2.1 CEO Business Educational Background and R&D Investment

Upper echelons theory views corporate outcomes as reflecting the individuals who make decisions. Firm policies depend on how CEOs interpret the environment they face (Hambrick, 2007; Hambrick and Mason, 1984). CEOs do not simply react to external information. They interpret signals from their own perspectives and act accordingly. This explains why firms facing similar conditions may adopt different policies (Adner and Helfat, 2003; Bertrand and Schoar, 2003; Driouchi et al., 2022). This perspective is particularly relevant for corporate R&D investment. R&D is among the most uncertain forms of corporate investment. It requires substantial upfront commitments and produces outcomes that are difficult to verify ex ante (Czarnitzki and Toole, 2011; Hottenrott et al., 2016). These features make R&D decisions highly dependent on managerial judgment (Custódio and Metzger, 2014; Malmendier and Tate, 2005). As a result, CEO characteristics can lead to systematic differences in R&D investment behavior. Firms differ not only in how much they invest, but also in how they finance those investments (Benmelech and Frydman, 2015; Custódio et al., 2019; Graham et al., 2013). The CEO does not merely implement a predetermined policy. The CEO shapes how the firm responds to external conditions.

Among observable CEO traits, educational background is particularly relevant because it reflects how managers are trained to process information, evaluate risk, and make resource-allocation decisions (Barker and Mueller, 2002; Darmadi, 2013; Gottesman and Morey, 2006; Gounopoulos et al., 2021; Kallias et al., 2023). A business-related educational background may support R&D investment for several reasons. First, business education can strengthen managers’ ability to evaluate investment opportunities and allocate capital across competing uses. Such capabilities are particularly relevant for R&D because the expected returns from R&D projects are difficult to assess ex ante and often materialize only over long horizons. CEOs with stronger training in valuation and capital allocation may therefore be better equipped to distinguish promising R&D projects from less attractive ones and to commit resources despite uncertainty (Dong et al., 2025; Graham et al., 2013).

Second, business education may facilitate the financing of R&D investment. R&D projects typically require sustained resource commitments before their economic benefits become observable, making access to and management of financial resources particularly important. Prior studies suggest that CEO educational background is related to firms’ financing capacity and investment decisions (Gounopoulos et al., 2021; Gupta, 2022; Gupta et al., 2021). If business-educated CEOs are better able to secure external capital and manage internal financial resources, they may be more willing and able to undertake R&D projects that would otherwise be difficult to finance. Korean evidence also suggests that CEO educational characteristics are associated with firms’ R&D investment behavior (Ha and Lee, 2014). Taken together, these arguments provide a theoretical basis for a positive relation between CEO business educational background and R&D investment: business education may equip CEOs with the evaluation, resource-allocation, and financing capabilities needed to undertake uncertain, long-horizon R&D projects.

However, a business-related educational background may discourage R&D investment. Greater emphasis on measurable performance and short-term outcomes can reduce willingness to commit resources to projects with uncertain and delayed returns (Christensen et al., 2007; Miller and Xu, 2019). Because R&D outcomes are difficult to evaluate and may generate short-term earnings pressure, business-trained CEOs may prefer investments with clearer and more immediate financial payoffs (Kim, 2025). Consistent with this view, Loukil and Yousfi (2022) find that CEOs with business degrees are less likely to engage in innovative activities.

Taken together, prior studies point to an ambiguous relation between business-related educational backgrounds and R&D investment. CEO business education may improve capital allocation and access to financing, yet it may also increase reluctance to invest in uncertain long-horizon projects.

2.2 Stock Overvaluation and R&D Investment

Prior research shows that stock overvaluation can influence real corporate decisions, particularly when firms rely on external financing or when managers have incentives to respond to firm valuations (Morck et al., 1990). One prominent explanation is the equity-financing channel. Overvaluation lowers the effective cost of external capital and allows firms to issue equity on favorable terms. This can enable firms to invest beyond what their internal resources would otherwise permit. Stein (1996) argues that stock prices can affect investment when financing frictions are present. Related studies show that investment becomes more sensitive to stock prices among equity-dependent firms and during periods of favorable financing conditions (Baker et al., 2003; McLean and Zhao, 2014; Warusawitharana and Whited, 2016).

