan 1960 8. Used cellulose triacetate as a layer in TFC flat sheet configuration North Star (predecessor of Filmtech) 1964 9. Proposed the terminology of “Hybrid filtration” as the earliest ultrafiltration concept Israel Desalination Company 1970 10. Developed a cellulose triacetate membrane Dow Chemical Company 1971 11. The permasep B-9 and B-10 permeator became the leading elements in the market Du-pont 1980s – 90s 12. Polyamide TFC-RO Filmtech (Dow Company) 1990s 13. Polymer nanocomposite membranes SAMSUNG SDI CO LTD 2005 Serial number . Invention/innovation . Inventor . Year . 1. Studies on osmosis (discovered the phenomenon that pig bladder can permeate water) Abbe Nollet 1748 2. Discovered osmosis and developed it Graham 1861 3. First inorganic membrane Traube 1876 4. Introduced electrodes in dialysis and furthered dialysis research Morse and Pierce 1903 5. Developed “artificial simulation bio selection osmosis membrane,” the principle of RO S. Sourirajan 1950 6. Demonstrated that cellulose acetate RO membranes were capable of separating salt from water Reid 1953 7. Asymmetric cellulose acetate membranes with relatively high water fluxes and separations Loeb and Suriranjan 1960 8. Used cellulose triacetate as a layer in TFC flat sheet configuration North Star (predecessor of Filmtech) 1964 9. Proposed the terminology of “Hybrid filtration” as the earliest ultrafiltration concept Israel Desalination Company 1970 10. Developed a cellulose triacetate membrane Dow Chemical Company 1971 11. The permasep B-9 and B-10 permeator became the leading elements in the market Du-pont 1980s – 90s 12. Polyamide TFC-RO Filmtech (Dow Company) 1990s 13. Polymer nanocomposite membranes SAMSUNG SDI CO LTD 2005 Sources: Johnson and Busch (2009), Development history of RO at http://aquavision.co.in/?page_id=185, https://en.wikipedia.org/wiki/Thin-film_composite_membrane, (Rao, 2011), USPTO and WIPO patent search. Table 5. Evolution of RO membrane Serial number . Invention/innovation . Inventor . Year . 1. Studies on osmosis (discovered the phenomenon that pig bladder can permeate water) Abbe Nollet 1748 2. Discovered osmosis and developed it Graham 1861 3. First inorganic membrane Traube 1876 4. Introduced electrodes in dialysis and furthered dialysis research Morse and Pierce 1903 5. Developed “artificial simulation bio selection osmosis membrane,” the principle of RO S. Sourirajan 1950 6. Demonstrated that cellulose acetate RO membranes were capable of separating salt from water Reid 1953 7. Asymmetric cellulose acetate membranes with relatively high water fluxes and separations Loeb and Suriranjan 1960 8. Used cellulose triacetate as a layer in TFC flat sheet configuration North Star (predecessor of Filmtech) 1964 9. Proposed the terminology of “Hybrid filtration” as the earliest ultrafiltration concept Israel Desalination Company 1970 10. Developed a cellulose triacetate membrane Dow Chemical Company 1971 11. The permasep B-9 and B-10 permeator became the leading elements in the market Du-pont 1980s – 90s 12. Polyamide TFC-RO Filmtech (Dow Company) 1990s 13. Polymer nanocomposite membranes SAMSUNG SDI CO LTD 2005 Serial number . Invention/innovation . Inventor . Year . 1. Studies on osmosis (discovered the phenomenon that pig bladder can permeate water) Abbe Nollet 1748 2. Discovered osmosis and developed it Graham 1861 3. First inorganic membrane Traube 1876 4. Introduced electrodes in dialysis and furthered dialysis research Morse and Pierce 1903 5. Developed “artificial simulation bio selection osmosis membrane,” the principle of RO S. Sourirajan 1950 6. Demonstrated that cellulose acetate RO membranes were capable of separating salt from water Reid 1953 7. Asymmetric cellulose acetate membranes with relatively high water fluxes and separations Loeb and Suriranjan 1960 8. Used cellulose triacetate as a layer in TFC flat sheet configuration North Star (predecessor of Filmtech) 1964 9. Proposed the terminology of “Hybrid filtration” as the earliest ultrafiltration concept Israel Desalination Company 1970 10. Developed a cellulose triacetate membrane Dow Chemical Company 1971 11. The permasep B-9 and B-10 permeator became the leading elements in the market Du-pont 1980s – 90s 12. Polyamide TFC-RO Filmtech (Dow Company) 1990s 13. Polymer nanocomposite membranes SAMSUNG SDI CO LTD 2005 Sources: Johnson and Busch (2009), Development history of RO at http://aquavision.co.in/?page_id=185, https://en.wikipedia.org/wiki/Thin-film_composite_membrane, (Rao, 2011), USPTO and WIPO patent search. Ever since the invention of commercially viable RO membranes, in the 1960s, it has been used with great enthusiasm and interest for drinking water purification for the water coming from varied kinds of sources.31 This enthusiasm often overlooked that the technology was developed primarily to make ocean water drinkable. With greater use, new technological innovations started flowing. Penate and Garcia-Rodriguez (2012) find that most of the technological advances and innovations in this technology attempted to reduce energy consumption, and harmful effects of scaling and fouling on membranes, and to obtain higher water flux membranes. These advances include developments in membrane materials and module design, process design, feed pretreatment, and energy recovery or reduction in energy consumption (Lee et al., 2011). Other innovations include the use of a solar photovoltaic system to power RO desalination and new membranes with new material types, such as the polyether–urea and polyamide–urea barrier, which have been reported to decrease microbial adhesion, and thus, fouling potential. Some other innovations in membranes include a cross-linked matrix of polymers and engineered nanoparticles designed to draw in water ions but repel nearly all contaminants including dissolved salts, organics, and bacteria. Less energy is needed to pump water through the membranes, and because they repel particles that ordinarily stick to the surface, the new membranes foul more slowly than conventional ones (Malaeb and Ayoub, 2011). Lee et al. (2011) argue that continuous process improvement and the consequent cost reductions have augmented RO’s commercial success. However, the above discussion also demonstrates that most of the innovations are directed toward increasing the efficiency of RO in terms of energy use, or life of membrane. Not much has taken place to reduce the rejection of water from RO,32 which is a serious environmental threat, both because it wastes water (scarce resource) and because this rejected water is highly concentrated with impurities. Indeed rejection of water amounts to around 30–40% of feed water in industrial units, and up to 80% in household systems (Perez-Gonzalez et al 2012; Field work).33 This trajectory of innovations in RO, which, in a sense, has neglected the dimension of water rejection, may be due to the fact that RO was initially