R&D investments require huge initial outlay and involve high uncertainty. But regardless, firms invest up to 25% of their revenue in R&D; for instance, Apple Inc. in 2016. In identifying the reasons for high investment, researchers analyze different firm characteristics and find its association with R&D spending. This study examines the behavioral aspect, i.e., CEO overconfidence and finds its significant positive influence on R&D expenditure. The financial crisis of 2007-2008 is considered and difference in dominance of CEO overconfidence is observed for pre and post-crisis period. This study controls for other determinants of R&D to avoid exogeneity in the results. Three hypotheses are formed and tested in the analysis. The first hypothesis links CEO overconfidence and R&D expenditure, the second hypothesis focuses on stock return volatility of overconfident CEOs and R&D spending and the third hypothesis links overconfident CEOs age and R&D. The sample is further divided among industries and all three hypotheses are tested separately. Financial and utility industry are excluded because of their skewed accounting fundamentals which provide biased results. For the overall sample, overconfidence is highly positively significant in explaining R&D, following higher stock return volatility and lower age of overconfident CEOs association with an increase in R&D. The influence of CEO overconfidence and sensitivity of CEO age increases in post-crisis period, while stock return volatility decreases. While analyzing the industry sample, overconfidence is insignificant for the construction industry in overall and pre-crisis period but significant in post-crisis period. Service industry demonstrates the highest influence of overconfidence followed by mining, manufacturing, and retail trade industry for overall, before crisis and after crisis period. In testing the second hypothesis, apart from manufacturing industry, others depict positive association of stock return volatility and R&D spending. However, the sensitivity decreases in post-crisis period. Finally, apart from mining and construction, all other industries demonstrate negative association of age of overconfident CEOs with R&D spending in overall analysis. When analyzing the pre-financial crisis period, only construction industry shows positive association; though all industries depict significant negative association in post-crisis period. To check the robustness of the study, CEO overconfidence is further measured considering the exercise of 100% in-the-money stock options. The impact of overconfidence under this measure significantly drops as the number of CEOs considered as overconfident decreases. The study concludes by mentioning some of the limitations and scopes for future research.