Diversifiction, Ricardian Rents, and Tobin’s q

(Montgomery and Wernerfelt 1988)

Group 1Meredith, Barclay, Woo-je, and Kumar

IntroductionEconomic analysis does not have a great deal to say about firm diversification and existing theory is largely untested

Prevailing arguments are based on excess capacity of production factors

Failures in the market may make diversification and efficient choice

This paper attempts to extend the prevailing theory by considering the heterogeneity of factors that prompt diversification and profit maximizing decisions of the firm

Sources of Ricardian Rent

Economic rents can result from collusive relationships with competitors, disequilibrium effects (luck) and unique factors (Ricardian rents)

Ricardian rents are ordinarily thought of as accruing to owners of unique factors

Firm operated by a good manager, owning attractively located land, patentsOwning factors subject to uncertain imitability

Sources of Ricardian RentSources of Ricardian Rent

Ricardian rents are only part of the story

Firms may also appropriate economic rent as trading partners of factor owners, and share the factor in question

Firms may employ a manager or supplier who creates switching costs

Diversification as a Way to Appropriate Ricardian Rents

Four important assumptions for author’s argument:

1. Assume that the firm can dispose of excess capacity (sell at price zero) without affecting the rest of its operations

2. Do not consider cases where there are natural economies of scope between 2 industries

3. Concentrate on firms that own or control economic rent-yielding factors

4. Conduct the analysis in a static model and evaluate the case of a singe diversification move in which a firm with excess capacity of a rent yielding factor considers a marginal expansion of its scope

Diversification as a Way to Appropriate Ricardian Rents

With respect to a marginal change in the scope of a firm, the givens are a set of factors and a list of markets to which they may be transferred and result in smaller or greater competitive advantages

Market in which the factor yields the highest economic rents are “closest”

The more the firm has to diversify, i.e., the farther from its current scope it must go, ceteris paribus

The larger will be the loss in efficiency The lower will be the competitive advantage conferred by the factor

Diversification as a Way to Appropriate Ricardian Rents

Marginal rents

Diversification Distance

More specific factors

Less specific factors

Hypothesized Relationship Between Diversification Distance and Marginal Rents for Different Degrees of Factor Specificity

Diversification as a Way to Appropriate Ricardian Rents

Give a specific set of factors, the optimal decision for a firm is to apply its excess capacity to the closest entry opportunity. The economic rent the firm can extract from the move depends on the specificity of the factors and the closeness of the new market

These conditions result in the following stylized relationships.

a) Firms with less specific factors and nearby entry opportunities will diversify narrowly and extract medium rents on average

b) Firms with more specific factors whose closest entry opportunity is “nearby” markets will diversify narrowly and extract high rents

c) Firms with less specific factors with “distant” opportunities will diversify widely and extract low rents

d) Firms with more specific factors and no nearby opportunities will not be able to diversify at positive marginal rents. These firms are likely to have very high average rents, although it is clear that their total profits would increase if they had the opportunity to diversify

Diversification as a Way to Appropriate Ricardian Rents

(a) Medium diversification,

medium average rents

(b) Narrow diversification, high

average rents

(c) Wide diversification, low

average rents

(d) No diversification very high average

rents

Prediction: As optimal diversification increases, average economic rents decline

Tobin’s q as a Measure of RentsTobin’s q as a Measure of Rents

Tobin’s q is defined as the ratio of market value to the replacement cost of the firm

q = M/Vp = 1 + (VI + VC + VR + VE)/VP (1)

M = market value of the firmVP = replacement value of physical assets

VI = value of intangible assets purchased by firm

VC = value of collusive relationships w/competitors

VR = capitalized Ricardian rents

VE = disequilibrium effects

Tobin’s q as a Measure of RentsTobin’s q as a Measure of Rents Estimate (1) using conventional proxies for VI, VC, VE to focus on the relationship between VR and multi-market activity

Relationship is not straight forward. If we denote d as diversification, s as specificity, and o as opportunities our theory is that

VR/VP (s,d) (increasing function s, decreasing ofd) (2) d (s,o) (decreasing function of s, increasing of o) (3) Problem: s and o are unobserved, industry dummies as instrument for d in

VR/VP (d) (4)This amounts to using the average industry level diversification, rather than each firm’s own diversification level as a proxy for s, and o

Data, Measures, and TestsData were gathered from multitude of sources, and from these data estimates of the following variables were constructed

Ai =firm i’s marketing expenditures (sales weighted)

Ri =firm i’s R&D (sw)Ci = concentration in firm i’s market (sw)Gi =growth of shipments in firm i’s markets (sw)Si = firm i’s market share (sw)Fi = firm i’s foreign sales (in percent) Vpi = replacement costs of firm i’s physical assets

Data, Measures, and TestsData, Measures, and Tests

Diversification measure (Di) requires more explanation, because goal is to differentiate between more and less similar diversification,and have chosen the concentric index of Caves et al. (1980)

Di =Σ mij Σ milril

j = 1

n n

l =1

Where mij is the percentage of firm i’s sales in industry j and rjl is zero if j and l have the same three digit code, 1 if they have different 3-digit codes but identical 2-digit codes and 2 if the have different 2-digit codes

Data, Measures, and TestsData, Measures, and Tests

Regression equationsq = β0 + β1A/Vp + β2R/VP+ β4C+ β5D+ β6F+ β7G +Є

q = β’0 + β’1A/Vp + β’2R/VP+ β’4C+ β’5Ď+ β’6F+ β’7G +Є

Where Ď indicates that D is estimated through the instruments

Because both sides are divided by VP, measurement error in this variable induces some problems.

Therefore, they follow Grilliches (1981) and take logs, using the x≈log(1+x) approximation

Results

Dependent Variable

I Log Vp

A/Vp R/Vp C S D Ď F G R2

q .908 (5.88)

2.71 (4.70)

8.98 (3.36)

-.006 (2.4)

1.01 (2.35)

-.145 (1.97)

.417 (1.84)

.007 (.091)

.293

q .888 (5.67)

2.52 (4.08)

9.33 (3.28)

-.006 2.17

1.13 (2.51)

-.186 (2.20)

.469 (1.93)

.008 (.086)

.296

Log M -.728 (1.83)

1.03 (35.4)

2.23 (4.12)

9.47 (4.05)

-.006 (2.57)

.846 (1.93)

-.035 (.538)

.273 (1.32)

.024 (.302)

Log M -.926 (2.12)

1.05 (32.0)

2.24 (3.93)

9.92 (3.99)

-.005 (2.27)

.714 (1.55)

-.139 (1.76)

.281 (1.29)

.011 (.133)

Table 1 Regression Results: Firm Value of Extent of Diversification

Discussion/ Conclusions

Using Tobin’s q authors have tested the hypothesis that large firms earn decreasing average rents as they diversify more widelyResults indicate that the farther a firm must go to use their factors, the lower the marginal economic rents they extractExplanation of negative relationship between market valuation and diversification

Faulty beliefs in rents diversification“Free cash Flow” hypothesis

Discussion/ ConclusionsDiscussion/ Conclusions

Limitations of study:

Several simplifying assumptionsPertains only to large firmsTest refer to average rents, not total profit

Assumptions allow the focus of few key implications of factor heterogeneity and are sufficient to explain the evidence