Multiple linear regression analysis of relationship between business results and strategy.

Pasic, Mugdim ; Sunje, Aziz ; Bijelonja, Izet 等

Abstract: In this paper results of the research of impact of strategy on business results are presented. In the developed mathematical model business result is defined as dependent variable, while eight strategic variables are defined as independent variables. The research sample consists of organizations from Bosnia and Herzegovina from wood processing industry that was observed as the reference industrial branch for this research.

Key words: Business result, strategy, multiple linear regression, variables.

1. INTRODUCTION

This research was based on a posture that related relationship exists between business result, as a dependent variable, and strategy as independent variable. As such, the construction of statistical model like multiple linear regression analysis can serve as a tool to verify or otherwise disprove the presence of relationships between interacting variables.

Studies in this areas show relationship between dependent and independent variables. Study (Gautam et al., 2003) shows that, in some circumstances, adopting the effectiveness of business processes as a dependent variable may be more appropriate than adopting overall firm performance as a dependent variable. Results are consistent with resource-based expectations, and they show that distinctive advantages observable at the process level are not necessarily reflected in firm level performance. Creation of a new business-model is considered to be one of the important factors of market growth (Yamaguchi et al. 2001). The creation of a new business-model means an appearance of new demand. Study (Evans, 1987) on some aspects of firm dynamics finds that firm growth, the variability of firm growth, and the probability that a firm will fail decrease with firm age. It also finds that a firm growth decreases at a diminishing rate with firm size even after controlling for the exit of slow-growing firms from the sample. Empirical research (Kang & Sorensen, 1999) based on the approach that shareholders are homogenous and that their influence on firm performance is directly proportional to the percentage of equity they hold has failed to produce definitive evidence.

Research (Pasic et al. 2006) examines influence of degree of wood processing on business result using regression method. The model is tested using analysis of variance and coefficient of determination. Study (Pasic et al. 2006) presents influence of type organization, ownership and number of employees on business result. The most influencing control variable is Type of Ownership, followed by Number of Employees and Type of Organization, and proper managing of these variables can significantly improve Business Result.

Data used in mathematical analysis for this research were obtained through empirical methodology of polling organizations using carefully designated Questionnaire. Sample data consists of 50 organizations from wood processing industry in Bosnia and Herzegovina. The questionnaire was designed in such a way that, beside general information about a company, it provides information about company's strategy and organization.

For obtaining information about company's strategy Design School Model was used (Mintzberg et al., 1998), and for information about company's performance Balanced Scorecard Model was used (Kaplan & Norton,. 1996). The Balanced Scorecard suggests that the organization is viewed from four perspectives, and to develop metrics, collect data and analyze it relative to each of these perspectives: the Financial Perspective, the Customer Perspective, the Business Process Perspective and the Learning and Growth Perspective. The View of the Design School Model of a strategy process, defined as one among ten schools of strategic thought, is based on strategy formation as a process of conception.

It is well known that strategy is a qualitative term. Proper measurement and the way of expressing a strategy in terms of numbers, is something challenging. Our approach to this problem was to make a questionnaire that will quantify examinee's answers. Answers to all questions are designed by Lickert's scale from 1 to 5 as depicted in Table 1.

2. PROBLEM APPROACH AND MATHEMATICAL MODEL

With regard to our intention to develop a model that will show how strategy impacts business result, we found that it is practical and common approach to use regression as a tool to describe and define this relationship.

Furthermore, as we consider business result as dependent variable, and eight strategic variables as independent explanatory variables, it is suitable to use multiple linear regression to predict the value of dependent variable. Variables, its symbols, type and the units of measurement are presented in Table 1.

After the multiple linear regression model is constructed, statistical tools such as ANOVA analysis, coefficient of determination, [R.sup.2], F-test, and t-test will be applied to prove or disapprove existence of a relationship. Multiple linear regression model is given by the following equation:

BR = [b.sub.0] + [b.sub.1][x.sub.1] + [b.sub.2] [x.sub.2] + [b.sub.3][x.sub.3] + [b.sub.4][x.sub.4] + [b.sub.5][x.sub.5] + [b.sub.6][x.sub.6] + [b.sub.7][x.sub.7] + [b.sub.8][x.sub.8] (1)

Where, coefficients from [b.sub.0] to [b.sub.8] represent regression coefficients, and [x.sub.1] to [x.sub.8] represent independent strategic variables, while BR represents dependent variable.

3. RESULTS AND INTERPRETATIONS

In order to get the regression coefficients and to apply statistical analysis to the proposed model we used SPSS 12.

The ANOVA analysis shown in Table 2. shows that a computed value for the F-ratio is 2,472. The corresponding table value for F-ratio is 2,17 at 0,05 level of significance, where degrees of freedom are [df.sub.1]=8 and [df.sub.2]=41. F-ratio and p value show that the linear dependency exists, and that it is significant and valid.

The coefficient of determination [R.sup.2] is 0,325, indicating that regression model explains 32,5 percent of the variability of business result by strategic variables.

