Swaps data: initial margin soars in Q1 2020
Model procyclicality drives wide variation in CCP IM hikes through Covid-19 volatility
The first quarter of 2020 was unprecedented in almost every way. Global lockdowns intended to control the Covid-19 pandemic tested market liquidity to the limit as financial markets experienced their widest ride in a decade.
A new set of quarterly CPMI-Iosco quantitative disclosures highlight the impact of whipsawing prices on clearing houses during that turbulent first quarter – and they make for an interesting read.
Larger and more frequent price changes result in higher volatility, which in turn feeds into initial margin (IM) requirements. Models produce higher amounts irrespective of whether they are Span, historical simulation, Monte-Carlo simulation or stress scenario based. That is what the data shows, yet differences in relative increases between products and clearing services are surprising.
Categorising clearing services into interest rate swaps, credit default swaps, and futures and options, we see that IM increased in the quarter for these by 23%, 46% and 66% respectively. Some of this difference reflects varying price volatility observed in each of these products. But the biggest contributor comes down to the procyclicality of the models themselves; so how much the IM increases in times of high volatility and decreases in periods of low volatility.
Comparing different clearing services with the same product class also throws up wide divergence. For interest rate swaps the quarterly changes range from 23% to 61% and for credit default swaps from 44% to 65%. For futures and options, we need to be more careful in drawing conclusions as the constituent clearing services are cross-asset, covering rates, equity, commodities and foreign exchange. While larger differences seem reasonable, a range of 20% to 100% for quarterly IM change exceeds expectation.
The magnitude of these relative differences is a topic that warrants more attention and discussion. Should there be more consistency in assumptions of procyclicality of margin? Is that a desirable goal? Is it possible given the diverse markets, models and the design choices for overall risk management? These are huge topics for future debate and discussion, for today, let’s look at the data and detail.
Cleared interest rate swaps
Starting with the largest over-the-counter product, interest rate swaps and quarterly trends.
Figure 1 shows:
- Total IM of $271 billion, up $51 billion or 23% in the first quarter from the prior quarter (quarter-on-quarter) and up $90 billion or 50% from a year earlier (year-on-year).
- LCH SwapClear with $199 billion IM on March 31, 2020, which is up $28 billion or 16.5% quarter-on-quarter and up $54 billion or 37% year-on-year in dollar terms.
- CME IRS with $40 billion, up 38% quarter-on-quarter and 76% year-on-year.
- Eurex IRS with $20 billion, up 61% quarter-on-quarter and 207% year-on-year.
- JSCC IRS with $12 billion, up 45% quarter-on-quarter and 93% year-on-year.
The cumulative total for these four clearing services of $271 billion IM on March 31, 2020 is a new record and the size of the jump is unprecedented in recent history. While some of this increase will be driven by a changed risk position from the start of the quarter, the bulk of the increase is likely to be from higher market volatility in driving an increase in the IM model requirement.
The relative quarterly increases vary dramatically – 16.5% at LCH SwapClear, 38% at CME, 45% at JSCC and 61% at Eurex. Unless the increase in risk positions cleared at LCH SwapClear was relatively smaller than at the other CCPs, which is unlikely, this suggests the LCH SwapClear IM model is much less procyclical than the other CCPs. Intuitively we may expect this to be the case as the IM model disclosures tell us that LCH SwapClear uses a 10-year lookback period, CME and JSCC are five years, while Eurex is just three years. However, both CME and Eurex IM Model disclosures also state the use of additional stress periods, while JSCC does not. With this in mind, it is understandable that JSCC’s margin increase would be much larger than LCH’s, which it is. It might also be reasonable to expect CME, Eurex and LCH to be in a similar ballpark, which they are not.
Cleared credit default swaps
Let’s turn next to credit default swaps, both index and single name.
Figure 2 shows:
- Ice Clear Credit the largest with $50 billion, up $15 billion or 44% quarter-on-quarter and $16 billion or 48% year-on-year.
- Ice Clear Europe with $11 billion, up 49% quarter-on-quarter and 67% year-on-year.
- LCH CDSClear with $5 billion, up 65% quarter-on-quarter and 51% year-on-year, in US dollar terms.
The cumulative total IM for these three clearing services was $67 billion on March 31, 2020, an increase of 46% from the prior quarter and 51% from a year earlier. Similar to interest rate swaps, the bulk of the IM increase is likely to be from higher market volatility driving an increase in the IM model requirement and credit spreads were particularly very volatile in this period.
Futures and options
Exchange traded derivatives, namely futures and options show similar increases in initial margin.
Figure 3 shows:
- CME Base the largest with $190 billion, up $82 billion or 75% quarter-on-quarter and $96 billion or 102% year-on-year.
