Firm Strategy and Outcome Uncertainty in R&D Firms

Abstract

We evaluate the impact of firm strategy on the variability of future performance for R&D firms and how firm strategy mediates the relation between R&D expenditures and firm outcome uncertainty of sales revenue, earnings, and operating cash flows. Following prior literature, we run exploratory factor analysis using resource allocations in the past towards intended strategy to measure the realized strategy pursued by firms. We find that, in R&D firms, differentiation strategy leads to lower variability of future sales, earnings, and operating cash flows. In contrast, cost leadership strategy leads to higher variability of future sales and operating cash flows, and lower variability of future earnings. Our study is the first, to our knowledge, to empirically document the impact of firm strategy of R&D firms on the variability of various future performance measures. Using mediation analysis, we further document that differentiation strategy negatively mediates the association between R&D expenditures and variability of sales revenue, earnings, and operating cash flows. While cost leadership strategy negatively mediates the association between R&D expenditures and variability of sales and variability cash flows, it positively mediates the association between R&D expenditures and variability of earnings.

1. Introduction

Corporate research and development (R&D) activities are generally considered to be riskier than most other investments, such as capital expenditures. Prior research has documented a positive association between firm outcome uncertainty and R&D expenditures (e.g., Kothari et al., 2002; Lev et al., 2021). In this study, we examine the effect of firm strategy on uncertainty of firm outcomes and the mediating role of firm strategy on the association between R&D expenditures and outcome uncertainty for a sample of R&D firms. Strategies are the investment and operating choices made by firms to achieve a competitive advantage. They play a critical role in firm performance and risk. A better understanding of their impact can shed light on the contexts in which capitalization (or not) of R&D expenditures could be more appropriate.

Following the framework in Porter (1980), we focus on two strategies that firms pursue: differentiation and cost leadership. A differentiation strategy is characterized by offering a product or service that has unique attributes that customers value and prefer over the products of competition. As a strategy built on products or services that are perceived to be unique and different from those of the competitors, it takes longer to be imitated, and hence would likely lead to more sustainable and stable performance. In contrast, firms adopting a cost leadership strategy achieve competitive advantage by becoming a low-cost producer in its industry, which may be achieved by pursuing economies of scale through improvement in process efficiencies, propriety technology, and/or preferential access to lower cost materials. Technological innovations enable the quick dissemination of best practices that enhance cost efficiency, and hence a cost leadership strategy is easily imitable (Porter, 1996, 2001).

We follow prior literature to measure the strategy variables by using archival financial statement data to assess the strategies that firms have carried out (Balsam et al., 2011; Banker et al., 2014). To measure an R&D firm’s strategic orientation, we run exploratory factor analysis using realized resource allocations in the past towards intended strategy reflected in six variables identified by Balsam et al. (2011) and Banker et al. (2014): the ratio of the selling, general, and administrative expenses to sales (SGA/SALE), the ratio of the research and development expenses to sales (RND/SALE), the ratio of sales to cost of goods sold (SALE/COGS), the ratio of sales to capital expenditures (SALE/CAPX), the ratio of sales to book value of property, plant, and equipment (SALE/PPE), and the ratio of the number of employees to total assets (EMP/ASSETS). A differentiation strategy is measured by the factor loading of the SGA/SALE, RND/SALE, and SALE/COGS ratios, and a cost leadership strategy is measured by the factor loading of the other three ratios, SALE/CAPX, SALE/PPE, and EMP/ASSETS.

Using a sample of 22,407 firm-year observations of R&D firms listed in Compustat between 1987–2018, we examine the effect of firm strategy on variability of sales, earnings, and cash flows in the five years following a given strategy, as well as the mediating role of firm strategy on the association between R&D expenditures and firm outcome uncertainty. We find that firms pursuing a differentiation strategy have lower variability of future sales, earnings, and cash flows over the subsequent five years. This finding aligns with the idea that differentiation enables a firm to maintain its competitive advantage by offering unique features that are difficult for competitors to replicate, which reduces market uncertainty. While the ongoing investments required to sustain these advantages can create uncertainty around future profitability and cash flows, as a firm following a differentiation strategy must continuously innovate to keep its products or services distinct to support premium pricing and customer loyalty, our results suggest that such a firm still achieves more stable future earnings and cash flows through their consistent innovation. The significant expenses entailed by a differentiation strategy for activities such as advertising and employee training to effectively communicate the value of these unique features to both current and potential customers may have well aligned with the revenue and cash inflows through consistent innovation efforts under this strategy.

Turning to cost leadership strategy, we find that it leads to higher outcome uncertainty measured by variability of future sales and variability of future cash flows. This finding supports the argument in prior literature that the best practices to enhance process and cost efficiency could be easily disseminated and imitated, causing the sales to fluctuate. In addition, since the success of a cost leadership strategy often relies on achieving a high sales volume, this strategy requires significant upfront investments in production and/or distribution (Kennedy, 2020), which would in turn increase variability of future earnings and cash flows. Contrary to our expectations, we find that firms pursuing a cost leadership strategy have lower (albeit with a small economic significance) variability of future earnings over the subsequent five years. Our study is, to our knowledge, the first to empirically document the impact of a firm’s strategy on the variability of future performance.

Using mediation analysis, we further document that the positive association between R&D expenditures and firm outcome uncertainty documented by the prior literature is mediated by both differentiation and cost leadership strategies. Specifically, we document that for R&D firms, both differentiation and cost leadership strategies negatively mediate the association between R&D expenditures and variability of sales revenue and cash flows. While pursuing a differentiation strategy negatively mediates the association between R&D expenditures and variability of earnings, pursuing a cost leadership strategy positively mediates this association. Our results also suggest that R&D incrementally influences the variability of future firm outcome, extending beyond its contribution as a key component of a firm’s strategic orientation. Our results have implications for managers, firm strategists, investors, researchers, and standard setters who are interested in understanding the differential implications of R&D expenditures on firm outcomes in the context of firm strategic orientation.

