As egreg explained, the space between the letters is a design decision. In most cases I'd stick with TeX's default spacing, but it's possible to remove the italic correction. The result is shown in the second line of the following image.

As you can see, the result is not always desirable: for the combination PW, the letters are much too close to each other. But for WA and Tf, the output looks better with the italic correction removed. Here's my implementation of \ric (remove italic correction):
\documentclass{article}
\newcommand*\ric[1]{\vphantom{#1}\smash{#1_{}\kern-\scriptspace}}
\begin{document}
$PS\ TN\ CD\ WA\ Tf\ PW$
$\ric{P}S\ \ric{T}N\ \ric{C}D\ \ric{W}A\ \ric{T}f\ \ric{P}W$
\end{document}
A short explanation of the code: the italic correction is not used when the character is followed by a subscript, but no superscript. However, a \scriptspace is automatically added to the width of the subscript. I used an empty subscript and a negative kern to compensate the \scriptspace. The rest of the code takes care of the unwanted additional depth caused by the empty subscript.
$\textit{AB}\textit{CD}$.in comparison has no space between the B and the C (apparently italic correction for\textitis deactivated in math mode). But forcing italic correction with$\texthit{AB\/}\textit{CD}$.gives the same output as$\mathit{AB}\mathit{CD}$. Is it because\mathitapplies italic correction with a multiletter argument? – Géry Ogam May 26 '15 at 17:48$\mathit{A}\mathit{B}\mathit{C}\mathit{D}$there's an added kern only after D (with a single character in the argument of\mathitTeX doesn't build a subformula, which it does with multiple characters). – egreg May 26 '15 at 17:52\mathrmand\textrm:$\mathrm{AV}\mathrm{AD}$.and$\textrm{AV}\textrm{AD}$.have a different output, but the difference is way subtler that I would expect (you have to zoom in to notice that the distance between A and V in the\textrmformula in less large than that in the\mathrmformula, while I would expect the distance in the\textrmformula to be the same as inAVAD.in text mode). – Géry Ogam May 27 '15 at 22:40Vhas a non zero italic correction. So in\mathrm{AV}\mathrm{AD}there is a kern after V, which is not added for\textrm{AV}\textrm{AD}in math mode. – egreg May 27 '15 at 22:42$\textrm{AV}\textrm{AD}$.doesn't display likeAVAD.(in text mode)? That is with a negative kern. – Géry Ogam May 27 '15 at 22:48$\textrm{AV}\textrm{AD}$makes two subformulas, which end up in two different\hboxes. No kern is thus applied. – egreg May 27 '15 at 22:51$\textit{Af}\textit{fD}$? (The two f are very close.) – Géry Ogam May 27 '15 at 23:11fis close to the second one because it extends past its nominal bounding box. – egreg May 27 '15 at 23:30