Finkelstein 1997

INTERINDUSTRY MERGER PATTERNS AND RESOURCE DEPENDENCE: A REPLICATION AND EXTENSION OF PFEFFER (1972)

Resource dependencies, (measured by interindustry economic transactions) -> merger patterns (Pfeffer’s 1972)

Hypotheses that seek to explain both longitudinal and cross-sectional variation in the strength of this effect

After applying more refined analytical methods to the data the explanatory power of resource dependence is greatly diminished

Variation in the strength of the resource dependence effect suggests some boundary conditions for the theory

* Introduction

Merger studies

1) Rational decision making (Salter and Weinhold, 1979; Chatterjee, 1986; Lubatkin, 1987; Singh and Montgomery, 1987) 

2) Emergent organizational processes (Duhaime and Schwenk, 1985; Jemison and Sit- kin, 1986; Haspeslagh and Jemison, 1991; Amburgey and Miner, 1992)

Constraint -> Influence

Organizational behavior is likely to be constrained by industry environments, in their effect both on

1) Competitive dynamics (Aldrich, 1979; Scherer, 1980) 

2) Norms of thought and behavior (Hirsch, 1985; Shearman and Burrell, 1987).

Point: The influence of industry environments on merger behavior

Resource dependence (Pfeffer 1972)

Merger: Absolve resource interdependencies 

854 large U.S. mergers completed between 1948 and 1969

The proportion of a manufacturing industry’s total mergers conducted with each other industry was significantly correlated with the proportion of economic transactions that occurred between those industries in 1947 # Industry level analysis

Subsequent work

1) Incorporating effects of industries’ relative market power 

2) Extending the analysis to a more detailed industry level

The industry-level environmental constraint in merger

Pfeffer and Salancik (1978)  

Focal industry concentration -> interdependence -> Merger

Burt (1980) 

a model of industry structural autonomy in a network of interindustry economic transactions

+ relative market power

Interorganizational constraint arising from transaction patterns -> Merger as a cooperative relation

Galbraith and Stiles (1984) 

An industry sector’s market power relative to that of its customer industries -> the types of mergers firms in the industry tended to engage in

Palmer et al. (1995) 

reported that firms in industries that posed the greatest constraint on other industries were more likely to engage in predatory mergers and less likely to undertake friendly mergers in the 1960s.

Three limitation

1) First, although archival merger data include a temporal dimension, each of these studies has relied exclusively on single cross-sectional analyses. -> Data on economic transactions are now available for a 45-year time period, not only allowing for analyses that are more complete and up-to-date, but also for tests of resource dependence that link transaction dependence and merger in a contemporaneous fashion

2) Second, the original study upon which much of this work is based—Pfeffer (1972)—is subject to a variety of design and methodological criticisms. -> More precise methods and the application of newer, more sophisticated network-based techniques to measure resource dependence

3) Finally, while much of this work has focused on establishing that resource dependence theory offers a reasonable explanation for the relationship between interindustry transactions and mergers, little attention has been paid to -> Sources of variation in the strength of this relationship, either longitudinally or cross-sectionally

Purpose

1) to reexamine the resource dependence explanation for interindustry merger patterns, replicating Pfeffer’s original study using first identical and then more restrictive methods

2) to extend Pfeffer’s basic findings by developing and testing hypotheses that seek to explain why there is longitudinal and cross-sectional variation in the relationship between interindustry transactions and mergers

MERGERS AND RESOURCE DEPENDENCE

Pfeffer (1972) 

Inter-industry economic transaction patterns for 1947 -> large mergers between organizations grouped by 2-digit SIC manufacturing industries from 1948 to 1969.

Finding: The proportion of a manufacturing industry’s total mergers conducted with each other industry was significantly correlated with the proportion of its total economic transactions that the industry carried out with the same other industry in 1947.

Proportions -> alternative explanations for mergers, including such macroeconomic factors as GNP and stock market volume (Golbe and White, 1988), are implicitly controlled for since the impact of these factors is felt on aggregate merger activity

Question

For economic transactions with buyers and suppliers

Pfeffer (1972: 384) essentially studied vertical mergers as a response to ‘symbiotic interdependence’, 

Mergers between firms in industries that were related through transactions were more common than mergers between firms in unrelated industries.

