Thursday, October 6, 2022

UK finance fable update

Update to my previous post on the UK treasury imbroglio: 

The Bank of England explains, saying about as much as I did. Pension funds levered up, lost money when treasury prices fell, needed to post collateral, and then started selling en masse. The explanation starts with lovely central-bankerese (my italics): 
Against the backdrop of an unprecedented [really? Literally never?] repricing [translation: fall in prices] in UK assets, the Bank announced a temporary and targeted intervention on Wednesday 28 September to restore market functioning in long-dated government bonds and reduce risks from contagion to credit conditions for UK households and businesses. 

It goes on to a hilarious graph to explain how you lose money when you borrow to lever up a portfolio:

Why is this so funny? Notice on the left hand side a gap between assets and liabilities, yet in the fourth bar there is a positive "capital" bar.  Accounting 101, assets = liabilities, including capital. I guess UK regulations operate differently. 

But enough fun, let's get to the point. In answer to my question, roughly "you're supposed to be this great gargantuan regulator, how could you miss something so simple?," the bank offers, deep in the report, this: 
The FPC has previously identified underlying vulnerabilities in the system of market- based finance, a number of which were exposed in the ‘dash for cash’ episode in March 20202. The Bank and the FPC strongly support and engage with the important programme of domestic and international work to understand and, where necessary, address those vulnerabilities.

The FPC conducted an assessment of the risks from leverage in the non-bank financial system in 2018, and highlighted the need to monitor risks associated with the use of leverage by LDI funds. 

It's so nice you've been studying this. But then double, how did this happen? 
Whilst the PRA regulates bank counterparties of LDI funds, the Bank does not directly regulate pension schemes, LDI managers or LDI funds. Pension schemes and LDI managers are regulated by The Pensions Regulator (TPR) and the Financial Conduct Authority (FCA). LDI funds themselves are typically based outside the UK. In this context, given our financial stability mandate, and as stated in the FPC’s November 2018 Financial Stability Report, the Bank has worked with other domestic regulators – including TPR and the FCA – on enhancing monitoring of the risks. That included working with TPR on a survey of DB pension schemes in 2019, and prompting work to improve DB pension liquidity risk management. Given that LDI funds are largely not based in the UK, this also underlines the importance of work on this topic being pursued internationally.
In short,  "where is your homework?" "Jimmy was supposed to write up the lab report." Really, though, if you were working on it, and knew about it, it's doubly shocking that nobody did anything about it. 
... it should also be recognised that the scale and speed of repricing leading up to Wednesday 28 September far exceeded historical moves, and therefore exceeded price moves that are likely to have been part of risk management practices or regulatory stress tests.   
The 30 year nominal gilt yield rose by 160 basis points in just a few days, having only had a yield of around 1.2% at the start of the year. On Wednesday 28 September the intraday range of the yield on 30 year gilts of 127 basis points was higher than the annual range for 30 year gilts in all but 4 of the last 27 years. In the 2018 assessment noted above, the FPC assessed the capacity of the biggest derivatives users among UK pension schemes to cover the posting of variation margin calls on OTC interest rate derivatives from up to a 100 basis point instantaneous increases in rates across all maturities and in all currencies. Other tests and risk management practices have similarly assumed a maximum of a 100 basis point move in such a short time period. 
In short, once again "it's a 100 year flood that nobody could have imagined." The sort of flood that seems to happen about once a year. And 100 basis points is an awfully round number.  
There has been significant progress, both domestically and internationally, on the regulation and monitoring of the non-bank sector in recent years. Much of this has been led by the Financial Stability Board, which set out its analysis of risks relating to non- banks and a program of work last year and is due to report on next steps in November.  Through the work of the FPC and the Bank more widely, as well as that of the FCA, the UK has been actively engaging with this programme. This episode underlines the necessity of this work leading to effective policy outcomes.
The eternal call for more research. 

A last thought. If some people are clearly selling in distress, why are others not buying? Why is your fire sale not my buying opportunity. Well, imagine if every Saturday, when garage sales start at 9 am, the Fed or Bank of England swoops in and buys up all the good stuff at inflated prices. There is then no point in cruising the garage sales any more, and then every time someone wants to sell anything the Fed or Bank of England has to swoop in and buy. 

Why do Chase, Citi, Goldman Sachs, or multi-strategy  hedge funds not keep some spare balance sheet capacity around to swoop in and buy in these events? Well, knowledge that the Fed and Bank of England will always swoop in first and cut off the profits might well be a part of the answer.  

