Agribusiness Review - Vol. 11 - 2003
Those Most Likely: Identifying Food Processing Firms With Export Potential
Dr Norman E. Philp
This paper suggests an instrument that can be used by export advisors and other export development practitioners to determine whether or not food and beverage processing firms that have not yet actively exported have, in fact, a high probability of ever doing so.
The instrument is developed from an empirically based, logistic regression model relating to the management and firm-specific determinants of a firm's export orientation. This model was able to predict the probability of a firm being an “active exporter” with an accuracy of over 84%.
The major management attitudes and attributes that appeared to strongly influence whether or not a firm became an active exporter were its management's willingness to commit resources to export development, their attitudinal commitment to export, their recognition of the significance of product price in the firm’s market competitiveness, the firm’s access to export-specific management skills, and whether or not the manager was tertiary educated.
Four of these constructs were then used to develop the suggested instrument.
Research published on the export behaviour of the firm has been rather long established and voluminous in quantity. Despite this, a definitive, multi-disciplinary, theoretical model of a firm's export behaviour has remained elusive. One might ask, however, whether this academically stimulating quest has always been in the best interests of export practitioners, i.e. the many private and government-employed export advisors and export managers, who await the new knowledge and have the task of applying it. It would even appear that export practitioners have developed their own body of experiential research, literature and guidelines (see Austrade and TradeNZ publications; undated) rather than to further rely on the output of academic research.
The current paper is a modest attempt to apply the results of a multivariate model of the export orientation of a sample of small to medium-sized food processing firms, to the development of an instrument that might be useful to the practitioner. The model is built upon constructs and variables that are now familiar within the extant literature. Hopefully, the paper might also contribute towards bridging the gap between academic research and practitioner needs.
The underlying research on the export behaviour of the firm centres on two fundamental questions;
a) what leads some firms and not others to seek out and export to foreign markets?
b) why are some exporting firms more successful in their export endeavours than others?
As the first question, basically, has a dichotomous dependent variable (i.e. a firm either is or is not an active exporter at a particular point in time) and, by necessity, involves a number of firms that have yet to establish an export performance record. It is thus different from the second question and this has caused some confusion in the literature. The later question calls for an accurate and composite interval measurement of export "success" for which the explanatory variables then seek to provide an answer.
Fortunately, the voluminous literature associated with the internationalisation of the business firm has no need for yet another detailed literature review. There have been, over the last 25 years or so, a number of very competent and published literature reviews that have analysed the few nascent theories associated with the topic area. They have also reported the diverse and oft conflicting findings of numerous empirical studies conducted either with or without any convincing theoretical base, and which sometimes have attempted their own integrative models to guide future research.
Several such review studies will be considered below. The first, by Bilkey (1978) covered some 43 studies on the export behaviour of firms conducted in 11 countries and accounts for the earliest of all known research in the field. The second, by Aaby and Slater (1989), reviewed a further 55 studies conducted between 1978-88 and proposed a model integrating the relationships between firm characteristics, firm competencies and strategy and their impact on export performance.
Such a model has had a strong influence on the empirical research conducted into the determinants of export performance over the last decade of the twentieth century. Another, and the most recent review, by Leonidou, Katsikeas and Piercy (1998) brings us almost up-to- date through a systematic analysis of 46 studies on the managerial characteristics that influence the export behaviour of the firm (although there is some overlap with the two previous reviews cited above).
Further work by Miesenbock (1988) and by Chetty and Hamilton (1993) are among the many others that serve as a useful review of this research, and the research associated with Cavusgil (1976, 1979, 1980, 1981, 1987, 1994) is, of course, fundamental.
Unfortunately, and despite the considerable time that has elapsed since the earliest research in this field, the most recent review by Leonidou, et al, (1998) concludes that this stream of research is still in its "exploratory stage of development" and that it lacks a "well-defined theoretical framework". Almost identical observations were made by Bilkey (1978) and Aaby and Slater (1989), decades before.
The internal, management-related and otherwise firm-specific influences on the decision to export or on export performance may never be adequately quarantined from the external environment in which managers and their firms operate. To seek out a perfectly general and definitive model of the firm's internationalisation process may indeed take us well beyond the reasonable boundaries set by all the previous international management and marketing researchers. It may take us beyond even the boundaries outlined by Porter (1990) and into the very heart of the underlying causes of the cultural diversity of nations.
