The terms were invented by evolutionists and remain in use today. E.g. see http://evolution.berkeley.edu/evolibrary/article/evo_01
The terms are not precisely defined and there tends to be some overlap at the species level with different definitions, so you can pick and chose your definition to suit your point of view.
One way of differentiating them is to say that microevolution corresponds to "a change in allele frequency in a population over time" while macroevolution is the addition of new genetic information and material to the genome. With these definitions you can see that speciation could be either micro or macroevolution depending on what causes it.
But what about a loss of genetic information? A change in allele frequency could include the case of it dropping to zero, the total loss of an allele, and once that happens it will not be restored by microevolution. We also know that organisms can sometimes lose genes. Lenski found this in his long term E Coli experiment. Such complete loss of genes does not fit comfortable in the usual definitions of micro or macroevolution and perhaps could more correctly described as devolution.
You can also see that you can have an endless succession of microevolutionary changes and they will never accumulate to a macroevolutionary change. No amount of change in allele frequency will ever produce a new gene. However a mutation could produce a new allele without producing a new gene; moving into the grey area.
So the question of what prevents microevolution events from accumulating to produce macroevolution is a nonsense. It's like asking how many poker hands you have to draw from a standard deck before a new (fifth) suit appears; it will never happen.
[edit] I suspect it would be bad for your health to lay down a hand of 5 aces of different suits (Spades, clubs, diamonds, hearts, and trumpets?), so that would not be a beneficial mutation.
Among the 22 comments received to date are several references to gene duplication. Since this is beyond the scope of the question asked I will leave that subject to another time.
[edit]
Microevolution vs Macroevolution: Two Mistakes
Friday, May 29, 2015 - 16:33, Kirk Durston
http://p2c.com/students/blogs/kirk-durston/2015/05/microevolution-vs-macroevolution-two-mistakes
"... let me propose the following definitions, which I will continue to use:
Microevolution: genetic variation that requires no statistically significant increase in functional information.
Macroevolution: genetic change that requires a statistically significant increase in functional information.
Both statistical significance and functional information are already defined in the literature. We also have a method to measure evolutionary change in terms of functional information, so we are ready to move on, avoiding the two mistakes discussed above."
"Macroevolution is very different from microevolution. The reason there are so many countless observations of variation/microevolution is that it requires no statistically significant levels of novel genetic information; it is trivially easy to achieve. "