Evaluation of the impact of R&D on EPS in the oil and gas industry.

Gondhalekar, Vijay ; Bhagwat, Yatin ; DeBruine, Marinus 等

ABSTRACT

Investment in Research and Development (R&D) creates intangible assets with a high earning potential. Several industries have a continuous R&D program due to the competitive nature of business. Thus R&D outlays may be viewed as fixed costs necessary for firms to achieve growth in size and earnings. The impact of R&D on earnings is measured by degree of R&D leverage, a term analogous to degrees of operating and financial leverage. This paper specifically tests the relationship between changes in R&D outlays and their impact on earnings for a sample of firms within petroleum drilling, refining, and oil and gas field services industry.

The results indicate that non-U.S. firms have higher R&D intensity compared to U.S. firms. Also R&D intensity is higher for firms within the petroleum-related service firms than for petroleum refining companies. Also the firms with lower R&D intensity have demonstrate higher degree of R&D leverage implying diminishing returns at the higher end of the R&D intensity spectrum. On average, every one- percent increase in R&D outlay results in a one-fourth of one-percent increase in earnings per share.

INTRODUCTION

The literature in finance and economics views R&D expenditure as an attractive means of investment in valuable intangible capital that has differing degrees of relevance in different economic sectors. This paper measures the impact of R&D investment on earnings in oil and gas exploration and petroleum refining industry - by estimating the degree of R&D leverage, a measure similar to degrees of operating and financial leverage. The degree of R&D leverage is useful for investors and analysts. If investors are optimistic about prospects for an industry they might favor companies with a high degree of R&D leverage as it may be a predictor of earnings growth. This paper is organized as follows. A brief literature review is provided in Section II. In Section III a theoretical discussion of the degree of R&D leverage is provided. Data, methodology, and results are discussed in Section IV. Finally, Section V summarizes the findings and provides suggestions for further research.

II. LITERATURE REVIEW

Several previous studies have demonstrated that there is merit in investigating the measurement and valuation of intangible capital. Trajtenberg (1990) (investigating optical scanners; Megna and Mueller (1991) (investigating distilled beverages, cosmetics, pharmaceuticals, semiconductors); and Shane and Klock (1997) (investigating semiconductors are some of the examples of industry studies valuing intangible capital. Different industries require different measures of intangible capital. Advertising expenditures are an important source of intangible capital in the consumer products industry. R&D expenditures are considered an important source of intangible capital for the pharmaceuticals industry. Patents are important for optical scanners, semiconductors and other products in the high-tech sector. Patents are the fruition of a successful R&D program.

Chan et. al. (1990), and Doukas and Switzer (1992) find statistically significant R&D announcement day returns, especially in the case of large high-tech industrial firms that devote substantial resources to R&D. Jose, Nichols, and Stevens (1986) and Morck, Schleifer, and Vishny (1988) report positive market value effects of R&D to samples of large firms. Chauvin and Hershey (1993) postulate that R&D spending appear to help investors form appropriate expectations concerning the size and variability of future cash flow. They find R&D investment to possess significant explanatory power in determining the market value of a firm after controlling for cash flows, growth, risk, and market share. Klock and Megna (1999) investigate the measurement and valuation of intangible capital in the wireless telecommunications industry. Advertising, R&D, radio spectrum licenses, and measures of installed customer base explain a statistically significant portion of the variation of Tobin's q.

Mandelker and Rhee (1984) have established the association between systematic risk, intrinsic risk, and degree of operating (DOL) and financial leverage (DFL). DOL and DFL magnify intrinsic business risk of common stock. Their study highlights the joint impact of DOL and DFL on the systematic risk of common stock. Technological breakthrough that requires new capital investment, shifting the company to a higher degree of operating leverage, may signal an offsetting shift in the degree of financial leverage. More recently, several companies belonging to the high-tech sector and some other industries issue almost no long-term debt. Also these firms invest heavily in R&D. R&D generate potential for growth in earnings. Operating leverage amplifies existing business risk by substituting fixed cost for variable cost in cost structure. Financial leverage creates and amplifies financial risk by substituting fixed return debt for variable return equity in capital structure. R&D leverage amplifies existing business risk by creating intangible assets that augment earning power.