His study implicitly assumed that interindustry economic transaction patterns do not change over time

Whether resource dependencies account for interindustry merger patterns in different time periods

Why?

‘Organizations under norms of rationality seek to place their boundaries around those activities which if left to the task environment would be crucial contingencies’ ( Thompson, 1967: 39).

By managing interdependencies through merger, organizations attempt to restructure their relationship with their environment, reducing uncertainty.

H1

Interindustry input-output tables, containing data on the value of goods an industry transacts with every other industry, quantify the pattern of resource exchange among industries (Burt, 1988).

Hypothesis 1: Interindustry merger patterns will be positively associated with interindustry transaction patterns.

Assumption: the objectively measurable environment of interindustry transactions reflects in a relatively transparent way the interdependencies that organizations in an industry perceive and act on.

Revision

A more straightforward test would involve a comparison of interindustry transactions and mergers in more contemporaneous time periods.

1) While Pfeffer generally relied on correlations to formally test hypotheses, multiple regressions are more appropriate (Davis and Powell, 1992).

2) Proportions produce heteroskedastic error terms (Cohen and Cohen, 1983), violating assumptions of the general linear model.

3) Differentiating between inter-industry mergers and intra-industry mergers

Whether the inclusion of intra-industry mergers in Pfeffer’s sample overstated the real strength of the resource dependence effect because of the clustering of data points he described.

4) The appropriate level of aggregation: Thus, although the present study adopts methods to reduce the aggregation problem evident in Pfeffer, data limitations require that it maintain an industry unit of analysis

5) more sophisticated methods to measure resource dependence.

1) investigate whether support for a resource dependence explanation for interindustry mergers can be maintained when methods more precise than those used in the original Pfeffer study are applied to the same data set,

2) isolate the effects of these methodological enhancements on the strength of the resource dependence relationship.

whether there is longitudinal and cross-sectional variation in the strength of the resource dependence explanation for interindustry merger patterns, and what some of the causes of such variation may be.

Sources of longitudinal variation in the strength of the resource dependence explanation

Two different explanations for why there may be variation in the strength of the resource dependence effect over time. 

- Time-boundedness of resource dependence

the relationship between interindustry transaction patterns and interindustry merger patterns will decrease over time. 

1) a saturation effect.

the relationship between transaction dependence and merger patterns may diminish over time because the strategy of absorbing uncertainty was successful (Gargiulo, 1993).

2) the opposite scenario. 

For some firms, attempts to reduce uncertainty from transaction dependence by engaging in vertical and horizontal mergers may be met with failure. Hence, to the extent that previous mergers did not reduce uncertainty between transacting firms, subsequent mergers may be seen as less valuable in actually ameliorating this uncertainty.

3) inertial (Amburgey and Miner, 1992) or institutional (Fligstein, 1990) reasons

resource dependence perspective: mergers are undertaken to reduce uncertainty emanating from transaction dependence

repetitive momentum and concerns for legitimacy (Amburgey and Miner, 1992; DiMaggio and Powell, 1983).

to the extent that there is persistence in merger activity, the relationship between interindustry transaction patterns and mergers will become weaker over time because merger decisions will gradually be driven less by concerns for reducing transaction dependence and more by these emergent processes that drive persistence 

Hypothesis 2: The strength of the relationship between interindustry merger patterns and interindustry transaction patterns will diminish over time.

- Antitrust enforcement of vertical and horizontal mergers

the role of the institutional environment

Anti-trust legislation and actions, including Supreme Court decisions, have played a key role in shaping merger activity by U.S. corporations since the late nineteenth century (Stigler, 1966; Eis, 1978; Fligstein, 1990; Shleifer and Vishny, 1990)

the Celler–Kefauver Act, which Congress passed in 1950. 

This legislation discouraged vertical and horizontal mergers, which were viewed as anti-competitive (Fligstein, 1990).

The Supreme Court (under Chief Justice Warren) voided the merger between Kinney Shoe and Brown Shoe Company in 1962—which was the court’s first decision on Celler–Kefauver—that vertical and horizontal mergers effectively became illegal.

