Literature has divided sources of technology for foreign subsidiaries into six. Each of them can be drawn considering the organizational boundary perspective, that is, whether the technology source originates inside the MNC or not. From the boundary MNC faces, internal and external sources of technology are separated. Internal sourcing are intra-unit sources (i.e., within the subsidiary limit) and intra-MNC source (i.e., sourcing from the headquarters or sister units). The three external sources are organizations in the host country, home country, and third-countries. Some studies have considered the traditional division of internal and external sources in their analysis of subsidiary technology sourcing (Hsu et al., 2016; Swan & Allred, 2009). As this study is explorative in nature, it will separate the sources into internal and external.

Literature has argued that internal sourcing affords the greatest control, more tacit knowledge can be shared, facilitates a tighter degree of coupling between MNC stakeholders, and provides learning benefits and cost-saving in resource transfer (Allred & Swan, 2014). However, the selection of internal sources (i.e., intra-MNC or intra-unit source) varies according to the role of the subsidiary.

External sources enforce international competitiveness as this is strictly connected to the dispersed knowledge of the market for technology. Studies concentrate on the distinction between home, host, and third-country external sources.

Host technology sources, specifically, can support the subsidiary in the creation process to develop new goods for the local market because these technologies are more aligned with local tastes and conditions (Zhang & Pearce, 2010). However, researchers assert that their application depends on the extent to which a subsidiary has developed a close relationship with local stakeholders, a rich local stock of knowledge (Berry, 2018), and operation experience in the host country (Miao et al., 2011; Zhang et al., 2018). In non-developed context, such as the Ibero-American region, it is rarely treated as potential knowledge resource of technology with a deep trust in local partners (Du & Williams, 2017). Therefore, we argue that home and third-country external sources are more often preferred than local sources for relevant production technologies.

As it is difficult to predict what outsourcing mode is preferred for a specific subsidiary strategic role, this can be better approached by the novelty of the external sources, whether it is known or new for the organization. Applying the concepts of the organizational learning literature of exploration and exploitation to outsourcing, we can draw on exploration of external sources (the use of new external sources) and exploitation of external sources (the use of known external sources (Rothaermel & Alexandre, 2009)). Outsourcing strategy in production technology acquisition varies according to the strategic role of the subsidiary.

The decision to internally source or outsource a production technology also depends on technological attributes, whether knowledge embedded in the technology is familiar with the MNC's core business or not, and its technological complexity.

Subsidiarity role is made effectively and benefits the group, when its operations are also defined interactively through an understanding of the technological trajectory of the parent-MNC network (Zhang et al., 2018). In this sense, the selection of internal sourcing in the subsidiary level must be related with a high familiarity with the technology so that the final creation would be in balance with the existing corporate technological knowledge. Additionally, internal personnel and resources can be more efficiently utilized when the production technology required is closely related to a group's core technology. Recent research on sourcing supports this (Cabrera & González, 2019). Alternatively, when subsidiaries recognize that production technology is not compatible with the MNC core business, it is more likely to be outsourced.

Regarding technology complexity, we establish that subsidiaries can source production technology with a varied level of complexity using external sources. However, internal sourcing is intended for a certain complexity compatible to the subsidiary's internal capacity (Murphree et al., 2022; Thakur-Wernz et al., 2020). Production technologies are often designed and developed around special resources such as skilled employees, machine tools, and CAD/ CAM/CAE software. This may constrain design on a more complex production technology or to a simpler one. As the allocation of these resources depends on the strategic roles of the subsidiary, higher complexity of technology is internally developed by major-role subsidiaries than minor-role units.

Our research builds upon the work of the abovementioned authors. The following proposition describes an empirically testable model depicted in Figure 1. Technological attributes combined with strategic roles are important in the selection of the mode in production technology sourcing by subsidiaries based in Ibero-America. We argue that technology familiarity determines the selection between internal and external sources and technology complexity divide internal sourcing modes when subsidiary roles are considered. Various technology complexities can be sourced by external sources, but their novelty is used differently by a subsidiary role. Then, we express the following proposition:

P. Ibero-American production subsidiaries internally source a production technology with certain technology complexity related to their strategic roles. However, when the technology familiarity is low, they will choose external sources, regardless of their roles and complexity of the technology.

Figure 1 Notes: the gray area are sourcing modes selected by subsidiaries with major strategic roles. While sourcing modes in white are selected by subsidiaries with a minor strategic role.

We used semi-structural interviews to capture projects on internal and external production technology sourcing. Data collection was carried out on field trips between late 2017 and early 2019, spending a complete day at each subsidiary. We relied on previous research to prepare the questions and adapt them to our context (e.g. Zhang et al. (2018)). As such, we pilot-tested the question list on a group of three subsidiary managers from the manufacturing industry based in Peru. The interviews with sample subsidiaries were done in three phases. We first started with introductory questions about the subsidiary role, functional scope, host market characteristics, intra-MNC communication, and strategic independence in production technology sourcing. Second, we asked the respondents to think about internal production technology with which they had spent a lot of money. Third, we asked them to tell the story about the decision (How was this decision made in the first place?') and the factors that affected decision-making (What was the factor that influenced the selection of this sourcing option?'). Next, we asked respondents to consider external production technology that required a significant investment from them. Finally, we asked them to tell the story and factors about the decision to use external sources. See the Appendix for the questionnaire.