The literature focuses on impact of intangible capital on firm value. However, from an investor's or an analyst's perspective, the value of a company is significantly affected by fulfillment of earnings growth expectations. Analysts are constantly estimating prospective quarterly earnings. Changes in stock valuations occur when the efforts resulting from intangible assets such as R&D are granted, albeit for a limited time, earning power by regulatory agencies. This paper establishes a relationship between the change in the level of inputs generated by R&D and the fruition of these efforts reflected by the growth in earnings. The next section lays out the analytical foundation for the empirical findings.

III. DEGREE OF R&D LEVERAGE: A THEORETICAL FRAMEWORK

Finance has borrowed the concept of leverage from physics. Leverage provides mechanical advantage in physics. In finance leverage implies amplification of earnings by application of inputs such as debt. Leverage measures such as DOL and DFL have been referred in an earlier section of this paper. The Degree of Total Leverage (DTL) is defined in terms of elasticity of earnings with respect to sales revenues and is equal to the product of DOL and DFL. The Degree of R&D Leverage (DRL) measures the earnings elasticity of R&D and represents an index number that measures the effect of incremental R&D expenditure on the net earnings per share. The effect of R&D can be dichotomized as follows. Firstly the effect of R&D on earnings before R&D and taxes is measured (denoted DRL1), and secondly the effect of earnings before R&D and taxes on net earnings is measured (denoted DRL2). DRL1 measures the pure operational elasticity of R&D by excluding the tax deductibility of R&D. DRL2 takes into account the impact of tax deductibility of R&D on the net income. The concomitant effect is measured by DRL that is operationalized in the following paragraphs.

(1) DRL1 = % [DELTA]EBRT/ %[DELTA]R

(2) DRL2 = % [DELTA]EPS/% [DELTA]BRT

(3) DRL = DRL1 (X) DRL2,

Where EBRT = Earnings before R&D and Taxes,

R = R&D expenditure

EPS = Earnings per share

[DELTA] = Change in a given variable

We can rewrite equation (3) and transform it into its logarithmic form. Hence, (4) DRL = %[DELTA]EPS/%[DELTA]R = [DELTA]EPS [sup.*][partial derivative]lnEPS/[partial derivative]EPS/[DELTA][R.sup.*][partial derivative]lnR/[partial derivative]R. DRL = [partial derivative]lnEPS/[partial derivative]lnR.

Equation 5 can be transformed as:

(6) lnEPS = [alpha] + [beta]lnR.

The slope of the model specified in equation (6) is an estimate of DRL. It measures the earnings elasticity of R&D. As such, the DRL denotes the percentage in earnings brought about by a percentage change in R&D expenditures.

IV. SAMPLE SELECTION AND DESCRIPTIVE STATISTICS

This study uses annual company data for the years 1987 through 1998 compiled by Standard and Poor's COMPUSTAT Services. R&D expenditures in absolute terms and in terms of intensity differ significantly among industries. Compared to manufacturing companies, petroleum-related companies spend considerably less of each sales dollar on development of new products and processes. According to Exhibit I, companies in the business services sector spend ten times as much of each sales dollar on R&D. However, since petroleum refining companies are also among the largest in terms of annual sales, companies in the business sector on average spend less than one-fifth as much on R&D. The communications sector reports the highest average annual R&D expenditures. High-tech companies also spend on fundamental research whereas other companies invest in applied research.

We selected all companies with the standard industry codes 1311 (crude petroleum & natural gas), 1382 (drilling oil and gas wells), 1389 (oil and gas field services) and 2911 (petroleum refining) for this study. The data include annual sales, R&D, and earnings per share before extraordinary items on an annual basis. Panel A of Exhibit II contains the data for the top 10 companies ranked by average annual sales. R&D intensity is defined as the ratio of R&D to Sales, and one company in this list--ELF Aquitaine--allocates a disproportionate amount of resources to R&D.