The impact of Celler –Kefauver was consequential until 1974 when the Supreme Court (under Chief Justice Burger) began to rule against the government by applying criteria other than those contained in Celler–Kefauver (Fligstein, 1990).

As a result, it was the years between 1962 and 1974 that were the most difficult ones for firms contemplating vertical or horizontal merger to reduce their transaction dependence on firms in other industries.

# Fun

Hypothesis 3: The relationship between interindustry transactions and interindustry mergers will be weaker in the years when the Celler– Kefauver Act discouraging vertical and horizontal mergers was strictly enforced (1962–74).

Sources of cross-sectional variation in the strength of the resource dependence explanation

1) Huge variation in the magnitude of these correlations: they ranged from 0.31 to 0.97. 

2) Hence, the strength of the resource dependence effect appears to vary substantially across industries.

Pfeffer and Salancik (1978): differences in patterns of sales and purchase interdependence were related to concentration levels.

1) Past reliance on mergers to reduce resource dependence

Although organizations can respond in different ways to environmental threats and opportunities, they are often heavily influenced by the shared experiences of other firms in the same industry (Huff, 1982).

 Firms in an industry tend to develop common repertoires or recipes that define what is seen as appropriate strategies and structures (Spender, 1987). As Hambrick (1982) pointed out, ‘a common body of knowledge appears to exist within an industry which is disseminated through media equally available to and used by executives within the industry’.

to the extent that firms in an industry develop ‘shared perceptual structures’ (Reger, 1990) or ‘industry recipes’ (Spender, 1987), they are more likely to follow established common procedures and strategies already in use in that industry.

# Other strategies are available (joint ventures, board cooptation, long-term contracts, etc.).

It is likely that some industries will tend to rely on mergers to reduce transaction dependence much more than others (Tolbert, 1985).

When firms consider merger to reduce transaction dependence, they will be more likely to engage in this strategy to the extent that they and other firms in the industry have traditionally done so.

Hypothesis 4: The greater the past reliance by firms in an industry on mergers to reduce transaction dependence, the more positive the relationship between interindustry transactions and interindustry mergers.

- Potential gains from reducing transaction dependence

The effects of transaction dependence are not experienced equally in all industries.

The logic of this argument suggests that firms in an industry will be more likely to engage in the merger as a response to resource dependence the more concentrated the industry.

Hypothesis 5: The greater the bidder industry concentration, the more positive the relation- ship between interindustry transactions and interindustry mergers.

Data and measures

DV: proportion of the total number of mergers of industry i that were with industry j between 1948 and 1969, where i and j represent one of the industries studied

IV: transaction dependence of the bidder industry—was measured as the proportion of industry i’s total transactions that were with industry j

CV: Potential alternative explanations for interindustry mergers.

1) the number of large firms (assets greater than $10 million) in the industry of the target firm. This variable controlled for the extent to which patterns of large mergers represent random merger activity.

2) prof- itable industries are likely to attract firms interested in acquisitions. 

-> the profitability of firms in target industries is a second control variable, measured as the average return on equity of firms in an industry.

3) the degree of concentration in target industries, reasoning that concentration ratios measure the height of barriers to entry in an industry, with high barriers reducing the likelihood of acquisition.

Correlational analysis

It is apparent from this table that the correlations are very similar between the original Pfeffer study and the present one, provid- ing the first direct replication of Pfeffer’s results.

Testing the robustness of resource dependence

1) using multiple regressions rather than correlations, 

2) transforming the dependent variable to remove heteroskedasticity, 

3) differentiating between inter- and intraindustry mergers, 

4) examining the sensitivity of results to different degrees of aggregation, and 

5) using network analysis to operationalize resource dependence.

- Inter- vs. intra-industry mergers

an association between transactions and mergers may arise in the latter

Pfeffer’s original hypothesis is nevertheless strongly supported by the results

- Aggregation

Pfeffer’s findings reflect the effects of aggregation but, when this aggregation problem is partially removed by employing a narrower unit of analysis, the resource dependence explanation for interindustry mergers still receives significant, though much less pronounced, support.

- Network measures of transaction dependence 

a resource dependence explanation for interindustry mergers is supported.