These data were complemented by technology observation, file notes collected in guided tours of production, and secondary data from innovation surveys and publicly available information such as news and web pages. The interviews and technology observation were led by the corresponding author, who is a mechatronic engineer with a more technological vision. Table 3 shows the summary of the different collection techniques applied in this study.

Table 2

MNC	Pencil manufacturer (A)		Home-appliance manufacturer (B)		Automation-tech. Manufacturer (C)
Role	Major	Minor	Major	Minor	Major	Minor
Identifier	A1	A2	B1	B2	C1	C2
Host Country	Brazil	Colombia	Spain	Peru	Brazil	Peru
City	Sao Carlos	Bogota	Navarra	Lima	Sao Paulo	Lima
Start of Operation	1930	1976	1988	1996	1968	2003

Souce: own elaboration.

Table 3

Type	City, Country	Date (month-year)	Duration (minutes)	Respondent
			25
Pilot interviews	Lima, Peru	Aug-17	23 28	General Manager
	Lima, Peru	Jun-18	30	General Manager
Pencil MNC Interviews	Bogotá, Colombia	Jul-18	62	Production Manager
	Sao Carlos, Brazil	Dic-18	68	Production Manager
	Lima, Peru	Sep-17	64	General Manager
Home-Appliance MNC Interviews	Lima, Peru	Oct-17	64	Production Manager
	Navarra, Spain	May-18	30	Production Manager
Automation Technology MNC Interviews	Lima, Peru Sao Paulo, Brazil Lima, Peru	Nov-18 Dic-18 Feb-19	42 87 42	General Manager Production Manager Production Manager
Summary
Interviews	12
General Managers	6
Production Managers	6
Technology Observation	12
Internally Sourced	6
Outsourced	6

Souce: own elaboration.

Qualitative data from interviews and file notes were transcribed. Data were coded with DEDOOSE, a qualitative data analysis software. Deductive analysis of the data guided by the theoretical model of the research was considered. Concepts suggested by the theoretical model create sub-categories and categories that were constantly compared with the data collected. The corresponding author coded the data independently and systematically. Subsequently, in a joint session, the classification of codes into the established categories were discussed by the research team until a consensus was reached. Throughout the research process, we used the memo-writing technique to keep us involved in the analysis. Table 4 shows the general categories and subcategories with example quotations from the interviewees.

Table 4

General Categories	Subcategories	Relevant quote examples from the Interviewees
		Major-role subsidiaries	Minor-role subsidiaries
Subsidiary attributes for	Internal capability	"We use the same resources as the headquarters. When it comes to development in Brazil, the parent company does not worry much because we have a large internal development capacity."	"We in the group have a system for managing the factory, but the way it's done here is very different from how it's done in Europe. Here, we work with very basic technologies, simple technologies. In other parts of the world, you may work with a lot of automation and robots."
sourcing	Internal embeddedness	"We actively participate in the group technology network, where we share knowledge with the goal of standardizing production technologies and processes."	"Our geographical position could be very important, but for the company, an international company that has factories around the world, there are managers who don't even know that the group has a factory in Colombia."
Subsidiary Sourcing	Internal Sourcing Modes	"There is a special machine that has been developed by the parent company, but it is very expensive because it is intended for mass production. I proposed making new equipment in Brazil for lower production volume, but I had to collaborate with them to develop it."	"Our only in-house development was a small machine that our maintenance staff developed with basic resources from our factory, but it was useful because we created several replicas"
Modes	Types of External Sources	"We approved to import this line from Germany. Not from the headquarters, but from a new supplier that works with the headquarters" (new external source)	"Some machines have been bought on the recommendation of the headquarter, German brands were bought, of the machines we have, to say a percentage, I think 90% have been bought on their recommendation" (known external source)

Souce: own elaboration.

We analyzed the technologies collected during the interviews by observation to evaluate their technological attributes. Twelve observations were carried out during the guided tour of the production plant at each subsidiary. We evaluated "technology familiarity" dichotomously. We labeled 'familiar' if the researcher observed that the knowledge underlying the technology was related to the core business of the MNC, and 'non-familiar' otherwise. Furthermore, technology complexity was measured by the complexity coding system method proposed by ElMaraghy (2005) and also used in technology sourcing studies (e.g. Cabrera & González, 2019). It captures the variety and amount of information in each technology and its components. The overall index is the sum of the structural, control, programming, and operation complex indexes. Therefore, the complex index that ranges from zero to six comprises low complex production technologies. Medium complexity with a range from 7 to 17. Finally, scores above 18 are described as high and advanced production technologies.