The review papers will be examined, not so much for the insights they provided for the selection of the managerial and firm-specific variables used in the accompanying model of export orientation, but for the extent to which their conclusions and recommendations were (or can be) directed towards assisting export advisors and managers. Although we might claim that all research findings are useful to practitioners, not all appear to be written with practitioner interests in mind.
Next, a logistic regression model of export orientation developed by the author is reported. This model seeks to distinguish a "non-exporter" from an "active exporter" (viz. one who has made an overt attempt to find an export market and has already made some export sales) on the basis of a number of managerially- related and firm-specific variables drawn from previous empirical studies.
On the basis of only five final variables (three of which were multi-item constructs/factors), the model was able to reliably differentiate an active exporter from a non-exporter with an accuracy of over 84%. However, it is the attempt to derive operationally useful outcomes from this research and from its results that the current paper wishes to highlight.
The model allows us to identify firms that currently are not exporting, even though the managerial attitudes and attributes that the research indicates they possess, suggests that they should be. An attempt is then made to develop an instrument that business and export advisors can use with some confidence to assess the probability that a firm, currently not exporting, may later develop into an active exporter. As such, it may well help a business/export advisor or an advisory body operating with limited resources to utilize these resources more effectively.
Throughout nearly all of the literature, the critical role played by management in initiating and succeeding in export activity is highlighted. In concluding their review, Aaby and Slater (1989) specifically identify that firms having a management firmly committed to export, having good management systems and plans and some export experience are likely to be successful. They further found that management which had an international vision, had favourable perceptions and attitudes to export, was willing to take risks and had the capability to engage positively in export activities was more likely to lead a firm to export success.
Even a decade earlier, Bilkey (1978) identified, from the 43 papers he reviewed, the following managerial attributes as being correlated with successful export initiation: management with an interest in and enthusiasm about export, with a positive foreign orientation, management which had studied a foreign language, had lived abroad, had confidence in the firm's competitive advantage (including price advantages), and which had positive but realistic attitudes to export's diverse advantages. His conclusion that "… the quality of management probably is the greatest single determinant of a firm's export success" (1978; p.43) hasn't been seriously challenged in the hundreds of studies that have followed.
In yet another critical review of the research, Cavusgil and Naor (1987), again draw attention to such factors as the type and extent of the manager's education, his/her knowledge of foreign languages, international orientation, risk-taking preferences, growth aspirations and general "open mindedness" to things foreign, as potential discriminators.
Management's preparedness to commit resources to exporting, for gathering foreign market information, making overseas visits, learning about export procedures and financing and management's perceptions of the overall attractiveness of exporting to the firm, were also cited from the works that they reviewed.
However, the firm's unique advantages with respect to its products, its product's price competitiveness, its technological orientation, marketing capability, resource availability and size were thought to be the most significant discriminators. Firm size has received considerable attention throughout the literature and has been extensively reviewed by Bonaccorsi (1992) and Calof (1994). These writers conclude that there is little association between the size of a firm and its export orientation. It is a variable, however, that one would want to include in any empirical model of the export behaviour of the firm.
The most recent review of the published research by Leonidou, Katsikeas and Piercy (1998) focuses very specifically on the managerial factors that facilitate or inhibit exporting and on the particular export dimensions that are influenced by these managerial factors. It also suggests some new lines of inquiry for this type of research to follow. Some 26 managerial characteristics with a possible effect on exporting were identified, although some were recognised as being very similar to others. However, only the factors "quality and dynamism" and the "innovative" conduct of the decision-maker add to those factors already mentioned above.
So much of this research has been (almost by definition) of an academic nature and conducted primarily for an academic, research-oriented audience. Our failed search for the "holy grail", viz. for a fully integrated and general model to be tested under conditions of almost absolute conceptual, methodological and statistical purity, has absorbed most of the very substantial effort and energy that has been directed to this line of inquiry for almost forty years.
Beside some brief advice directed towards those responsible for "public policy", very few of the research papers published in this area of inquiry appear to focus on how their results might be used to assist the field-based business and export advisors. Yet, it is these, perhaps more than anyone else, who are the practitioners most able to stimulate and assist existing managers in their export endeavours.