Table 3. shows estimated values of regression coefficients, corresponding standard errors, t-value and p value. Results of the t-test indicate that regression coefficients from [b.sub.0] to [b.sub.8] are not statistically significant at 0,05 level of significance (t-value table for sample size greater than 30 is 1,96 at 0,05 level of significance). Comparison with p value brings to the same conclusion. So, the final multiple linear regression equation showing linear relationship between business result and strategy is given by:

BR = -169820,7 + 13762,55[x.sub.1] + 14249,39[x.sub.2] + 11424,47[x.sub.3] + 2046,80[x.sub.4] -6303,99[x.sub.5] + 34443,91[x.sub.6] + 26932,6[x.sub.7] -31778,29[x.sub.8]

All coefficients in equation (2) are obtained from Table 3, while independent variables [x.sub.i] a explained in Table 1. Figure 1. shows scatter plot of actual versus predicted values of dependent variable BR (business result).

[FIGURE 1 OMITTED]

4. CONCLUSION

This research proves that there is a relationship between business result and strategy. Importance and usefulness of this model is based on its simplicity and suitability for general use in predicting company's business result with respect to the strategy, in this industry branch, because the model was constructed without any stratifications and data filtrations against different company's characteristics.

In the future research data should be collected from these and some new companies in this branch, in order to improve, refine and confirm our model. Future plans can also include development of multiple nonlinear regression model which might improve significance of regression coefficients as well as coefficient of determination.

5. REFERENCES

Evans S. D. (1987). The Relationship Between Firm Growth, Size, and Age: Estimates for 100 Manufacturing Industries. Journal of Industrial Economics, Vol. 35, No. 4, (June, 1987) pp. 567-581, ISSN 0022-1821

Gautam R.; Barney J. & Waleed M. (2003). Capabilities, business processes, and competitive advantage: Choosing The Dependent Variable In Empirical Tests Of The Resource-Based View. Strategic Management, Vol. 25, No. 1, (December, 2003) pp 23-37, ISSN 0143-2095

Kang D. L. & Sorensen A. B. (1999). Ownership Organization and Firm Performance. Annual Review of Sociology, Vol. 25, No. 1, (August, 1999), pp. 121-144, ISSN 0360-0572

Kaplan R. S. & Norton D. P., (1996). The Balanced Scorecard, Harvard Business School Press, ISBN 0-87584-651-3, Boston, USA.

Mintzberg H.; Lampel J. & Ahlstrand B., (1998). Strategy Safari: A Guided Tour Through The Wilds of Strategic Management, The Free Press, ISBN 0-684-84743-4, New York, USA

Pasic, M., Sunje, A. & Karic, E. (2006). Regression Analysis of Relationship Between Degree of Wood Processing and Business Result, Annals of DAAAM for 2006 & Proceedings of the 17th International DAAAM Symoposium, Katalinic, B. (Ed.), pp 291-292, ISSN 1726-9679, Vienna, November 2006, DAAAM International, Vienna

Pasic, M., Sunje, A. & Kadric, E. (2006). Analysis of Influence of Type of Organization, Ownership and Number of Employees on Business Result, Proceedings of the 10th International Research/Expert Conference TMT 2006, pp. 501-505, ISBN 9958-617-30-7, Barcelona-Lloret de Mar, September 2006, TMT, Barcelona-Lloret de Mar

Yamaguchi T., Gemba K. & Kodama F. (2001). Quantitative Analysis of Business-Model, Portland International Conference on Management of Engineering and Technology, Vol. 1, No. 1, pp. 1-34, ISBN: 1-890843-06-7, Portland, OR, USA, July-August, 2001, IEEE, Portland

Table 1. Variables definitions and units of measurement.

 Type and unit of
Variable measurement

Business Result BR BAM (KM)
Do You have a clear vision for next [x.sub.1] Ordinal, Lickert's
 five years? scale from 1 to 5
Do You have written mission [x.sub.2] Ordinal, Lickert's
 statement? scale from 1 to 5
How do You grade your knowledge about [x.sub.3] Ordinal, Lickert's
 wood processing industry? scale from 1 to 5
How do You recognize your business [x.sub.4] Ordinal, Lickert's
 chances? scale from 1 to 5
How do you estimate strengths and [x.sub.5] Ordinal, Lickert's
 opportunities of your company? scale from 1 to 5
Do You estimate company's weaknesses [x.sub.6] Ordinal, Lickert's
 and how do You eliminate them? scale from 1 to 5
How do You create competitive [x.sub.7] Ordinal, Lickert's
 advantage? scale from 1 to 5
How is your company oriented against [x.sub.8] Ordinal, Lickert's
 customer segmentation? scale from 1 to 5

Table 2. ANOVA table for multiple linear regression.

 Sum of Mean p
 Df Squares Square F value [R.sup.2]

Regression 8 1,806E+11 2,257E+10 2,472 0,028 0,325
Residual 41 3,743E+11 9,13E+09
Total 49 5,549E+11

Table 3. Multiple linear regression coefficients, t and p value.

 Standard
 Coefficients Error t value p value

[b.sub.0] -169820,7 119108,79 -1,426 0,162
[b.sub.1] 13762,55 16,736,408 0,822 0,416
[b.sub.2] 14249,39 10,717,251 1,33 0,191
[b.sub.3] 11424,47 29,394,226 0,389 0,7
[b.sub.4] 2046,80 18,176,709 0,113 0,911
[b.sub.5] -6303,99 28918,02 -0,218 0,829
[b.sub.6] 34443,91 24,131,162 1,427 0,161
[b.sub.7] 26932,6 21,054,404 1,279 0,208
[b.sub.8] -31778,29 28,468,697 -1,116 0,271