It has been proposed that the MACHOs in our galaxy could be clumped in globular cluster--like associations or RAMBOs (robust associations of massive baryonic objects) (Moore \& Silk 1995). Here we investigate the effect such clustering has on the microlensing of stars in the Large Magellanic Cloud. We find that the lensing in a one square degree field could be dominated by just a few clusters. As a result the lensing properties vary widely depending on the position and velocity of those clusters which happen to lie between us and the LMC. Moreover we find a large variance in time scale distributions that suggests that the small number statistics could easily be dominated by events in the tails of the unclustered distribution ({\it e.g.} by long periods). We compare our results with the MACHO collaboration's data and find that a ``standard'' halo made entirely of MACHOs is not strongly disfavored if the clusters have masses of $10^6 M_{\odot}$. For less massive clusters such a halo is not as likely. For $10^4 M_{\odot}$ clusters the microlensing statistics are essentially unchanged from the unclustered case. It may be possible to detect very massive clusters from the distribution of events in time-scale and space. We provide some example time- scale distributions.

uuencoded, gzipped PostScript, 14 pages