Friday, September 30, 2022

A familiar finance fable in UK bonds

Guy Adams in the Daily Mail has an intriguing story of what's going on in UK bond markets.  It's intriguing because it's so utterly familiar. And it reveals that all the masses of regulation and armies of regulators aimed at preventing exactly this sort of thing from happening again and again have failed again. 

UC pensions take in contributions when people are young, invest them, and then pay out fixed amounts when people get old. They hold large quantities of government bonds, currently 1.5 trillion pounds per Adams. That's a good strategy: if you have fixed payments to make, invest in risk free assets that provide fixed payments, and ignore the mark to market. But it fell apart in a classic way. 

Leverage

As often is the case, however, they didn't have enough assets to pay out the promises. So... Lever up! Pensions used their government bonds as collateral, borrowed money, and invested that money in more government bonds or, to a lesser extent, in stocks or other investments. 

So now rather than a portfolio that matches liabilities (pension payments) with assets (government bond coupons), you have a portfolio that is making a bet on declining interest rates and rising stock prices. 

How many times have CEOS and university presidents set (average) rate of return targets, and just levered up to meet them? Risk 101. 

It paid off. For a while. 

 in the ten years to January, while stocks recovered from the financial crisis, the proportion of pension schemes which were under-funded fell from 80 per cent to around 20 per cent.

During the Covid crisis in 2020, the yield of gilts (then regarded as a ‘safe haven’) fell to almost zero, meaning that their value rose substantially.

Yet, as ever in the world of investing, greater rewards tend to carry greater risk. And with pension funds becoming more and more comfortable using LDIs to dramatically ‘leverage’ their holdings – at times using them to purchase £3 of gilts for every £1 they actually invested – they would inevitably become sensitive to changes in market conditions. ...

I.e. rising interest rates.  This is a huge bet. The pensions are borrowing short term to buy long-term bonds. Small changes in long-term interest rates can imply huge changes in long-term bond prices. 

 By early this week, gilts which had been worth £3.50 in January were valued at as little as 50p. 

Now wait a minute, you say, interest rates haven't gone up that much. But long term bond prices are very sensitive to small changes in interest rates, especially when interest rates are low. For example, a 30 year zero coupon bond paying 100 pounds is worth 100 pounds when the interest rate is zero. If the interest rate goes up to 2%, that is the same thing as the price going down to 100/(1.02)^30 = 55.2 pounds, half the original value. Borrowing to invest in long-term bonds is a very risky strategy on a mark to market basis.

Selling options

When interest rates are zero, they can't go down any more. So, you can see that leveraging in to long-term bonds in a time of low interest rates has a quite skewed distribution of returns. You can make a little bit of money most of the time, if interest rates stay low or go down a hair more. But when interest rates rise, you can lose a lot of money all of a sudden. This is a tried and true method of boosting fund performance. For a while. "Write out of the money puts." "Short volatility." "Write insurance." "Collect pennies in front of a steam roller." Then of course the grim reaper comes and it all falls apart massively. 

Collateral 

Well, wait, you ask. Who cares about mark to market? The market value of the pension fund's liabilities -- the pensions -- has also gone down. The mismatch is  bad, but not that bad. 

Oh yes it is. When the value of the government bonds used as collateral falls by half, the fund has to come up with cash to cover the mark to market losses on their collateral:  

This [loss] in turn meant that pension funds were required to hold less cash as collateral for the bonds they kept in LDIs, allowing more money to be returned to their pot.

Again, in plainer English. You have 1,000 pounds of bonds. You use that as collateral to borrow 1000 pounds, and invest in another 1000 pounds of bonds. The bond prices fall to 500 pounds. Now you have to come up with 500 pounds cash, right now, to replace the missing collateral on your loan. You must sell the  500 pounds of bonds you bought immediately.  

For pension funds which held LDIs, this caused immediate problems. The UK gilts they were using as collateral on other investments were suddenly worth an awful lot less and cash was needed to plug the gap. Unfortunately, the only way they could raise cash was to sell large quantities of gilts, which in turn had the effect of driving the price even lower.

This is a risk-management failure that happens over and over. Even seemingly good hedges or ``arbitrage''  strategies fall apart because you have to post extra collateral when markets move against you. 

Spiral

Of course, that meant they needed even more loot. And so a sort of vicious circle was created. ‘I have called it the death spiral,’ Mackenzie said last night. ‘That’s what we were in. A sort of volatility vortex. It became self-fulfilling. Pension funds were essentially eating themselves.’