A second explanation emphasizes catering incentives. Elevated valuations may give managers stronger incentives to justify optimistic market expectations and sustain high stock prices. Managers may therefore adjust corporate policies in ways that investors view favorably, even when such actions are not fully supported by fundamentals. Jensen (2005) argues that managers may overinvest during high-valuation periods. Polk and Sapienza (2009) show that stock overvaluation can induce overinvestment when managers respond to stock prices. Other studies document that firms are more likely to undertake valuation-sensitive actions during periods of overvaluation (Baker and Wurgler, 2004; Gao, 2010; Jin, 2022).¹⁾

These mechanisms are also relevant to R&D investment. Easier access to external capital may support R&D projects that would otherwise be delayed or underfunded (Dang and Xu, 2018; Li, 2026). At the same time, R&D can serve as a visible signal of future growth (Platikanova, 2026). Managers may increase R&D spending when such investment is likely to be viewed as consistent with optimistic market expectations (Dong et al., 2021; Duong, 2024; Shen et al., 2021). Taken together, prior studies suggest that stock overvaluation affects both the opportunity to invest and the incentive to do so. The net effect on R&D investment depends on how managers interpret valuation conditions and translate them into investment decisions.

2.3 Testable Hypotheses

The preceding discussion suggests that both CEO business educational background and stock overvaluation shape R&D investment decisions. CEO business educational background influences how managers evaluate investment opportunities, manage financing, and respond to capital market conditions. Stock overvaluation, in turn, changes both the resources available for investment and the incentives that guide managerial behavior. These forces jointly determine how firms adjust R&D investment.

The first possibility is that CEO business education is positively associated with R&D investment. As discussed above, business-educated CEOs may be better equipped to evaluate uncertain investment opportunities, allocate financial resources, and obtain financing for long-horizon projects. These capabilities can make R&D investment more feasible and attractive despite its uncertain and delayed payoffs. Under this view, business-educated CEOs are more willing and able to commit resources to R&D even in the absence of stock overvaluation.

Stock overvaluation may reinforce business-educated CEOs’ tendency to invest in R&D. One channel operates through financing conditions. When firms are overvalued, access to external equity becomes less costly and financing opportunities expand (Baker et al., 2003; McLean and Zhao, 2014). Business-educated CEOs, who are trained in valuation and capital allocation, may be more likely to recognize these opportunities and channel additional resources into R&D investment. A second channel reflects market responsiveness. Greater attention to capital markets may lead business-educated CEOs to align investment decisions with market expectations (Sun et al., 2021). This channel can further strengthen their tendency to increase R&D investment. Taken together, these arguments suggest that stock overvaluation amplifies the positive relation between CEO business educational background and R&D investment.

Hypothesis 1 (Amplification). Stock overvaluation strengthens the positive relation between CEO business educational background and R&D investment.

Stock overvaluation may also weaken the conservative tendency toward R&D investment associated with CEO business educational background. Under normal conditions, business-educated CEOs may be reluctant to invest in R&D because such investment requires large upfront commitments and is difficult to evaluate ex ante (Czarnitzki and Toole, 2011; Hottenrott et al., 2016). These concerns are likely to be more salient for CEOs who emphasize financial discipline and short-term accountability (Kim, 2025; Miller and Xu, 2019). Overvaluation can partially reduce this reluctance through two channels. First, by lowering the effective cost of external equity, overvaluation weakens one financial reason for avoiding R&D (Baker et al., 2003; McLean and Zhao, 2014; Warusawitharana and Whited, 2016). Second, because R&D can serve as a visible signal of future growth, an excessively conservative R&D policy may become less consistent with valuation-related market expectations when firms are overvalued (Dong et al., 2021; Duong, 2024; Shen et al., 2021). Thus, even if business-educated CEOs remain less inclined to invest in R&D, the negative association between CEO business educational background and R&D investment should be weaker under overvaluation.

Hypothesis 2a (Mitigation). Stock overvaluation weakens the negative relationship between CEO business educational background and R&D investment.

This mitigation effect can take a stronger form. Here, the relevant question is not whether overvaluation strengthens an already positive relation, but whether it is strong enough to overturn an initially negative relation. Under this scenario, overvaluation does not merely reduce business-educated CEOs’ reluctance toward R&D. It changes the direction of their R&D response. Favorable financing conditions make R&D less costly to undertake, while valuation-related incentives increase the payoff to visible investment in future growth. Prior studies suggest that overvalued firms face incentives to justify optimistic market expectations and sustain high valuations through corporate policies that investors view favorably (Gao, 2010; Jensen, 2005; Jin, 2022; Polk and Sapienza, 2009). When these forces dominate the baseline conservative tendency, business-educated CEOs may move from reduced reluctance to active R&D investment. Accordingly, stock overvaluation may reverse the negative relation between CEO business educational background and R&D investment.

