I am no longer directly involved in markets, and I certainly I have no expertise in the overhaul of derivatives plumbing that happened after the Financial Crisis. But based on my experience in the Financial Crisis and what I have heard about the LTCM crisis, I urge caution with respect to explanations of financial crises. Unless there is a comprehensive inquest by the authorities, even the most detailed journalistic investigation is likely to be a partial explanation. During a financial crisis, all the dead bodies come to the surface, and everything is going wrong at the same time. Meanwhile, experts in particular areas love to over-emphasise their area of expertise. (I had a lot of contact with swap strategies, so that is my beat, but I will repeat — interest rate swap issues might not explain everything.)
Since people often do not read to the end of articles, I will summarise my guess of what has happened: the mark-to-market losses on interest rate swaps would create margin calls and a need to post collateral. The swaps would not be 100% of the collateral, but given the chunky size of the positions, they would have been a good portion of the problem. The need to post margin created a cash drain on some entities, that was worrying enough for the Bank of England to step in.
Liability-driven investment appears to be one of the management consultant buzz phrases that has vaulted into the headlines. Although particular strategy structures might be new, the premise is solid: pretty much all fixed income management after the 1990s is liability-driven.
Insurance companies need to hold fixed income assets to match the structure of their actuarial liabilities (which depends upon the policy mix — annuities are a very different liability than fire insurance).
Pension funds manage their assets versus the actuarial liability of projected pension payments.
Fund managers manage their portfolios versus the implicit liability of their benchmark.
Hedge funds manage their portfolios against some hurdle rate, typically a short rate plus a spread.
Sovereign reserve currency funds are one of the few large actors with the greatest discretion, but they are probably hemmed in by guidelines. Macro funds also have discretion in taking duration exposure — but their balance sheets are tiny versus the previously listed entities.
Pension Liability Management With Swaps
If we go back to the 1980s, pension management was amateurish and dangerous for pensioners. Pension funds held assets that were only vaguely related to their promised payments. They could get away with this because their workforces were young, and growing. However, the relentless headcount reductions, ageing of the workforce, and the shrinkage of firms that offered defined benefit pensions meant that there were no longer a wave of contributions of youngsters to make payments to oldsters.
Pension failures and accounting shenanigans led to a tightening of the regime, and the U.K. had the strictest rules. Pension funds had to have assets that generated long-term safe cashflows, and so they were crowded into long-dated bonds in at least part of their portfolio. The long end of the U.K. gilt curve (short for “gilt-edged security”) was described as the “freak show of the international capital markets” in the 1990s in a fixed income pricing textbook that is sitting somewhere on my shelf.
Chronic short-termism by modern management means that the private sector cannot produce truly safe long-dated bonds (other than regulated utilities). But buying long-dated central government bonds poses a problem: government securities are treated as having the lowest potential returns of any asset class. So using them to match liabilities is “expensive” in that it lowers the prospective returns on the portfolio. This implies greater contributions to meet the same payment commitments.
Investment banks came to the rescue: use interest rate swaps! Interest rate swaps are economically equivalent to borrowing to buy a bond. Using them allows a fund to implicitly use leverage against their portfolio: they get the cash flows from synthetic long-term bonds, but they “borrow” against those bonds to invest in assets with higher prospective returns (e.g., equities).
Aside: What Is A Swap?
A pension fund that would want a £100 million “synthetic” 30-year bond would enter into a swap contract that implies the following cash flows.
The pension fund receives the fixed coupon cash flows that would be generated by a 30-year bond, with a coupon rate the variable to be negotiated — the swap rate. For example, if the swap rate was 4%, the pension fund would receive fixed payments of (about) £4 million per year (with the exact payments determined by wacky bond market conventions). This is a “synthetic bond.”
The pension fund pays the floating interest rate on a £100 million loan (quarterly?), using a public short-term interest rate fix (historically, typically LIBOR). This is a “synthetic loan.”
The £100 million principal payments between the synthetic bond/loan at the 30-year point cancel out.
By convention, the swap rate is normally negotiated so that the net present value of the two payment legs are equal, so it costs £0 to enter into the contract. This means that the pension fund can replace a £100 million 30-year gilt position with a swap, and something more exciting, like £100 million worth of shares in a Chupacabra ranch in Arizona that a manager heard about at a conference in New York.
By itself, the strategy is entirely reasonable, and one of the large pension funds in Canada followed it and was highly successful. The problem is that the strategy generates liquidity management concerns, as discussed next.