For a few short hours, early on Wednesday, the market became effectively paralysed. That forced the Bank of England to take dramatic action, stepping in as a sort of ‘buyer of last resort’ and agreeing to plough about £65billion into gilts over the coming weeks.

Now wait a minute, you say, especially if you got your degree at the University of Chicago. If people are selling in a panic and driving prices down, why doesn't Goldman or Blackrock stop playing ESG games for 5 minutes and scoop up the bonds? Well, they don't, at least fully, and at the very high frequency we're taking about. Collateral has to be posted daily.  (Just why they don't is a very good question.) 

Creative finance 

The scheme was called "Leveraged Liability-Driven Investments," and involved a lot of fees. I presume the fees and complexity are necessary to get around regulations that said funds have to hold government bonds to match their payouts and not take risks. Yes, we're holding those bonds. We just incidentally borrowed against them to buy something else, and unwind the regulation. 

The genius of these products was that, despite performing solidly, they looked extremely conservative – at least on paper. Government bonds are among the most reliable investments. The world of pensions is designed to be dull. While the sheer volume of gilts being packaged into these products was unprecedented, regulators and the vast majority of observers saw nothing in the way of red flags.

But really, is it that hard to see though?  We teach every MBA the lesson of mortgage-backed securities held by "special purpose vehicles" that issued overnight debt and how that fell apart. 

Prophets ignored

Perhaps regulators can complain that nobody could see it coming. But as usual at least some people did see through the scheme. 

One day in May 2019, a Dutch financier named Hans Van Zwol hit ‘send’ on a document outlining what he suggested was a ‘terrible’ threat to the global economy....

... the bespectacled fund manager had become concerned about a range of complex products known in the trade as ‘Leveraged Liability-Driven Investments’ or LDIs. 

Yet neither the industry nor regulators seemed to act. Indeed, as recently as July, an article by a writer in the Financial Times noted that: ‘LDI managers claim that their activities pose no systemic risk and I read the Bank of England financial policy committee’s silence as agreement.’

Events of recent days would, of course, prove them (and the Government) quite wrong. Negligent, even. 

Well, prophets are only recognized ex-post. But regulators cannot say that it was impossible to see. 

The coverup

I was a loud critic of the Dodd-Frank approach to regulation, arguing for lots of equity rather than continued high leverage and the fantasy that regulators would really really see asset risks building up next time. That has obviously failed once again. But at least they had the decency to notice something went wrong, to say the words "moral hazard" and to do something about it. 

The Fed engaged in a massive bailout in 2020, and nobody is peeping a word about it other than congratulating the Fed for saving the world once again. Why did the world need saving, in all the ways that Dodd-Frank and stress tests were supposed to prevent? Silence. 
As the dust settles, and taxpayers count the cost, there will naturally be calls for a prompt investigation into how pension funds were allowed to invest billions in such high-risk products. Smooth financiers, who for years profited from these exotic deals, will shrug their shoulders while politicians deflect and regulators seek others to blame.

History, meanwhile, suggests no one will be properly held accountable. And therein lies the real scandal. For as so often when financial markets implode, the outrageous fact is that this was well and truly a crisis foretold.

As with the FDA and the CDC we have entered the CYA era of government, in which nobody can admit mistakes or reform institutions. 

Over and over

I italicized the headings to emphasize how common this story is. Over and over, plus ça change, plus c'est la même chose. And now the same crew, having missed debt financed mortgage backed securities in special purpose vehicles (2008), having missed banks loading up on Greek debt (2010), having completely missed that a pandemic or other supply shock might cause trouble (2020), having missed that energy prices might rise or war might break out (2022), is off on the quest to stress test banks for climate change. 

Update

Great coverage by Robert Armstrong at FT Unhinged. Armstrong adds the use of derivatives to goose the portfolios, and points out that the 1% or more interest rate moves happened in a matter of days, really pushing the collateral channel. He also asks why central banks don't directly control long-term rates rather than, or rather than just, short term rates. It's a good question. 

Great coverage here by Allison Schrager at Bloomberg. HT Casey Mulligan.




19 comments:

  1. In the U.K., L.D.I. was encouraged by the regulators, according to reports this week in The Wall Street Journal. In those same WSJ reports, it is said that L.D.I. is not used as much in the U.S.

    Different country approaches, e.g., U.K. vs. U.S., will limit contagion to an extent in much the same way contagion was limited to ASEAN members and S. Korea during the Asian banking crisis of the 1990s.