I repeat. The model that financial institutions can lever up, but omniscient regulators will foresee risks, is irretrievably broken. If you need more research on this one, we'll never get there. We are firmly in the regime of lever to your heart's content, the Fed and BoE, and thus taxpayers, will offer a free put option on losses. Until, someday soon, they too run out of bailout money. 
 
I should change the tone of this post -- let us not blame the BoE and Fed. They are smart, they are human, they are doing their best. It's just that the best that these great people can do, with 15 years since 2008 to do it, is not enough for this most simple task. And woefully inadequate for the much larger task they have set themselves. 


18 comments:

  1. Oooh boy! finally a chance to dust off my probability theory for the first time in a decade. Let's see, four of the last 27 years would be 15% chance in a particular year (assuming independence). They've been working on this since 2018, so that's 4 years, plug that into the old binomial distribution and you get basically a 50/50 chance that an event like this would happen by now.

    ReplyDelete
    Replies
    1. As to : "In short, "where is your homework?" "Jimmy was supposed to write up the lab report." Really, though, if you were working on it, and knew about it, it's doubly shocking that nobody did anything about it."
      I betcha' rank and file analysts in the system saw this. But as to being doubly shocking...does anybody not believe that the people in charge are muffled either by their fear or/of their "superiors"? Who wants to admit bad news until it is too late?

      Delete
    2. Has your probability theory been peer reviewed by experts?😜

      Delete
  2. Yup, and the same was true in 2008. The regulators were in the room watching everything happen in real time. They had full access to the books and records and the personnel, including top management. And they just watched it happen.

    ReplyDelete
  3. On point. See:
    https://www.imf.org/en/Blogs/Articles/2022/10/04/how-illiquid-open-end-funds-can-amplify-shocks-and-destabilize-asset-prices

    ReplyDelete
  4. The LDI Fund strategy was predicated on continuation of the status quo low interest rates across the maturity spectrum. The bond market yields changed with the FOMC's aggressive FED FUNDS rate hikes. If you were watching or participating in the bond market, as I have been on a very modest scale, the rise in long term yields has far outpaced the rise in yields at the short end.

    It's very likely that the fund managers were in the process of unwinding their positions when the new cabinet shocked the financial markets with proposed tax cuts and consumer energy subsidy programmes. The news caused the jump in yields in long term government bonds and unhinged the slow measured winding down in LDI positions. It became rout of significant proportions.

    Call it a rookie error on the part of the Liz Truss cabinet that nearly tipped the U.K. market into a full-blown crisis. Not on the scale of Lehman Bros. in 2008, but serious enough.

    As to the regulators, the description shows the disadvantages of decentralized regulation charaterised by management silos. There's co-operation but not urgency. Each silo has part of the picture, but no one silo has the complete picture.

    In the "Before" section of the chart, the grey bar under the label "LDI FUND" shows "Capital". This for the LDI FUND itself, not the pension fund as a whole.

    The unfunded liabilities of the fund, being the difference between assets and the present value of future pension obligations, is not shown on a IFS basis in the chart, i.e., Assets = Liabilities, but on a projection or forecast basis where the unfunded liabilities would be disclosed in a footnote to the financial statements. One can argue the merits of that style of presentation but to accuse the pension fund management of not comprehending basic principles of accounting is far-fetched. The British may have lost their empire at the behest of the U.S., but they haven't lost their mastery of accounting principles.
    But the remarks were good for a grin.

    ReplyDelete
  5. I don't understand the inability of gov officials to assess risk. Every trading platform follows FINRA rules for maintaining enough equity in customer accounts so as to avoid firm insolvency.
    Trader deposits $12,000 in a margin account
    Margin loan $12,000
    Minimum required equity in the account: margin account value = margin loan/ 1-.25.
    e.g. 12000/(1-.25) = 16,000.
    If the customer falls below $16,000, they are liquidated at the market or told to increase deposits to meet minimum maintenance requirements.
    I don't think we are splitting the atom here.

    ReplyDelete
    Replies
    1. FINRA = U.S. regulator. There are U.K. equivalents in one or two regulatory bodies. Margin in U.S. Treasury bills and notes is 10%. Requirements vary by country. It was the margin maintenance requirement that prompted the sell-off in long-term U.K. 'gilts' that pushed the yield to maturity rate up so suddenly. The rapid yield rise caused knock-on effects that threatened an orderly market adjustment.

