015-0395
TITLE
The Sequence of Cumulative Capabilities: A Comparison of Three Industries in Thailand
Sakun Boon-itt[1]
ThamasatBusinessSchool
2 Prachan, Rd Pranakorn
Bangkok, THAILAND 10200
E-mail:
Chee Yew Wong
Logistics Institute
HullBusinessSchool
Hull, UK HU6 7RX
E-mail:
POMS 21st Annual Conference
Vancouver, Canada
May 7 to May 10, 2010
The Sequence of Cumulative Capabilities: A Comparison of Three Industries in Thailand
Abstract
Purpose – This research compares models (sequences) of cumulative capabilities among automotive, electronics and food industries in Thailand, and further explains how different industrial contexts shape different models of cumulative capabilities.
Design/methodology/approach – Based on survey data, structural equations modelsfor each of the three industries are examined and compared.
Findings – Quality is the base for all other capabilities for all three industries. The research found different models (sequences) of cumulative capabilities for the three industries: delivery becomes the next capability for the automotive industry; production flexibility for the electronics industry and production cost for the food industry.
Research limitations/implications – Results confirm the effects of industrial differences, suggesting further research into industry contexts in terms of the rate new product introduction (clock-speed), demand characteristics, product life-cycle, industry maturity, and competitive priorities. The results should not be interpreted in a prescriptive manner. Similar research replicated in other countries may yield different results.
Practical implications – Recognise the relative importance of various manufacturing capabilities in different industries and different sequences of cumulative capabilities for different industries.
Originality/value – Clarify the effects of industrial contexts in shaping different models of cumulative capabilities.
Keywords Manufacturing capability, operations strategy, automotive industry, food industry, electronics industry
Paper type Research paper
Introduction
The operations strategy literature has long recognised that various manufacturing capabilities (i.e. quality, cost, flexibility and delivery) may reinforce each other to form a certain pattern of cumulative capabilities. Some literature also suggests that the sequence of cumulative capabilities may differ from one industry to another because the importance of various capabilities in different industries is not the same (Noble, 1995; Flynn and Flynn, 2004). For example, quality and product introduction speed may form the basis of the cumulative capabilities for industries competing in more competitive and rapidly changing industries, where product life cycles are shorter (Flynn and Flynn, 2004). Furthermore, in a more mature industry, where cost becomes an order qualifier, cumulative of multiple capabilities is often essential for success (Hill, 1988; Schmenner and Swink, 1998). To-date different models of cumulative capabilities have been suggested or discovered but the reasons for their differences are not adequately explained.
Recognising the potential contextual effects of industry differences, this research is set out to compare models of cumulative capabilities among automotive, electronics, and food industries in Thailand. This research posits that quality forms the base of the models of cumulative capabilities for all these three industries, but the next capabilities in the models of cumulative capabilities vary depending on industrial contexts. We hypothesise that delivery, production flexibility, and production cost are the next capabilities for the automotive, electronics and food industries respectively. To test these hypotheses, all together 151, 102, and 115 usable responses respectively from these three industries are analysed in three separate structural equation models. Path analyses of the three models support the above assumptions, which confirm that different industries have different models (sequences) of cumulative capabilities.
This research contributes to the operations management literature in several aspects. First, unlike most previous cross-sectional studies, which combine data from various industries, this research builds three separate structural equations models for each of the three industries. In this research, cross-sectional analysis of dataset from multiple countries is avoided because it can only provide indirect empirical measures of dynamic changes of capabilities and it can never indicate chronological sequences (Gröler and Grübner, 2006). Second, this research applies structural equation modelling because it is able to conclude a total model of cumulative relationships between capabilities (Gröler and Grübner, 2006). Other methods (i.e. correlation, stepwise regression, multiple regression, common method variance, and path analysis) commonly applied to study cumulative model are not able to establish such a total model and therefore only provide us with a partial understanding of the sequence of cumulative capabilities.
