Methodological Issues in Researching Internationalisation Strategies of Australian Firms[1]
Geoffrey Lewis
Visiting Professor, Melbourne Business School
21 Osborne Street, Hackney, S.A. 5069
Tel (618) 8362 0621
Tatiana Minchev
Doctoral Candidate, Flinders University of South Australia
16 Dotterel Drive, Semaphore Park S.A. 5019
Tel. (618) 8449 8215
SCHOOL OF COMMERCERESEARCH PAPER SERIES: 01-3ISSN: 1441-3906
Abstract
The central issue of research into internationalisation of Australian companies is under what circumstances geographical diversification improves the financial performance of a firm.
Significant methodological problems related to operationalising internationalisation and performance are confronted. Of these, difficulties associated with the collection of firm-level data in Australia are the most pressing. A mechanism needs to be established for the collection of such data relevant to internationalisation. This is consistent with calls for greater disclosure by Australian companies in general.
1.Introduction
At the heart of research into the patterns and processes of internationalisation of Australian companies is the issue of under what circumstances geographical diversification improves the financial performance of a firm. Despite the theoretical and empirical advances, many authors note the increased inconsistencies between the findings of different studies on the geographical diversification-performance relationship (Grant et al., 1988; Sullivan, 1994; Hitt, Hosskisson & Kim, 1997; Gomes & Ramawasamy, 1999, Delios & Beamish, 1999; Geringer, Tallman & Olsen, 2000).
Although the process of firms’ internationalisation is a well-researched area, studies in the Australian context are limited and focus on either industry-level policy (The Global Challenge, 1990: Yetton et al., 1992; Industry Commission Report No 53, 1996), the behavioural characteristics of exporting firms (Barrett, 1986; Welch, 1993) or the behaviour of multinationals in Australia (Bora, 1998). The overwhelming majority of research on the international diversification-performance paradigm comes from the USA and Europe (Rugman, 1983; Michael & Shaked, 1986; Grant, 1987; Grant et al.; 1988, Geringer et al., 1989; Olusoga, 1993; Sullivan, 1994; Tallman & Li, 1996; Hitt et al., 1997; Gomes & Ramawasamy, 1999).
The study of the relationship between internationalisation and performance of Australian companies is of considerable practical importance, as many of these firms have failed to match their record of success in domestic markets with performance in overseas markets. Well-known examples include TNT, Burns Philp, Westpac, CSR and ANI (Ferguson & James, 1998).
2.Area of research interests
Discussion of previous research
Early research (Stopford & Wells, 1972; Rumelt, 1974) does not specifically explore the important relationship between geographical diversification and the performance of an international firm. Most of the subsequent studies, as Kim et al. (1993) observe, are still dominated by Rumelt’s product diversification-based approach. Therefore, they fail to capture both the diversity of foreign markets served by a firm and the dispersion patterns of a firm’s activities around the globe.
Empirical studies which were focussed on international diversification and performance, and patterned on product diversification research began to appear in the late 1980s. A concise review of the state of knowledge in this area is given in Lewis & Minchev (2000). More comprehensive reviews of the product and international diversification and performance literature are provided by Dess, Gupta, Hennart & Hill (1995), Delios & Beamish (1999), Gomes & Ramaswamy (1999), Geringer, Tallman and Olsen (2000), and Palich, Cardinal & Miller (2000). Despite more than a decade of research into the area and increasingly more sophisticated statistical techniques, the precise nature of the relationship between geographical diversification and performance remains problematic. For example, Sullivan (1994) cites six previous studies that report a positive, six an intermediate and five a negative relationship. Geringer et al. (1989) establish a deterministic relationship between financial performance and internationalisation, and theorise about the existence of a certain "threshold of internationalisation". Another issue is whether the relationship is linear or quadratic (Grant et al., 1988; Gomes & Ramaswamy, 1999). Furthermore, causation of the relationship is ambiguous, i.e. whether multinational diversification promotes profitability or whether profits generated at home are financing overseas investments (Grant, 1987).
To our knowledge, there are few recent Australian studies that address this topical issue at a firm level. A notable exception is a study by Lewis & Jarvie (1997) that show that there is no consistent relationship between the degree of diversification and performance of the 63 Australian companies studied. Their study, because it uses Rumelt’s (1974) methodology for comparability, does not incorporate geographical diversification. The other study by Mangos & OBrien (1999), who studied 50 Australian public firms, establishes an inverse U-shaped curvilinear relationship between international diversification and economic performance. International diversification beyond the low- to moderate levels (where the Foreign Sales to Total Sales Ratio was >35%) tends to depress performance.