      Not rocket science, perhaps, but the effect can be more devastating to financial markets. Cf., Lehman Bros. 2008.

      Delete
    2. Eagle Eye, thanks for replying. Equity required by FINRA exceeds the margin loan by $4000. (33% over liability) Many clearing firms require 30% that calculates to $5143 excess equity. (43%). In volatile markets , the requirement of increases to 35%. Using the formula, 10% requirement only leaves $1333 over margin,(11%).
      An hypothetical. 12 Ten year zero coup gilts come at 3%. PV price =
      12000/(1.03)^10 = 8929. Assume Pension pledges 8929 in collateral and borrows 8929.
      Maintenance margin = 8929/.9 = 9921.
      Interest rate increases to 4%. Bond price = 8109. Pension's equity is below maintenance by 1812. Leverage kills when struck by black swan events. It isn't rocket surgery, but a little first year algebra is useful.


      Delete
    3. LDI is not so simple. "Liability driven investment", "LDI", for pension funds pertains to (1) reducing unfunded liabilities over the course of time so that "Assets = Liabilities" at the target time horizon of the defined benefit pension plan; and, (2) reducing the balance sheet impact of variation in assets and in liabilities in the pension fund sponser's financial statements.

      Cambridge Associates describes the LDI objectives, modes, and risks in their paper found by navigating to: https://www.cambridgeassociates.com/insight/pension-risk-management/

      Fitch Ratings Limited (London) describes the expected transition arising from the volatility of the past week and a half in their paper which is found by navigating to:
      https://www.fitchratings.com/research/insurance/ldi-fallout-will-spur-pension-funds-to-seek-life-sector-arrangements-05-10-2022

      There are four principal strategies on which LDIs are based, according to Cambridge Assoc.:
      (a) Duration matching; (b) Segmented duration matching; (c) Matching cash flows; and (d) Overlay.

      The action in the London market since 9/28/22 is concerned with the fourth LDI strategy--"Overlay".

      Cambridge Assoc. describes this strategy.
      "An overlay strategy uses derivatives (and sometimes leverage) to supplement a plan’s physical fixed income portfolio and to provide a more capital-efficient way to increase asset duration and inflation-hedging benefits. This is the primary manner in which a plan can increase its allocation to the hedging portfolio without decreasing the allocation to the growth portfolio, which implicitly reduces the expected return on plan assets. Overlay strategies also present an opportunity to use port-able alpha/beta platforms and synthetic equity."

      This description should be sufficient to inform us that this is not a plain vanilla bond buying on margin risk portfolio investment strategy.

      Delete
    4. Eagle, thank you for the detailed and elegant explanation. But If I'm a managing partner and senior risk manager at a hedge fund, and I was for many successful years, I am going to ask one simple question. What does a 100 bps or 200 bps increase in long dated bonds do to my bond portfolio. Pretty plain vanilla and the results were so disastrous, the Bank of England had to buy 65 billion of government bonds. The first question I asked my traders: What is the most we can lose on this position. Leverage in heavy tails is worrisome

      Delete
  6. And what would this crisis look like? - "on the part of the Liz Truss cabinet that nearly tipped the U.K. market into a full-blown crisis."

    ReplyDelete
  7. John is correct, its amazing we have to watch re-runs of the same issue time and again. The solution is simple, tax debt issuers when they borrow money and put it into a bailout fund. This should extend to derivatives. Governments should also have constitutional limits placed on their ability to run structural deficits as well, but that's wishful thinking. Instead we get endless regulations and bloated regulatory agencies.

    ReplyDelete
  8. "Why is this so funny? Notice on the left hand side a gap between assets and liabilities, yet in the fourth bar there is a positive "capital" bar. Accounting 101, assets = liabilities, including capital. I guess UK regulations operate differently. " - The difference between asset and liabilities gets consolidated to the sponsor's balance sheet. The focus of the BoE here is on the pension fund assets and liabilities in isolation.

    ReplyDelete
  9. I am struggling to make sense of the price moves at the long end of the gilts market. Prices almost doubled following the initial BoE move, but since then have pretty much dropped right back over the last 4 or 5 days. Amounts actually sold were way lower than the facility announced. One interpretation is the big jump up in prices cancelled a lot of the collateral calls, but it's starting to hit the fan again now. The whole LDI market is supposed to be all about not trading duration, but the temptation to do exactly that in response to the liquidity calls, or drop out of the hedges all together in the face of upcoming QT must be huge. Real test for clearing and the IBs. Not just the pension funds.