Hypothesis 2b (Reversal). Stock overvaluation reverses the negative relationship between CEO business educational background and R&D investment, such that business-educated CEOs exhibit higher R&D investment than non-business CEOs among overvalued firms.

Figure 1 summarizes the theoretical framework and the three competing hypotheses developed above. The upper panel illustrates how stock overvaluation may modify the relation between CEO business educational background and R&D investment through the potential equity-financing and catering channels, while the lower panel summarizes the amplification, mitigation, and reversal hypotheses.

3.1 Sample and Data

This study uses a sample of Korean listed firms from 2011 to 2023. The sample period begins in 2011, when Korean listed firms adopted Korean International Financial Reporting Standards, which enhances the comparability of financial statements across firms. The sample ends in 2023 based on the availability of financial and firm-level data. Financial firms (two-digit Korean Standard Industrial Classification (KSIC) codes 64–66) are excluded because their accounting structures and regulatory environments differ substantially from those of non-financial firms.

Information on firms’ R&D investment, stock prices, and financial statement variables is obtained from DataGuide 6. Information on CEO educational background is hand-collected from firms’ annual reports filed with the Financial Supervisory Service and supplemented with data from the TS-2000 database provided by the Korea Listed Companies Association. When information on CEO educational background is not available from these sources, we further supplement it through searches on major online portals, including Naver, Daum, and Google. Observations with missing values for variables are removed. The final sample consists of 2,397 unique firms and 21,618 firm-year observations.

3.2 Variables

3.3 Research Design

We estimate the following baseline regression model:

RDI denotes R&D investment, BUSINESS is an indicator variable for CEO business educational background, and OVER denotes stock overvaluation. X represents a vector of control variables. We include firm and year fixed effects ($\mu_{i}$ and $\lambda_{t}$) to control for time-invariant firm characteristics and time trends. $\varepsilon_{i,t+1}$ is the error term. Standard errors are clustered at the firm level.

The coefficient of interest is β₃, which captures how stock overvaluation modifies the relation between CEO business educational background and R&D investment. Specifically, β₃ reflects the difference in the effect of BUSINESS between overvalued firms (OVER = 1) and non-overvalued firms (OVER = 0). The conditional effect of BUSINESS on RDI is given by β₁ when OVER = 0, and by β₁ + β₃ when OVER = 1. Thus, β₃ represents how the effect of CEO business educational background changes when firms are overvalued.

The interpretation of β₃ depends on both the sign of β₁ and the conditional effect of CEO business educational background among overvalued firms, β₁ + β₃. Under Hypothesis 1 (Amplification), stock overvaluation strengthens the positive relation between CEO business educational background and R&D investment. This pattern requires both β₁ and β₃ to be positive. Under Hypothesis 2a (Mitigation), stock overvaluation attenuates, but does not reverse, the negative relation between CEO business educational background and R&D investment. This pattern requires a negative β₁, a positive β₃, and a non-positive β₁ + β₃. Hypothesis 2b (Reversal) represents a stronger case in which stock overvaluation reverses the negative relation, such that business-educated CEOs increase R&D investment when firms are overvalued. This pattern requires a negative β₁, a positive β₃, and a positive β₁ + β₃.

Table 1 summarizes how the estimated coefficients correspond to the testable hypotheses. All three hypotheses can involve a positive interaction coefficient, but they differ in the baseline relation between CEO business educational background and R&D investment and in the conditional relation among overvalued firms. Thus, the empirical separation between Hypothesis 2a and Hypothesis 2b requires testing whether β₁ + β₃ is statistically greater than zero.

3.4 Summary Statistics

Table 2 presents the distribution of CEO business education and the descriptive statistics for the variables used in the main analysis. Panel A shows that, among the 4,975 CEOs identified in our sample, 2,178 have at least one business-related degree. Of these CEOs, 1,590 hold a business-related bachelor’s degree, 1,685 hold a business-related master’s degree, and 29 hold a business-related doctoral degree. These degree-level categories are not mutually exclusive because a CEO may hold business-related degrees at more than one level.