If it wasn't for the very admirable and encouraging fact that the results of this research are now beginning to "filter down" through the better college text books to the export managers of the next generation, we might conclude that much of the work really has been in vain.
Of the major literature reviews cited above, Aaby and Slater (1989) and Leonidou, et al,(1998) do not attempt to give any practitioner advice at all. Cavusgil and Naor (1987) conclude with implications for public policy and recommend that export promotion efforts can be more fruitful if firms selected for assistance are carefully segmented and that export assistance programs should vary in nature to provide the types of assistance best suited for a particular group of firms. They further recommend that "government agencies" through conferences and well-targeted promotional campaigns should attempt to enhance managers' expectations of exporting and educate them about government assistance programs designed to reduce the risks of exporting.
Only Bilkey (1978), in the earliest of all reviews, discussed specific ways in which the research might be able to assist the practitioner. He reports quite favourably of research designed to "profile" exporting and non-exporting firms as a means of identifying potential exporters among firms that are not yet exporting. Furthermore, he recommends that properly developed export profiles could be used by government export promotion agencies, banks, export agents, etc, to identify non-exporters with high export potential. Such instruments would allow the limited resources available for export promotion to be concentrated on high export potential firms. A similar idea will now be developed in the following sections of this paper.
An empirical model, focussing primarily on the management-oriented and firm- specific variables that the literature had suggested as being able to distinguish a firm that had actively engaged in export activity from one that had not, was then developed. Details of this model, the study sample, the research methodology, together with a defence of all analytical procedures can be found in Philp (1997, espec. Chapter 6).
Basically, a questionnaire was administered to approximately 700 chief executives of food and beverage- processing firms with products thought to have some export potential. All firms were located in the non-metropolitan (regional) areas of New South Wales and Victoria, Australia. Some 244 eligible and complete responses were then collected and analysed.
The questionnaire used ( predominantly) a multi-item, Likert-scale approach to collect data on management's commitment to exporting, management's perceptions relating to the benefits arising from export and to export specific risk, and management's perceptions of various firm-specific advantages relating to the firm's overall competitiveness. In addition, it sought information on the firm's access to export-specific management skills, the firm's size and 8 characteristics or attributes relating to the managers themselves (e.g. age, education level, overseas travel and work experience, foreign language fluency, etc).
Where appropriate, factor analysis (principal-axis method with eigenvalues set at 1) was then used to reduce the large number of individual items (variables) into a smaller number of factors or constructs. Again, factor structures and measures of construct validity, the items pertaining to the derived factors and the extent to which they loaded onto the various factors are detailed in Philp (1997; Chapters 3 and 6).
Items robust enough to stand alone outside the derived factors, and other variables (often dummy variables) relating to management characteristics and firm size, were also entered into the model. Where reliable factors were identified, extracted and confirmed, the resultant regression factor scores (RFS) became the explanatory variables in the model.
Because of the dichotomous nature of the dependent variable (an "active exporter"=1, versus a "non-exporter"= 0), a logistic regression model was formulated and then estimated using the SPSS for MS WINDOWS LOGISTIC REGRESSION (Release 6.0) statistical program. A desirable feature of logistic regression is that we can estimate directly the probability that a firm will be "an active exporter" from the parameter estimates of the model.
The following, constitute the explanatory variables in this logistic regression model. Kaiser-Meyer-Olkin (KMO) measures of sampling adequacy and Cronbach a scores of the factors/constructs are also indicated:
The results of the estimation of the model containing all of the above variables are given in Table 1. The full model was able to predict with an accuracy of nearly 87% (average) that a firm was either an active exporter or a non-exporter, against only 51.5% for a model containing nothing but the constant.
In determining which particular variables contributed most to the prediction, several statistics can be used. The Wald statistic and its associated p values indicates if the coefficients of the explanatory variables are significantly different from 0. The R statistic indicates the partial correlation between the dependent variable and each independent variable.
The higher the R values, the greater the contribution of the variable to the model. However, in drawing out the particular variables that the model indicates as making the most significant contribution, we rely on the likelihood-ratio test (-2LogLR). The model was then subjected to a process of backward stepwise selection using the criteria that the significance level of the LogLR of any variable remaining in the model be kept at better than 0.10.
