This dissertation consists of three essays in the areas of Macroeconomics and Firm Dynamics, examining the role of resource (mis)allocation on business dynamism and market outcomes. The first essay examines the role of input specificity in affecting firm-to-firm trade dynamics and how business dynamism and a sticky input market interact. I construct a novel panel data of supply chain relationships with an additional layer of input specificity measures on each linkage across US firms and point out the foreclosure effect of input specification contracts. I then propose a Schumpeterian model where each firm endogenously chooses whether to produce a specific input for an exclusive customer firm at the expense of giving up larger customer base. Bringing the micro-evidence to the equilibrium model, I show that the input specificity channel is quantitatively important and can exaggerate the downturn in business dynamism. I show that shutting down the foreclosure effect can lead to a 10% and 25% increase in growth and entry rates. In the second essay, I study the role of local vintage capital market in driving co-locations of entrants and capitalists and shaping the spatial disparities in business dynamism and productivity. I combine the US-based venture capital (VC) investment records with local business vintage capital supply across metropolitan statistical areas (MSAs). I find that the VC investment elasticity with respect to local vintage capital is about 9% - 14%. Based on the motivating empirical evidence, I integrate vintage capital induced co-location of entrants and VC investment into a theoretical framework. I highlight a straightforward mechanism where a vintage capital market with abundant supply can lower the capital cost and thus increase firms' profits, attracting VC investment, encouraging entrepreneurship and further leading to a selection-induced agglomeration effect. I show that a larger city intensifies such allocative forces and thus amplifies the agglomeration effect, making the city further attractive to VC investment and entrants relative to others. The third essay, which is joint work with Qinshu Xue, examines how firm-to-firm partnership shapes a firm's innovation strategy and its implications for business dynamism. My coauthor and I combine the firm-to-firm partnership information with the patent data filed by the US publicly listed firms from 2003 to 2013. We find that the degree of firm-to-firm partnership engagement along the supply chain is positively associated with stronger incentives in conducting exploitative innovations, leading to less knowledge spillover than exploratory innovations do. To mitigate the potential selection bias, we further adopt a double debiased machine learning technique to allow for a more flexible relationship between partnership variables and control variables and confirm the positive correlation. We then propose a quality-ladder framework where firms optimally choose to either continue exploitative innovations by leveraging possible collaborations with other firms or stop them and turn to exploratory innovations for its product. Our model predicts that partnership engagement can induce more technologies that are too sophisticated to generate spillover and thus lead to a lower entry rate and precludes growth.