Panel B presents the summary statistics for the variables used in the main analysis. The mean value of RDI is 0.033, indicating that, on average, R&D expenditure accounts for 3.3% of total sales. The mean value of BUSINESS is 0.443, indicating that 44.3% of firm-year observations are led by CEOs with business educational backgrounds. The mean value of OVER is 0.333, which is consistent with the classification of overvalued firms based on the top tercile of firm-specific misvaluation.⁴⁾ The average of the natural logarithm of CEO age (CEOAGE) is 4.020, and the average CEO ownership (CEOOWN) is 2.5%. Regarding firm characteristics, the mean values of FIRMSIZE and FIRMAGE, both measured in natural logarithms, are 26.117 and 3.241, respectively. The mean values of return on assets (ROA), leverage (LEV), cash holdings (CASH), capital expenditure (CAPEX), and sales growth (GROWTH) are 0.025, 0.420, 0.110, 0.042, and 0.120, respectively. The average foreign ownership (FOR) is 6.3%, while the average ownership by large shareholders (LARGE) is 39.8%.⁵⁾

4.1 Baseline Regressions

Table 3 reports the baseline regression results. In the regression analyses, RDI is multiplied by 100 for ease of interpretation. Column (1) includes only the main variables, column (2) adds CEO characteristics (CEO age (CEOAGE) and CEO ownership (CEOOWN)), and column (3) includes the full set of control variables. Across all specifications, the coefficient on BUSINESS is negative and statistically significant.

In column (3), the coefficient on BUSINESS is −0.469 and significant at the 5% level, indicating that business-educated CEOs are associated with R&D intensity that is 0.469 percentage points lower among non-overvalued firms. This finding is consistent with the view that business-educated CEOs may be more cautious about R&D investment (Miller and Xu, 2019), especially when firms are not overvalued and R&D involves uncertain outcomes and short-term performance pressure. The coefficient on BUSINESS × OVER is 0.775 and significant at the 1% level. This estimate indicates that the association between CEO business educational background and R&D intensity is 0.775 percentage points higher among overvalued firms than among non-overvalued firms.

To assess the effect of CEO business educational background among overvalued firms, we report the linear combination (β₁ + β₃) of the coefficients on BUSINESS and BUSINESS × OVER at the bottom of the table. The linear combination is positive and statistically significant in all specifications. In Column (3), the linear combination is 0.306 and significant at the 5% level. This finding indicates that, among overvalued firms, business-educated CEOs are associated with a 0.310 percentage point higher R&D intensity than non-business CEOs. Thus, the negative relation between CEO business educational background and R&D investment is not merely attenuated under overvaluation. It becomes positive.⁶⁾

Taken together, the baseline results support Hypothesis 2b. Business-educated CEOs are associated with lower R&D investment when firms are not overvalued, but the association becomes positive among overvalued firms.

4.2 Endogeneity

4.3 Robustness Checks

We conduct several robustness checks to examine whether the baseline reversal pattern is sensitive to alternative variable definitions, sample restrictions, and fixed-effect structures. Table 7 reports the results. Columns (1) and (2) examine whether the baseline results are sensitive to the measurement of R&D investment. In Column (1), we use RDA (%), measured as R&D expenditure scaled by lagged total assets and multiplied by 100. In Column (2), we use LnRD, measured as the natural logarithm of R&D expenditure. In both specifications, BUSINESS remains negative, BUSINESS × OVER remains positive and statistically significant, and the linear combination of the two coefficients remains positive and statistically significant. These results indicate that the reversal pattern is not driven by the choice of sales-scaled R&D intensity as the baseline dependent variable.

Column (3) restricts the sample to firms with positive R&D expenditure. This test addresses the possibility that the baseline results are driven by firms that report zero R&D expenditure. The results remain consistent with the baseline findings. The coefficient on BUSINESS is negative and statistically significant, the coefficient on BUSINESS × OVER is positive and statistically significant, and the linear combination, β₁ + β ₃, is positive and statistically significant. This finding suggests that the reversal pattern also holds among firms that engage in R&D investment.

Column (4) redefines OVER using the highest quintile of firm-specific misvaluation, providing a more restrictive definition of overvaluation (e.g., Alzahrani and Rao, 2014). The results remain unchanged. BUSINESS is negative and statistically significant, BUSINESS × OVER is positive and statistically significant, and the linear combination is positive and significant at the 1% level. This result indicates that the baseline conclusion is not sensitive to the tercile cutoff used to define overvalued firms.