    ReplyDelete
    Replies
    1. The market in London is adjusting for a higher interest rate environment. The LDI segment is raising collateral at the request of asset managers hired by the pension funds to manage the funds' LDI investing strategies. From the WSJ (Oct. 12/22) "U.K. Pension Funds Rush to Raise Cash as Central-Bank Deadline Looms", Julie Steinberg reporting.

      ...“asset managers are requiring a bigger cushion than they did historically up to two weeks ago,” said Ben Gold, head of investment at XPS Pensions Group PLC a U.K. pensions consultant.

      "Mr. Gold said the asset managers he deals with, including Legal & General Investment Management, are now asking clients to bump up their collateral buffer so they can withstand as much as a three-percentage-point rise in gilt yields. Thirty-year gilt yields have surged this year from 1.1% to about 4.9%.

      "Likewise, Kerrin Rosenberg, chief executive officer of Cardano Investment, which both advises on and manages LDI portfolios, said most funds wanted by Friday to be able to withstand a rapid, 2- to 3-percentage-point rise in bond yields without having to sell other assets.

      “We are planning to be at the conservative end of the range,” Mr. Rosenberg said. Before the market meltdown, the amount of collateral people were more likely to have on hand would serve as a buffer against a rise of one or two points, he added.

      What are they selling? "Stocks, credit and real-estate fund investments have all been sold to replenish collateral. The offloading has rippled through areas such as the market for collateralized loan obligations." In other words, the funds are selling equities, not 'gilts'.

      We see two effects of the rise in the administered interest rates working on the financial markets. Firstly, the level shift of the yield curve across the maturity spectrum (excepting the very short end where one-month bills and three-month bills are strongly influenced by the central bank overnight interest rates). Secondly, the readjustment of investment strategies as margin calls rise on falling collateral values.

      What effect did the fiscal side moves have? The 'shock' of the new U.K. cabinet's move to lower tax rates and provide consumer subsidies to insulate households and firms from the extreme price moves in the cost of energy in the U.K. was an unexpected development on the fiscal side and a change in direction for the Conservative government that had hitherto been moving in the direction of higher taxes and greater restrictions on fossil fuel sources of energy. Evidently, that was enough to trigger a sell-off in 'gilts' last month.

      The notion that the fiscal authority and the monetary authority must always move in lock-step found its paragon in former Bank of England Governor Mark Carney this week. He criticized the Truss cabinet for making the job of the Bank of England to control inflation much harder because cabinet's fiscal policy stance is "expansionary" whereas the BoE is endeavouring to implement a "contractionary" monetary policy to rein in "inflationary expectations". Raises the question, "Who is governed, and who is the governor?" According to Carney, it is the government that must dance to the Governor's tune and not the other way around. Some democracy, eh?

      Delete
  10. I don't think this story is quite right. The problem is that pension funds (and others) have a long duration of liabilities, so they mostly have to invest in long bonds and stocks. But with zero interest rates (and negative real rates), you can't reach your return targets using bonds. So, you either need to move more toward stocks (risky! so we'll do PE instead), contribute more to offset the low returns (ACK - don't even think about that), or get creative in other ways. LDIs were the result of that creativity - get the maximum duration benefit for your dollar invested at (nearly) zero. But like many financial strategies (like hedging with futures), LDIs work better in theory than in practice.
    With yields so low, I can't see how it could work to lever up for return - although I don't know UK pension funds' borrowing costs versus gilt yields but the spread couldn't have been that wide, no? A quick check online suggests that before the GFC ultra-long gilts yielded 4-4.5%. By 2016, it was 1.5-2%. End of 2019, 1% where they mostly stayed (or lower) until this Spring (a low of 40bps in May 2020). Pension Funds probably weren't borrowing lower than that. I think they were after duration more than return. And arguably they did match their overall asset and liability durations. In theory...

    ReplyDelete
  11. Since 2008, I have wondered if the long-term solution to the liquidity-contagion-death spiral-bailout problem lies in a change to the legal "operating system." Limits or pauses on obligations to make debt payments (and margin calls) during periods when the macro system shows signs of general crisis would a) prevent this sort of turmoil and panic selling and b) incentivize lenders to protect themselves in the first place, knowing that stressed borrowers could not be forced into panic sales to make payments and the government could not be panicked into bailouts due to liquidity problems. Ideally, there would be some objective, non-manipulable criteria for when such a temporary jubilee would be imposed, but even a common-law-like standard for "market force majeure" might work.

    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.