The results of this reduced model are given in Table 2. The accompanying histogram of predicted probabilities has a good concentration of correct estimates at each end of the histogram, again indicating that the model fits the data particularly well.
As shown in Table 2 and in order of the relative contribution of the variables to the model, an active exporter could be distinguished from a non-exporter on the basis of:
Significantly, this research again indicated that the size of the firm was a poor indicator of whether or not it was an active exporter.
The five variables listed above were able to predict the probability of a firm being an active exporter (as defined) as against being a non-exporter, with an average accuracy exceeding 84%.
Although the results indicated have been derived from a fairly powerful model, using a sample that has enabled many factors associated with the external environment of the firm to be held reasonably constant, they only confirm much of what has been reported in other research over the last forty years. The intention of this paper is to take these finding one step further and to indicate how they might be used to assist practitioners associated with export development.
The research indicates clearly some five managerial attitudes and attributes that allow us to predict with considerable confidence that a firm is, or is not, an "active exporter". The intention is to incorporate this information into an instrument that a trade consultant/advisor might then use to pick out firms that are not currently exporting, but which have the greatest probability of doing so in the future.
This might then allow trade promotional bodies with limited time and resources to concentrate their time and efforts on firms that are most likely to enhance the national export effort, rather than persist with firms that appear to lack the necessary managerial attitudes and attributes for successful export initiation.
First, the various factors found to be most significant in the managerial model were simplified by re-running reliability tests on the items included in each factor and eliminating any item whose impact was considered to be marginal.
Then, each item set (factor) was measured as a simple aggregated scale, based on the 5 point Likert response given for each item. TERTED was dropped because it could only be measured on a Yes/No (1/0) basis. The logit model was then re-run using these simple aggregated scales as the four independent variables. Such a model was still found capable of estimating the probability that a firm would or would not be an exporter with an average accuracy of around 84%.
Next, the mean, mode and median aggregated score for each of these four variables (item sets), for all the firms that were predicted in the model to be active exporters with a probability of 0.75 or higher, and which were also exporters by observation, were determined. Also calculated were the same mean, mode and median scores for the firms predicted to be non-exporters (and which were) with a probability of being an exporter of 0.25 or less.
The liberty was also taken to modify the wording of the items in the instrument to better reflect the fact that the instrument would be directed towards a firm that we know has not yet exported, and wish to get some indication that it might. An additional item was also included for the price competitiveness variable for much the same reason.
The suggested instrument is presented as Table 3. No aggregated total score for the whole instrument is included, only aggregated scores for each of the four discriminating factors (i.e. resource commitment, attitudinal commitment, etc).
The former would require further assumptions about the relative "weights" of the four factors. Aggregate scores have been suggested for each separate factor scale (item set). If these score are attained then the research indicates that the firm has a high probability (at least 50-50 and supposedly a 75% chance) of emerging as an active exporter. Such scores are recorded in the square brackets [ ].
The score that suggest a firm has a very low probability of becoming an active exporter are, likewise, indicated in round brackets, ( ). A firm that is able to score in the suggested range for each of the four item sets thus has a very good chance of emerging as an active and, hopefully, successful exporter.
Table 3: An Instrument for Determining Firms with Export Potential
Please indicate the extent to which you agree or disagree with the following statements by circling the most appropriate response.
Scale: 1=Strongly Disagree; 2=Disagree; 3=Neither Agree or Disagree; 4=Agree; 5=Strongly Agree.
A. Resource Commitment
[20 or >] (10 or <) SUM out of 25
B. Attitudinal Commitment
[13 or >] (7 or <) out of 15 SUM out of 15
C. Lack of Export Specific Skills
[12 or <] (16 or >) SUM out of 20
D. Price Competitiveness
[4 or > each] (3 or < each) SUM out of 10
The accuracy of this instrument should not be overestimated. It has not yet been tested and normed in the field and we do not know the extent to which it is sample specific. However, if we are to await the final and definitive work on why some firms are able to develop into active exporters whilst other apparently cannot, we may never make the contribution to applied knowledge that underlies much of the justification for the research we do.
