The problem only arose when the second was changed from being 1/86400th of a solar day to being defined by atomic clocks. The old definition worked well for civil purposes but was by the 1950s impossible to accurately co-ordinate between researchers.
We have to make adjustments when coordinating anyway since local time anywhere outside sea level needs relativistic corrections to derive global time, so might as well also periodically introduce another adjustment factor based on changes in the rotation period of the earth…