The outline for the remainder of the paper is as follows. Section 2 provides the literature review and develops the hypotheses. Section 3 presents the research design and results of the paper. Section 4 concludes with discussions.

2. Materials and Methods

2.1. Literature Review

Companies with sustainable competitive advantages tend to outperform their rivals. Strategic choices guide firms in making investment and operational decisions aimed at achieving such advantages. These strategies are typically established early in the firm’s lifecycle and require significant commitments of physical, financial, and intellectual resources. As a result, a firm’s strategic direction often remains consistent over time (Snow & Hambrick, 1980; Bentley et al., 2019). Numerous frameworks in the literature explain how businesses seek to secure a competitive edge within their industries. Porter (1980) outlines two key approaches firms can adopt to gain an advantage: cost leadership and differentiation. Success can be achieved by effectively pursuing either strategy, a concept widely embraced in both academic research and practical applications. (Dess & Davis, 1984; Porter, 1985, 2000, 2001; Miller & Dess, 1993; Allen, 2007; Balsam et al., 2011; Asdemir et al., 2013; Banker et al., 2014).

A differentiation strategy revolves around offering products or services with unique attributes that resonate strongly with customers, making them favor these offerings over competitors’. To achieve differentiation, firms typically invest significantly in research and development, branding, marketing, and other intangible assets that enhance the perceived value of their products. This strategic focus allows the firm to command premium prices, resulting in higher profit margins per unit.

In contrast, a cost leadership strategy emphasizes gaining a competitive edge by being the lowest-cost producer in the industry. This advantage is often achieved through economies of scale, process optimization, proprietary technologies, or preferential access to cost-effective resources. By focusing on cost minimization, these firms accept lower unit margins in exchange for high sales volumes, leveraging their low prices to attract a broad customer base.

However, firms attempting to pursue both differentiation and cost leadership strategies simultaneously often fail to excel in either, a phenomenon described by Porter (1980, 1985) as being “stuck in the middle”. Such firms struggle to establish a clear market position, as they lack a distinct competitive advantage and fail to cater effectively to any specific customer segment. This strategy dilution results in inefficiencies and reduced profitability.

Empirical evidence supports the notion that clarity in strategic focus drives superior performance. Studies by Dess and Davis (1984) and Thornhill and White (2007) demonstrate that firms dedicated to a well-defined generic strategy—whether differentiation or cost leadership—consistently outperform those without a clear strategic direction. These findings underscore the importance of committing to a single, coherent strategy to achieve long-term success in competitive markets.

It is worth noting that there are other ways of classifying business strategies. For example, Miles and Snow (2003) introduce three strategic types of business organizations that may exist within industries: prospectors, defenders, and analyzers. March (1991) describes business strategies as making explicit and implicit choices between exploration and exploitation. Treacy and Wiersema (1995) describe business strategies in terms of becoming a leader in one of the three value disciplines: operational excellence, product leadership, and customer intimacy. While the labels for business strategies differ across the various typologies, a common feature of all the proposed strategy classifications is that they most clearly identify companies that operate at one end or the other of a strategy continuum (Langfield-Smith, 1997).

R&D is a crucial component of innovation, a key factor in developing new competitive advantages, and an important productive input for a significant number of firms (Aboody & Lev, 2000). The EU Industrial R&D Investment Scoreboard (Scoreboard), which comprises of the world’s top 2500 R&D investors, reports that the 2022 Scoreboard firms increased their investments in research and development significantly by 14.8% during 2022, in contrast to a decline of 2.2% in 2021. The level of R&D spendings was above the pre-pandemic levels, exceeding the trillion-dollar mark for the first time. The U.S. and Chinese companies included in the Scoreboard in 2022 increased their R&D spendings by 16.5% and 24.9%, respectively. During this time, their EU counterparts saw an increase in their expenditures by 8.9%. Scoreboard attributes the enduring growth in global industrial R&D spendings for the twelfth consecutive year to the strategic significance of these investments. The Conference Board’s CEO challenge survey reported that managing innovation was consistently ranked one of the top five global challenges a decade ago (Mitchell et al., 2012, 2013, 2014). In the latest 2023 survey, CEOs identified recession, inflation, and disruptions related to COVID-19 as their foremost concerns (Mitchell et al., 2023). CEOs highlight the speeding up of innovation and digital transformation as one of their four overarching strategies for addressing a recession in their immediate plans.

While R&D is most often associated with innovation, R&D firms could pursue either a differentiation or cost leadership strategy. Firms invest in product R&D to improve the quality of existing products and create new products, and in process R&D to lower the cost of making existing products. There is evidence that the share of process R&D tends to rise with the size. Additionally, it has been observed that firms devote an increasing share of their R&D efforts to process R&D over the life cycle of a product. Klepper (1996) dubs this as one of the defining characteristics of the product life cycle model of industry evolution in industries that have rich opportunities for both product and process R&D. In addition, the composition of buyers of a product changes over a firm’s life cycle. Initial buyers of a product usually care primarily about the features or the quality of the product and are willing to pay for improvements in product quality. Over time, however, a product is usually sold increasingly to consumers who care more about the price over the quality of the product.

While R&D is considered as an asset by investors (Hirschey, 1982; Lev & Sougiannis, 1996) and is strategically important, it is shown to be much riskier than investments in fixed assets. Concerned with the uncertainty of benefits from R&D, the U.S. Generally Accepted Accounting Principles (GAAP) requires firms to expense R&D other than certain software costs. Consistent with this concern, Kothari et al. (2002) find that R&D expenditures contribute to uncertainty in future earnings four times more than capital expenditures. This relationship is more pronounced in R&D-intensive industries (Amir et al., 2007). In addition, firms that have more productive R&D (process-related R&D) exhibit less volatility in future operating performance (Pandit et al., 2011), which is still higher than the performance variability stemming from capital expenditures. Future earnings variability generated by R&D is smaller for larger firms (Ciftci & Cready, 2011) and firms with high R&D intensity (Ciftci et al., 2011). For bondholders, when credit risk associated with R&D-intensive firms is considered, the risk and uncertainty of R&D dominate the future benefits (Shi, 2003). Furthermore, excess returns to R&D-intensive firms might be due to the risk associated with R&D expenditures (Chambers et al., 2002). Given the importance of strategic orientation in determining the nature and cycle of R&D, how firm strategies shape the relation between R&D and uncertainty of firm outcomes warrants further investigation.