Panel B of Exhibit II contains the data for the top 10 companies ranked in terms of R&D expenditures. Not surprisingly, ELF Aquitaine leads the list, but a relatively smaller company - Schlumberger LTD - outspends all the others in terms of R&D expenditures. Almost sixpercent of its annual sales dollars is devoted to R&D. It can be inferred from Panel B that R&D intensity is higher for the petroleum-related service industry companies than for petroleum refining companies.

Within the oil and gas industry sectors, companies differ in how much of their sales dollars are spent on R&D. Exhibit III lists 20 companies that make up the bulk of those industry sectors and have data for the most recent years. The companies are ranked in descending order of R&D intensity. Those in the top half tend to be from outside of the U.S., while the bottom half is comprised of U.S. companies. Ever since the FASB issued Statement #2 in 1972, many have claimed that the U.S. reporting requirements for R&D expenditures have negative economic implications for U.S. competitiveness in research and development, and the distribution in this table appears to support that claim.

V. METHODOLOGY AND RESULTS

This study employs a log-linear model:

lnEPS = [alpha] + [beta]lnR

This relationship has been derived in Section III above. The coefficient of the independent variable is the measure of degree of R&D leverage. R&D studies primarily have used various forms of regression techniques to measure relationship between R&D effort and firm value. Researchers in the field of industrial organization, in general, measure the long-term relationship between R&D effort and value of innovation. Studies in the finance discipline focus on the announcement effect of R&D programs on firm value. In this section prior methodologies are discussed.

Jose, Nichols, and Stevens (1986) test the contributions of diversification, promotion, and R&D to the value of the firm using data from Federal Trade Commission's Line of Business Survey. They construct a number component of firm product diversification determined by lines of business and a size component of firm product diversification determined by the firm's sales originating in the line of business. Further they define variables to measure firm promotional effort, industry promotional effort, firm R&D effort, and industry R&D effort. OLS regression of Tobin's q on a combination of variables measuring diversification, advertising, and R&D suggests that diversification has a significant and positive influence on the value of the firm. Limited support is found for the hypothesis that higher individual firm promotional investment, both absolute and relative to the industry norm, reduces the value of the firm. Industry level R&D intensity is significant, and it lowers the value of firms within that industry, ceteris paribus. This is most likely because high R&D intensity signals a high probability of product obsolescence. The insignificance of the ratio of firm level R&D to the industry level indicates that spending more than the industry norm neither helps nor hurts a firm. However, significant deviation from the industry benchmark in either direction makes a firm appear more risky.

In industrial organization, one explanation for dramatic differences in profit rates across firms and industries is the failure of conventional accounting methods to adjust for intangible capital stocks. To test this hypothesis Megna and Mueller (1991) calculated advertising and R&D stocks for the pharmaceutical firms. Using firm profit as the dependent variable, regression estimates on firm and the competitor calculated tangible and intangible capital stocks as independent variables did not show support for this hypothesis. The adjustments in measurement of capital stock with the inclusion of intangible capital such as R&D and advertising did not eliminate the wide dispersion in profit rates.

Chauvin and Hershey (1993) are primarily concerned with the role played by advertising and R&D as determinants of the current market value of the firm measured by Tobin's q ratio and relative excess valuation (EV/S). Tobin's q ratio is defined as the market value of the firm normalized by the replacement cost of tangible assets and relative excess valuation is measured as the market value of common minus book value of stockholders' equity, all normalized by sales. Using COMPUSTAT data over the three-year period 1988 through 1990 the following model is estimated using Ordinary Least Squares Estimates (OLS) for the manufacturing and non-manufacturing firms. In the manufacturing sector, more than one-half of the variation in market values can be attributed to variation in cash flows, growth, market share, advertising and R&D. In the non-manufacturing sector, roughly one-third of the variation in market values can be similarly attributed.

Klock and Megna (1999) construct stocks of intangible capital for firms in the wireless communications industry using advertising, R&D expenditures, and spectrum license coverage. They regress Tobin's q on these stocks using linear and semi-log models. They found the licensing variable and advertising to possess significant explanatory power. Colinearity between R&D and licensing variable also existed.

