Should the SEC self-finance?

I haven't seen this issue receive much attention on the usual blogs (Yves Smith is one exception).  Here is one argument for self-finance:

The Obama administration has requested long overdue increases in both budget and staff for the S.E.C., and has plans to add as many as 374 employees. Those increases are vital, but because they’re dependent on Congress, there is no guarantee that they will be sustained.

Instead, the commission should finance itself – much as the Federal Reserve and the Federal Deposit Insurance Corporation do today through fees on banks. These two pivotal financial regulatory agencies thus have the flexibility to adjust their own staff.

And it appears easy:

Such a self-financing system would not mean higher fees; the commission collects far more in fees from corporate filings and stock market trading than it gets from Congress. But those fees go back into the federal coffers. In 2007, the S.E.C. brought in $1.5 billion, almost twice its 2007 budget.

This seems like a short-run improvement, but the idea nonetheless makes me a bit nervous.  What will it look like in practice, ten or fifteen years from now?  Was reliance on fees in every way beneficial for the FDA?  Admittedly, self-finance is one pathway to higher levels of finance, but the two issues are conceptually distinct and we might prefer to implement the appropriate level of finance directly through Congress.  I fear that in the longer run self-finance means that the SEC never wishes to see the financial sector shrink.  (Of course maybe it's not going to shrink anyway.)  A related question is what kind of internal controls the SEC would need to maintain its own fiscal discipline and prevent overspending, backed by an excess raising of funds and fees.  So whether self-finance is a good idea probably depends on what you are comparing it to.  A final question, and not a small one, is whether you think the SEC should be more independent from Congress.

Here is the SEC's own case for self-finance.  Here is a 2002 GAO study of the idea, very useful and it also discusses other cases of self-finance among regulatory authorities.

I don't have any strong conclusion here other than "maybe."