In many cases, if you encounter zero reactance or zero impedance, this often represents bus ties or breakers. However, for the IEEE 14-bus test system, that is not the case.
As you can see from the numbers, the reactance is quite high. Obviously a breaker or a short cable can not have zero resistance and a high reactance. Reactors used for reactive power compensation are usually modeled at the buses and not at the branches. It could, as Li-aung suggests be a series reactor used for limiting fault currents. However, it seems to be a bit large for that purpose. Also, I believe these are usually used between breakers and loads, not between different buses in a meshed grid.
Although it's not a physically correct representation, these numbers represent transformers. Transformers do of course have resistance, but compared to the reactance this is close to negligible (at least when it comes to power flow).
From To R X B
Bus Bus pu pu pu
4 7 0.0 0.20912 0.0
4 9 0.0 0.55618 0.0
5 6 0.0 0.25202 0.0
As you can see from the single line diagram below, the branches with zero resistance are transformers.
A common problem in power flow calculations is when impedance values are set to zero. This is because the current through a branch in general is given by the voltage difference divided by the impedance. When the impedance is zero, this gives an infinite current, thus an unfeasible power flow solution. It is common practice to set the impedance values to a very small value in such cases (for instance R = 0, X = 0.0001 pu). In AC systems, zero resistance is often not a problem, but zero reactance is, as this will cause the decoupled load flow, and the DC load flow algorithms to fail.