For those uninitiated in the technicalities of NIPA and national accounting logic in general, this comment may seem strange. What's the issue with the Financial sector running a "nonzero" balance? As you'll see though, there are a number of ways where such an approach comes into conflict with the traditional treatment of the NIPA accounts and need readjustment if the approach I would like to take is to be seen as consistent.
As it Happens Ruggles and Ruggles, in their excellent book "National Accounting and Economic Policy: The United States and UN Systems" tackle precisely this issue. The crux of the issue Mason refers to is summarized at the beginning of the chapter:
"The approach of the United Nations System of National Accounts (SNA) to privately funded pensions and insurance is essentially a neo-classical one.Apart from the costs of operation, private pension contributions and life insurance premiums are considered to be a form of household saving, part of the accumulation of wealth by households that should appear as a category of assets on the household balance sheet. But publicly funded schemes- social security arrangements-- are not treated in this way. Entitlements under public programs are not credited to households until such time as the benefits are actually received. "
Thus, in one treatment there are private transfer payments and in the other only the government makes transfer payments. To make the difference between these two approaches clear, the two sets of T accounts below may be of some use. One Note: here subscripts (e.g. HH, MM, C etc) denote the sector making payment
As you can see from the above, by definition Institutional Investors are in financial balance at all times. Any shortfall or excess of cash flow is counteracted by an offsetting change in their obligation to households. Thus what we would otherwise consider to be the financial net worth of institutional investors is counted as part of household wealth.
From the point of view of flow of funds analysis, there are a number of problems with this type of approach. For one one's claim on "future pension obligations" is not transferrable and thus is not a financial asset as such. The fact that I'll eventually have a certain income in some undetermined future has a very different impact on current spending programs than having assets one can sell or pledge. Second, "excess" cash flow into pension funds, insurance companies etc don't guarantee me or even the household sector as a whole future payments. The conditions under which these institutions pay out are very strongly defined in contracts and aren't very sensitive to current financial surpluses. Someone looking at household wealth with these numbers included will get a very misleading picture of what's going on. Ruggles and Ruggles argue that treating them as transfer payments and thus separating out Institutional Investors net worth from household wealth is the best way of approaching this issue and I agree
So to finally get around to answering Mason's question in brief: transfer payments between private sectors change saving rates in such a way as to make Institutional Investor financial surpluses and financial deficits a consistent accounting concept.
UPDATE: I forgot to add the sectoral balance equation with private transfers
Ruggles, Nancy D., and Richard Ruggles. National accounting and economic policy: the United States and UN systems. Edward Elgar Publishing, 1999.