Late cycle financial innovation: Are private credit funds the new MBS CDOs?

The FT produced an interactive graphic last month exploring the many ways that banks are providing funding to the world of private equity, and asked the question: “could the private equity industry pose a risk to the wider financial system?”

FT interactive graphic

In my view this question should be placed in the context of what we learned from the Great Financial Crisis. First, one should remember that from 2004-08 just as US mortgages were being used as the raw material for mortgage-backed securities (MBS) and for collateralized debt obligations (CDOs) made up of MBS tranches (MBS CDO), the leveraged loans that support private equity deals were being bundled into collateralized loan obligations (CLOs). In the denouement of the crisis, however, it was the mortgage-based products that blew up, and the CLOs for the most part did okay. There were two reasons for this: first, the CLOs mostly securitized senior debt not subordinated debt like MBS CDO and this meant that they had much less risk (as will be explained below), and, second, the Federal Reserve’s low-for-long interest rate policy rescued the leveraged loan borrowers who were for the most part able to refinance their way out of their obligations. In short, while 2008 saw a reckoning in the mortgage market, there was no reckoning in the leveraged loan market.

The fact that there was no readjustment in the leveraged loan market despite the bubbly lending from 2004 through 2007, may be a sign that now that interest rates have readjusted to higher-for-longer, private equity may finally be facing the first real downcycle for the product. If this is the case, then we should look carefully for the financial industry to engage in a replay of the run-up to the 2008 crisis by designing products that will facilitate the transferal of upcoming losses to the bagholders: institutional investors, unsavvy hedge funds, governments of every shape and size, and a few banks with poor risk controls.

To understand this process of using financial innovation to transfer anticipated losses to unsavvy bagholders, let us review what happened in the US mortgage market in the run up to the 2007-08 financial crisis. Fixed income products were the vehicles for this process, precisely because they pay a fixed, bond-like return, and thus if you can convince the purchaser to buy into a leveraged product or something with equity-like risk, the purchaser is buying downside only leverage. That is, the purchaser is bearing significant risk of loss, for a fixed return that was frequently less than one or two hundred basis points over LIBOR. As one might imagine these products are very useful for a seller that is seeking to lay off risk.

Let’s review the 2004-08 financial products in more detail. Mortgage backed securities (MBS) hold as assets a bundle of mortgages and sell interests in that bundle. Typically the liabilities of the MBS are tranched, so that they are comparable to the liability structure of a corporation. There is an equity tranche, typically making up about 3-5% of the structure, that bears any losses on the mortgages first. Mezzanine or subordinated debt tranches bear losses next (making up 8-20% of the structure). Note that the uppermost subordinated debt tranches could in practice be rated AAA. Thus, there was a ‘super-senior’ AAA tranche, making up 75 to 90% of the structure, that will only incur losses after all the other tranches are wiped out.

At the peak of the US bubble, a second stage of securitization took place when the (relatively high risk) mezzanine tranches of MBS were packaged into securities and sold to investors as MBS CDO. Even the super-senior AAA-rated tranche of these products was wiped out – transferring losses to a wide range of investors who had been looking to put their money in ‘safe’ debt. (The synthetic CDO which effectively provided insurance to MBS investors who didn’t like the risks they were running anymore – such as Goldman Sachs – was another way of transferring the losses.)

A schematic understanding of how securitization was used to disseminate losses in the MBS bust after 2007 is useful.

1. Concentrated risk that pays low, fixed-income returns
   Mechanism: a subordinated debt product
   All subordinated tranches of a securitization (such as MBS) bear concentrated risk and therefore are equity-like. This is because only the super-senior AAA tranche of the securitization benefits from the diversification in the MBS. Because the subordinated tranches, even those rated AAA, are protecting the senior tranche from loss, they face the risk of all the loans in the securitization package that sit above them in the liability structure. Thus, even though the value of the equity plus subordinated tranches only comprise, say, 13% of the value of the whole package, this thin segment has to absorb all realized losses on the whole package. Thus, when loss rates on the underlying loans turn out to be high, it becomes likely that the subordinated tranches will face losses of 100%.
2. Cliff risk
   Mechanism: packaging of subordinated debt into a new product
   Whenever subordinated tranches are securitized (e.g. MBS CDO) the concentrated risk embodied in the subordinated tranches means that the two-tier securitization has ‘cliff risk’ (see Nomura 2005 ‘CDOs-Squared Demystified’). Thus, in the case of the MBS CDO, if the general default rate exceeds a threshold level, such as 20%, even the super-senior AAA rated tranche of the securitization will lose 100% of its value.

Overall the investors in all of these ‘fixed-income’ products, except for the super-senior AAA tranche of a single-tier securitization, were buying downside only leverage, or in other words equity-like risk with the returns of a bond. In short, these are products that concentrate risk. Second-tier securitization (and the cliff risk associated with it) was only possible because there were some very large market participants that were willing to bear the risk of the super-senior AAA rated tranche. Examples from 2008 include Citibank, UBS, and AIG.