Yet there are several reasons to expect multinational corporations (MNCs) to have high levels of profitability. These include returns on intangible assets such as technological know-how and brand names, market power resulting from scope and scale economies, a wider range of investment opportunities, competitive retaliation, lower factor costs, and lower systematic, or beta, risk (Grant, 1987, Rivera, 1991, Kim et al., 1993). In addition to these, the diversity of environments in which a MNC operates “exposes the MNC to multiple stimuli, allows it to develop diverse capabilities, and provides it with a broader learning opportunity than is available to a purely domestic firm” (Bartlett & Ghoshal, 1995).
Research currently being undertaken by the authors will attempt to fill in the gap in the body of Australian strategic management research by answering the following questions:
- How the internationalisation of Australian public companies influenced their performance, and
- Whether Australian companies have indeed increased shareholder value as a result of internationalisation.
3.Methodological issues
General methodological problems
In this research, significant methodological problems are confronted. Of these, difficulties associated with the collection of firm-level data in Australia are the most pressing.
One of the obvious reasons for the reported conflicting results is lack of a robust methodology to measure the degree of internationalisation. Hence the difficulty in comparability of results as different studies use different measures of internationalisation, such as the ratio of Foreign Sales to Total Sales, Foreign Assets to Total Assets and Overseas Subsidiaries to Total Subsidiaries. These difficulties are compounded by the use of different measures of performance, such as Return on Sales, Return on Assets and Rate of Return.
Grant et al. (1988) argue that findings on the relationship between diversity and profitability are highly susceptible to the choice of profitability measures, time period, control variables, choice of data and method of analysis. Other authors (Hoskisson et al., 1993) also attribute much of the confusion regarding the relationship between internationalisation and performance to measurement deficiencies.
Another factor that may account for the conflicting findings is the composition of the sample, whereby the sampling frame is heavily biased toward companies in a single industry, rather than being cross-sectional. The differences in the time frame may also have a significant influence, reflecting either a stable or uncertain economic environment under which a particular study was undertaken. A further difficulty is associated with collecting and interpreting data on the firm’s internationalisation strategies and performance due to confidentiality and differences in management accounting practices across firms and countries (Woodcock et al., 1994). Lastly, as noted by Varadarajan & Ramanujam (1987) and Bartlett & Ghoshal (1989), managers’ ability to handle diversity may have undergone significant changes as a result from their experience with diversification.
Tallman & Li (1996) suggest that further research should be carried using new larger samples and different approaches. In their view, a longitudinal approach might better capture the complexities of the firms’ organisational structures, which cannot be studied using secondary data.
International strategy research has traditionally been based on the dichotomy between case studies, or fine-grained methodologies, and database surveys (such as COMPUSTAT and PIMS), or coarse-grained methodologies (Harrigan, 1983). While fine-grained methodologies lack generalisibility, replicability and statistical rigour, they have the benefits of attention to important details and access to multiple viewpoints. On the other hand, coarse-grained methodologies are generalisible, comparable and often suited to cross-sectional studies of the market. However, they fail to incorporate the nuances in a firm’s strategy.
Harrigan (1983) suggests that the gap should be bridged by devising a hybrid, or ‘medium-grained’, methodology. This methodology is characterised by multiple sites, multiple data sources (eg. published data, databases, archival materials, field interviews), and intricate sample designs (incorporating testable hypothesis into sampling design). This approach is also consistent with Hill's (1994) view that the diversification-performance research needs to be undertaken within organisations and use internal company data.
Measures of geographical diversification
The measurement of diversification per se is not straightforward. The measurement of geographical diversification is even less straightforward, which is evidenced by terminological confusion and lack of a single accepted measure of internationalisation. Thus, at least 7 definitions of geographical diversification can be established from the literature:
- International diversity having two dimensions, multinationality and country scope (Tallman & Li, 1996);
- International involvement (Johanson & Valhne, 1977, Rivera, 1991);
- Degree of internationalisation, or DOI (Grant, 1987 and Sullivan, 1994);
- Degree of multinationality (Rugman, 1983) and multinationality (Gomes & Ramaswamy, 1999);
- Degree of multinational diversity (Grant, 1987) and international diversity (Geringer et al., 2000);
- Global market diversification measure (Kim, 1989); and
- Geographic scope (Delios & Beamish, 1999).