Chan, Martin, and Kensinger (1990) find share price responses to 95 announcements of increased R&D made in the Wall Street Journal between June 1979 and June 1985 spending to be significantly positive on average, even when the announcement occurs in the face of earnings decline. High-technology firms that announce increases in R&D spending experience positive abnormal returns on average, whereas announcements by low-technology firms are associated with negative abnormal returns. Further, higher R&D intensity than the industry average leads to larger stock-price increases only for firms in high-technology industries.

Doukas and Switzer (1992) also study share price responses to R&D expenditure for a sample of 87 announcements by 45 firms made in 1984. The results reveal significant differences in the abnormal returns for firms operating in different marketing environments. Specifically, firms in industries characterized by high (low) seller concentrations, announcements of increases in planned R&D expenditures are associated with significant positive (negative) excess (abnormal) stock returns. Cross-sectional analysis of the cumulative standardized abnormal returns (SCAR) for varying event windows is conducted by regression of SCAR on market concentration and R&D spending increase factor. The latter is defined as the product of R&D intensity and the percentage increase of announced R&D spending over previous year. Both independent variables have significant explanatory power. The results, according to Doukas and Switzer, are consistent with the Schumpeterian view on the relationship between market structure and innovation that predicts a differential market response to R&D that depends on the firm's market concentration.

This relationship is estimated for all eligible companies in selected industry sectors and is reproduced in Exhibit IV.

The estimated models, intercept, and slope terms are highly significant with appropriate signs. The R2 value of .22 for the overall sample signifies that the model explains more than one-fifth of the variation in EPS. The results indicate that on average during the period 1987 through 1998 for all eligible companies, a 1% increases in R&D expenditure results in a .27% increase in the EPS. The results are obtained separately for the following industry sectors: chemicals and allied products, petroleum refining, rubber and plastic products, industrial machinery, measuring and analytical instruments, communications, and business services.

Within 2-digit SIC industries, the communications sector shows the strongest relationship between R&D expenditures and EPS: for every 1% change in R&D the communications industry's EPS increases by .35%. Although the petroleum-related industries show a much weaker relationship between R&D expenditures and EPS, the positive sign on the coefficients implies that the earnings from operations are enhanced by increases in R&D.

Ranked by R&D intensity within the petroleum-related sector, the top 10 companies show a weaker relationship (DRL estimate = .27; R2 = .10) relative to the bottom 10 companies (DRL estimate = .31; R2 = .21). This somewhat surprising result is reflected in Exhibit V may imply that an optimum R&D level exists for each firm and that firms with high R&D intensity may be receiving diminishing returns.

VI. CONCLUSION

The impact of R&D expenditure on company value, size, and variability of cash flows has been studied widely. This study develops and measures the earnings elasticity of R&D coined as Degree of R&D leverage. Using a log-linear model, it estimates the effect of R&D expenditures on earnings per share for various industries, including several petroleum-related industry sectors. The results indicate that a one-percent change in R&D expenditure results in a quarter-percent change in earnings on average. However, differential impact of R&D is observed across industry categories. Availability of protection of intellectual property rights, existence of competition, and the ability to translate innovation into commercial viability may be some of the reasons for the differences.

The results support the hypothesis that on average earnings growth would suffer from a reduction in R&D expenditures. This needs to be studied further by analysis of DRL1 and DRL2 values defined in equations 1 and 2. Further work is being done to measure the degree of R&D leverage for some specific four-digit SIC codes. The study will provide investors with a simple methodology to measure the impact of R&D on EPS.

Readers must understand some of the limitations of this study. Firstly, the R&D expenditure may be considered discretionary and will be one of the first expenditures cut by management when the market's earnings expectations need to be met because under GAAP (FASB #2) this immediately improves current EPS. Under this view, EPS expectations drive R&D expenditures for the current period. Modeling this would make R&D expenditures a function of EPS. Secondly, the impact of R&D on earnings may take place on a lagged basis. R&D may be viewed as an investment which payoff will benefit the EPS of several future periods. Under this view, R&D expenditures in the current period drive EPS expectations for multiple future periods. Pooling time series and cross-sectional observations may be the next step for getting additional insights. Nevertheless for firms with continuous R&D outlays, a static measure such as degree of R&D leverage provides us with a general relationship between R&D outlays and earnings growth in the Petroleum Refining and Exploring industries.