Third, though differences between countries (regions) have already been identified as a potential contingent factor affecting the relationships between manufacturing capabilities (Hall and Nakane, 1990; Noble, 1995; Flynn and Flynn, 2004) the effects of the industrial contexts are still unclear. Noble (1995) argued that industrial contexts may affect the sequences of cumulative capabilities because different types of industries compete with different sets of competitive priorities. Based on a study of three North American industries, Noble (1995) concluded that the process industry performed low on productivity but competed on delivery; the metal fabrication and assembly industry focused on quality, dependability and cost (high productivity cluster); and the high-technology industry competed on flexibility and innovation. Noble (195) found no significant difference among different industries from the Korean sample. However, in a more recent study of five countries, Flynn and Flynn (2004) found limited support for the hypothesis that there are differences in the pattern of cumulative capabilities between electronics, machinery, and transportation component manufacturing industries. With these mixed results the effects of industrial contexts are still inconclusive, and therefore further research is required.
Models of cumulative capabilities
When developing an operations strategy, managers need to decide which manufacturing capabilities to prioritise for improvement. They may apply the trade-off model, which argues that one manufacturing capability can only be improved at the expense of other capabilities (Skinner, 1974). An alternate view to that of the trade-off model is that one capability would enhance another; they become cumulative (Ferdows and De Meyer, 1990). Ferdows and De Meyer (1990) further argue that when capabilities are developed in a cumulative manner, it is likely to be more lasting and less fragile than if it were developed at the expense of other capabilities. The operations management literature also suggests that cumulative capabilities can be achieved by improving various capabilities according to a particular sequence. For example, a study of Japanese manufacturers concluded that manufacturers must qualify for a minimum level of abilities on quality, dependability and cost improvement before they can offer flexibility; failing to achieve these “base” capabilities can end up with a tragic or chaos condition (Nakane, 1986).
Quality as the base of cumulative capabilities
Even though the operations management literature has reached little consensus about the sequence of cumulative capabilities, most empirical studies concluded that quality should be the base of all other capabilities for most manufacturing industries (i.e. Nakane, 1986; Noble, 1995; Ferdows and De Meyer, 1990; Corbett and van Wassenhove, 1993; Swink and Way, 1995; Rosenzweig and Roth, 2004; Gröler and Grübner, 2006; Amoako-Gympah and Meredith, 2007). In line with the above literature, this research considers quality as the base of the other cumulative capabilities. The relationships between quality and other capabilities (delivery, flexibility and cost) are thus formulated as follows.
In this research, a product is considered to be of high quality when it meets customers needs (market-based quality) without generating in a lot of rework or waste during the process of production (conformance or process-based quality). Conformance or process-based quality is essential because when there is effective quality control the production process will become reliable, less variable and more predictable; such a capability is essential to warrant on-time and reliable delivery (Ferdows and De Meyer, 1990; Noble, 1995; Fawcett et al., 1997). Furthermore, manufacturers with less quality problems will be able to reduce delivery lead time because they will spend less time and resources to rework or handle rejects (Flynn et al., 1995; Milgate, 2000). Several empirical studies have confirmed that an enhanced product (conformance) quality positively and directly influences improvements in delivery reliability (Flynn and Flynn, 2004; Gröler and Grübner, 2006). Based on these arguments, the following hypothesis H1 is formulated.
H1. Quality has a direct positive impact on delivery capability
The relationship between quality and production flexibility has also been widely examined. One of the reasons a manufacturer suffers from low production flexibility is that much of its production capacity and resources areoccupied by activities that handle poor quality or produce buffer stocks, instead of being allocated to produce the right products for the right customers whenever they are needed. Theoretically, an improvement in conformance quality will reduce the uncertainties of customer requirements and timing of material supply quality (Wacker, 1996). These improvements will improve the accuracy of production scheduling and subsequently increase not only delivery reliability but also volume and lead-time flexibilities (Corbett and van Wassenhove, 1993). Also, poor quality reduces speed (Ferdows and De Meyer, 1990) and speed is an essential enabler of volume and lead-time flexibility. Furthermore, in order to offer many product variants to meet different customer needs, production flexibility (e.g. range or product mix) becomes even more critical. Many operations managers with a trade-off mindset would argue that offering more product variants will lead to poorer quality or increased cost. However, for example, the Yamazaki machine tool factory in the UK was able to offer four times more models in the third of the time normal to the industry, while the quality of their products “matched or beat” the high Japanese standard (Jones et al., 1988). One of the explanations for this success is that lower process variability, as a result of a higher level of product (conformance) quality, leads to greater flexibility in offering a wider variety of products. Based on these arguments, as well as existing empirical evidence (i.e. Gröler and Grübner, 2006) this research formulates the following hypothesis H2.
