This paper examines the theory of saving under liquidity constraints, focusing on how such constraints affect consumption behavior. It argues that when consumers are impatient and labor income is independently and identically distributed over time, assets act as a buffer stock, protecting consumption against income fluctuations. However, if income is positively autocorrelated, assets are less effective at smoothing consumption and require greater sacrifice. In the extreme case where income is a random walk, liquidity-constrained consumers cannot save and must consume their income immediately. The paper also explores how microeconomic income processes differ from their macroeconomic aggregates, with individual incomes often being more volatile and subject to idiosyncratic shocks. This can lead to negative serial correlation in income growth, which may result in buffering behavior similar to that observed in simple models without growth. However, aggregate savings can still emerge if some component of income growth is common to all consumers. The analysis shows that liquidity constraints significantly influence saving and asset accumulation behavior. When income is stationary and independently distributed, assets serve as a buffer stock, and consumption can be smoothed without borrowing. However, with positively autocorrelated income, more assets are needed to smooth consumption, and the effectiveness of saving decreases. The paper also considers non-stationary income processes, such as logarithmic random walks with drift, and shows that these can generate aggregate saving behavior that aligns with observed data. However, the presence of liquidity constraints limits the ability to smooth consumption, and savings tend to be procyclical, rising during economic downturns and falling during booms. Overall, the paper highlights the importance of liquidity constraints in explaining saving behavior and shows that they can account for many stylized facts in the data, even when traditional life-cycle models fail to do so. The analysis underscores the role of precautionary saving and the sensitivity of consumption behavior to the stochastic process generating income.This paper examines the theory of saving under liquidity constraints, focusing on how such constraints affect consumption behavior. It argues that when consumers are impatient and labor income is independently and identically distributed over time, assets act as a buffer stock, protecting consumption against income fluctuations. However, if income is positively autocorrelated, assets are less effective at smoothing consumption and require greater sacrifice. In the extreme case where income is a random walk, liquidity-constrained consumers cannot save and must consume their income immediately. The paper also explores how microeconomic income processes differ from their macroeconomic aggregates, with individual incomes often being more volatile and subject to idiosyncratic shocks. This can lead to negative serial correlation in income growth, which may result in buffering behavior similar to that observed in simple models without growth. However, aggregate savings can still emerge if some component of income growth is common to all consumers. The analysis shows that liquidity constraints significantly influence saving and asset accumulation behavior. When income is stationary and independently distributed, assets serve as a buffer stock, and consumption can be smoothed without borrowing. However, with positively autocorrelated income, more assets are needed to smooth consumption, and the effectiveness of saving decreases. The paper also considers non-stationary income processes, such as logarithmic random walks with drift, and shows that these can generate aggregate saving behavior that aligns with observed data. However, the presence of liquidity constraints limits the ability to smooth consumption, and savings tend to be procyclical, rising during economic downturns and falling during booms. Overall, the paper highlights the importance of liquidity constraints in explaining saving behavior and shows that they can account for many stylized facts in the data, even when traditional life-cycle models fail to do so. The analysis underscores the role of precautionary saving and the sensitivity of consumption behavior to the stochastic process generating income.