conceived to treat ocean water, where rejection of water is not an issue. Nor the rejected water, which is put back into ocean, enhances any environmental hazards there.34 This path of innovation continued even after RO found use at the households, in non-coastal areas, where salinity is not a major concern. The problem of rejection becomes more acute in household scenarios, compared to industrial scenarios, due to low pressure with which water is passed through RO membrane. Moreover, insofar, as piped water supply is subjected to municipality tax, rejection of water remains costly to the households. One form of innovation very commonly observed is combination of various other purification technologies, like carbon adsorption, UV, UF, NF, etc., along with RO membrane. As we see in Section 3 Table 1 that different technologies serve different purposes in the highly complex process of drinking water purification, the manufacturers of water purifiers started combining many technologies. Interestingly, RO features in almost all these combinations, and the combined filter is known by the name of RO purifier. It is important to note also that the reasons cited for combining these technologies also revolve around concerns with RO membrane. To elaborate, in the early years, RO purifiers had an RO membrane, along with carbon adsorption. These technologies, however, reportedly suffered short life of RO membranes because of fouling (growth of microbes on the membrane) or choking of membrane. To reduce the pressure on RO membrane, other kinds of membranes like UF, NF, with slightly higher pore sizes, are added just before the RO filter. An UV tube is added before or after the RO membrane to remove microbial pollutants from the water. This way, the life of an RO membrane is lengthened by reducing the chances of fouling. These combinations enhance the legitimacy of the innovation mentioned above, where the portion of feed water untreated by RO (to maintain pH and sweet taste) is treated with other technologies, and, thereby, reducing possibilities of microbial contamination. Scientifically, it may make sense to combine various technologies, and make a purification device universally applicable, which would work in varied source water quality. However, such a device proves to be cost-effective only when consumers are perfectly mobile. Indeed, RO has been popularized by the requirements of the US navy, and many innovations in it have undertaken to facilitate mobility of the US army, as documented in the USPTO.35 For general households, however, this is not pervasive in reality. Moreover, we may recall that at the household level, RO entails significant installation costs, making such movements far from being costless.36 In the Indian context, the innovations in RO have taken the form of combining various methods of purification with RO in a single purification device. For this the country depends entirely on imported components (e.g., pumps, membranes, UV lamps, etc.). Recall, however, these combinations are supposed to enhance the effectiveness of the RO membrane whose main contribution is to demineralize water. Interestingly, regulations for water quality do not specify any lower limit for minerals and dissolved solids in water, reflecting, perhaps, a lack of consensus among the scientists on their disease-causing potential.37 It is, however, well-recognized that RO makes water acidic and sour in taste, by indiscriminately removing minerals. Consumers, on the other hand, prefer sweet water, often at the cost of impurities (WHO, 2008). As mentioned above, municipality-supplied piped water is often treated for microbes and harmful heavy metals.38 At the same time, RO cannot treat microbes and bacteria effectively that may contaminate piped water due to leakages in the pipe network. In such a scenario, a particular kind of innovation has helped RO’s adoption by appealing to sensory dimension of taste. This innovation, patented by the IPO in 2005, is intended “to provide of reverse osmosis based water purifier which would not have the limitation of producing totally de-mineralised health affecting water.” The steps by which the said innovation fulfills this objective are sought to achieve through the following steps (emphases added):39 “(a) Sediment filtration means preferably 5.00 microns; (b) activated carbon filter (c) passing a part of treated water from (b) above partly at high pressure through reverse osmosis membrane to demineralize raw water by reverse osmosis and; (d) mixing the demineralised water through said reverse osmosis membrane with the remainder of the treated water from (b) at controlled rate;(e) filtration means to free the thus mixed water of (d) above of harmful bacteria and viruses selectively involving UF membrane and/or UV lamp.” Due to this process, the innovation has earned the name of TDS controller. Note that only a part of feed water treated through sediment and activated carbon filtration is passed through the RO. The rest is only treated by UF/UV for microbes and bacteria. To put it differently, the entire feed water is treated by many filtration processes (sedimentation, active carbon filtration, and UV/UF) but by RO. The patent document, however, “claims” to categorize this innovation as “a reverse osmosis based drinking water purifier …” (ibid, emphases added). In other words, RO is portrayed to be the mainstay of the purification devise, despite not being the main purification method. We get back to this issue later in the next section to discuss the underlying mechanism of feedback and self-reinforcement. 5.3 The self-reinforcement and information contagion From Tables 1 and 3, we see presence of a variety of other purification technologies sharing the goodness of RO but not its undesirable properties. Yet, a choice is made in favor of RO. To analyze the consumer’s perspective, we first need to understand the quality of source water used by Indian households. In our survey, the source water varied from public supply of piped water to private tankers to underground water. Since we do not have sufficient data on water quality in all these places, we have to go by a proxy measure for their quality. We use consumer’s perception of water quality as the proxy. We note that 75% of our respondents (97 in total) found water supplied in their household to be “good” for drinking and cooking (Table 6). Table 6. Source of water and quality parameters Source of water . Percentage . Quality parameter . Percentage (very bad to bad) (1 and 2 on a 5-point scale) . Percentage (satisfactory to very good) (3–5 on a 5-point scale) . Public supply (pipeline) 66.22 Smell 5.84 94.16 Public supply (water tankers) 7.75 Taste 15 85 Private supply (pipeline) 7.75 Clarity 15.56 84.44 Private supply (water tankers) 6.98 Hardness 25.42 74.58 Groundwater (direct boring) 24.81 Usability (cooking/drinking) 22.61 77.39 Source of water . Percentage . Quality parameter . Percentage (very bad to bad) (1 and 2 on a 5-point scale) . Percentage (satisfactory to very good) (3–5 on a 5-point scale) . Public supply (pipeline) 66.22 Smell 5.84 94.16 Public supply (water tankers) 7.75 Taste 15 85 Private supply (pipeline) 7.75 Clarity 15.56 84.44 Private supply (water tankers) 6.98 Hardness 25.42 74.58 Groundwater (direct boring) 24.81 Usability (cooking/drinking) 22.61 77.39 Source: Field survey. Table 6. Source of water and quality parameters Source of water . Percentage . Quality parameter . Percentage (very bad to bad) (1 and 2 on a 5-point scale) . Percentage (satisfactory to very good) (3–5 on a 5-point scale) . Public supply (pipeline) 66.22 Smell 5.84 94.16 Public supply (water tankers) 7.75 Taste 15 85 Private supply (pipeline) 7.75 Clarity 15.56 84.44 Private supply (water tankers) 6.98 Hardness 25.42 74.58 Groundwater (direct boring) 24.81 Usability (cooking/drinking) 22.61 77.39 Source of water . Percentage . Quality parameter . Percentage (very bad to bad) (1 and 2 on a 5-point scale) . Percentage (satisfactory to very good) (3–5 on a 5-point scale) . Public supply (pipeline) 66.22 Smell 5.84 94.16 Public supply (water tankers) 7.75 Taste 15 85 Private supply (pipeline) 7.75 Clarity 15.56 84.44 Private supply (water tankers) 6.98 Hardness 25.42 74.58 Groundwater (direct boring) 24.81 Usability (cooking/drinking) 22.61 77.39 Source: Field survey. Despite finding water good enough for drinking and cooking, 73 of those 97 individuals use some kind of purification device at home. Table 7 gives the list of various purification technologies used by our respondents. Table 7. Frequency distribution of purification technology adoptions by respondents Method of purification . Numbers (percentages)— current (120) . Numbers (percentages)— previous (92) . Boiling/chemicals/no method (no technologya) 25 (20. 83%) 54 (58. 69%) Purifier devices (some technology) (95) RO (and RO-based technologies) 54 (56. 84%) 11 (11. 95%) Non-RO technologies (41) UV 9 (21. 95%) 6 (6. 52%) Off-line (mixed) 32 (78. 04%) 22 (23. 91%) Method of purification . Numbers (percentages)— current (120) . Numbers (percentages)— previous (92) . Boiling/chemicals/no method (no technologya) 25 (20. 83%) 54 (58. 69%) Purifier devices (some technology) (95) RO (and RO-based technologies) 54 (56. 84%) 11 (11. 95%) Non-RO technologies (41) UV 9 (21. 95%) 6 (6. 52%) Off-line (mixed) 32 (78. 04%) 22 (23. 91%) Note: a Considering that boiling and chemical treatment was used occasionally by the users (and not on a regular basis), and they are analyzed along with the “no method used” category. Table 7. Frequency distribution of purification technology adoptions by respondents Method of purification . Numbers (percentages)— current (120) . Numbers (percentages)— previous (92) . Boiling/chemicals/no method (no technologya) 25 (20. 83%) 54 (58. 69%) Purifier devices (some technology) (95) RO (and RO-based technologies) 54 (56. 84%) 11 (11. 95%) Non-RO technologies (41) UV 9 (21. 95%) 6 (6. 52%) Off-line (mixed) 32 (78. 04%) 22 (23. 91%) Method of purification . Numbers (percentages)— current (120) . Numbers (percentages)— previous (92) . Boiling/chemicals/no method (no technologya) 25 (20. 83%) 54 (58. 69%) Purifier devices (some technology) (95) RO (and RO-based technologies) 54 (56. 84%) 11 (11. 95%) Non-RO technologies (41) UV 9 (21. 95%) 6 (6. 52%) Off-line (mixed) 32 (78. 04%) 22 (23. 91%) Note: a Considering that boiling and chemical treatment was used occasionally by the users (and not on a regular basis), and they are analyzed along with the “no method used” category. It is also important to note here that the maximum number of RO users (74.04%) have adopted this technology during the past 10 years. In total, 32 users switched to RO from non-RO, in our sample, during this time.40 This phenomenal switch has the possibility of pushing RO, to use the words of Arthur (1989), “far enough ahead” of non-RO technologies to get into a situation of lock-in. We obtain further insights into this process by analyzing the adoption behavior of consumers having piped supply and non-piped supply separately. Around 70% of our respondents have piped water supply, of which 43% use RO. The share of RO among individuals with non-piped supply is comparatively less, standing at 37%. In addition, 69% of those who decided to switch to RO in the recent years have piped water supply at home. We see that preference for RO is higher among people with piped water supply. This is counterintuitive, given that piped water supplies treat for pollutants (heavy metals) which RO filters aspire to treat.41 Before concluding along this line, however, one needs to analyze the reason, and rationality, which guide adoption of RO by the individuals. For this purpose, we analyze the sources of information employed by our survey respondents. Note that majority of our respondents belong to “internet-literate” group for whom the costs of accessing information about purification technologies can be assumed to be negligible. In this particular case, there are several guidelines by WHO regarding safe drinking water as well as the household methods for water purification, which are available online. Newspaper reports42 as well as blogs/Web sites discussing the advantages and disadvantages of the various water purification technologies43 are all available in the public domain. When asked about the reasons for their choice of RO water purifiers, 37.04% of the respondents cited RO’s effectiveness to remove chemicals, 25.93% cited its ability to remove “hardness” (dissolved solids). In fact, the three most significant reasons attributed to the current use of RO were it being “technologically advanced,” being “capable of removing chemicals,” and “removing hardness.” Moreover, there are high correlations among these reasons. The correlation coefficient between those who think RO is technologically advanced and those who prefer it for effectively removing chemicals is 0.37 (significant at 1% level). Those who think it effectively removed bacteria and microbes also have significant overlap with those who think it is technologically advanced (r = 0.37, significant at 1% level) and those who think it can effectively remove chemicals (r = 0.5, significant at 1% level). On the other hand, the current users of RO do not see it cheaper (r = −0.28, significant at 5% level) or easy to maintain (r = −0.17, significant at 10% level).44 Insofar being “inexpensive” and having “low maintenance cost” reveal cost advantages, the choice of RO by our sample households is clearly not guided by cost consideration. They purchase RO, despite its higher costs. From the first set of factors, one may be tempted to argue that RO gives better quality of purification. However, the above Tables 2 and 5 show that RO is not a suitable let alone the best technology to treat