As a general rule, non-stationary time series variables should not be
used in regression models, to avoid the problem of spurious regression. If ݕ௧ and ݔ௧
are nonstationary ܫሺ1ሻ variables, then their difference or any linear combination of
them is expected, such that, ݁௧ ൌ ݕ௧ െ ߚଵ െ ߚଶݔ௧ to be ܫሺ1ሻ as well. However, there is
an important case when ݁௧ ൌ ݕ௧ െ ߚଵ െ ߚଶݔ௧ is a stationary ܫሺ0ሻ process. In this case
ݕ௧ and ݔ௧ are coinegrated. Cointegration implies that ݕ௧ and ݔ௧ share similar
stochastic trends and since ߝ௧ is stationary they would not be much different.