The next step was to perform cross-case analysis, in which we compared single cases and looked for similarities, differences, and patterns. Table 5 summarizes the data on the collected cases and details the sourced technology name, sourcing mode, a brief of the decision-making, technology familiarity, and complexity index. 3.4 Ensuring trustworthiness.

To ensure that Yin's (2018) trustworthiness criteria have been met, we used four separate techniques. The first one was data triangulation through four different collection techniques: interviews to general managers of the Peruvian subsidiaries that supported us to find the relevant production subsidiaries, interviews to the production managers of the sample subsidiaries to collect the projects, data from technology observation, and secondary data.

The second technique was peer examination, in such a way that we asked two of our colleagues that were not part of the research team to examine the interpretation process of the data. We also applied audit trail as another trustworthiness strategy, data collection strategies, analytic choices, core reflections, and decisions were entered in many memo-writings. Finally, we spent a complete day at the subsidiary to fulfill the 'Prolonged engagement' technique. It allowed for sufficient time to observe the technology and build trust with the informants and correct some misinterpretations.

Findings suggest that the headquarters recognized the importance of major role units by supporting their internal sourcing projects. The three internal sourcing projects for A1, B1 and C1 showed that they pursued internal exploration by co-developing production technologies with the headquarters and sister subsidiaries (i.e., intra-MNC source). In turn, minor-role subsidiaries embarked solely on internal development using their own internal resources (i.e., intra-unit source).

In the cases of outsourcing, both types of subsidiaries requested support to the MNC about production-technology suppliers. However, there were differences.

Regarding major-role units, home and third-country suppliers were used. Headquarters supported the unit in the supplier procurement. Data showed that headquarters shared new suppliers with the unit that were not part of the stablished staff of suppliers. This suggests that the head office recognized the capacity of the unit to interact with new suppliers and expand the MNC supplier network. This excerpt from A1 evidenced it:

"We approved to import this line from Germany. Not from the headquarters, but from the suppliers that supply the headquarters. They are the same guys, the same guys that source the headquarters, but to Brazil. The headquarters knew what was happening".

Minor-role units used home and third-country suppliers as external sources. They sourced technology to improve performance indicators such as production volume, flexibility, and customization. The minor-role unit had short relationships with the supplier because the technology was readily available in the market to be purchased and installed in the factory without any modifications. Furthermore, minor-role units were supported by the headquarters in the procurement of the supplier. The unit received not only the supplier's name but also the model and characteristics of the technology, indicating that suppliers were known. The manager of C2 described it as follows:

"Germany is very, very standardized in the processes, what machines, what models are already standardized. If I want to leave, probably and Germany won't find me out, cool, but if they find me out, they will bother. Everything is very standardized, no, no, I can't start inventing things".

Subsidiaries decide to internalize when the knowledge underlaying the technology is related to the MNC's core business, regardless of the strategic role. Although, major-role subsidiaries use the knowledge from MNC repositories for internal exploration by co-developing production technologies with the headquarters and sister subsidiaries. This excerpt showed the involvement of the headquarters in the internal sourcing project of the C1 unit:

"There is a special machine that has been developed by the parent company, but it is very expensive because it is intended for mass production. I proposed making new equipment in Brazil for lower production volume, but I had to collaborate with them to develop it."

Additionally, it is observed that these minor-role subsidiaries use resources as part of the subsidiary's heritage that was transferred by the MNC from the beginning of operations and used to run the business in the host market. These resources were readily available in the unit rather than emerging from a formal R&D. One instance was expressed by the respondent of C2 about his internal project:

"We had the components and the knowledge to make the technology because this is our business, we knew what we needed":

On the contrary, when production technology was not familiar, minor-role subsidiaries chose outsourcing. However, contrary to our expectation, two major-role subsidiaries did not follow this reasoning. Outsourcing by major roles was different in comparison to minor role units because the technologies were not readily available in the market to purchase them directly. Then major-role units needed to co-develop special-purpose technologies with the suppliers. In this sense, technology familiarity was a prerequisite to the knowledge transfer to the suppliers to accomplish the goal of adapting the technical features to production demands.

Subsidiaries internally sourced production technology with certain technology complexity related to their strategic roles. Major roles units internally developed entire lines of production. Again, regarding internal sourcing, MNC stakeholders such as the sister and headquarters were used as a source of production technology. Sourcing knowledge from the headquarters and sister units were useful in the development of complex lines of production.

By contrast, minor-role units did not receive any attention from the headquarters in their internal sourcing projects. Furthermore, their projects were the first and the only ones within their factories as opposed to major-role units where projects triggered subsequent developments, and the continued accumulation of knowledge related to production technologies. The three internal cases in A2, B2, and C2 subsidiaries were small applications characterized as low technological developments carried out less formally by members of the engineering unit and production personnel.

In cases of outsourcing, both types of subsidiaries sourced production technologies with different levels of complexity. Hence, the pattern was not identified.