econ forecasting midterm

cross-sectional data

Data collected on multiple individual entities (e.g., consumers, firms, countries)
across a single point in time.

time-series data

Data collected on a single entity across multiple consecutive time periods (e.g., daily,
monthly, quarterly, annual).

panel data

Data tracking multiple entities across multiple time periods (longitudinal data).

trend (mt)

A long-term systematic change or directional movement in the level of a series that does not
appear to be periodic.

seasonal variation (st)

A repeating, cyclical pattern of fluctuations linked to fixed natural periods (e.g.,
months of the year, quarters, days of the week)

cyclical variation

Repeating patterns and multi-year fluctuations that do not correspond to a fixed
calendar period (e.g., economic business cycles)

random/error component (zt)

The irregular, unsystematic remainder component left after trend and
seasonality are removed

deterministic trend

A trend that follows a predictable path over time.

stochastic (random) trend

A trend whose trajectory shifts at unpredictable times

additive decomposition model

Appropriate when the magnitude of the seasonal variation does NOT
depend on the level of the trend (i.e., remains constant over time)

multiplicative decomposition model

Appropriate when the magnitude of seasonal variation grows or
shrinks proportionally with the level of the trend

centered-moving average

Used to estimate a non-linear trend mₜ

stationary in the mean

The mean of the series is constant and invariant across all time periods (no trend
or shift)

stationary in the variance

The variance of the series is constant and invariant over time

leading variable

A variable whose directional changes reliably precede the target series by one or more
time periods (optimal if there is long lead time and close correlation)

associated variable

A variable that moves contemporaneously with the target series but is significantly
easier or faster to forecast

discrete white noise (wt)

A sequence of independent, identically distributed random variables with mean
0, constant variance σ², and zero autocorrelation at all non-zero lags

random walk model

Each observation equals the previous value plus an unpredictable white noise shock

first-order differencing

A random walk is non-stationary in both mean and variance, but taking first
differences transforms it into stationary white noise