Third, judging from the magnitude of stan- dardized coefficients and R2 in reported regressions, it appears that the inclusion of intraindustry mergers greatly inflates the strength of the resource dependence effect (by a factor of 2.5 when 20 industries are studied and a factor of more than 3 when 51 industries are studied).

Finally, although Pfeffer and others have suggested that network constraint measures hold more promise than simple proportions (Pfeffer, 1987; Palmer et al., 1995), the results reported here suggest that the use of proportions over network measures does not inflate results as much as does reliance on either intraindustry mergers or aggregated industry samples.

Summary

First, the significant association between transaction dependence and interindustry mergers first reported in Pfeffer has been repli- cated.

Second, however, and in contrast to the robust results reported by Pfeffer, the strength of the resource dependence effect became quite muted in this study as successively more precise methods were applied.

  

EXTENSION OF PFEFFER (1972)

Hypothesis 2 

This hypothesis predicted that the strength of the relationship between interindustry transaction dependence and interindustry mergers would di- minish over time. However, as the results in Table 7 indicate, no such pattern was observed. In fact, some of the strongest findings emerged for the three most recent time periods: 1977–81, 1982 – 86, and 1987 – 92. Thus, the data suggest that Hypothesis 2 be rejected.

Hypothesis 3 focused on the role of antitrust enforcement

Hypothesis 3 predicts a weaker resource dependence effect in the four time peri- ods between 1958 and 1976, and especially between 1963–66 and between 1967–71 — panel years that were fully within the Celler–Kefauver enforcement period. 

Overall, these results appear to provide partial support for Hypothesis 3.

Hypothesis 4

Hypothesis 4 predicted that the resource depen- dence explanation for interindustry mergers would be more positive in those industries where there was an already established relationship between industry constraint and merger in the previous time period. 

This finding is consistent with Hypothesis 4 and offers an explanation for the cross-sectional variation in the resource dependence explanation for interindustry mergers apparent in Pfeffer’s data.

 

Hypothesis 5 

We argued that because firms in highly concentrated industries would have the most to lose from industry constraint, the associ- ation between transaction dependence and interin- dustry mergers would be moderated by bidder industry concentration

As is evident from this regression, the coefficient for the inter- action term was positive and significant (p 􏰀 0.05), indicating that the relationship between transaction dependence and merger is indeed stronger when bidder industries are more concentrated. Thus, Hypothesis 5 is supported. 

DISCUSSION

1) transaction dependence was significantly related to interindustry mergers in five of the eight time periods studied

2) transaction dependence accounted for only a small percentage of the variance in interindustry mergers

3) the

Limitations

1) because we focused on the industry level of analysis, internal organizational processes and decisions that might affect merger activity were not studied.

2) by extending the investigation to the input–output sector level of 51 industries, more fine-grained analysis that avoided the same degree of aggregation as in Pfeffer was possible.

3) only examined dependence among manufacturing industries, leaving unexamined the effects of transaction dependence on service industries.

4) the number of mergers between two industries in any given year was generally quite small, and transactions data were only available periodically from the Survey of Current Business, ruling out the possibility of using annual data for an even more fine-grained analysis.

CONCLUSIONS

1) the basic resource dependence hypothesis on the relationship between interindustry transactions and mergers was supported

+ interindustry transactions patterns are not as predictive of interindustry merger patterns as previously thought

2) the boundary conditions of the theory, at least with respect to its relevance for explaining interindustry merger patterns

2-1) Firms were less likely to respond to transaction dependence through merger in those time periods when state action in antitrust enforcement appeared strongest, suggesting that the institutional environment may act as a constraint on the application of uncertainty-reducing strategies. 

2-2) In addition, differences in the strength of the resource dependence effect across industries were also observed.

3-1) established industry patterns of response -> Strategy

3-2) variability in the extent to which firms in some industries may benefit from reducing constraint -> Strategy

+ would be interesting to investigate whether these factors play similar roles in affecting whether firms attempt to reduce other types of uncertainty

Future

1) More complex empirical formulations may be required in subsequent tests of resource dependence theory

Not only may firms respond to a resource constraint with multiple strategies, they may be experiencing several different types of resource dependencies simultaneously.

2) The need for firm-level analysis, more appropriate measures of resource dependence, and studies that consider the consequences of uncertainty-reducing strategies as well.