REFERENCES

Chan S. H., J. D. Martin, and J.W. Kensinger, "Corporate Research and Development Expenditures and Share Value," 1990, Journal of Financial Economics, (August), 255-276.

Chauvin, Keith and Mark Hirschey, 1993, "Advertising, R&D Expenditures and the Market Value of the Firm," Financial Management, (Winter), 128-140.

Doukas, John and L. N. Switzer, 1992, "The Stock Market's View of R&D Spending and Market Concentration," Journal of Economics and Business, (May), 95-114.

Jose M., L. M. Nichols, and J. L. Stevens, 1986, "Contributions of Diversification, Promotion, and R&D to the Value of Multiproduct Firms: A Tobin's q Approach," Financial Management, (1986), 33-42.

Klock, Mark and Pamela Megna, 1999, "Measuring and Valuing Intangible Capital in the Wireless Communications Industry," Presented at the Financial Management Association Meetings held in October 1999 in Orlando, FL.

Mandelker Gershon and S. Ghon Rhee, 1984, "The Impact of the Degrees of Operating and Financial leverage on Systematic Risk of Common Stock," Journal of Financial and Quantitative Analysis, (March), 45-57.

Megna, Pamela and Dennis Mueller, 1991, " Profit Rates and Intangible Capital," Review of Economics and Statistics (November) 632-642.

Morck, Shleifer and Vishny (1988)

Shane, Hilary and Mark Klock, 1997, "The Relation between Patent Citations and Tobin's Q in the Semiconductor Industry," Review of Quantitative Finance and Accounting (9), 123-138.

Tratgenberg, Manuel, 1991, "A Penny for Your Quotes: Patent Citation and the Value of Innovations," Rand Journal of Economics, 172-187.

Vijay Gondhalekar, University of Michigan-Flint

Yatin Bhagwat, Grand Valley State University

Marinus DeBruine, Grand Valley State University

Exhibit I
Distribution of R&D expenditures
Select 2-digits SIC industries

Industry description Industry Company Industry
(2-digit SIC Code) sales sales (m$) R&D
 (m$) (m$)

Oil and gas (13) 419,025 4,506 7,169

Chemicals and allied 4,542,707 2,698 289,785
products (28)

Petroleum refining (29) 7,252,167 32,232 48,664

Rubber and plastic products 322,511 921 6,598
(30)

Industrial machinery and 3,857,144 1,545 200,444
computer equipment (35)

Electronic equipment (36) 5,438,839 2,021 304,849

Communications (48) 2,467,770 13,485 80,641

Business services (73) 1,092,524 616 85,370

All industries 40,065,987 2,724 1,439,048

Industry description Company R&D Sample
(2-digit SIC Code) R&D (m$) intensity size

Oil and gas (13) 77 1.71% 93

Chemicals and allied 172 6.38% 1,684
products (28)

Petroleum refining (29) 216 0.67% 225

Rubber and plastic products 19 2.05% 350
(30)

Industrial machinery and 80 5.20% 2,497
computer equipment (35)

Electronic equipment (36) 113 5.61% 2,691

Communications (48) 441 3.27% 183

Business services (73) 48 7.81% 1,774

All industries 98 3.59% 14,710

Exhibit II
Top 10 leading petroleum-related companies
Panel A: in terms of annual sales

 Annual Annual R&D R&D Sample
Company name sales expenditures intensity size

Royal Dutch/ Shell Grp 99,829 666 0.67% 12
Exxon Corp 99,772 573 0.57% 12
BP Amoco PLC 57,034 388 0.68% 12
Mobil Corp 56,378 258 0.46% 12
Veba AG 44,114 203 0.46% 3
ELF Aquitaine 39,140 934 2.39% 8
Texaco Inc 36,442 185 0.51% 12
ENI S P A 34,952 281 0.80% 3
Chevron Corp 32,146 209 0.65% 12
Petroleas de Venezuela 27,351 76 0.28% 3