    Why would Blackrock, etc., jump in to U.K. gilts? Blackrock's ESG funds are limited to its non-Index ETFs and its actively managed mutual funds.

    Margin is limited to 90% in U.S. Treasuries. Can't say what it is in the U.K. markets, but it is hard to believe that it would ever be 100%.

    You have a typo in the 1st sentence of the 2nd para.

    Mark-to-market is required for trading stock. Securities held to maturity are not required to be marked down to market values. But the choice of which accounting and tax rule to adopt for securities designated as held to maturity is at the option of the fund management.

    ReplyDelete
    Replies
    1. Less common in the Us, but there are definitely some US pensions that do it. I know that MSERS does something similar. Not sure if the leverage is structured exactly the same, but they run 40% leverage, I believe using treasuries. See their annual report. https://www.mosers.org/docs/default-source/funding/annual-reports/2021-acfr/2021-annual-report-financial.pdf?sfvrsn=b6e17ec1_13
      Was flabbergasted when the CIO mentioned this on stage at a conference

      Delete
    2. MSERS has adopted an "Overlay" strategy to manage the pension fund's estimated unfunded liabilities. The net position as at June 2021 was 63% of actuarial determined liabilities. I exclude the Judicial Plan from these remarks. Investment expense is $98 mm, losses on derivatives is $20 mm. Growth in net fiduciary position is attributed to investments in equities, hedge funds, etc. 2021 was a banner year in equity markets. One would be interested to see if the strategy being followed in fiscal 2021 resulted in a large decrease in the net fiduciary position in the year ending June 2022. Fiscal year 2023 will be of even greater interest, in terms determining hedging potential given the price declines this year in both fixed income and equities.

      Delete
  2. "[T]he same crew (...) is off on the quest to stress test banks for climate change. "

    To expect that they will proceed with care is almost delusional at this point. Still, I'm a little bit puzzled by the international coordination on that front as it's something I would have expected in the United States more than in Europe. I can imagine the gridlock in Congress to motive the left-leaning intellectual class to engage in "legislation by other means": leveraging existing legislation to enforce new demands seems like the obvious course of action from their perspective.

    That the same discussions occur in Europe and in Canada is kind of a surprise to me. Over here, Trudeau has had enough power with the socialists (NDP) on his side to basically do whatever the hell he wants on climate and Germany basically screwed its own energy security over to please the environmentalists. Why aren't they just legislating changes without the excuse? Maybe it sounds scarier to repeal financial regulation than climate regulation, so they feel it would be safer to work that way.

    In any event, I'm not sure I buy "climate as a systemic financial risk" so I'm leaning towards "those people don't buy it either."

    ReplyDelete
  3. This English problem arises only because of the promise to pay out fixed nominal pensions (or pensions indexed to inflation.) In other countries the retiree takes on the rate of return of return risk, which should be manageable over the relevant time period, which is several decades.

    ReplyDelete
  4. Thanks for this sane analysis and your last paragraph (Over and over) underscores the current crisis on lack of accountability in government and unforeseen risks despite the resources devoted to bank supervision. Please do the same for the fiscal expansion dreams which have hit the hard wall of inflation... Kwaku

    ReplyDelete
  5. You can also feel the insidious effects of moral hazard coming out of this. No one wants to face Mark to market in a panic so of course there's x post justification for the Fed or the central government to bail out the pensions. And how dare anyone take away pensions from hard-earned university professors

    ReplyDelete
  6. A simple question. What choices are there really to get out of the hole? (I'm limiting the valid answers to those that are socially and politically acceptable which eliminates: *Austerity, *Default without bailout, *Higher taxes (al-la post WW2), *Global debt jubilee. Without some or all of these, there will be more of the same Ponzi-ism with different names and twists. Over and over. It may even be too late for those solutions

    ReplyDelete
  7. Amazing how the markets are always so efficient that supposedly no one can see it coming, yet - mirabile visu - just inefficient enough that a little more regulation is ALWAYS the answer. How uncanny.

    ReplyDelete
  8. While not able to see the exact transactions entered in to, I find it surprising that with regard to interest rate derivatives no collars were placed around these bets which of course would limit upside, but would have limited downside losses.
    More interesting perhaps is the amount of leverage public and private pension funds are allowed to have. They usually have requirements for diversification and max investment in any given asset class. Boards of these pensions are equally culpable with any degree of accountability of regulators it would seem.

