LOGIT REGRESSION ANALYSIS

Where the prior analysis presents information on the type and location of strategic alliances in the data, it does not address the key questions posed in the study: (1) do strategic alliances have a relationship with firm performance?, and if so (2) in what ways do strategic alliances have a relationship with firm performance.

To investigate these questions, a logit regression model was developed based on the sample. Each firm was classified as either a high performer or a low performer based on average annual growth in revenues between the years 1990 and 1994. Forty-three percent of the sample was found to have an average annual growth rate in revenues greater than 50%. This then became the discriminator for determining which firms would be classified as high performers and which firms would be classified as low performers. Because some respondents did not answer questions regarding revenue performance in the questionnaire, the total sample size in the logit regression analysis was reduced to 167. Of these high technology firms, 71 were classified as high performers. When the sample was segmented into the three classifications of firm age where new firms are less than six years of age, young firms are 7-12 years of age, and old firms are above 12 years of age, the number of firms in the sample were 63, 47, and 57 respectively.

Each type and location of strategic alliance was dummy coded based on the presence or absence of a strategic alliance. Hence a firm that has one strategic alliance that is international would receive a code of 1 (presence) as would a firm that has 10. Absence of a strategic alliance in a category received a code of 0. In addition, a measure of total strategic alliances that a high technology firm has was included in the logit regression model. Two measures of strategic alliance diversity based on location and type were created. The measures of diversity was created using the following formula:

D = ( sai2 )/SA2)

where sa is the number of observations or strategic alliances, within the ith category, and SA is the total number of strategic alliances established by a high technology firm across all categories. The maximum score would be 1.0 with a low score less than one but greater than 0. The lower the score the more diverse the strategic alliances a firm has. The higher the score, the more similar in type or location the high technology has in terms of their strategic alliances. Two separate measures were created, one representing the diversity of strategic alliances by location and a second representing the diversity of strategic alliances by type.

Stepwise regression procedures were used. Hence, only those measures that had a significant impact on performance were included in the model. Significance criteria was established at .10 to qualify inclusion in the logit regression model. Table 6 displays the results form the analysis.

 

TABLE 6
Logit Regression
Strategic Alliances As Predictors Of Sales Growth
50% + OVER FIVE YEARS = HIGH

  All Firms
(N=167)
Hi=71
New Firms
(N=63)
Hi=31
Young Firms
(N=47)
Hi=16
Old Firms
(N=58)
Hi=25
Technology License -.68(b) -.87(a)   1.42(b)
Marketing   1.96(c)    
City -.96(a) -2.23(d)   -1.49(b)
United States 1.05(b)      
International   3.29(d)   .70(b)
Total Number   .07(a) .20(b) .06(a)
Diversity in Type     -3.02(c)  
Diversity in Location -2.11(c)     -1.08(b)

Significance: a .1; b .05; c .01; d .001

The preliminary results of the logit regression analysis suggests that in the total sample, technology licenses have a negative relationship with high performance of high technology firms (b = -.68). A negative relationship is also recorded between firms with alliances in the city of Milwaukee and high performance (b = -.96) and a negative relationship where firms were found to have similarity in strategic alliances by location and high performance (b = 2.11). This latter finding suggests that high performing high technology firms are those with strategic alliances that are dispersed across multiple locations rather than being situated in a few locations.

Within subsamples of high technology firms separated by age, we find that separate logit regression models identify measures of strategic alliance that are associated with their performance. For new high technology firms (six years of age or less), strategic alliances that are technology licenses (b = -.87) and firms with strategic alliance partners located in the city of Milwaukee (b = -2.23) are primarily low performers in terms of average annual revenue growth rate. Several types and locations of strategic alliances, on the other hand, appear to be associated with high firm performance. Marketing strategic alliances for new high technology firms (b = 1.96) and strategic alliances that have an international partner (b = 3.29) are associated with high performance. The results also indicate that new high technology firms with greater numbers of strategic alliances are higher performers than those firms that have lesser numbers of strategic alliances (b = .07).

Total number of strategic alliances were also found to be positively associated with performance for those high technology firms classified as young, or between the ages of 7-12 (b = .20). Young high technology firms that have greater diversity of strategic alliances by type also have higher levels of performance than those firms that have similar types of strategic alliances (b = -3.02).

Older high technology firms, those 13 years old or greater, were found to have greater levels of performance when they have technology licenses as a form of strategic alliance (b = 1.42). Other forms of strategic alliances that are positively associated with revenue performance are international (b = .70), total number (b = .06), and diversity in location (b = -1.08). The only significant relationship between strategic alliances and firm performance for older high technology firms was for city (b = -1.49) suggesting that when the strategic alliance partner is located in the same city or the largest city in their state of location, performance is low.

 

IMPLICATIONS

While the analysis of the data is preliminary, there is evidence of several important relationships between strategic alliances and high technology firm performance. Technology licenses as a type of strategic alliance appears to benefit older high technology firms, but not younger ones. Marketing strategic alliances are quite beneficial to new high technology firms though no significant evidence is available on their relationship to young and old high technology firms.

Having a strategic alliance partner in the same city or the largest city of the state in which the firm is located is perhaps not advantageous in regards to performance. For the entire sample, and for firms that are new and old, the relationship was negative. While a causal relationship cannot be established in this instance, the results suggest that high performing high technology firms are not prone to these types of strategic alliances. On the other hand, having a strategic alliance partner located in the United States (outside of the Great Lakes region); and international does associate with higher performance in some cases. Those cases being when the high technology firm is new or old.

Total number of strategic alliances were found to be associated positively with performance for firms that are new, young, and old. Hence, one can conclude that high performing firms are likely to have higher numbers of strategic alliances without consideration for type or location. Having a diverse set of strategic alliance types was found to be beneficial for young high technology firms. Diversity in location was found to be beneficial for all firms in the sample, and in particular for older firms.

Coupled with the descriptive evidence on strategic alliances, it appears that new and old high technology firms are more inclined to achieve benefits from strategic alliances. On the average, they engage in them more frequently and appear to reap benefits. Young firms, or those between the ages of 7 and 12, are inclined to have fewer strategic alliances, though there is evidence that greater numbers of strategic alliances is associated with higher performance.

Speculation may be that new firms, having strong innovative competencies, lack marketing and distribution channels in order to achieve higher levels of sales and therefore performance. In fact, partnering with technology licenses may be disadvantageous. On the other hand, older high technology firms may have competencies in marketing and distribution channels, but lack innovative competencies. Indeed, in examining responses from CEOs of older high technology firms regarding the levels of innovation in their firm, we find that older firms are much less innovative than new or younger firms (see Table 1). While the data does not identify the age of the strategic partner, it may be true that older firms are seeking new firms to acquire innovation whereas new firms are seeking older firms to acquire marketing and distribution channels.

An important question then is why do young firms have less activity in the establishment of strategic alliances. It may be that as new firms grow and age, they are better equipped to infuse capital into the firm to internally develop marketing and distribution channels. This would make the need for strategic alliances less desirable as it would allow the firm to obtain greater controls over its operations and products. However, the data does not have the capability to affirm this conclusion.

 

Previous Page | Main Menu | Next Page


1997 Babson College All Rights Reserved
Last Updated 4/27/97 by Germaine Wong

To sign-up for the Center for Entrepreneurial Studies' publication lists,
please register with the
Entrepreneurship WebTeam.