2.2. Hypothesis Development

Differentiation strategy, which is achieved through unique products or services, helps firms attain a more sustainable competitive advantage since such attributes cannot be easily replicated by rivals (Grant, 1991). A differentiation strategy typically involves innovations and distinctive marketing strategies that are challenging to replicate rapidly. While the response to pricing is almost immediate, responses to innovation through R&D take a longer period (Asdemir et al., 2013). While the length of the time period during which a firm would be able to sustain its advantages depends on how long it takes for its competitors to respond, we expect that, in general, differentiation can help a firm reduce product market uncertainty in a short-to-medium time window following the adoption of such a strategy. Prior studies have shown that R&D expenditures are positively associated with volatility of subsequent sales (Lev et al., 2021). Since R&D often plays a key role in shaping and driving an organization’s vision for the future, it is important to understand how differentiation affects the relation between R&D expenditures and firm outcome uncertainty.

A crucial part of developing an effective R&D strategy is aligning its outcomes with the overall business strategy. When there is a disconnect between the business and R&D, and the organization is not structured to leverage or act on the insights generated by R&D, these strategic efforts can fail. Pisano (2012) outlines three vital purposes of a well-defined and -executed strategy: consistency, coherence, and alignment. Consistency refers to the idea that competitive advantage results from a series of decisions and actions over time, with a strong strategy providing a framework for making consistent choices that lead to a desired outcome. Coherence highlights the need for integration within complex organizations, where daily decisions regarding resource allocations are often made in disparate locations and can lack coordination. A clear strategy ensures that these tactical decisions align with overarching goals. Lastly, alignment emphasizes that organizations perform best when their strategies resonate with the external environment and overall business context. For R&D units, it is essential to align their strategies with the broader organizational strategy to foster effective collaboration and success. In other words, the R&D and corporate strategy teams must collaborate closely (Brennan et al., 2020). Corporate strategy aligns R&D with company priorities, while R&D explores technical possibilities. Both strategies must sync, addressing key questions such as company goals and the role of R&D. The strategy for an R&D organization can be categorized into four main components: architecture, processes, people, and portfolio (Pisano, 2012). The architecture of an R&D system encompasses key decisions regarding its organizational and geographical structure. Processes include both formal and informal methods for conducting R&D activities. People are a critical component, significantly influencing the R&D system’s effectiveness. The portfolio refers to the desired distribution of resources across various R&D projects and the criteria used to prioritize and select these projects.

A well-defined and well-executed corporate strategy can reduce the uncertainty of future performance caused by R&D expenditures by ensuring that the R&D strategy is consistent, coherent, and aligned with the company’s overall goals.

We posit that a differentiation strategy mitigates uncertainty in future sales and that the positive impact of R&D spending on sales variability is likely mediated by a differentiation strategy, which works by aligning R&D investments with the overall business strategy. In other words, when R&D efforts are closely integrated with the company’s broader strategic goals, the resulting innovations and unique products can drive sales variability in a more favorable way. The hypotheses are summarized below.

H1a: 

R&D firms pursuing a differentiation strategy have lower variability of future sales.

H1b: 

Differentiation strategy mediates the effect of R&D expenditures on the variability of future sales of R&D firms.

The methods of a cost leadership strategy to achieve operational efficiency can typically be replicated (D’Aveni, 1994). Therefore, competitive advantage achieved through such strategies will be temporary, and profitability over the long term does not persist (Eisenhardt & Brown, 1998; Eisenhardt & Martin, 2000). A cost leadership strategy which is based on generic solutions related to operational efficiency is more prone to imitation by competitors, resulting in comparative cost advantages that will diminish over time (Asdemir et al., 2013). Cost leadership is not likely to yield a sustainable competitive advantage, especially if the sources of it are developed by suppliers and sold on the open market (Barney, 2002). Being first with a new process only provides a firm with a temporary cost advantage because imitation is inevitable (Murray, 1988). In this light, firms pursuing a cost leadership strategy are expected to have a higher variability of future sales, given the uncertainty of being imitated within a short period. We posit that a cost leadership strategy increases uncertainty in future sales and also, similarly to differentiation strategy, mediates the positive effect of R&D expenditures on variability of sales. Formally put:

H1c: 

R&D firms pursuing a cost leadership strategy have higher variability of future sales.

H1d: 

Cost leadership strategy mediates (positively or negatively) the effect of R&D expenditures on the variability of future sales of R&D firms.

Unique strategic advantages of a firm are replicated and even improved upon by competitors over time. This is because rival firms can often gather information about their competitors’ innovation activities and make an effort to benefit from these activities. Competitor and competitive information are generally available to all firms, and new techniques diffuse rapidly (Barney, 1986). Surveys show that information on product development decisions is generally in the hands of rivals within twelve to eighteen months (e.g., Mansfield, 1985). The information on the detailed nature and operation of the new products or processes developed by a firm is obtained by at least some of its rivals within a year. Hence, a firm’s ability to achieve sustained competitive advantage depends how successful it can deter attempts to replicate its key strategy by competitors (Ghemawat, 1995). Such efforts could incur investments with high risk and cost, leading to elevated variability of future earnings and cash flows. For example, stability of sales under a differentiation strategy may require constant investments to achieve technological innovations in order to sustain the unique features, and hence, the premium pricing of a firm’s products. The investment could also be reflected in enhanced advertising and training activities that promote effective communication with existing and potential customers about the value of the uniqueness of a firm’s products or services. Since research and development encompasses the activities that companies engage in to innovate, introduce new products and services, or enhance their existing offerings, R&D firms, through pursuing differentiation strategy, would achieve stability of future earnings and cash flows. Additionally, a differentiation strategy helps firms sustain and potentially enhance financial performance in the future (Banker et al., 2014). Therefore, we posit that differentiation reduces the variability of a firm’s future profitability and cash flows, and that the relation between R&D expenditures and earnings variability (or cash flow variability) is mediated by firm strategy.