These measures are operationalised by different indices, while different authors may have their own interpretation of a single index. Furthermore, while some authors adhere to the dichotomy of the domestic versus foreign markets, other segment global markets into several zones. These indices are summarised in Table 1 below.
Measures of performance
Measurement of performance also presents significant challenges. In the majority of previous studies, two types of financial performance measures are used:
- accounting-based performance measures, which are most common, and
- market-based performance measures, which are becoming increasingly popular.
As Hoskisson et al. (1993) observe, accounting and market-based measures represent different approaches to performance. Nevertheless, there is sufficient evidence to suggest that past performance as expressed by accounting-based measures is a good predictor of future market performance suggesting a close correlation between accounting-based and market-based measures (Keats & Hitt, 1988). Several authors (eg. Grant, 1987; Kim et al., 1993, Habib & Victor, 1991) defend the use of accounting measures on the premise that managers and external analysts use these as a measure of efficiency of top managers. On the other hand, market-based measures more directly measure the performance of a firm in terms of returns to shareholders (Lewis & Jarvie, 1997). Dubofsky & Varadarajan (1987) argue that market-performance measures are positively correlated with each other but are negatively correlated with the accounting measures. Lewis & Minchev (2000) suggest that economic value added (EVA) may be a more appropriate measure of performance. This is because the underlying motive for overseas expansion, at least for a proportion of firms, is the increase in the amount of profits, not profitability per se. The reason EVA is not widely used in diversification research is that it is very time-consuming to assemble as well as difficult to access.
Table 1 – Summary of geographical diversification and performance measures
Study / Measure of geographical diversification / Measure of performanceRugman (1972) /
- FSTS, where FS includes exports plus sales of subsidiaries
- Number of countries
K=value of shareholders’ equity, or net worth
Kim & Lyn (1986) /
- FSTS
- No of foreign affiliates
Michel & Shaked
(1986) / FSTS / Rate of return using Sharpe, Treynor and Jensen measures
Grant (1987) / FSTS / RONA, profit, ROE, ROS
Geringer, Beamish & DaCosta (1989) / FSTS / ROS, ROA
Markusen (1989) / Stong correlation between multinationality and the following proxies for intangible assets:
- R&D expenditure
- Adverstising expenditure
- Life span of the firm
- Accumulated production experience
Kim (1989) and Kim et al. (1993) /
- Unrelated diversification (2-digit SIC codes)
- Global market diversification (dispersion of operations across 7 zones)
- Global related diversification (4-digit SIC-codes)
Habib & Victor (1991) /
- Product/service diversity (2-digit SIC codes)
- FSTS
Hoskisson et al. (1993) /
- ROA, ROE, ROS
- Sharpe, Treynor, Jensen
- No of employees, leverage, R&D intensity
Sullivan (1994) / DOI=FSTS+FATA+OSTS+PDIO+TMIE
Olusoga (1993) /
- FSTS
- Market concentration
- Market diversification
Sambharya (1995) / Multiple measures:
- Wrigley/Rumelt
- Entropy SIC-based
- NSD and BSD
- FSTS
- FATA
- FSUBS
Tallman & Li (1996) /
- FSTS
- No of countries
Hitt, Hoskisson & Kim (1997) / Entropy measure (FS weighted by a global market region)
Product diversification: SIC-based / ROA, ROS and ROE
Qian (1997) / International diversification: entropy measure (subsidiaries) / ROA, ROE
Gomes & Ramaswamy (1999) / Composite index incorporating:
- FSTS
- FATA
- Country scope
Ratio of operating cost to sales
Mangos & OBrien (1999) / FSTS
Geographical scope (No of geographical segments) / ROA
Delios & Beamish (1999) / Geographic scope:
- No of FDIs
- No of countries in which FDIs occurred
Geringer et al. (2000) / FSTS
ESTS / ROA, ROS, sales
Explanation of terms:
BSDBroad Spectrum Diversification / PDIOPsychic DistanceESTSExport Sales to Total Sales / ROAReturn on Assets
FDIForeign Direct Investment / ROEReturn of Equity
FSTSForeign Sates to Total Sales / RONA Return on Net Assets
FATAForeign Assets to Total Assets / ROSReturn on Sales
NSDNarrow Spectrum Diversification / SDStandard Deviation
OSTSOverseas Subsidiaries to Total Subsidiaries / TMIETop Management’s International Experience
Most of the measures of internationalisation in the literature are unidimensional and operationalised primarily by the FSTS ratio. Therefore, two studies, by Sullivan (1994) and Gomes & Ramaswamy (1999), which use multi-dimensional indices, warrant special attention. Sullivan's multi-factor measure of DOI incorporates three attributes – performance, structural and attitudinal – and is operationalised via five measures: FSTS, FATA, OSTS (ratio of overseas subsidiaries to total subsidiaries), PDIO (psychic distance)[2] and TMIE (top management’s international experience). Critics of Sullivan’s DOI index (Ramaswamy, Kroeck & Renforth, 1996) correctly point out that it focuses on the later stages of internationalisation, i.e. foreign direct investment. A further methodological deficiency is the use of FATA: many authors agree that asset valuation may differ depending on the accounting practices in host countries. Ramaswamy, Kroeck & Renforth (1996) are particularly critical of the attitudinal dimension (PDIO and TMIE). Indeed, while exploring the role of psychic distance in the internationalisation of Australian firms, Fletcher (1997) establishes that managers perceive New Zealand as seven times more distant than the USA (indices of 0.7 and 0.1, respectively), which is a somewhat counterintuitive result.