Panel B: in terms of R&D expenditures

 Annual Annual R&D R&D Sample
Company name sales expenditures intensity size

ELF Aquitaine 39,140 934 2.39% 8
Royal Dutch/ Shell Grp 99,829 666 0.67% 12
Exxon Corp 99,772 573 0.57% 12
Schlumberger LTD 7,020 420 5.98% 12
BP Amoco PLC 57,034 388 0.68% 12
ENI S P A 34,952 281 0.80% 3
Mobil Corp 56,378 258 0.46% 12
Chevron Corp 32,146 209 0.65% 12
Veba AG 44,114 203 0.46% 3
Total Fina S A 27,423 200 0.73% 8

Exhibit III
R&D intensity for petroleum-related companies

 Annual Annual R&D R&D Sample
Company name sales expenditures intensity size

Schlumberger LTD 7,020 420 5.98% 12
ELF Aquitaine 39,140 934 2.39% 8
Halliburton Co 7,116 127 1.78% 12
Burmah Castrol PLC 4,542 68 1.50% 7
Imperial Oil LTD 6,877 63 0.92% 12
ENI S P A 34,952 281 0.80% 3
Total Fina S A 27,423 200 0.73% 8
BP Amoco PLC 57,034 388 0.68% 12
Royal Dutch/ Shell Grp 99,829 666 0.67% 12
Chevron Corp 32,146 209 0.65% 12
Atlantic Richfield Co 16,452 103 0.63% 12
Exxon Corp 99,772 573 0.57% 12
Texaco Inc 36,442 185 0.51% 12
Mobil Corp 56,378 258 0.46% 12
Petro-Canada 3,875 16 0.41% 8
Repsol SA 13,820 34 0.25% 10
USX Corp 16,783 39 0.23% 9
Conoco Inc 18,737 42 0.22% 3
Occidental Petroleum Corp 12,484 23 0.18% 12
Sunoco Inc 8,873 11 0.12% 12

Exhibit IV Effects of R&D on EPS
Comparative results and Industry effects

Industry description
(2-digit SIC Code) Intercept R&D [R.sup.2]

Oil and gas (13) -.8792589 .1989165 .1531
 (-5.506) (4.057)

Chemicals and allied -1.3926820 .3033467 .2954
products (28) (-32.631) (26.553)

Petroleum refining (29) -1.0892170 .2943434 .1929
 (-5.591) (7.302)

Rubber and plastic -.8736472 .3516804 .3228
products (30) (-14.310) (12.881)

Industrial machinery and -1.2578790 .2443277 .1831
computer equipment (35) (-42.880) (23.652)

Electronic equipment (36) -1.3863860 .2560270 .2044
 (-49.234) (26.280)

Communications (48) -1.3278960 .3576900 .3413
 (-8.035) (9.685)

Business services (73) -1.7117120 .2337250 .1449
 (-51.510) (17.325)

All industries -1.2544790 .2731555 .2267
 (-102.166) (65.665)

(2-digit SIC Code) F Sample size

Oil and gas (13) 16.46 93

Chemicals and allied 705.07 1684
products (28)

Petroleum refining (29) 53.31 225

Rubber and plastic 165.91 350
products (30)

Industrial machinery and 559.41 2497
computer equipment (35)

Electronic equipment (36) 690.65 2691

Communications (48) 93.80 183

Business services (73) 300.16 1774

All industries 4311.83 14710

Exhibit V
Effects of R&D on EPS for petroleum-related companies
Top vs. Bottom by R&D intensity

Petroleum-related Sample
companies Intercept R&D [R.sup.2] F size

Bottom R&D intensity -.928864 .313284 .2184 24.31 89
 (-3.241) (4.930)

Top R&D intensity -1.117161 .269114 .1022 10.36 93
 (-2.421) (3.218)

Combined groups -.728199 .227378 .1216 24.92 182
 (-3.166) (4.992)