H2. Quality has a direct positive impact on production flexibility
Plenty of previous research has confirmed that enhanced cost competitiveness can be achieved by investment in a quality programme (Crosby, 1979, Deming, 1982; Juran et al., 1974; Garvin, 1987; Skinner, 1986; Gupta and Campbell, 1995; Flynn et al., 1995). Furthermore, Ferdows and De Meyer (1990) argue that the quality-cost relationship does not work in the opposite direction - an increase in cost efficiency does not seem to improve quality. This is because any reduction in labour and/or material cost must be originated from (but not a consequence of) the improvements in process or innovation in material owing to a quality programme or any similar improvement initiative (Ferdows and De Meyer, 1990). This argument also furthersupports our assumption that quality is the base of other cumulative capabilities. Essentially, when quality control becomes effective, the production process will produce less rejects and therefore less rework is required, subsequently resulting in lower cost of poor quality (Crosby, 1979; Deming, 1982; Gupta and Campbell, 1995; Flynn et al., 1995). Furthermore, quality management techniques such as the quality deployment function, Taguchi method, and design for manufacturing design are often used to improve the features of a product as well as the costs of production (Taguchi and Clausing, 1990; Lockamy and Khurana, 1995). Based on these arguments, and existing empirical evidence (i.e. Gröler and Grübner, 2006; Amoako-Gympah and Meredith, 2007), this research formulates the following hypothesis H3.
H3. Quality has a direct positive impact on production cost
This research posits that the above three hypotheses (H1, H2 and H3) form the base of the models of cumulative capabilities applicable for most industries. Though there is already a lot of empirical evidence which supports this assumption, this research has not ignored the contradictory results reported by Flynn and Flynn (2004). They found that quality was the base capability for Korean manufacturers but not for Italian and German manufacturers. Experience in quality management and the theory of performance frontier are among the explanations used to explain such a contradictory finding. As explained by Flynn and Flynn (2004), the Italian and German manufacturers, in their samples, were operating at a relative high level of competition and perhaps at performance frontiers; therefore they could not solely rely on quality to improve other capabilities. Instead, manufacturers from developing countries such as Ghana, as argued by Amoako-Gympah and Meredith (2007), will rely heavily on quality improvement to improve other capabilities. Since this research focuses on relatively less matured industries in Thailand, quality is expected to form the base for all other capabilities.
The next capability after quality
There is a need to determine the next capability to be improved after determining quality as the base. Which capability should be improved next - delivery, flexibility, or cost? Flynn and Flynn (2004) found that cumulative capabilities may involve more delivery for some industries but cost for other industries. That means delivery, production cost and perhaps production flexibility may become the next capability of the models of cumulative capabilities, depending on industrial and other contexts. When each of these three capabilities is considered as the base of the next (upper) level of the models, we will find three possible models of cumulative capabilities. Figure 1 illustrates the three possible models and their respective supporting hypotheses.
< Insert Figure 1 about here >
Figure 1 also illustrates another primary assumption of this research i.e. the models of cumulative capabilities do not follow a simple and serial sequence (such as quality>dependability>flexibility>cost, proposed by some literature), but they follow a divergent sequence in such a way that quality will have positive impacts on all three other capabilities (delivery, cost, and flexibility) while one of these three capabilities will form the base for the other two capabilities. This research also suggests that the choice of the next capability depends on differences in industrial characteristics. Based on different characteristics for automotive, electronics and food industries in Thailand, we further establish hypotheses (H4, H5 and H6) about the relationships among delivery, cost and flexibility for each of these industries.