A pension fund is flush with assets, it should have no problems meeting the individual cash flows on the bond as they arise. The problem for the pension fund is that it needs to post collateral on the position, which increases as the net present value of the swap falls.
The net present value of the position is equal to the discounted cash flows on the swap. For the pension fund, it is equal to the value of the synthetic bond, less the value of the synthetic loan. The synthetic loan will have a value that is roughly equal to £100 million regardless of market conditions (there is a small amount of value wiggle), but the “bond” price will move up and down with the market swap rate like a 30-year bond (yield up, price down).
As swap rates rise, the pension fund would need to post more collateral — loosely speaking, a margin call.
Although this might sound scary, interest rate risks are easy to model. With modern risk management software, the collateral needs — as well as the drop in the value of collateral — could be accurately for any interest rate shock scenario. You pick your worst case interest rate scenario, and build your liquidity management around that scenario.
Aside: Cross-Currency Swaps
I have seen assertions that cross-currency swaps were involved. Given the history of currency swap involvement in financial crises, this would be unsurprising. However, a sane application of currency swaps in liability management would be qualitatively different.
It is entirely possible that a U.K. pension fund looked at the GBP swap yield curve, and said “nope.” They might have looked greedily at the yields at the long end of the USD yield curve. So instead of holding a £100 million 30-year gilt, they want to buy a 30-year Treasury.
The problem with just buying an U.S. Treasury outright is that this is matching USD cash inflows versus GBP liabilities. This does not work for liability matching. So the pension fund would use a cross-currency swap to hedge the USD into GBP.
Since I am lazy and want to make life easier for readers, I will assume that the exchange rate when they initiated the transaction was that 1 GBP was worth 2 USD. They enter into a cross-currency basis swap that is economically equivalent to lending £100 million GBP in exchange for borrowing $200 million USD, with both (synthetic) loans paying a floating rate plus a spread (the cross currency basis swap quoted spread). This allows the pension fund to buy $200 million piece of a 30-year Treasury with the £100 million GBP that would have been used to buy a gilt. There are floating rate payments to be made/received, but the swap cash flow that matters is the need to exchange $200 million USD for £100 million GBP in 30 years.
(Note that the exchange rate on ending transaction is the same as the entry level: the floating rate payments compensate for interest rate differentials. This is different than a currency forward, where the forward rate offset on a given date is driven by the fixed rate differentials as well as the cross currency basis swap spread.)
Well, what happens if the spot exchange rate drops to £1 GBP = $1 USD (which is a 50% depreciation in the pound for people like me who have to think about what forex quotes mean)? Since the spot exchange rate is £100 million GBP for $100 million USD, the swap contract implies $100 million future loss. Although discounting reduces the magnitude of the loss (I think…), this is still a big bucket o’ money to lose on the derivatives contract.
Of course, the whole purpose of the exercise was that it was a hedge. The fund owns a $200 million in U.S. Treasurys, and the increase in principal value in GBP terms matching the loss on the cross-currency basis swap.
Although this might look fine on a balance sheet, the problem is that we have two distinct instruments. The cross-currency basis swap loss implies a need for more collateral, and the risk is that the USD-denominated bond might not be available as collateral. Even worse, Chairman Powell has been torching the value of long-dated U.S. Treasuries in USD terms, and cross-currency basis swaps are floating-floating instruments, so they sit out the fixed interest carnage. (A fixed-floating currency swap would be a better hedge, but the floating cross-currency basis swap market is the main wholesale funding market.)
If the above scenario happened to any extent, the story is slightly different. The funds were not technically leveraged, instead they had overseas investments that were hedged back into GBP. The problem was that the drop in the GDP would increase the need to post collateral for hedges, adding to the funding stresses within the GBP markets.
Although it is clear that somebody blew up, absent a comprehensive inquest by the authorities, we do not know whether (almost) everybody blew up. Everybody blowing up is the traditional cause of fixed income derivatives crises, and face it, the Brits are sticklers for tradition. This article explains why rising interest rates and a falling pound would cause major drains that would either start or exacerbate a funding market crisis.
Outside the GBP markets, the hope is that there is a greater variety of positions. For example, when I worked in finance in Canada, only a couple of the major players were heavily into using swaps for liability-driven investment. The largest players (including my old employer) used swaps, but on a discretionary basis — so they had the capacity to lean against squeezes, and the balance sheet capacity to endure squeezes on their positions.