I think of politics on three levels: 1) abstract normative level, 2) ideal institutions level, 3) non-ideal, “second best” level. In day-to-day analysis, people, myself included, move between these levels without making note of it, which can lead to considerable confusion when trying to figure out what label to attach to someone politically.
So for me, on the abstract normative level, I am an egalitarian. I think equality (e.g. of capabilities) is the principle that should guide how we construct our institutions. I also subscribe to the Rawlsian view that deviations from equality may be justified, but only if it leaves those on the bottom of the disequalizing move better off than they would otherwise be. But then I also subscribe to G.A. Cohen’s skepticism that substantial disequalization is as necessary as many make it out to be.
On the ideal institutions level, I am a market socialist. I have no particular interest in markets or even necessarily social ownership of capital, but I think these institutions best cohere with egalitarianism, at least of all of the sets of institutions that are understandable to me (in the sense that I can envision how they will successfully function). But I could be moved off of this if someone could convince me another set of (functional) institutions were better suited to egalitarian principles.
On the non-ideal, “second best” level, it’s very day to day. The basic notion of “second best” considerations is that you have to stipulate (at least in the short term) that you cannot get all the institutions to align with the ideal institutions you think best cohere with your abstract normative principles. So that means you have to assume that at least X, Y, and Z will not be what you want them to be. But if X, Y, and Z are not what you want them to be, that can have cascading effects on whether A, B, and C are actually worth doing. Just because A, B, and C would be the best under ideal institutions, that does not mean they are the best under non-ideal institutions. This is very messy analysis, needless to say.