microbes. This mismatch between the survey response and scientific evidence mentioned in Table 5 prompts us to analyze the source of information, for consumers of RO technology, further. We know that advertisement and promotional activities form a major source of information about products today, where the biases and fraud cannot be ruled out. These biases are more for credence goods whose quality is difficult to ascertain through either search or experience, making consumers vulnerable to wrong claims. These wrong claims may remain unverified for a reasonable amount of time, unless a disease breaks out.45 For a water purification device, as we have argued earlier, quality may be assumed to consist of two different (and may be mutually exclusive) criteria: sensory aspects (tastes and turbidity) and the ability to prevent waterborne diseases. Sensory dimensions like “taste” can be reasonably assumed to be “experience goods.” Our household surveys and the interviews of the manufacturers of water purifiers bring out clearly that such a preference exists among consumers, and the preference for TDS-controlled sweet water of RO gets strengthened through word of mouth. In fact, many households report to us that without an RO filter, many guests would not drink water in their houses. In a typical Indian household, where offering (and accepting) water in a household is the first step of hospitality, such outcomes become socially unbearable, which have forced many households to shift to RO. Having an RO purifier, at home, thus made them socially more acceptable. Our interviews with the manufacturers reveal how important this dimension has been for rapid diffusion of the “RO with TDS controller” among urban households, especially in Delhi. Our sample respondents, however, could not convincingly ascertain, even after experience, whether the use of RO has reduced the instances of waterborne diseases in their households. Around 56% of respondents either said “can’t say” or opted for “no” to this question. Note that Payment (1989) finds that bacterial colonization in storage units of household RO systems may weed out any health advantages of RO purifiers. Similarly, Payment et al (1993) did not find any clear advantage of RO purified water over tap water, once contamination due to stagnation in pipe was controlled for. Although there is no reliable information available on the number of RO producing firms in India, our interviews with technology manufacturers put the number close to 5000. Note here that most of the manufacturers are small enterprises and microenterprises, who rely more on social networks and word of mouth than any organized form of promotional activity. A sizeable number of them are also of fly-by-night nature, who do not have enough incentives to build long-term relationship with the consumers. Probability of biased information increases in such scenarios (Darby and Karni 1973). These manufacturers mostly promote their products by offering high incentives to neighborhood retail stores.46 The more established manufacturers, on the other hand, use advertisements in televisions and newspapers to promote their products. However, they also have incentives in downplaying the critical sides of a technology especially when it is a “credence” good. On the other hand, one may argue that such incentives are lower when the consumers can also cross-check a claim through a variety of information sources, which could be the case for our Internet-literate respondents. Indeed newspapers in India have carried many news items highlighting criticality of RO technologies. Devarhubli and Jha (2011) for instance claim that RO water is deficient in vitamin B12. Besides, scientific reports easily accessible through Internet, for instance, Kozisek (2004, referred above) indicate the disadvantages of drinking water devoid of key minerals like calcium and magnesium, which also makes water acidic. Ironically, our respondents rely more on sources like advertisements in televisions and newspapers and information from storekeepers not only to choose RO47 but also to conclude about its technological superiority (see Table 8).48 According to Arthur (1994) such rather informal word-of-mouth process of information dissemination is categorized as “information contagion.” This behavior also may be interpreted to fall into the category of popularity rules, which, once again, lead to path-dependent scenarios. Table 8. Probit estimation: various sources of information about technological superiority of RO Dependent variable (RSN_TECH_ADV) . Coefficient (z values) . PC_TV_PAPR 0.337** (3.33) SOURC_STR_KPR 0.549* (1.66) SOURC_SCI_PAPR 0.692 (1.61) Wald χ2 19.50*** Number of observations 123 Dependent variable (RSN_TECH_ADV) . Coefficient (z values) . PC_TV_PAPR 0.337** (3.33) SOURC_STR_KPR 0.549* (1.66) SOURC_SCI_PAPR 0.692 (1.61) Wald χ2 19.50*** Number of observations 123 Notes.: 1. RSN_TECH_ADV= dependent variable: RO is technologically advanced. Independent variables: PC_TV_PAPR= principle component of advertisements in television and news papers, SOURC_STR_KPR = source of information is storekeeper, SOURC_SCI_PAPR= source of information is scientific paper. 2. *** denotes significance at 1% level, ** denotes significance at 5% level, * denotes significance at 10%level. Table 8. Probit estimation: various sources of information about technological superiority of RO Dependent variable (RSN_TECH_ADV) . Coefficient (z values) . PC_TV_PAPR 0.337** (3.33) SOURC_STR_KPR 0.549* (1.66) SOURC_SCI_PAPR 0.692 (1.61) Wald χ2 19.50*** Number of observations 123 Dependent variable (RSN_TECH_ADV) . Coefficient (z values) . PC_TV_PAPR 0.337** (3.33) SOURC_STR_KPR 0.549* (1.66) SOURC_SCI_PAPR 0.692 (1.61) Wald χ2 19.50*** Number of observations 123 Notes.: 1. RSN_TECH_ADV= dependent variable: RO is technologically advanced. Independent variables: PC_TV_PAPR= principle component of advertisements in television and news papers, SOURC_STR_KPR = source of information is storekeeper, SOURC_SCI_PAPR= source of information is scientific paper. 2. *** denotes significance at 1% level, ** denotes significance at 5% level, * denotes significance at 10%level. 