Column (5) replaces the indicator measure of overvaluation with the continuous firm-specific misvaluation component (FSE). The coefficient on BUSINESS remains negative, while the coefficient on BUSINESS × FSE is positive and statistically significant. Thus, the relation between CEO business education and R&D becomes more positive as FSE increases, indicating that the baseline interaction result is not driven by the discrete cutoff used to define overvalued firms.¹¹⁾

Column (6) includes industry-by-year fixed effects. This specification absorbs time-varying industry-level shocks that may affect firms within the same industry-year. The coefficient on BUSINESS remains negative, whereas the coefficient on BUSINESS × OVER remains positive and statistically significant. The linear combination is also positive and statistically significant. Thus, the reversal pattern persists even after controlling for time-varying industry-level factors.

4.4 Mechanism Tests

Having established the reversal pattern across the baseline, endogeneity, and robustness analyses, we next examine the mechanisms through which stock overvaluation changes the relation between CEO business educational background and R&D investment. The baseline results show that business-educated CEOs are associated with lower R&D investment when firms are not overvalued, but higher R&D investment when firms are overvalued. This pattern may arise through different channels.

One possible channel is the equity-financing mechanism. Stock overvaluation may provide firms with favorable equity financing opportunities (Baker et al., 2003; McLean and Zhao, 2014; Warusawitharana and Whited, 2016), and business-educated CEOs may be more likely to recognize and use these opportunities to support R&D investment. Another possible channel is the catering mechanism. Stock overvaluation may increase incentives to respond to market expectations (Gao, 2010; Jensen, 2005; Jin, 2022; Polk and Sapienza, 2009), and business-educated CEOs may respond by increasing visible growth-oriented investments such as R&D. These mechanisms are not necessarily mutually exclusive.

To distinguish these mechanisms, we jointly consider financial constraints and investor horizons. Financial constraints capture the extent to which favorable equity financing opportunities are valuable to the firm. If the reversal pattern is driven by the equity financing mechanism, the positive conditional effect of CEO business educational background among overvalued firms should be more pronounced among financially constrained firms. Investor horizons capture the extent to which firms are exposed to investors with shorter trading horizons, which may intensify valuation-related market pressure. If the reversal pattern is driven by the catering mechanism, the positive conditional effect should be more pronounced among firms with short-horizon investors. We therefore divide the sample into four groups based on financial constraint status and investor horizon.

We measure financial constraints using a modified KZ index following Baker et al. (2003). Because the conventional KZ index includes Tobin’s Q, it may mechanically overlap with our stock overvaluation measure. We therefore exclude Tobin’s Q from the index and calculate the modified KZ index as follows:

In Eq. (3), CF denotes cash flow, DIV denotes cash dividends, C is cash holdings, A denotes total assets, and LEV is leverage. Firms with modified KZ index values above the annual median are classified as financially constrained, and the remaining firms are classified as financially unconstrained. Following Alzahrani and Rao (2014), we measure investor horizon using stock turnover. Specifically, the turnover ratio is calculated as the average monthly ratio of shares traded to shares outstanding over the fiscal year. A higher turnover ratio indicates that shares change hands more frequently and is therefore interpreted as reflecting a shorter investor horizon. Firms with turnover ratios above the annual median are classified as short-horizon investor firms, and the remaining firms are classified as long-horizon investor firms.

We first examine financial constraints and investor horizons separately before examining their joint classification. If the equity-financing mechanism is the primary explanation, the reversal should be more evident among financially constrained firms, for which favorable equity-financing conditions are particularly valuable. By contrast, if the catering mechanism is more relevant, the reversal should be more evident among firms with short-horizon investors, where pressure to respond to market expectations is likely to be stronger. We then combine the two dimensions in a 2 × 2 classification to examine whether the investor-horizon pattern persists across financial constraint status.

Table 8 reports the mechanism test results. Panel A presents the separate subsample analyses. Columns (1) and (2) divide the sample by financial constraint status, while Columns (3) and (4) divide the sample by investor horizon. In both the constrained and unconstrained subsamples, BUSINESS is negative and statistically significant, while BUSINESS × OVER is positive and statistically significant. The interaction coefficients are also very similar in magnitude, at 0.703 and 0.713, respectively. The conditional effects of BUSINESS among overvalued firms, β₁ + β₃, are 0.248 for constrained firms and 0.244 for unconstrained firms. Thus, the relation between CEO business education and R&D under overvaluation does not appear to be concentrated among financially constrained firms.

