Key difference between 3D and 4D solutions In many papers in theoretical physics especially in the more advanced theories, I notice that physicists solve problems in 3 dimensions (2 spatial+ 1 temporal). 
In some specific papers (I can't seem to recall exactly which ones), some models solved in three dimensions gave finite solutions, whereas 4D (3+1) were either hard to solve or gave divergent results.
I don't seem to understand why adding a 3rd spatial dimension should cause problems if 2 spatial dimensions gave no problems. Intuitively, it seems straightforward matter to extrapolate 4D solutions from 3D. Then what is the essential difference between the 2?
 A: I don't know which examples you are talking about but there are a couple reasons to study toy models in 2+1 dimensions. The quickest answer would be that when doing QFT, having one less spatial dimension helps with convergence problems of loop integrals. 
The difference adding an extra dimension causes in QFT is probably more pronounced in going from 1+1 to 2+1 dimensions. In 1+1 dimensions, we have only boosts, no rotations, so the structure of the symmetries is much simpler. In addition, in 1+1 dimension there is only forward and backward scattering, which evades an assumption of the Coleman-Mandula theorem (analyticity of S-Matrix) and essentially allows for exactly solvable S-matrices. You lose all of this in moving to 2+1 dimensions.
Probably one of the more important examples of a 2+1 toy model would be 3D gravity. It turns out that gravity in 3 dimensions is very different from gravity in 4 dimensions. Essentially, in 3 dimensions the Riemann curvature tensor is completely determined by the Ricci tensor, so that 
$R_{\mu \nu}=0 \implies R_{\mu \nu \rho \sigma}=0$. The Einstein equation says $R_{\mu \nu}\propto(T_{\mu \nu} -\frac{1}{2}Tg_{\mu \nu})$ so that vacuum solutions in 3D satisfy 
not only $R_{\mu \nu}=0$ but also $R_{\mu \nu \rho \sigma}=0$.
This means that 3D pure gravity does not have local propagating degrees of freedom. This is a dramatic simplification that allows you to get much more mathematical control over the theory and lets you test ideas about quantization. Going to 4D gives you local gravitational degrees of freedom and makes the theory much more intractable.