6. Summary and conclusion The objective of the article was to explore the nature of incremental innovations, of a currently evolving technology, in the presence of preference heterogeneity to understand the relationship between incremental innovation and path dependence better. The existing literature concludes that incremental innovations, post-adoption of an inferior technology, make it superior to its competing technologies over time. In most occasions, however, the studies have taken up cases which have been settled in distant past, and competing technologies have already been driven out of the market. As a result, data regarding their comparative innovation paths remain inadequate. These studies also do not take into consideration the demand-side factors, which can influence the innovations of path-dependent technologies. This is important, particularly, in credence goods situation where consumers’ experiences alone do not reveal enough information about quality and efficiency of a technology. We find that RO technology is fast-gaining market share in India, despite being inferior to other available technologies. Popularity rule, driven by social preference for sweet water, influences its path of technological change, which is, however, not making the technology superior in terms of its disease-preventing capability compared to its competing technologies, especially when the source water is piped supply of water. From the supply side, innovations are aiming at reducing stress on RO membrane by pretreating the feed water through various other purification technologies. While these combinations do add to the longevity of the membrane, the combined devices are costlier too. It is, therefore, difficult to conclusively judge their cost-efficiency. These costs are pure private costs, which only add up if we include social costs like damages to environment done through rejecting a significant portion of feed water, that too with higher concentration of harmful contaminants. These concerns are, apparently, getting overridden by the information contagion which visualizes RO as the panacea to all water quality problems. The recent historical controversy on presence of high pesticide in groundwater may have acted as a key historical contingency shaping the spiral of such perception in India. When contrasted with the innovation paths of RO at the global level, we observe that innovations in RO, globally, has been shaped by the requirement of army movement in areas where reduction in salinity may be a prerequisite to achieve the desired level of pH, i.e., brackish water). Wastage of water and its disposal are not the prime concerns in treatment of sea water. Guarding oneself against frequent change in source water quality is a prime concern in such cases. None of these applies, at least to the same degree, to the households in Urban India. These innovations, therefore, fail to make it a superior technology in the context of a large section of urban household consumption of drinking water. Insofar as its adoption is guided by the socially preferred parameter of taste, when other economically more efficient technologies are available, it follows the third-degree path dependence. However, these social parameters are not only influencing its initial entry into the market but also the path of its subsequent innovations, yet being unable to make it superior to its competing technologies. Innovations, here seem to be deepening the reasons for path dependence, rather than removing them. Therefore, the path dependence in this case seems stronger than what Liebowitz and Margolis might have had in mind. This insight has been possible to obtain only by including the demand-side factors in analysis. Arthur (1989) claims that lock-in would be difficult in the presence of heterogeneity of preference. Our analysis, however, hint at deepening of path dependence, leading possibly to lock-in, even when preference heterogeneity exists. We find that incremental innovations, when guided by information contagion, can deepen path dependence, and may cause lock-in. This seems to be true, at least, for credence goods where preferences cannot adequately reflect quality and efficiency of a product. Acknowledgment: We gratefully acknowledge the constructive critical comments by the anonymous reviewers of the journal. Pradosh Nath and Aviram Sharma provided valuable insights during the early phase of this research. The first author acknowledges the research fellowship obtained from the University Grants Commission, India. Footnotes 1 " Cowan (1990) being an exception. 2 " Firms’ choice of actions, especially in the current techno-economic paradigm, is strongly shaped by various industry or product-specific standards not leaving much scope to exhibit diversity. Consumers, on the other hand, are not guided by such compulsory regulations and exhibit more heterogeneity of preference. 3 " In the well-known cases of VHS vs. Betamax, QWERTY do incorporate the users in their analysis. However, these studies do not analyze the paths of innovations in the technologies once a technology was adopted. 4 " Note, however, that conceiving “a market” for purification technologies is difficult, as their applicability is often limited by nature of source water. 5 " http://www.techsciresearch.com/2482 (accessed November 29, 2014). 6 " http://articles.economictimes.indiatimes.com/2012-06-02/news/31984263_1_water-purifiereureka-forbes-rs-7k-crore (accessed March 31, 2013); it is noteworthy that other purification technologies like chlorination, sand filtration, and candle filters existed long before RO came in the Indian market. UV purifier too came in 1984, while RO came in 1999 (Super Brands Report, 2009). 7 " It wastes 35–85% of water input in it (Perez-Gonzalez et al., 2012; NDSU report, 2009). 8 " It is also to be noted here that RO requires a specified water pressure below which it does not work properly. Its market share in terms of value (45%, against 23% by UV) (ADI Media, 2012) and in terms of installed capacity, it is behind UV only by 4% (2012), as against 9% (in 2010) (ADI media, 2010; Marwas, 2010). 9 " This is indeed another distinguishing feature of our study. All the previous studies have analyzed cases where the choice between competing technologies is mutually exclusive (David, 1985; Cowan, 1990; Puffert, 2002). 10 " Only 46.6% of the households in India have access to drinking water within their premises. A far lower, 43.5% of the households have access to tap water (Census data for the year 2011, as cited in the Economic Survey 2016-17) (http://indiabudget.nic.in/es2015-16/echapvol2-09.pdf) Accessed 15 July 2016. 11 " At times, coevolution may not imply path dependence and lock-in. See, for instance, Norgaard (1994) where pest, pesticides, and pesticide policies coevolved, but brought, in the process, very radically different approaches into policy-making and technology selection. Bergek and Onufrey (2013) also admit not to capture self-reinforcement dynamics, without which path dependence is difficult to ascertain. 12 " Some studies focus on behavioral aspects of path dependence, which analyzes how habits formed over long period, with regard to the use of certain technologies (Barnes et al., 2004) may lead to lock-in (Quitzau, 2007). 13 " Of course, there may be regret about these choices at a later stage (Arthur, 1994; 116). 14 " According to this rule, an agent chooses the product with higher mean value in his sample (Lane and Vescovini 1996). 15 " Subjects choose the product that shows highest performance in the sample. 16 " Subjects choose the highest minimum performance in the sample to avoid a bad outcome as far as possible. 