The results differ more clearly when the sample is divided by investor horizon. In the short-horizon subsample, BUSINESS is negative and significant, whereas BUSINESS × OVER is positive and significant. The corresponding conditional effect of BUSINESS among overvalued firms is 0.444 (t = 3.52). In the long-horizon subsample, BUSINESS remains negative and significant, but BUSINESS × OVER is statistically insignificant, and the corresponding conditional effect is only 0.079 (t = 1.05). These results provide more direct evidence that the reversal pattern is associated with investor horizon rather than financial constraint status.

Panel B further examines the two dimensions jointly. The results reinforce the pattern in Panel A. The conditional effect of BUSINESS among overvalued firms is positive and statistically significant for financially constrained firms with short-horizon investors (0.273) and for financially unconstrained firms with short-horizon investors (0.362). By contrast, the corresponding effects are statistically insignificant for constrained firms with long-horizon investors (0.129) and unconstrained firms with long-horizon investors (0.086). Thus, the reversal is observed in both financial constraint groups when investor horizons are short, but not when investor horizons are long.

Taken together, the separate and joint subsample analyses point to investor horizon, rather than financial constraint status, as the dimension more closely associated with the reversal pattern. If favorable equity financing were the primary mechanism, the reversal would be expected to be concentrated among financially constrained firms. Instead, it appears across both constrained and unconstrained firms when investor horizons are short. This pattern is difficult to reconcile with a pure equity-financing explanation and is more consistent with a catering-based interpretation.

Additional analysis in Table A4 of the Appendix shows that business-educated CEOs are more likely to issue equity under overvaluation, but equity issuance itself is not significantly associated with subsequent R&D investment. This finding suggests that business-educated CEOs may use favorable equity market conditions, but such financing behavior does not appear to explain the reversal pattern in R&D investment.

4.5 Further Tests

CEO ownership status and Chaebol affiliation may provide alternative explanations for the documented relation between CEO business education and R&D investment. Prior Korean evidence suggests that owner-managers and professional managers differ in their R&D investment behavior. For example, Kim and Song (2011) find that firms managed by professional CEOs exhibit significantly lower R&D intensity than firms managed by owner CEOs. In addition, the Chaebol structure represents an important feature of the Korean corporate environment. Ryu and Cho (2021) argue that the distinctive ownership and control structure of Korean Chaebols can affect firms’ R&D expenditure behavior beyond governance characteristics at the individual-firm level. We therefore examine whether the baseline reversal varies systematically with CEO ownership status and Chaebol affiliation.

Following the Korean literature, we identify owner CEOs using the CEO’s relationship with the firm’s largest shareholder. Specifically, OWNERCEO equals one if the identified CEO is reported as the largest shareholder or a related party of the largest shareholder, and zero otherwise. This approach is consistent with prior Korean research that uses information on the largest shareholder, relatives, and related parties to distinguish owner-managers from professional managers (Kim and Kim, 2025). CHAEBOL equals one if the firm belongs to a designated business group in year t, and zero otherwise. We estimate separate three-way interaction models for OWNERCEO and CHAEBOL and include all corresponding lower-order interaction terms. This specification allows us to test directly whether the positive interaction between BUSINESS and OVER varies with CEO ownership status or Chaebol affiliation.

Table 9 reports the results. Column (1) examines CEO ownership status. The coefficient on OWNERCEO is positive and statistically significant, whereas BUSINESS × OWNERCEO is insignificant. More importantly, BUSINESS × OVER remains positive and significant, whereas BUSINESS × OVER × OWNERCEO is statistically insignificant. Thus, there is no evidence that the reversal in the relation between CEO business education and R&D under overvaluation differs systematically between owner and professional CEOs.

Column (2) examines Chaebol affiliation. The coefficient on CHAEBOL is positive and statistically significant, and the coefficient on BUSINESS × CHAEBOL is also positive and significant. However, the coefficient on BUSINESS × OVER × CHAEBOL is statistically insignificant. Thus, although Chaebol affiliation is associated with the baseline relation between CEO business education and R&D, it does not significantly alter the reversal associated with overvaluation. Taken together, these results provide no evidence that the documented reversal differs systematically by CEO ownership status or Chaebol affiliation, suggesting that the baseline finding is unlikely to be attributable solely to these ownership and governance characteristics.