- Options Clearing Corporation with $91 billion, up 70% quarter-on-quarter and 112% year-on-year.
- Eurex Clearing with $76 billion, up 108% quarter-on-quarter and 92% year-on-year in US dollar terms.
- Ice Clear Europe F&O with $57 billion, up 20% quarter-on-quarter and 22% year-on-year in US dollar terms.
- Ice Clear US F&O with $31 billion, up 64% quarter-on-quarter and 82% year-on-year.
- JSCC ETP with $19 billion, up 118% quarter-on-quarter and 120% year-on-year.
- SGX DC, HKEX HKCC, ASX CLF with 21%, 33% and 13% quarter-on-quarter increases.
The cumulative total IM for these clearing services was $487 billion on March 31, 2020, an increase of $195 billion or 66% from the prior quarter. This is a much larger increase than either interest rate swaps and credit default swaps, which were up 23% and 46% respectively.
The relative quarterly changes between the futures and options CCPs are also interesting, ranging from 118% at JSCC ETP to 22% at Ice Clear Europe F&O.
Both of these are probably explained by the sharp increases in IM for futures and options in March, as highlighted in the BIS bulletin covered in my article, procyclical margins in the time of Covid-19, due to the higher procyclical calibration of these IM models as compared with interest rate swaps.
コンテンツを印刷またはコピーできるのは、有料の購読契約を結んでいるユーザー、または法人購読契約の一員であるユーザーのみです。
これらのオプションやその他の購読特典を利用するには、info@risk.net にお問い合わせいただくか、こちらの購読オプションをご覧ください: http://subscriptions.risk.net/subscribe
現在、このコンテンツを印刷することはできません。詳しくはinfo@risk.netまでお問い合わせください。
現在、このコンテンツをコピーすることはできません。詳しくはinfo@risk.netまでお問い合わせください。
Copyright インフォプロ・デジタル・リミテッド.無断複写・転載を禁じます。
当社の利用規約、https://www.infopro-digital.com/terms-and-conditions/subscriptions/(ポイント2.4)に記載されているように、印刷は1部のみです。
追加の権利を購入したい場合は、info@risk.netまで電子メールでご連絡ください。
Copyright インフォプロ・デジタル・リミテッド.無断複写・転載を禁じます。
このコンテンツは、当社の記事ツールを使用して共有することができます。当社の利用規約、https://www.infopro-digital.com/terms-and-conditions/subscriptions/(第2.4項)に概説されているように、認定ユーザーは、個人的な使用のために資料のコピーを1部のみ作成することができます。また、2.5項の制限にも従わなければなりません。
追加権利の購入をご希望の場合は、info@risk.netまで電子メールでご連絡ください。
詳細はこちら コメント
オペリスク・データ:バイナンスの大規模ハッキングを受け、損失が急増
また、米国の銀行はAMLに関する新たな罰金の波に直面しており、トルコの銀行は労働市場を操作しています。データ提供:ORX News
予測市場は、制度的な信頼性テストに合格できるのか?
専門家によると、規制対象の企業は、利用を可能にするために、信頼性の高い事象判定手法と商品横断的な証拠金管理が必要だとのことです
信用リスクモデリングにおけるAIには、なぜハイブリッドのアプローチが最も適しているのか
AIはモデルが四半期報告書以上の情報を把握するのを支援していますが、CROによると、ガバナンスが完全な自動化を妨げているとのことです
ポッドキャスト:アリエ・シェーター氏とハリオム・タツァット氏によるAIの説明可能性に関する対談
バークレイズのクオンツたちが、メカニスティックな解釈可能性を用いたAIの監視手法について語ります
スイスフラン建てローンの損失は、EUの消費者金融業界にとっての灰色のサイである
mBankのCROであるマレク・ルシュティン氏は、ポーランドの銀行への影響が、法的リスクのモデリングにおける盲点を浮き彫りにし、それがより広範な影響を及ぼしていると述べています。
オペリスクデータ:EFG、20年にわたるマネーロンダリングで巨額の損失を被る
また、ディープフェイクによる音声クローンがインテサの部門を騙し、仮想通貨関連のハッキング事件が相次いでいます。データ提供:ORX News
ナビエ・ストークス方程式からキロ単位の論文まで:AIがクオンツ研究に与える影響
AIが大量の成果を生み出す中、金融業界では、その出所、信頼性、重要性を評価するための信頼できる手段が必要とされています
AIを正しい軌道に乗せ続けるには、継続的な検証が鍵となる
意思決定ごとの検証を行うことで、ユーザーがAIの判断を信頼できるようになります