17 " Subjects choose the product that most informants in the sample already have chosen. 18 " Indeed, governmental interventions in terms of regulation have been justified for such goods (Anania and Nistico, 200487). 19 " In response to Liebowitz and Margolis (199587), David (1997) makes clear that “the concept of path dependence itself carries no necessary implications whatsoever in regard to the existence or non-existence of market failure.” 20 " http://www.delhi.gov.in/wps/wcm/connect/doit_djb/DJB/Home/About+Us (accessed December 29, 2014). 21 " Are not connected to the pipeline water supply and do not need electricity to operate. 22 " Are connected to the pipeline water supply and require electricity to operate. 23 " http://www.adi-media.com/PDF/TVJ/annual_issue/011-Water-Purifiers.pdf (accessed March 31, 2013). 24 " http://articles.economictimes.indiatimes.com/2012-06-02/news/31984263_1_water-purifiereureka-forbes-rs-7k-crore (accessed March 31, 2013). 25 " We have, at times, complemented the data obtained from USPTO by other online data sources by the World Intellectual Property Organization and the European Patent Office. However, the USPTO database remains our main reference point, in view of the fact that this database is widely used as a source of data for the frontier of technology. 26 " http://patft.uspto.gov/netahtml/PTO/search-adv.htm (accessed July 16, 2016). 27 " http://www.delhi.gov.in/wps/wcm/connect/doit_djb/DJB/Home/About+Us (accessed July 20, 2016). 28 " Reserve Bank of India classifies Indian habitats into six categories. Using these categories, Pay Commissions have classified cities into three tiers. Tier 1 cities are the largest urban agglomerates having population above 100,000. Tier 2 cities have population ranging from 50,000 to 99,999. Tier 3 cities have population size between 20,000 and 49,999. See https://en.wikipedia.org/wiki/Classification_of_Indian_cities for details (accessed July 17, 2016). 29 " The relevant share from the offline survey was 16%. 30 " Ion exchange could be another option at large-scale operations. 31 " The earliest patent documents (USPTO, 1968) which talked about providing potable drinking water through RO purifier was for use in sea water and also in areas where there was scarcity of potable drinking water. 32 " However, we found that WIPO patent number WO/2016/175703, dated November 3, 2016, talks about “A Device which can salvage the waste water from Reverse Osmosis Water Filter up to 100%, in others words, ‘Zero Waste’, without using electricity and electrical pump, and regardless of whether hot water pipeline and hot water tank are available.” 33 " Note here that two requirements for RO process to occur are (i) semipermeable membrane (RO membrane) and (b) external or applied pressure (HP pump). In the absence of the optimum water pressure, the membrane does not function properly and may yield poor results both in terms of quality and quantity of water. Water pressure at households is often much less compared to industrial units, leading to higher rate of rejection. 34 " The earliest patent documents on RO date back to 1966–1968. These documents reveal that the RO purifiers were developed to desalinate sea water and to provide potable water in areas where it was otherwise very difficult to get water worth consumption. See, for instance, USPTO, 1968, Patent number 3365061. 35 " See, for instance, US patent number US5244579A. www.google.com/patents/US5244579 (accessed January 5, 2015). 36 " Sharma and Bhaduri (2013) in their research on water consumptions by Delhi’s mobile consumers found that excessive mobility often motivates individuals to consume packaged water. 37 " Barring of course disease-causing TDS like arsenic and fluoride. 38 " In India, municipalities also use river or surface water for treatment, which do not have much dissolved solids. 39 " See Indian Patent number: 199716. 40 " In contrast, only two individuals have switched from RO to non-RO technologies. 41 " Indeed, groundwater in various parts of India has high levels of harmful heavy metals like arsenic and fluoride, against which RO often is seen as an effective private solution. The rampant use of RO in such places, however, comes at the cost of tremendous public health hazards in the absence of any comprehensive disposal mechanism for RO-rejected water, which carries higher level of harmful heavy metals back to land or aquifers. See, for instance, http://indianexpress.com/article/lifestyle/health/ro-filters-a-threat-to-public-health/ (accessed July 17, 2016). 42 " http://www.waterfyi.com/featured/w-h-o-ya-gonna-believe/ (accessed July 20, 2016). 43 " http://waterfilters.mercola.com/whole-house-water-filters.aspx (accessed July 20, 2016). 44 " Here, we need to attach a caveat that our sample size is small to be taken as representative of a country like India. Statistical results based on such a sample may be hard to generalize. We, however, offer two explanations for its larger acceptability. First, as we have explained already, we have been able to include almost all relevant social groups, demographic profile, and geographic area for our study. Second, there was a report published in all leading Indian newspapers (http://indianexpress.com/article/lifestyle/health/ro-filters-a-threat-to-public-health/;http://economictimes.indiatimes.com/magazines/panache/are-reverse-osmosis-filter-systems-a-threat-to-public-health/articleshow/47426210.cms) a year ago pointing out these issues about RO technologies. Our Internet browsing reveals that the report has been one of the most read report on science technology issue in India. It also enjoyed more than 2000 Facebook likes and more than 1000 Facebook shares (https://www.facebook.com/TimesofIndia/posts/10153357523587139) separately for two leading dailies in India, Times of India and Indian Express. To us, such a wide engagement with the topic and Facebook likes by the readers reveal their general agreement with the views of the article, which are very similar to what has been found in our survey. In addition, there are also many critical comments one may observe on the Web sites of these newspapers about the article. To us, once again, it reflects that same “information contagion” and perception of consumers about efficacy of RO technology for water purification. Once again, we find very similar responses in our survey. 45 " This is why regulations are argued to be essential for credence goods. 46 " This scenario is similar to medicine, where small, run of the mill producers capture market through offering incentives to pharmacists. See Ray (2004) for an account of this practice in various parts of India. 47 " Our analysis of simple frequency distribution shows that 44 of 54 users of RO made their adoption decision based on information from these sources. 