Automotive industry (Model A)
Quality has always been an order qualifier in the automotive industry (Curkovic et al., 2000). Thus, quality should become the first capability to be improved and form the base of the model of cumulative capabilities for the automotive industry. Other than quality, just-in-time (JIT) delivery has been identified as the second (and perhaps equally critical) competitive weapon of successful automakers (Schonberger, 1982; Goyal and Deshmukh, 1992). JIT is characterized by small batch-size production with relatively low defects and process variability (Schonberger, 1982; Ohno, 1988). JIT cannot exist in manufacturing plants with poor conformance quality; instead, quality is the prerequisite for implementing JIT (Zipkin, 1991). Furthermore, it is world-class quality and JIT delivery that allowed Japanese automakers to overtake in the competitive landscape of automotive markets (Schonberger, 1982; Womack et al., 1990; Zipkin, 1991). Even facing low-cost competition from China and sophisticated competitors from Korea and the United States, Japanese automakers have not changed their strategic focus on quality and JIT (Daniel et al., 2009). Instead, successful automakers rely heavily on conformance quality and reliable JIT delivery to improve cost competitiveness and flexibility. Since Japanese automakers are still the leaders of the industry, many other countries (including Thailand) attempt to imitate them in many aspects, including engaging in the same sequence of building up cumulative capabilities i.e. first quality and then JIT delivery (model A in figure 1).
Model A represents the case where quality becomes the base capability and delivery becomes the base at the next (upper) level of the model of cumulative capabilities. This is by far the most accepted model in the literature (i.e. Nakane, 1986; Noble, 1995; Ferdows and De Meyer, 1990; Swink and Way, 1995; Rosenzweig and Roth, 2004; Gröler and Grübner, 2006). As suggested by Sakakibara et al. (1997) and Funk (1995), manufacturers operating at a high delivery speed are able to improve flexibility of their operations because less time is required to respond to different demands or adjust to changing requirements. Furthermore, when manufacturers reduce variance in their delivery processes and the predictability of the production and distribution systems (delivery reliability), production (volume) flexibility will be enhanced (Flynn and Flynn, 2004; Rosenzweig and Roth, 2004). Often, a reduction in lead time and process variance means higher productivity and lower inventory, therefore leading to a reduction in production cost. In addition, a relationship between delivery and cost has been found in the literature by Narasimhan and Jayaram (1998). Based on these arguments the following two hypotheses are proposed for the automotive industry:
H4a. Delivery capability has a direct positive impact on production flexibility
H5a. Delivery capability has a direct positive impact on production cost
The next challenge is to explain the relationship between flexibility and cost for the automotive industry. Theoretically a high level of production flexibility will enable automotive manufacturers to reduce inventory costs while achieving a high service level. This is the competitive advantage of JIT. A number of studies have found that without a high level of flexibility the industry will simultaneously suffer from a high level of inventory, stock-out, and excessive overtime especially in a uncertain environment (Gerwin, 1993; Pagell and Krause, 1999; Koste and Malhorta, 2000). With this theoretical argument hypothesis H6a is established as follows.
H6a. Production flexibility has a direct positive impact on production cost
Hypothesis H6a has actually been tested empirically and remains inconclusive in the literature. A meta-analysis of previous empirical studies found mixed results on the flexibility-cost relationship (White, 1996). According to the empirical study of Gröler and Grübner (2006), flexibility and cost are not mutually exclusive and a trade-off relationship (negative impact) seems to exist between these two capabilities. Since this study applied cross-sectional data it is difficultto compare with the automotive industry. Furthermore, this finding needs to be examined carefully rather than simply accepting that trade-off occurred. According to Schmenner and Swink (1998), trade-off becomes necessary when a firm reaches its performance frontier, and negative relationship between two capabilities may occur due to time-slack between the effects of improvement of a capability on another capability. Since we do not expect to find industries with performance frontiers in Thailand, time-slack may become one of the other potential conditions which might lead to non-significant or even a negative flexibility-cost (H6a) relationship.