On Tuesday, saddle up for what could prove a marathon mark-up of GSE-reform legislation. In it, Fannie and Freddie are to be blown to smithereens. However, the third housing GSE – the Federal Home Loan Bank System – will be dramatically empowered if drafts of the measure are any guide. The FHLBs are getting all the goodies because they are seen as a refuge for community bankers frightened to death of a private mortgage-securitization market. But, a close look at the FHLBs shows their growing dependence on high-risk insurance companies. This poses serious policy issues Congress would do well to address before it has to follow the messy clean-up of Fannie and Freddie with an equally-costly one for the FHLBS.

The Office of Finance’s most recent data show that MetLife has climbed into the top-ten of System borrowers. Its $15 billion outstanding isn’t all that much in the System’s total $499 billion advance book. More interesting, though, is the fact that insurance-company borrowers now hold slightly over twelve percent of system advances – making them a big borrower in aggregate.

What’s so worrisome about FHLB draws for big insurance companies? Leaving aside the ostensible housing mission of the System, it poses two serious safety-and-soundness and perhaps even systemic worries: does each of the Home Loan Banks lending to insurance companies have stable earnings to limit concentration risk as well as the credit and interest-rate controls in place to protect it from a borrower default since – unlike in a bank failure – it can’t count on collateral when insurance-company borrowers go bye-bye.

Unlike little-old community banks and even some very big ones borrowing from the FHLB, insurance companies are generally not using the System to fund mortgage finance.   They thus will come and go as opportunistic investment needs require. If they go fast, that will put several-fragile Banks under acute earnings pressure. If they grow only larger because market stress forces them into the System’s arms, the risk could magnify. As noted, there’s no prior lien – that is, a call on all assets ahead of the bankruptcy estate – when a System borrower isn’t an insured depository. The FHLBs have prided themselves for decades on never taking a loss, but that’s only because they clean out every bank before the FDIC comes in – check out Indy Mac for a very costly case in point.

That insurance borrowers are opportunistic and, thus, risky, is demonstrated by a hard look at a recent major case in point: Two Harbors, a $17 billion REIT with a captive insurance company. Two Harbors is publicly-traded, but it’s closely linked to Pine River, a large hedge/PE company.   One of the benefits of these reinsurance companies – particularly to hedge funds – is that the captive funnels capital into the parent that funds its ventures without the redemption risk inherent with the parent’s investors. As long as the reinsurance company meets whatever capital requirements Bermuda demands of it, funds can happily cross the blue water back to New York. The reinsurer can also invest in mortgages supported by servicing rights housed by the parent – Two Harbors just bought a bunch of these from Flagstar – creating not only a lot of profit potential in the mortgage space, but correlation risk reminiscent of the ill that befell muni-bond insurers who insured mortgage positions they also held as investments. With only the capital constraints of whatever regulation covers the reinsurance company and how much leverage investors in the parent will stomach, this mortgage model has terrific profit potential but, of course, a heck of a lot of high-flying risk.

But, even if all the credit-risk pieces fall the right way, the REIT/insurance model needs funding, and here’s where the FHLB System comes in. REITs and hedge funds in this space are major players in the repo market. They ducked a bullet when the new Basel III leverage rules decided against a punitive capital charge for banks providing the securities financing necessary to play in this space, but they are still facing an array of challenges as the FRB continues it crusade to curtail repo risk. The Fed is determined to demand high margins from repo counterparties, margins most of these non-banks will struggle to post.

So, what else to do? In Two Harbors’ case, it’s to ask the Federal Home Loan Bank of Des Moines for a $1 billion line, a line it got late last year with the usual strings – e.g., collateral quality, a shut-off valve – attached to it. These strings are, though, loose and thus can’t pull back expended FHLB funds, although of course the Bank can seize collateral if a REIT or hedge fund can’t downstream funds to the captive-insurer/borrower.

Or, it can try to. Without a prior lien, it will have to stand in line. FHFA – the System’s regulator – has long talked of forcing the FHLBs back to their basic business of supporting community banks with simple funding structures that pose little risk. Until it does, though, the System is taking on more risk from new borrowers with no meaningful improvement in capital or prudential controls. Sound familiar?