The article explores a new way to analyze time series data by separating trends and cycles. By using a special model, researchers can identify common patterns in economic data. They found that when cycles have a specific pattern, they can accurately predict trends. Testing this method on U.S. income and consumption data showed that temporary changes in income have a bigger impact on spending than on earnings. Overall, this approach improves predictions of future trends in economic data.