5.1 Research Implications

This study extends the upper echelons literature by showing that CEO educational background does not have a uniform effect on R&D investment. Rather, its effect depends on the valuation conditions under which CEOs make investment decisions. This finding suggests that CEO business education should be understood not only as a proxy for managerial ability or preference, but also as a factor that shapes how CEOs interpret and respond to valuation-related market signals. In this sense, CEO educational background functions as a context-dependent managerial attribute rather than as a fixed predictor of R&D investment.

The findings also have implications for how the relation between stock misvaluation and corporate investment should be interpreted. Prior studies show that overvaluation can affect real investment through financing opportunities or catering incentives. Our results suggest that this relation should not be viewed as a purely firm-level response to valuation conditions. Rather, the investment response to overvaluation may depend on the characteristics of the CEO who interprets valuation signals and translates them into investment decisions. Thus, overvaluation does not mechanically translate into R&D investment. Its effect on real investment should be understood in conjunction with managerial characteristics.

The findings also have practical implications for how R&D investment is interpreted. Increases in R&D spending are often viewed as evidence of stronger innovation commitment. Our results suggest that this interpretation may be incomplete when firms are overvalued. In such settings, R&D investment may reflect a response to valuation-related market pressure, particularly when firms are exposed to short-horizon investors. This distinction is important because increases in R&D investment among firms led by business-educated CEOs under overvaluation should not be interpreted simply as the result of improved financing capacity. They may also reflect responses to market expectations that become more salient in short-horizon investor environments. This does not imply that R&D investment under overvaluation is necessarily inefficient or value-destroying. Rather, it suggests that such investment should be interpreted in relation to the valuation environment and investor horizon in which it occurs.

The findings further suggest that investors, boards, and innovation evaluators should look beyond the amount of R&D spending. Investors and analysts should consider whether R&D increases among overvalued firms reflect sustained innovation plans or market-responsive behavior. Boards should assess the continuity and technological rationale of R&D projects, rather than focusing only on the amount of spending. More broadly, innovation evaluation frameworks should distinguish between R&D investment driven by long-term technological objectives and R&D investment induced by market pressure.

5.2 Limitations and Future Research

This study has several limitations that provide opportunities for future research. First, the analysis focuses on R&D investment rather than innovation outcomes or innovation quality. Although the results show that business-educated CEOs increase R&D investment when firms are overvalued, this study does not directly examine whether such investment leads to greater innovation output, higher-quality innovation, greater R&D efficiency, or long-term firm value. This limitation is important because R&D spending captures resource commitment, but not necessarily the economic or technological value generated by that commitment. Future research could examine whether the R&D increases documented in this study translate into patents, citation-weighted innovation outcomes, R&D efficiency, or subsequent firm performance.

A further limitation relates to the measurement of stock overvaluation. This study relies primarily on the misvaluation decomposition approach of Rhodes-Kropf et al. (2005) to identify firm-specific overvaluation. Although this approach is widely used and allows us to separate firm-specific valuation deviations from broader valuation components, overvaluation can also be measured using alternative frameworks, such as residual income models, fundamental cash-flow-based valuation measures, or forecast-based intrinsic value estimates (e.g., Bonaimé et al., 2014; Dong et al., 2021; Jin, 2022; Shen et al., 2021). Due to data limitations, this study does not fully examine whether the results hold across these alternative measures. Future research could use richer valuation data to test whether the reversal pattern documented in this study is robust across different approaches to measuring stock overvaluation.

Finally, although this study conducts several analyses to mitigate endogeneity concerns, including dynamic panel estimation, instrumental variable estimation, and propensity score matching, the findings should be interpreted cautiously in causal terms. CEO appointment decisions, valuation conditions, and R&D investment may still be jointly influenced by unobservable firm characteristics or market expectations. In addition, the mechanism tests provide evidence consistent with the catering mechanism, but they do not directly observe managerial intentions or the specific expectations imposed by investors. Future research could use exogenous shocks to valuation conditions, investor horizons, or CEO appointments to provide sharper causal evidence. Future research could also use textual disclosures, conference calls, or R&D project-level information to examine more directly how CEOs justify R&D investment decisions under stock overvaluation.