The recent study by Gomes & Ramaswamy (1999) uses a different index of multinationality, incorporating sales, assets and countries of operation and thus capturing different facets of foreign involvement.
Sectoral composition of the study
The majority of the previous studies explore the internationalisation-performance paradigm for the manufacturing sector. Services (with the exception of banking and financial services) are less studied despite the growing demand for services worldwide coupled with strong financial performance of leading service firms[3].
As Enderwick (1989) suggests, there are marked differences between service and non-service companies, including degrees of internationalisation, growth strategies, size and specialisation. According to Dunning (1993), most service firms (with the exception of finance, investment banking, up-market hotels and some types of consultancy) do not need to pursue internationalisation strategies to compete. Other researchers (Ellis & Williams, 1995), although generally agreeing with this proposition, note that services are likely to follow manufacturing in becoming increasingly subject to international competition.
Dunning (1993) acknowledges the following differences in internationalisation between the service and other sectors:
- Services are less geographically concentrated (as contrasted, for example, with the oil and minerals industries);
- Service firms have low foreign production ratio due to lower tradeability of services; and
- Since many services are not capital-intensive, assets are not an appropriate measure of internationalisation. Internationalisation is thus better captured by sales, net output or employment.
Research into internationalisation strategies of resources companies is even more scarce.Most of prior Australian studies on internationalisation are narrowly focused in that they examine manufacturing firms, thus excluding service and resource-based industries. Given the importance of resources[4] and services in the Australian economy, the emphasis on manufacturing could be misleading in terms of the overall experience of internationalisation of Australian companies.
Operationalisation of a firm's internationalisation strategy
In the absence of a unified methodology for measuring the degree of internationalisation, the variables listed in Table 2 are believed to adequately reflect the complexity of the internationalisation construct. These are established from a comprehensive review of the relevant literature and the authors’ own judgement. The choice of measures is based on the premise that the internationalisation process can be best conceptualised when decomposed into discrete stages corresponding to the following internationalisation strategies:
- Exporting;
- Licensing/franchising;
- Joint venturing; and
- Establishment of a wholly owned subsidiary.
Most of the empirical evidence, and hence the variables used to operationalise firms’ international involvement, tend to emphasise strategies requiring the establishment of foreign subsidiaries. Therefore, it is critical to include those variables that address other strategic options, such as joint venturing and licensing/franchising.
Furthermore, a single measure of internationalisation, such as the widely used Foreign Sales to Total Sales, is intrinsically unreliable. It is thus worthwhile to include other indices, such as Foreign Assets to Total Assets, Number of Overseas Subsidiaries to Total Subsidiaries and Number of Overseas Employees to Total Employees.
Table 2 - Operationalisation of a firm’s internationalisation strategies
Variables / Reason for selection- Exports to a host country and from a host country by an overseas subsidiary
- Foreign sales and total sales, distribution of foreign sales by country
- Profits (EBIT) by country
- Foreign assets and total assets, distribution of foreign assets by country
- Number of employees overseas and total employees
- Number of overseas subsidiaries and total subsidiaries
- Countries in which a company has subsidiaries
- R&D overseas and R&D total
- Licence/franchise fees paid and received by country
- Number of franchisees/licencees
- Number of joint ventures by country
- Joint ventures’ sales, and asset commitments by country
If it were possible to get access to these data, then multiple measures of firm internationalisation could be factor analysed. There is no a priori expectation that internationalisation will be a uni-dimensional construct. The practical problem, however, is more of gathering data than of analysis.