    ReplyDelete
  9. " if you have fixed nominal payments to make, invest in risk free assets that provide fixed nominal payments, and ignore the mark to market.".

    I believe your insertion of the word "nominal" was a critical component of an issue that is often ignored. Please correct me if you think I'm off base, but the mismatch between real liabilities and nominal bonds poses another practical problems for UK pensions (tongue twister!). Long-term inflation protected GILTs, the proper match were trading at under 2% real yields at the beginning of 2022. Matching real liabilities was just too expensive for most plans. After hitting a positive 2% last week, they are back down to 0...better but still costly.

    As an aside, LDI may not be as frequently used in the US, but it it used: https://www.nisa.com/primers/primer-on-liability-driven-investing/ if your interested in how those strategies are marketed. Again, they employ derivatives to achieve leverage, but also ignore the mismatch between pension liabilities that rise with inflation and bond investments that don't.

    So far US pensions have dodged the implications of that bullet. Like the UK, the US back-up in long-term nominal rates is less due to a change expected inflation and almost fully due to a dramatic increase in long-term real yields. Since late 2021, real yields on long dated TIPs have surged from a negative 50bps to nearly positive 2%. It seems like a golden opportunity for pensions (and individuals!) to get back onsides.

    ReplyDelete
  10. When I still advised sovereign debt management agencies on their derivatives use, I would emphasize the need to consider cash flow at risk (CFaR) as well as value at risk (VaR). They can both result from "mark-to-market" as discussed in the blog post, but typically the VaR perspective is better understood. CFaR only becomes an issue when you acknowledge that financing is not infinitely available, and is always harder to get the more you need it.

    I find the blog post (and comments) would benefit from some clarity on the distinction between the two risks. On a mark-to-market basis, the combined balance sheets of the UK pension funds may have been hedged or immunized against interest-rate movements in terms of VaR. (One of the goals of LDI.)

    The large market movements exposed CFaR. If pension funds needed to generate cash (or cash-like instruments) to provide as collateral against margin calls from their derivative LDI positions while their overall financial were healthy, they could in principle have entered repo transactions rather than outright sales of Gilt holdings. But increased capital requirements mean that banks have scaled back their activity in that market. Is that way pension funds resorted to outright sales in a falling market? Could the Bank of England have intervened by providing repo funding to pension funds instead of making outright open-market purchases or does that fall outside of the BoE's mandate?

    ReplyDelete
    Replies
    1. Temporary Expanded Collateral Repo Facility (TECRF) announced: https://www.bankofengland.co.uk/markets/market-notices/2022/october/temporary-expanded-collateral-repo-facility-market-notice-10-october-2022

      Delete
  11. Question. What was trading at 300 pounds towards the end of last year, and briefly touched 60 last week? A 40 year UK index linked bond. That's a 80% fall in value over the period of less than a year. Twice as big as the worse stock market falls we've seen historically in developed markets. But its 'only' a liquidity crisis because pension funds have been told their liabilities fall in parallel. John, you've been writing about the fiscal theory of inflation, and inflation expectations. You might want to look closely at the long end of the inflation linked curves in the US and the UK.

    ReplyDelete
    Replies
    1. It’s a good question. Sure, inflation linked guilts were the proper hedge to pension liabilities , but you can’t ignore price. At negative 2% earlier this year they were just too expensive as a hedge.

      Delete
  12. And if you want to look at regulation failures, I recommend you take a look at the interest rate stressed in the Solvency II standard formula. Hey are supposed to be 1 in 200 year stresses but they've been blown out the water. Now Banks have had a free pass on duration mismatches under the Basel rules, since before the financial crisis. But how do you think 30 year mortgages are holding up in value against deposits?

    ReplyDelete
  13. Another question. If you let the monetary genie out the bottle at 300 pounds per bond, and try to put it back but only get 60 pounds per bond, how much of the genie are you going to get back in the bottle?

    ReplyDelete
  14. Great post and thanks for sharing an article on bail bondsmen. Jail is an overwhelming and frightening experience. A bail bond allows you to leave jail until your court date. For this to happen, a state-licensed bail bondsman, backed by an insurance company, guarantees the court full payment if the defendant fails to appear for their court date. If you are searching for immigration bail bonds, then GOODFELLAS BAIL BONDS, “FORGET ABOUT IT” is the best choice for you. A few months ago I have taken the services from them & it was great.

    ReplyDelete

Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.