Similarly, firms pursuing a cost leadership strategy need significant investments in production and/or distribution (Kennedy, 2020). Improved financial performance from following a cost leadership strategy in the current year fades over time, and cost leadership does not facilitate firms to maintain future performance as effectively as the differentiation strategy. Therefore, we posit that cost leadership strategies increase the variability of a firm’s future earnings and cash flows. In addition, the relation between R&D expenditures and earnings variability (or cash flow variability) is mediated by firm strategy. Formally stated:

H2a: 

R&D firms pursuing a differentiation strategy have lower variability of future earnings and cash flows.

H2b: 

Differentiation strategy mediates (positively or negatively) the effect of R&D expenditures on variability of future earnings and cash flows of R&D firms.

H2c: 

R&D firms pursuing a cost leadership strategy have higher variability of future earnings and cash flows.

H2d: 

Cost leadership strategy mediates (positively or negatively) the effect of R&D expenditures on variability of future earnings and cash flows of R&D firms.

The following figures illustrate the main hypotheses. Figure 1 depicts Hypothesis 1a, 1b, 1c, and 1d. Figure 2a illustrates Hypothesis 2a and 2b, while Figure 2b depicts Hypothesis 2c and 2d.

2.3. Mediation Model

New product development through product R&D is at the core of a differentiation strategy, while process improvement through process R&D is essential to a cost leadership strategy. Hence, R&D expenditures are components of the broader business strategy a firm pursues. In the previous section, we hypothesize that the strategies an R&D firm pursues affect variability of future sales, earnings, and cash flows, and that firm strategies mediate the relation between R&D expenditures and uncertainty of firm outcomes.

We use mediation analysis to explore how firm strategy mediates the relationship between R&D and firm outcome uncertainty. Mediation analysis is utilized to examine the process underlying an established relationship between independent variables and a dependent variable (MacKinnon, 2008). To test our hypotheses and explore the mediating impact of firm strategies, we estimate the following models (firm subscript is suppressed):

Differentiation_t = b₀ + b_1t RND_t + b_2t CAPX_t + b_it Control Variables_t + error_t

(1a)

CostLeadership_t = c₀ + c_1t RND_t + c_2t CAPX_t + c_it Control Variables_t + error_t

(1a′)

Uncertainty_{t+1, t+5} = d₀ + d_1t Differentiation_t + d_2t CostLeadership_t + d_3tADV_t
+ d_4t LEVERAGE_t + d_5t Log(MVE)_t + error_t

(1b)

Uncertainty_{t+1, t+5} = g₀ + g_1t Differentiation_t + g_2t CostLeadership_t + g_3t RND_t + g_4t CAPX_t
+ g_5t ADV_t + g_6t LEVERAGE_t + g_7t Log(MVE)_t + error_t

(1c)

We include industry and year fixed effects in our models. In the above regressions, RNDt is defined as the research and development expenditures in a fiscal year t. CAPXt is the capital expenditures, and ADVt is the advertising expense. All three variables are deflated by the beginning book value of equity. Uncertainty_{t+1, t+5} is the standard deviation of sales revenue (StdSale_{t+1, t+5}), earnings before extraordinary items (StdEarn_{t+1, t+5}), or operating cash flows (StdCash_{t+1, t+5}) for years t + 1 to t + 5. Each observation of sales revenue, earnings, and operating cash flows is deflated by book value of equity at the beginning of the fiscal year. These three Uncertainty variables capture different aspects of firm risk and firm outcome uncertainty. Sales revenue provides a measure of the realized demand for the firm’s products and services, and future sales variability measures product-market uncertainty. Future earnings variability and cash flow variability measure overall uncertainty in business operations. Investors’ and creditors’ expectations about returns depend on their assessment of the amount, timing, and uncertainty of future earnings and net cash inflows to the entity (FASB SFAC No. 8). Hence, they comprise elements of competition risk (Lev et al., 2021).

Our measures of the strategies that an R&D firm pursues to achieve competitive advantage are based on Porter’s (1980, 1996) typology: differentiation (Differentiation) and cost leadership (CostLeadership). Strategy could be defined as general intentions or defined schemes, as well as realized outcomes (Mintzberg, 1987; Banker et al., 2014). Consequently, Mintzberg (1987) defines strategy as a pattern. Since strategic decisions involve making resource allocations and subsequently impact the reported numbers (Banker et al., 2014), a firm’s intended strategy could be largely deduced from information reported in its financial statements. Accordingly, following prior literature, we use archival financial statement data to assess the strategies that firms have carried out (Banker et al., 2014). Although numerous previous studies have used perceptual measures obtained through surveys to assess intended strategy, issues of perceptual biases have been cited in the literature (Reger & Huff, 1993; David et al., 2002). Using resource allocations in the past towards intended strategy does not suffer from perceptual biases. We run exploratory factor analysis using six variables identified by Balsam et al. (2011) and Banker et al. (2014) to measure an R&D firm’s strategic orientation.

