This dissertation consists of three chapters which contribute to theoretical and quantitative understanding of business cycles. The first essay studies firm dynamics over the business cycle. I present evidence from the United Kingdom that more rapidly growing firms are born in expansions than in recessions. Using administrative records from Census data, I find that this observation also holds for the last four recessions in the United States. I also present suggestive evidence that financial frictions play an important role in determining the types of firms that are born at different stages of the business cycle. I then develop a general equilibrium model in which firms choose their managers’ span of control at birth. Firms that choose larger spans of control grow faster and eventually get to be larger, and in this sense have a larger target size. Financial frictions in the form of collateral constraints slow the rate at which firms reach their target size. It takes firms longer to get up to scale when collateral constraints tighten; therefore, businesses with the largest target size are affected disproportionately more. Thus, fewer entrepreneurs find it profitable to choose larger projects when financial conditions deteriorate. Using Bayesian methods, I estimate the model using micro and aggregate data from the United Kingdom. I find that financial shocks account for over 80% of fluctuations in the formation of businesses with a large target size, and TFP and labor wedge shocks account for the remaining 20%. An independently estimated version of the model with no choice over the span of control needs larger aggregate shocks in order to account for the same data series, suggesting that the intensive margin of business formation is important at business cycle frequencies. The model with the choice over the span of control generates an empirically relevant and non-targeted collapse in the right tail of the cumulative growth distribution among firms started in recessions, while the model without such a choice does not. The paper also discusses implications for micro-targeted government stimulus policies. The second essay is based on joint work with Makoto Nakajima. We investigate cyclicality of variance and skewness of household labor income risk in a unified estimation framework using PSID data. We make five findings. First, we find that head’s labor income exhibits countercyclical variance and procyclical skewness. Second, cyclicality of hourly wage is muted, suggesting that head’s labor income risk is mainly coming from volatility of hours. Third, the second earner lowers the cyclicality of both volatility and skewness of labor income risk. Fourth, government taxes and transfers reduce fluctuations of income risk even further, making it nearly invariant with respect to business cycle. Finally, among heads with strong labor market attachment, cyclicality of labor income volatility becomes weaker, while cyclicality of skewness remains. The third essay explores the aggregate implications of countercyclical heteroscedasticity of TFP shocks. First, by drawing on data of publicly traded U.S. firms, I find that both the investment rate and the probability of investment spikes decrease as macroeconomic volatility rises, with this effect weakening as firms age. Second, to rationalize empirical patterns, I build an uncertainty regime-switching general equilibrium model of investment with capital adjustment costs and firms’ entry and exit. In the model, the option value of waiting for younger businesses is larger than for mature firms, since the former are further away from their target size than the latter. As a result, young businesses respond stronger to heightened uncertainty as compared to older ones. Finally, in contrast to existing literature, I show that the entry/ exit margin can be quantitatively important, especially during high uncertainty episodes such as the Great Recession.