The situation when the value of a stochastic time-series variable is not independent of its value in previous periods. If x(t) is a time-series variable, observed at times 1, 2…t, and y(t) is its trend value at time t, define z(t) = x(t) - y(t). There is zero serial correlation if the expected value E[z(t). z(t - 1)] = 0. If E[z(t). z(t - 1)] > 0 there is positive serial correlation; if E[z(t). z(t - 1)] < 0 there is negative serial correlation. Many economic variables such as the rate of inflation or the unemployment rate show strong positive serial correlation.

• local minimum
• location
• lock-out
• locomotive principle
• logarithm
• logarithmic scale
• logistic curve
• log-linear function
• log-normal distribution
• log-rolling
• Lomé Convention
• London Inter Bank Offered Rate
• London International Financial Futures and Options Exchange
• long-dated security
• long position
• long rate
• long run
• long-run average cost
• long-run marginal cost
• long-run Phillips curve
• long-term interest rate
• long-term unemployment
• Lorenz curve
• loss
• loss leader