The ratio of the selling, general, and administrative expenses to sales (SGA/SALE) represents a firm’s investments to differentiate its products and services from its rivals (Berman et al., 1999; David et al., 2002; Miller & Dess, 1993; Thomas et al., 1991). The ratio of the research and development expenses to sales (RND/SALE) represents a firm’s commitment to creating new products and services, improving the existing ones, and developing innovative processes. The ratio of sales to cost of goods sold (SALE/COGS) represents either a firm’s ability to charge premium prices for its differentiated products or services (Kotha & Nair, 1995; Nair & Filer, 2003), or its efficiency in managing costs (Hambrick, 1983; Berman et al., 1999). The ratio of sales to capital expenditures (SALE/CAPX), the ratio of sales to book value of property, plant, and equipment (SALE/PPE), and the ratio of number of employees to total assets (EMP/ASSETS) represent a firm’s investment into resources to excel in operational efficiency (Berman et al., 1999; Hambrick, 1983; Kotha & Nair, 1995; Miller & Dess, 1993; Nair & Filer, 2003).

We calculate the average of the previous five years of observations for each of the above ratios only for the first firm-year in our sample. Since firm strategy is chosen early on in the life of the firm, we did not calculate these averages on a rolling basis for every firm-year observation. For example, for a firm that is in our sample in 1986 first, we calculate the average of each ratio during the years 1981–1985 only. For another firm that is in our sample in 1991 first, we calculate the average of each ratio during the years 1986–1990 only. The results of the exploratory factor analysis implemented indicate that SGA/SALE, RND/SALE, and SALE/COGS ratios load together on one factor, which we identify as Differentiation_t. The other three ratios, SALE/CAPX, SALE/PPE, and EMP/ASSETS, load together on another factor, which we identify as CostLeadership_t. We also carry out the exploratory factor analysis after calculating the ratios, using the average of the previous five-year observation on a rolling five-year basis. For example, for the year 1986, we calculate the average of each ratio during the years 1981–1985, and for the year 1987, we calculate the average of each ratio during the years 1982–1986. The variables loading for Differentiation_t and CostLeadership_t remain the same, and the results of the analysis using factor scores based on the five-year rolling average remain qualitatively similar to the earlier results, using the factor scores with the average of the ratios only for the first firm-year in our sample.

3. Results

3.1. Sample and Descriptive Statistics

3.1.1. Sample

We obtain financial data from the Compustat Annual Industrial and Annual Research files for the period 1986–2018. For each year t from 1986 to 2018, we retain all observations with non-missing data for the following variables: RND_t, research and development expense (Compustat data XRD), with a zero amount reported not treated as a missing value; CAPX_t, capital expenditures (Compustat data CAPX); ADV_t, advertising expense (Compustat data XAD), with missing values set to zero; Log(MVE)_t, the market value of equity, measured as the natural logarithm of the product of the fiscal-year-end price and common shares outstanding [log(PRCC_F*CSHPRI)]; LEVERAGE_t, the sum of long-term debt (data item DLTT) and debt in current liabilities (data item DLC), divided by the sum of long-term debt and the market value of equity; Earn_t, the earnings before extraordinary items and discontinued operations available for common shareholders (data item IBCOM); Sale_t, sales revenue (data item SALE); Cash_t, cash flows from operations (data item OANCF); operating income after depreciation (data OIADP); and total Assets (data item AT). Our sample period starts from 1986 as the first year when operating cash flows data became available is 1987. The sample period ends in 2018 because variability of future performance is calculated using data for five consecutive years. The final sample consists of a total of 22,407 firm-year observations.

Research intensive firms are likely to go through mergers and acquisitions and not survive the five-year future period required to calculate variability of future benefits (Pandit et al., 2011). This would lead to survivorship bias. To avoid such a bias, following Amir et al. (2007), for a given R&D firm, if data are unavailable for the subsequent five years, we calculate the variable using the data from the current year and the subsequent four years. If this variable is still missing, we use the lagged, current, and subsequent three years of data to calculate the variability measures.

3.1.2. Descriptive Statistics

Table 1. Descriptive statistics.

Variable *	N	Mean	Std Deviation	Q1	Median	Q3
StdEarnt+5	22,407	0.081	0.073	0.033	0.058	0.100
StdCasht+5	22,407	0.107	0.102	0.047	0.077	0.100
StdSalest+5	22,407	0.583	0.883	0.170	0.321	0.128
Differentiationt	22,407	0.276	1.126	−0.505	−0.057	0.629
CostLeadershipt	22,407	0.112	0.523	−0.033	0.271	0.781
RNDt	22,407	0.079	0.097	0.008	0.047	0.428
CAPXt	22,407	0.135	0.131	0.050	0.097	0.114
ADVTt	22,407	0.036	0.088	0.000	0.000	0.173
Log (MVE)t	22,407	−0.196	2.001	−1.611	−0.105	0.029
Leveraget	22,407	0.173	0.175	0.025	0.126	1.244

* Firm subscripts are omitted.

presents the descriptive statistics for our sample. Firms, on average, spend 7.49% of their book value on R&D expenditures. Average spending for capital expenditures (CAPX) is 13.5% of book value. The average (median) standard deviation of future earnings (StdEarn) is 0.081 (0.058), while the average (median) standard deviation of future cash flows (StdCash) is 0.107 (0.077). The average (median) standard deviation of future sales revenue (StdSales) is 0.583 (0.321), much higher than those of earnings and cash flows.

3.1.3. Correlations

Table 2. Correlations *.