One must caution that there may well be firms that score outside the suggested limits on each set, and which might, eventually, also become active and successful exporters. However, by concentrating resources on those firms that have all or most of the critical managerial attitudes and attributes of active exporters, a practitioner can seek to determine what other factors are currently holding them back.
For instance, a firm may lack finance or knowledge of financial assistance available to exporters; it may have a risk concern that could be handled by EFIC; it may have difficulty in understanding or complying with local or overseas health/ packaging/ promotional regulations; or, any other of a myriad of specific or perceived constraints which an advisor can help them tackle.
Although such constraints may seem no less inhibitive than a lack of commitment, a lack of export-specific management skills, or inadequate appreciation of the importance of price competitiveness in export activity, the underlying research indicates that such constraints are more easily overcome by those firms that have already actively exported.
Aaby, N. & Slater, S. (1989). Management influences on export performance: A review of the empirical literature 1978-88. International Marketing Review, 6(4), pp. 7-26. Andersen, O. (1993). On the internationalisation process of firms: A critical analysis. Journal of International Business Studies, 24(2), pp. 209-231.
Andersen, O. (1993). On the internationalisation process of firms: A critical analysis. Journal of International Business Studies, 24(2), pp. 209-231.
Austrade. (undated) A Guide to Exporting, Export Update,
Austrade. (undated) Are You Ready to Export? (single sheet checklist)
Barrett, N. & Wilkinson, I. (1985). Export stimulation: A segmentation study of the exporting problems of Australian manufacturing firms. European Journal of Marketing, 19(2), pp. 53-72.
Bilkey, W. (1978). An attempted integration of the literature on the export behaviour of firms. Journal of International Business Studies, 9(Spring-Summer), pp. 33-46.
Bonaccorsi, A. (1992). On the relationship between firm size and export intensity. Journal of International Business Studies , 23(4), pp. 605-635.
Calof, J. L. (1994). The relationship between firm size and export behaviour revisited. Journal of International Business Studies, 25(2), pp. 367-387.
Cavusgil, S. & Naor, J. (1987). Firm and management characteristics as discriminators of export marketing activity. Journal of Business Research, 15(3), pp. 221-235.
Cavusgil, S. & Nevin, J. (1981). Internal determinants of export marketing behaviour: An empirical investigation. Journal of Marketing Research, 18(February), pp. 114-119.
Cavusgil, S. & Zou, S. (1994). Marketing strategy-performance relationship: An investigation of the empirical link in export market ventures. Journal of Marketing, 58(1), pp. 1-21.
Cavusgil, S. (1976). Organizational determinants of firms' export behaviour: An empirical analysis. PhD dissertation, The University of Wisconsin, Madison, Wisconsin.
Cavusgil, S. (1980). On the internationalisation process of firms. European Research, 8, pp. 273-281.
Cavusgil, S. (1994). A quiet revolution in Australian exporters. Marketing News, 28(11), pp. 18.
Cavusgil, S. T., Bilkey, W. J. & Tesar, G. (1979). A note on the export behaviour of firms: Exporter profiles. Journal of International Business Studies, 10(Spring - Summer), pp. 91-97.
Chetty, S. & Hamilton, R. (1993). Firm-level determinants of export performance: A meta-analysis. International Marketing Review, 10(3), pp. 26-34.
Johanson, J. & Wiedersheim-Paul, F. (1975). The internationalisation of the firm: Four Swedish cases. Journal of Management Studies, October, pp. 305-322.
Leonidou,L.,Katsikeas, C. & Piercy, N. (1998). Identifying managerial influences on exporting: Past research and future directions. Journal of International Marketing, 6 (2), pp.74-102.
Leonidou, L. & Katsikeas, C. (1996). The export development process: An integrative review of empirical models. Journal of International Business Studies, 27(3), pp. 517-551.
Miesenbock, K. (1988). Small business and internationalisation: A literature review. International Small Business Journal, 6(1), pp. 42-61.
Philp, N. (1998). The export propensity of the very small enterprise (vse). International Small Business Journal, 16(4), pp. 79-93.
Philp, N. (1997). Why some firms successfully export processed food and beverages. Report for the Rural Industries Research & Development Corporation, RIRDC Research Paper 97/33, 146pp.
Porter, M. E. (1990). The Competitive Advantage of Nations. New York, Free Press.
Tradenz. (undated) Guide to Exporting.