48 " Scientific papers, which discuss the desirable quality parameters of drinking water, are relevant here. Note that RO is desirable under many medically challenging situations like purification of water for dialysis, where mineral content is completely undesirable. References Abernathy W. J. , Utterback J. M. ( 1978 ), ‘ Patterns of industrial innovation ,’ Technology Review , 64 ( 7 ), 254 – 228 . OpenURL Placeholder Text WorldCat ADI Media ( 2012 ), ‘Annual issue, Water purifiers,’ An ADI Media Publication, adi-media.com, April 2012, TV Veopar Journal. http://www.adi-media.com/PDF/TVJ/annual_issue/011-Water-Purifiers.pdf (accessed 12 February 2013). Anania G. , Nistico R. ( 2004 ), ‘ Public regulation as a substitute for trust in quality food markets: what if the trust substitute cannot be fully trusted? ,’ Journal of Institutional and Theoretical Economics , 160 ( 4 ), 681 – 701 . Google Scholar Crossref Search ADS WorldCat Arthur W. B. ( 1989 ), ‘ Competing technologies, increasing returns, and lock-in by historical events ’. The Economic Journal , 99 ( 394 ), 116 – 131 . Google Scholar Crossref Search ADS WorldCat Arthur W. B. , Lane D. A. ( 1993 ), Information contagion . Structural Change and Economic Dynamics , 4 ( 1 ), 81 – 104 . Google Scholar Crossref Search ADS WorldCat Arthur W. B. ( 1994 ), Increasing Returns and Path Dependence in the Economy . University of Michigan Press : Ann Arbor . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Barnes W. , Gartland, Stack M. ( 2004 ), ‘ Old habits die hard: path dependency and behavioral lock-in ,’ Journal of Economic Issues , 38 ( 2 ), 371 – 377 . Google Scholar Crossref Search ADS WorldCat Bergek A. , Onufrey K. ( 2013 ), ‘ Is one path enough? Multiple paths and path interaction as an extension of path dependency theory ,’ Industrial and Corporate Change , 23 ( 5 ): 1261 – 1297 . Google Scholar Crossref Search ADS WorldCat Bhaduri S. , Sharma A. ( 2012 ), ‘Public understanding of participation in regulatory decision-making: The case of bottled water quality standards in India,’ Public Understanding of Science published online 24 July 2012. Bhaduri S. , Sharma A., Talat N. ( 2015 ), ‘ Growth of water purification technologies in the era of ‘regulatory vacuum’ in India ,’ Current Science , 108 ( 8 ), 1421. OpenURL Placeholder Text WorldCat Cowan R. ( 1990 ), ‘ Nuclear power reactors: a study in technological lock-in ,’ The Journal of Economic History , 50 ( 3 ), 541 – 567 . Google Scholar Crossref Search ADS WorldCat Cowan R. , Hulten S. ( 1996 ), ‘ Escaping lock-in: the case of the electric vehicle ,’ Technology Forecasting and Social Change , 53 , 1 – 124 . Google Scholar Crossref Search ADS WorldCat Cowan R. , Gunby P. ( 1996 ), ‘ Sprayed to death: path dependence, lock-in and pest control strategies ,’ The Economic Journal , 106 ( 436 ), 521 – 542 . Google Scholar Crossref Search ADS WorldCat Cusumano M. , Mylonadis Y., Rosenbloom R. ( 1992 ), ‘ Strategic manoeuvring and mass-market dynamics: the triumph of VHS over Beta ,’ Business History Review , 6 , 51 – 94 . Google Scholar Crossref Search ADS WorldCat Daniels B. , Mesner N. ( 2010 ), ‘Drinking water treatment systems, drinking water facts,’ Water Quality Extensions, Utah State University. http://extension.usu.edu/smac/files/uploads/dw_systems_jan2011.pdf (accessed 16 July 2013). Darby M. R. , Karni E. ( 1973 ), ‘ Free competition and the optimal amount of fraud ,’ Journal of Law and Economics , 16 ( 1 ), 67 – 88 . Google Scholar Crossref Search ADS WorldCat Devarhubli C. , Jha P. K. ( 2011 ), Is RO slowing you down?, DNA Syndication, 12 December 2011. http://dnasyndication.com/dna/article/DNAHM49968 (accessed 10 July 2013). David P. A. ( 1985 ), ‘ Clio and the economics of QWERTY ,’ The American Economic Review , 75 ( 2 ), 332 – 337 . Papers and Proceedings of the Ninety-Seventh Annual Meeting of the American Economic Association (May, 1985). OpenURL Placeholder Text WorldCat David P. A. ( 1994 ), ‘ Why are institutions the ‘carriers of history’? Path dependence and the evolution of conventions, organisations and institutions ,’ Structural Change and Economic Dynamics , 5 ( 2 ), 205 – 220 . Google Scholar Crossref Search ADS WorldCat David P. A. ( 1997 ), ‘Path dependence and the quest for historical economics: one more chorus of the ballad if QWERTY,’ Discussion Papers in Economic and Social History, Number 20, November 1997, University of Oxford. David P. A. ( 1999 ), ‘At last, a remedy for chronic QWERTY-skepticism!,’ All Souls College, Oxford. Presented at European Summer School in Industrial Dynamics (ESSID), held at l’Institute d’Etudes Scientifique de Cargèse (Corse), France, 5-12 September 1999. David P. A. ( 2000 ), Path Dependence, Its Critics and the Quest for ‘Historical Economics . All Souls College; Stanford University , Oxford . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Dobusch L. , Schubler E.. ( 2012 ), ‘ Theorizing path dependence: a review of positive feedback mechanisms in technology markets, regional clusters, and organizations ,’ Industrial and Corporate Change , 22 ( 3 ), 617 – 647 . Google Scholar Crossref Search ADS WorldCat Economic Survey ( 2016 -2017), Finance Ministry, Government of India. http://indiabudget.nic.in/vol1_survey.asp Foray D. ( 1997 ), ‘ The dynamic implications of increasing returns: technological change and path dependent inefficiency ,’ International Journal of Industrial Organization , 15 , 733 – 52 . Google Scholar Crossref Search ADS WorldCat Foreman-Peck J. ( 1996 ), ‘ Technological Lock-in ,’ and the Power Source for the Motor Car (No. _007). Economics Group , Nuffield College, University of Oxford . OpenURL Placeholder Text WorldCat Johnson J. , Busch M. ( 2009 ), Engineering Aspects of Reverse Osmosis Module Design . Lenntech BV : Netherlands . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Katz M. L. , Shapiro C. ( 1994 ), ‘ Systems competition and network effects ,’ The Journal of Economic Perspectives , 8 ( 2 ), 93 – 115 . Google Scholar Crossref Search ADS WorldCat Kay N. M. ( 2013 ), ‘ Rerun the tape of history and QWERTY always wins ,’ Research Policy , 42 , 1175 – 1185 . Google Scholar Crossref Search ADS WorldCat Kay N. M. ( 2013 ), ‘ Rerun the tape of history and QWERTY always wins: response to Arthur, Margolis, and Vergne ,’ Research Policy , 42 , 1195 – 1196 . Google Scholar Crossref Search ADS WorldCat Kozisek F. ( 2004 ), Health risks from drinking demineralised water, Rolling revision of the WHO Guidelines for drinking-water quality. World Health Organization Geneva. Lane D. , Vescovini R. ( 1996 ), ‘ Decision rules and market share: aggregation in an information contagion model ,’ Industrial and Corporate Change , 5 ( 1 ), 127 – 146 . Google Scholar Crossref Search ADS WorldCat Lee K. P. , Arnot C., Mattia D. ( 2011 ), ‘ A review of reverse osmosis membrane materials for desalination—Development to date and future potential ,’ Journal of Membrane Science , 370 , 1 – 22 . Google Scholar Crossref Search ADS WorldCat Liebowitz S. J. , Margolis S. E. ( 1995 ), ‘ Path dependence, lock-in, and history ,’ Journal of Law, Economics, & Organization , 11 ( 1 ), 205 – 226 . OpenURL Placeholder Text WorldCat Liebowitz S. J. , Margolis S. E. ( 1999 ), Path Dependence (online). http://ecsocman.hse.ru/data/018/784/1216/0770book.pdf (accessed 4 June 2012). Madaeni S. S. ( 1999 ), ‘ The application of membrane technology for water disinfection ,’ Water Research , 33 ( 2 ), 301 – 308 . Google Scholar Crossref Search ADS WorldCat Malaeb L. , Ayoub G. M. ( 2011 ), ‘ Reverse osmosis technology for water treatment: state of the art review ,’ Desalination , 267 , 1 – 8 . Google Scholar Crossref Search ADS WorldCat Marwas A. G. ( 2010 ), ‘Water & water treatment in India, Market opportunities for Swiss companies,’ OSEC. Meyer T.G. ( 2012 ), Path dependence in two-sidedmarkets: A simulation study on technological path dependence with an application to platform competition in the smartphone industry, Ph.D. thesis, Freie University, Berlin. Narduzzo A. , Warglien M. ( 1996 ), ‘ Learning from the experience of others: an experiment on information contagion ,’ Industrial and Corporate Change , 5 ( 1 ), 113 – 126 . Google Scholar Crossref Search ADS WorldCat Nelson P. ( 1970 ), ‘ Information and consumer behaviour ,’ Journal of Political Economy , 78 ( 2 ), 311 – 329 . Google Scholar Crossref Search ADS WorldCat Norgaard R. B. ( 1994 ), Development Betrayed. The End of Progress and a Coevolutionary Revisioning of the Future . London . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Page S.E. ( 2006 ), ‘ Path dependence ,’ Quarterly Journal of Political Science , 1 ( 1 ), 87 – 115 . Google Scholar Crossref Search ADS WorldCat Payment P. ( 1989 ), ‘ Bacterial colonization of domestic reverse-osmosis water filtration units ,’ Canadian Journal of Microbiology , 35 ( 11 ), 1065 – 1067 . Google Scholar Crossref Search ADS PubMed WorldCat Payment P. , Franco E., Siemiatycki J. ( 1993 ), ‘ Absence of relationship between health effects due to tap water consumption and drinking water quality parameters ,’ Water Science and Technology , 27 ( 3-4 ), 137 – 143 . © International Association on Water Quality. OpenURL Placeholder Text WorldCat PBI ( 2013 ), ‘Quality and safety evaluation of potable water purified by different technologies,’ Punjab Biotechnology Incubator for M/s Indo Canadian Clear Water Technologies, December 2012-February 2013. Penate B. , Garcia-Rodriguez L. ( 2012 ), ‘ Current trends and future prospects in the design of seawater reverse osmosis desalination technology ,’ Desalination , 284 , 1 – 8 . Google Scholar Crossref Search ADS WorldCat Perez-Gonzalez A. , Urtiaga A. M., Ibanez R., Ortiz I. ( 2012 ), ‘ State of the art and review on the treatment technologies of water reverse osmosis concentrates ,’ Water Research , 46 , 267 – 283 . Google Scholar Crossref Search ADS PubMed WorldCat Puffert D. J. ( 2002 ), ‘ Path dependence in spatial networks: the standardization of railway track gauge ,’ Explorations in Economic History , 39 , 282 – 314 . Google Scholar Crossref Search ADS WorldCat Puffert D. ( 2008 ), "Path dependence", EH.Net Encyclopaedia, edited by Robert Whaples. February 10, 2008. http://eh.net/encyclopedia/article/puffert.path.dependence (accessed 15 February 2013). Quitzau M. B. ( 2007 ), ‘ Water-flushing toilets: Systemic development and path-dependent characteristics and their bearing on technological alternatives ,’ Technology in Society , 29 , 351 – 360 . Google Scholar Crossref Search ADS WorldCat Rao A. P. ( 2011 ), Handbook on Water and RO . Health point Services (India) Private Ltd : New Delhi . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ray A. S. ( 2004 ), Medicines, Medical Practice and Health Care in India in the Era of Globalisation: Political Economy Perspectives, the Independent Commission on Development and Health in India, New Delhi. Rosenberg N. ( 1976 ), Perspectives on Technology . Cambridge University Press , Cambridge . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Rosenberg N. ( 1982 ), Inside the Black Box: Technology and Economics . Cambridge University Press , Cambridge Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Sharma A. , Bhaduri S. ( 2013 ), Consumption conundrum of bottled water in India an STS perspective ,’ Bulletin of Science, Technology and Society , 33 ( 5–6 ), 172 – 181 . Google Scholar Crossref Search ADS WorldCat Super Brands Report ( 2009 ), Aquaguard: leaders in water purification systems. http://www.superbrandsindia.com/images/brand_pdf/consumer_3rd_edition/Aquaguard.pdf (accessed 30 June 2013). US EPA ( 1999 ), 25 years of the safe drinking water act: History and trends, United States Environment Protection Agency, EPA 816-R-99-007, 1999. Vergne J.-P. , Durand R. ( 2010 ), ‘ The missing link between the theory and empirics of path dependence: conceptual clarification, testability issue, and methodological implications ,’ Journal of Management Studies , 47 , 4 . Google Scholar Crossref Search ADS WorldCat Vergne J.-P. ( 2013 ), ‘ QWERTY is dead; long live path dependence ,’ Research Policy , 42 , 1191 – 1194 . Google Scholar Crossref Search ADS WorldCat WHO ( 1996 ), Guidelines for Drinking-Water Quality, 2nd ed.-Vol. 2-Health Criteria and other Supporting Information . World Health Organization : Geneva . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC WHO ( 2003 ), pH in Drinking-water, Background document for development of WHO Guidelines for Drinking-water Quality. WHO. World Health Organisation , Water Quality and Health Strategy 2013-2020. World Health Organization ( 2004 ), Guidelines for drinking-water quality (Vol. 1) , World Health Organization , Geneva . World Health Organisation ( 2008 ), Guidelines for drinking-water quality [electronic resource]: incorporating 1st and 2nd addenda, Vol. 1, Recommendations—3rd ed. Geneva. http://www.who.int/water_sanitation_health/dwq/fulltext.pdf World Health Organisation ( 2011a ), Guidelines for Drinking-water Quality, Fourth Edition, Geneva. http://whqlibdoc.who.int/publications/2011/9789241548151_eng.pdf World Health Organisation ( 2011b ), Evaluating household water treatment options: health-based targets and microbiological performance specifications, Geneva. http://www.who.int/water_sanitation_health/publications/2011/evaluating_water_treatment.pdf © The Author 2017. Published by Oxford University Press on behalf of Associazione ICC. All rights reserved.
TI - Incremental innovations, information contagion, and path dependence: the case of drinking water purification technologies in urban India
JF - Industrial and Corporate Change
DO - 10.1093/icc/dtx010
DA - 2017-12-01
UR - https://www.deepdyve.com/lp/oxford-university-press/incremental-innovations-information-contagion-and-path-dependence-the-8dYXSHyalq
SP - 1089
VL - 26
IS - 6
DP - DeepDyve
ER -