	StdEarnt+5	StdCasht+5	StdSalest+5	Differentiationt	Cost Leadershipt	RNDt	CAPXt	ADVTt	Log(MVE)t	Leveraget
StdEarnt+5		0.494 <0.0001	0.473 <0.0001	0.303 <0.0001	−0.062 <0.0001	0.194 <0.0001	0.129 <0.0001	0.053 <0.0001	0.038 <0.0001	0.174 <0.0001
StdCasht+5	0.517 <0.0001		0.564 <0.0001	−0.125 <0.0001	0.124 <0.0001	0.166 <0.0001	0.226 <0.0001	0.148 <0.0001	−0.095 <0.0001	0.228 <0.0001
StdSalest+5	0.494 <0.0001	0.583 <0.0001		−0.273 <0.0001	0.225 <0.0001	0.035 <0.0001	0.248 <0.0001	0.104 <0.0001	−0.049 <0.0001	0.294 <0.0001
Differentiationt	0.034 <0.0001	−0.150 <0.0001	−0.376 <0.0001		−0.554 <0.0001	0.203 <0.0001	−0.102 <0.0001	−0.034 <0.0001	0.272 <0.0001	−0.291 <0.0001
CostLeadershipt	−0.067 <0.0001	0.157 <0.0001	0.340 <0.0001	−0.553 <0.0001		−0.239 <0.0001	0.082 <0.0001	0.155 <0.0001	−0.266 <0.0001	−0.100 <0.0001
RNDt	0.172 <0.0001	0.072 <0.0001	−0.062 <0.0001	0.279 <0.0001	−0.362 <0.0001		0.042 <0.0001	0.000 0.9999	0.175 <0.0001	−0.097 <0.0001
CAPXt	0.147 <0.0001	0.203 <0.0001	0.299 <0.0001	−0.125 <0.0001	0.121 <0.0001	−0.054 0.0034		0.165 <0.0001	0.092 <0.0001	0.276 <0.0001
ADVTt	0.018 0.0202	0.077 <0.0001	0.048 <0.0001	−0.014 0.0654	0.130 <0.0001	−0.073 <0.0001	0.057 <0.0001		0.058 <0.0001	0.032 <0.0001
Log (MVE)t	0.039 <0.0001	−0.132 <0.0001	−0.052 <0.0001	0.265 <0.0001	−0.272 <0.0001	0.166 <0.0001	0.148 <0.0001	0.050 <0.0001		−0.111 <0.0001
Leveraget	0.148 <0.0001	0.184 <0.0001	0.313 <0.0001	−0.310 <0.0001	−0.093 <0.0001	−0.106 <0.0001	0.289 <0.0001	−0.055 <0.0001	−0.020 0.0088

* Upper (lower) part of the table presents the Pearson (Spearman) correlations. p-values are provided under correlation values.

provides the univariate correlations among the regression variables. We present Pearson (Spearman) correlations above (below) the diagonal. Differentiation is positively correlated with future variability of earnings (StdEarn) and negatively correlated with future variability of cash flows (StdCash) and sales (StdSales). CostLeadership is negatively correlated with future variability of earnings (StdEarn) and positively correlated with future variability of cash flows (StdCash) and sales (StdSales). While RND is positively correlated with Differentiation, its correlation with CostLeadership is negative. In addition, Differentiation and CostLeadership are both negatively correlated with CAPX. Finally, Differentiation and CostLeadership are negatively correlated.

3.2. Regression and Mediation Results

Table 3. Mediation analysis—variability of future sales.

Model:
Differentiationt = b0 + b1t RNDt + b2t CAPXt + bit Control Variablest + errort					(1a)
CostLeadershipt = c0 + c1t RNDt + c2t CAPXt + cit Control Variablest + errort					(1a′)
Uncertaintyt+1, t+5 = d0 + d1t Differentiationt + d2t CostLeadershipt + d3tADVt + d4t LEVERAGEt + d5t Log(MVE)t + errort					(1b)
Uncertaintyt+1, t+5 = g0 + g1t Differentiationt + g2t CostLeadershipt + g3t RNDt + g4t CAPXt + g5t ADVt + g6t LEVERAGEt + g7t Log(MVE)t + errort					(1c)
	Coeff	t-Stat	p-Value	99% CI
RND
g1	−0.061	−466.31	<0.0001
g2	0.110	476.18	<0.0001
g3	1.392	550.91	<0.0001
b1	4.430	991.50	<0.0001
c1	−0.331	−106.55	<0.0001
b1g1 (Mediation of Differentiation)	−0.2682	−434.42	<0.0001	−0.2698	−0.2666
c1g2 (Mediation of CostLeadership)	−0.0362	−105.91	<0.0001	−0.0371	−0.0353
CAPX
b2	−0.652	−111.71	<0.0001
c2	−0.476	−147.38	<0.0001
b2g1 (Mediation of Differentiation)	0.0394	109.51	<0.0001	0.0386	0.0404
c2g2 (Mediation of CostLeadership)	−0.0523	−138.38	<0.0001	−0.0532	−0.0513

reports the results from the mediation analysis of Hypothesis 1 on variability of future sales. We present the regression results with standard deviation of future sales (StdSales) as a measure of outcome uncertainty. Our results suggest that RND has a significant and positive association with Differentiation in Model 1a and a significant and negative association with CostLeadership, providing initial validation for the two strategy variables to serve as mediators in the regression system. Differentiation is negatively associated with StdSales. The coefficient estimate g₁ is −0.061 and significant at the 1% level, supporting our hypothesis that a differentiation strategy reduces uncertainty in product market captured by variability of future sales. The bootstrapping results for the mediation tests also provide support for Hypothesis 1b that differentiation strategy mediates the positive relation between R&D expenditures and variability of sales. The product of coefficients, b₁g₁, is −0.2682 and significant at the 1% level. And zero is not in the relevant interval at the 99% confidence level, which indicates that the indirect effect tested above is significantly different from zero at the 1% level (Preacher & Hayes, 2004). By emphasizing unique product attributes and creating value that resonates with customers, firms can reduce some of the unpredictability linked to their investments in innovation. However, this strategy does not entirely eliminate the inherent risks and variability tied to R&D activities, as factors such as market reception, technological advancements, and competitive responses continue to play a role in shaping outcomes. On the other hand, CAPX has a significant and negative association with Differentiation in Model 1a, and similarly, a significant and negative association with CostLeadership. Differentiation strategy positively mediates the positive relation between capital expenditures and variability of sales. The product of coefficients, b₂g₁, is 0.0394 and significant at the 1% level.