Because it is not clear why anyone should agree to a bond-like return for such high-risk products, the structured finance industry that created these products collapsed after 2008. Thus, we do not expect to see these products re-emerge. The question this post wants to ask, is whether the same basic principles for transferring risk to bagholders may be applied to the private equity space.

So let’s think of private equity portfolio companies using the MBS framework. Each company is a package of business activities that currently is structured with equity, subordinated debt, and senior debt investors.

Institutional investors are already exposed to the losses in PE because they are limited partner (LP) investors in private equity funds. In this role as investor, they typically (as I understand it) share pari passu with each other and with the GP investor who manages the fund. In other words there are no tranches. There are then two principle ways that the risk of a private equity downcycle can be transferred to the bagholders.

I. Concentrate the risk borne by LPs

Now that there is a substantial risk of upcoming bad performance, the goal is to get the LPs to bear more of the losses, so the challenge is to think of ways to make the risk of the LP’s equity position more concentrated. The solution is to make the LP (and GP) interest subordinate to additional claims on the assets that sit in the PE fund.

This can be done by borrowing at the PE fund level and giving the lender a senior interest in the fund’s assets. The GP can use all (or some) of the assets in a fund as collateral for borrowing by the fund. This is a net asset value or NAV loan. When a fund takes out a NAV loan, instead of owning interests in different companies that can be expected to perform independently, losses incurred on one company may have to be covered by gains on a more successful company in order to pay off the NAV lender. Now (possibly without even noticing it) the LP’s pari passu interest in all the fund’s assets has been converted into a subordinated interest in those assets, as the NAV lender has a prior claim. In short, the risk of the LP’s equity position has been concentrated.

Observe that the GPs will have to bear the increase in risk on their equity positions too. The question however is whether the GPs are taking other countervailing positions that will benefit them in the event that their equity stakes go to zero. To be clear, I am not aware of evidence of this, but given that apparently most limited partner investment agreements include a waiver of GP conflicts of interest (Yves Smith 2013), it certainly seems possible that some GPs will take investment positions that will protect them in the event that the sector faces significant losses.

II. Create a product that packages subordinated debt into ‘cliff risk’

Where can the ‘cliff risk’ of the MBS CDO show up? Private credit funds are the most likely candidate. Like private equity funds, LPs invest alongside GPs in private credit funds. A private credit fund can lend directly to a private equity fund in the form of a NAV loan, or it can lend directly to private equity portfolio companies. Given that it is the most senior lending that banks are willing to provide to portfolio companies, private credit funds that focus on lending directly to portfolio companies are likely to end up owning a bundle of subordinated debt obligations. That is, private credit funds are probably taking up the segments of the lending market that is hard to place, just like MBS CDOs used to do. Recall also that the equity in many of the debtor companies have been used as collateral for NAV loans, so that these debt obligations cannot be treated as independent: one company’s default can have knock on effects on a sister company. Thus, while the private credit funds should perform better than the equity in PE funds, because they hold claims that are senior to the equity, they are also very much at risk of incurring significant losses in the event that bankruptcy rates on private equity portfolio companies are unusually high.

Furthermore, the LP interest in the private credit fund is often a subordinated interest, because of borrowing at the fund level. Thus, instead of bearing risk pari passu across all of the assets in the private credit fund, the risk is concentrated as there is a senior creditor with an interest in the fund. Borrowing by the private credit fund is one way for this to take place. There are, however, also more complicated structures: for example, Blackstone has used the assets in its private credit fund, BCRED, to raise money using a collateralized loan obligation (CLO). The credit fund investors then hold the equity interest in the CLO. Thus the LP investors in Blackstone’s BCRED private credit fund now hold an interest in the assets that is subordinated to the other CLO investors. They hold concentrated risk. For this reason, in the event that portfolio companies start to default at high rates, private credit fund investors are likely to face losses or even lose their entire investment.

Furthermore, if it turns out that a bank has been convinced to fund the super-senior tranche of the BCRED CLO or a similar product, then the analogy with MBS CDO will be complete. While it would not surprise me if a few banks get into trouble by misreading the risks in this sector, it appears to me that the LPs are at much greater risk of significant, or even devastating losses.

These then are my questions: Are the LPs being set up to be the bagholders in a private equity downcycle? Have they been paying attention? Do they understand how the risks of their ‘equity’ investments are being concentrated by fund borrowing? Do they understand that private credit funds may end up behaving like MBS CDOs – just like the senior creditors in the underlying MBS were the only ones who were not wiped out, it may be that only the senior creditors in the underlying portfolio companies will be left with a valuable claim?