CostLeadership is positively associated with StdSales. The coefficient estimate g₂ is 0.110 and significant at the 1% level, consistent with the prediction that a cost leadership strategy is associated with greater uncertainty in product market captured by variability of future sales. The bootstrapping results for the mediation tests confirm that cost leadership strategy mediates the positive relation between R&D expenditures and variability of future sales negatively. The product of coefficients, c₁g₂, is −0.0362 and significant at the 1% level. The findings suggest that a cost leadership strategy helps lessen the uncertainty in future performance stemming from R&D expenditures. By focusing on operational efficiency, economies of scale, and cost minimization, firms pursuing this strategy are better positioned to absorb the risks associated with investments in research and development. These efficiencies create a more stable financial foundation, allowing the firm to navigate performance fluctuations more effectively. However, while cost leadership reduces some of the unpredictability, it does not entirely eliminate the inherent variability tied to R&D activities. Factors such as market demand, competitor innovation, and the success of cost-driven technological advancements still introduce elements of uncertainty that the strategy alone cannot fully address. On the other hand, cost leadership strategy also negatively mediates the positive relation between capital expenditures and variability of sales. The product of coefficients, c₂g₂, is −0.0523 and significant at the 1% level.

As discussed before, we utilize two proxies for uncertainty of overall performance, namely, the standard deviation of future earnings and the standard deviation of future operating cash flows, to test Hypotheses 2.

Results from the analysis of Hypothesis 2 are summarized in

Table 4. Mediation analysis—variability of future earnings.

Differentiationt = b0 + b1t RNDt + b2t CAPXt + bit Control Variablest + errort					(1a)
CostLeadershipt = c0 + c1t RNDt + c2t CAPXt + cit Control Variablest + errort					(1a′)
Uncertaintyt+1, t+5 = d0 + d1t Differentiationt + d2t CostLeadershipt + d3tADVt + d4t LEVERAGEt + d5t Log(MVE)t + errort					(1b)
Uncertaintyt+1, t+5 = g0 + g1t Differentiationt + g2t CostLeadershipt + g3t RNDt + g4t CAPXt + g5t ADVt + g6t LEVERAGEt + g7t Log(MVE)t + errort					(1c)
	Coeff	t-Stat	p-Value	99% CI
RND
g1	−0.001	−64.06	<0.0001
g2	−0.001	−19.95	<0.0001
g3	0.177	663.34	<0.0001
b1	4.694	748.22	<0.0001
c1	−0.044	−10.27	<0.0001
b1g1 (Mediation of Differentiation)	−0.0051	−63.58	<0.0001	−0.0053	−0.0049
c1g2 (Mediation of CostLeadership)	0.0000	4.43	<0.0001	0.0000	0.0000
CAPX
b2	−1.021	−155.29	<0.0001
c2	−0.551	−137.31	<0.0001
b2g1 (Mediation of Differentiation)	0.0011	57.44	<0.0001	0.0011	0.0012
c2g2 (Mediation of CostLeadership)	0.0003	19.31	<0.0001	0.0003	0.0004

and

Table 5. Mediation analysis—variability of future cash flows.

Differentiationt = b0 + b1t RNDt + b2t CAPXt + bit Control Variablest + errort					(1a)
CostLeadershipt = c0 + c1t RNDt + c2t CAPXt + cit Control Variablest + errort					(1a’)
Uncertaintyt+1, t+5 = d0 + d1t Differentiationt + d2t CostLeadershipt + d3tADVt + d4t LEVERAGEt + d5t Log(MVE)t + errort					(1b)
Uncertaintyt+1, t+5 = g0 + g1t Differentiationt + g2t CostLeadershipt + g3t RNDt + g4t CAPXt + g5t ADVt + g6t LEVERAGEt + g7t Log(MVE)t + errort					(1c)
	Coeff	t-Stat	p-Value	99% CI
RND
g1	−0.005	−266.21	<0.0001
g2	0.002	61.65	<0.0001
g3	0.234	678.71	<0.0001
b1	4.421	948.25	<0.0001
c1	−0.257	−105.48	<0.0001
b1g1 (Mediation of Differentiation)	−0.0224	−253.17	<0.0001	−0.0226	−0.0222
c1g2 (Mediation of CostLeadership)	−0.0005	−52.68	<0.0001	−0.0006	−0.0005
CAPX
b2	−0.654	−138.18	<0.0001
c2	−0.299	−90.11	<0.0001
b2g1 (Mediation of Differentiation)	0.0033	120.10	<0.0001	0.0032	0.0034
c2g2 (Mediation of CostLeadership)	−0.0006	−48.55	<0.0001	−0.0007	−0.0006

. Table 4 presents the regression results using variability of future earnings as a measure of firm outcome uncertainty, while Table 5 focuses on variability of cash flows.

In Table 4, the results provide support for Hypothesis 2a and 2c that both differentiation and cost leadership strategies lead to less uncertain future earnings. In Model 1c, the coefficient of Differentiation (g₁) is −0.0001 and significant at the 1% level, while the coefficient of CostLeadership (g₂) is −0.0001 and significant at the 1% level. The bootstrapping results for the mediation effect of differentiation strategy on the relation between R&D expenditures and earnings variability shows that the product of coefficients, b₁g₁, is −0.0051 and significant at the 1% level. The bootstrapping results for the mediation effect of cost leadership strategy shows that the product of coefficients, c₁g₂, is 0.0000 and not economically significant, albeit being statistically significant at the 1% level. Since zero is not in the relevant interval at the 99% confidence level, it indicates that the mediation effect tested is significantly different from zero at the 1% level. Taken together, for R&D firms, there is negative mediation of differentiation strategy on the effect of R&D expenditures on earnings variability and economically insignificant mediation of cost leadership strategy. The results on capital expenditure are also presented in Table 4. The mediation effect of differentiation strategy shows that the product of coefficients, b₂g₁, is 0.0011 and significant at the 1% level. The mediation effect of cost leadership strategy for capital expenditures regressions shows that the product of coefficients, c₂g₂, is 0.0003 and significant at the 1% level.

In Table 5, we use standard deviation of future cash flows as an alternative proxy for overall uncertainty in firm outcome. First, R&D expenditures exhibit a positive association with variability of future cash flows with the presence of the strategy mediators. The coefficient of RND (g₃) in Model 1c is 0.234 and significant at the 1% level. Differentiation is negatively associated with StdCash in Model 1c, as the coefficient estimate g₁ is −0.005 and significant at the 1% level. The bootstrapping results for the mediation tests shows that differentiation strategy negatively mediates the positive relation between R&D expenditures and variability of cash flows. The product of coefficients, b₁g₁, is −0.0224 and significant at the 1% level. CostLeadership is positively associated with StdCash, with a coefficient estimate (g₂) of 0.002, significant at the 1% level. In addition, the bootstrapping results for the mediation tests show that cost leadership strategy mediates the positive relation between R&D expenditures and variability of cash flows negatively. The product of coefficients, c₁g₂, is −0.0005 and significant at the 1% level. The results for capital expenditures and differentiation strategy are similar to those presented in Table 4. The mediation effect of cost leadership strategy for capital expenditures shows that the product of coefficients, c₂g₂, is −0.0006 and significant at the 1% level, indicating that cost leadership strategy mediates the positive relation between capital expenditures and variability of cash flows negatively.

Our results in Table 4 and Table 5 provide new insights into previous research, which show that a differentiation strategy is positively linked to market beta and negatively associated with the stability of future return on assets (Banker et al., 2014). While Banker et al. (2014) found a marginally significant association between cost leadership strategy and future return on assets stability, they did not find a significant association with market beta. We document a significantly positive association between cost leadership strategy and variability of operating cash flows. We further establish that R&D leads to volatility of future earnings, and operating cash flows beyond what is explained by a firm’s strategic orientation.

Taken together, we show that the strategy pursued by an R&D firm has implications for both competition risk and overall business risk. R&D firms pursuing a differentiation strategy have lower variability of future sales, lower variability of future earnings, and lower variability of future operating cash flows. Results with cost leadership are slightly different. To the degree a firm pursues cost leadership strategy, it has greater variability in future sales and cash flows and lower variability in future earnings. We also provide evidence that both differentiation and cost leadership strategies negatively mediate the association between R&D expenditures and variability of sales revenue and variability of cash flows. When it comes to variability of future earnings, its association with R&D expenditures is only mediated by differentiation strategy in an economically significant way.

3.3. Variable Definitions

StdEarn_{t+1, t+5} = the standard deviation of earnings before extraordinary items for years t + 1 to t + 5.

StdCash_{t+1, t+5} = the standard deviation of operating cash flows for year t + 1 to t + 5.

StdSale_{t+1, t+5} = the standard deviation of sales revenue for year t + 1 to t + 5.

Differentiation_t = the factor loading of Selling, General, and Administrative Expense to Sales; Research and Development Expenses to Sales; and Sales to Cost of Goods Sold ratios.

CostLeadership_t = the factor loading of Sales to Capital Expenditures, Sales to Plant, Property and Equipment, and Number of Employees to Total Assets ratios.

RND_t = the research and development expenditures in year t deflated by the beginning book value of equity.

CAPX_t = the capital expenditures in year t deflated by the beginning book value of equity.

ADV_t = the advertising expense in year t deflated by the beginning book value of equity.

Log(MVE)_t = the natural logarithm of the market capitalization of equity at the end of year t.

LEVERAGE_t = the ratio of long-term debt to the sum of market value of equity and long-term debt, both at the end of year t.

4. Discussion

In this study, we evaluate the impact of firm strategy on the variability of future firm outcomes for R&D firms and how firm strategy mediates the relation between R&D expenditures and firm outcome uncertainty. We consider three future firm outcomes: sales revenue, earnings, and operating cash flows, and calculate performance variability as the standard deviation of realized performance from five future years. We find that for R&D firms, differentiation strategy leads to lower variability of future sales, variability of future earnings, and variability of operating cash flows. In contrast, cost leadership strategy leads to higher variability of future sales and operating cash flows for R&D firms. Contrary to our expectations, a cost leadership strategy leads to lower variability in future earnings. This can be attributed to several advantages that cost leaders possess. Their lower production costs provide stable profit margins, insulating them from market fluctuations. The ability to adjust prices to stimulate demand offers a buffer against revenue downturns. Furthermore, streamlined operations and economies of scale contribute to more predictable cost structures and fewer operational shocks. Finally, the focus on price-sensitive, high-volume markets and less reliance on risky innovation also contribute to more stable revenue streams and earnings. To our knowledge, our study is the first to empirically document the impact of firm strategy on the variability of future performance. Using mediation analysis, we further document that both differentiation and cost leadership strategies serve as mediators in the association between R&D expenditures and firm outcome uncertainty. Specifically, both differentiation and cost leadership strategies negatively mediate the association between R&D expenditures and variability of sales revenue, as well as variability of cash flows. Differentiation strategy negatively mediates the association between R&D expenditures and variability of earnings, while cost leadership positively mediates the association. The evidence suggests that R&D incrementally influences the variability of future performance, extending beyond its contribution as a key component of a firm’s strategic orientation.

Our study contributes to the literature on R&D expenditures and highlights the important role a firm’s strategic orientation may play in shaping the risks and uncertainties associated with R&D expenditures and hence its relationship with an R&D firm’s future competition and performance risks. Given the valuation implications of R&D activities documented by prior studies (Hirschey & Weygandt, 1985; Sougiannis, 1994; Lev & Sougiannis, 1996), future research can explore the differential impact of firm strategy on market assessment of risk of R&D expenditures.