4.6.6 Advanced Extensions

Basic Prerequisites: 04 · Margin & Funding Spirals and 05 · Failure Modes.

---

1. Intuition & Practical Objective

The two immediate extensions of the folder are (a) the endogenous liquidity spiral - where the impact coefficient itself becomes a function of aggregate deleveraging, turning the geometric amplification into a genuine bifurcation - and (b) the pricing of liquidity as a systematic risk factor, which turns liquidity from a cost footnote into a determinant of expected returns.

Why these first? The geometric amplification $1/(1-k)$ of 04 is the partial-equilibrium spiral: it takes $\kappa$ as given. Real systemic events are general-equilibrium: deleveraging by many agents raises $\kappa$ (liquidity begets illiquidity), so the amplification factor itself rises with the size of the unwind - the origin of multiple equilibria, fragility, and phase transitions (bifurcations) in liquidity. The second extension, the liquidity-adjusted CAPM of Acharya & Pedersen (2005), explains why investors demand compensation for holding assets whose liquidity fails in bad times - the pricing counterpart to the risk management.

---

2. Mathematical Ground Truth & Derivations

2.1 The endogenous impact coefficient and the bifurcation. In the partial-equilibrium spiral, $\kappa$ is a constant. Systemically, market depth falls as the crowd unwinds - an inverse relation often summarised as $D(D_{\text{agg}})$ with $\partial D/\partial (\text{aggregate selling})<0$. Writing the market-impact coefficient from a square-root impact law (Bouchaud et al. 2009),

$$\Delta p=\sigma Y\sqrt{\frac{Q}{V}},Y\sim \mathcal{O}(1),$$

the coefficient $Y/(V)$ is not a constant: as the stressed volume $V$ shrinks and volatility $\sigma$ rises, impact per unit sold rises. Substituting $k(L,\text{stress})=L\kappa (\text{stress})$ into $S_{\text{total}}=S_0/(1-k)$, the system crosses from a stable fixed point ($k<1$, finite liquidation) to divergence ($k\ge 1$) as stress rises - a saddle-node/tipping structure. This is the formal content of “liquidity spirals”: the economy has complementarities (your selling makes my selling more damaging), hence multiple equilibria and the possibility of a sudden jump between them.

2.2 The amplification multiplier - exact. For the geometric feedback where each round’s impact causes a further sale $L\kappa \times$ the previous round,

$$S_{\text{total}}=\sum _{j\ge 0}(L\kappa {)}^j{S}_0=\frac{S_0}{1-L\kappa },|L\kappa |<1,$$

with amplification $1/(1-k)$ and a pole at $k=1$. The multiplier is the clean analogue, for liquidity, of the money multiplier and the Kahn–Keynes investment multiplier: a small exogenous shock is leveraged by the feedback between leverage and price impact.

2.3 Liquidity-adjusted pricing (Acharya–Pedersen 2005; Foucault Ch 9). The gross-return premium $R\simeq r+s/h$ (Foucault eq. 9.6) prices the level of illiquidity. But the priced risk is the covariance of illiquidity with market returns. Acharya–Pedersen’s liquidity-adjusted CAPM (Foucault eq. 9.18) has four betas:

$$\mathbb{E}(R_i)-r={\beta }_1{\lambda }_{\text{mkt}}+{\beta }_2{\lambda }_{\text{illiq commonality}}+{\beta }_3{\lambda }_{\text{return–illiq hedge}}+{\beta }_4{\lambda }_{\text{illiq–return}},$$

where ${\beta }_4$ prices the asset being liquid when the market is falling - the deepest failure. Empirically (Foucault Ch 9) the unexplained spread is ~$1.1\%$/yr and is dominated by ${\beta }_4$: what investors pay to avoid is not illiquidity per se but illiquidity that arrives exactly with market stress.

2.4 The search/OTC channel (Duffie–Gârleanu–Pedersen 2005). For non-exchange assets, market liquidity and funding liquidity are jointly determined by a search-and-bargaining friction (Foucault Ch 9): the spread depends on dealer bargaining power $\varphi$, meeting intensity $\psi$, and investor types, e.g. $S=(1+z)c/[2(r+2\psi )+(1-2\psi )\varphi (1-z)]$. This is the rigorous bridge to OTC markets (corporate bonds - Bao et al. 2011; CDS) where “market liquidity” is literally a search cost.

2.5 Systemic-risk metrics built on liquidity. Regulators and academics convert these ideas into measures: the Amihud illiquidity ratio aggregated to market level; liquidity co-movement (Chordia–Roll–Subrahmanyam 2000; Hasbrouck–Seppi 2001); fire-sale / deleveraging spillover measures; and the BCBS LCR/NSFR plus the margin-procyclicality workstream (setting margin on through-the-cycle rather than point-in-time risk to damp the spiral). The unifying principle: measure the covariance, not the level.

---

3. Computational Implementation - the amplification multiplier and its pole

We compute the total forced liquidation for a range of leverage $L$ and impact feedback $\kappa$, verify it against the closed form $1/(1-L\kappa )$, and locate the divergence. Stdlib only.

The simulation reproduces the closed form $S_0/(1-k)$ to the printed precision for every converging case. The message is in the last row and in the shape, not any single number: at $k=0.1$ the amplifier is a mild $\times 1.11$; at $k=0.5$ it is $\times 2$; and at $k=1$ the iteration runs without bound - the total deleveraging required is finite only if the feedback is bounded away from 1. Since $\kappa$ rises with stress in the general-equilibrium model (§2.1), a system calibrated at $k=0.5$ in normal times can cross $k=1$ in a crisis with no change to individual leverage: the fragility is in the feedback, not in anyone’s balance sheet.

---

4. Failure Modes & First-Principles Breakdowns

1. Partial equilibrium mistaken for the whole. The closed form $1/(1-k)$ assumes $\kappa$ fixed. Taking $\kappa$ constant across a stress scenario deletes the very nonlinearity (systemic amplification) the scenario exists to measure. Stress $\kappa$ and $L$ jointly.
2. The pole is a warning, not a forecast. Beyond $k=1$ the model produces divergence - which in reality becomes a policy intervention (circuit breakers, central-bank liquidity, margin holidays) or a market closure. A model that reports “infinite loss” is reporting “your assumptions have broken”; the right output is the distance to $k=1$, not the loss at it.
3. Pricing model risk in the L-CAPM. The four Acharya–Pedersen betas require estimating time-varying, illiquid risk factors - the hardest estimation in asset pricing. ${\beta }_4$ (illiquidity–return) dominates the premium but is the least stably estimated; report sensitivity (Foucault Ch 9; Pastor–Stambaugh 2003).
4. Square-root vs linear impact model risk. The functional form of $\kappa (Q)$ changes the multiplier’s level (not its pole). Do not report a single amplification number without the impact-law caveat; the qualitative tipping survives, the quantitative claim does not (see Stress Testing & Scenario Analysis).
5. Systemic metrics are lagging. Liquidity co-movement and Amihud aggregates are computed from realised trades; they spike after the spiral has begun and give little warning. Early-warning work (margin-use, repo-haircut dispersion, dealer inventory) is more forward-looking but coarser.
6. Regulatory procyclicality remains. LCR/NSFR reduce the probability of a funding spiral but can amplify one if all banks meet them simultaneously (the buffer is a common-mode liability). Margin-procyclicality rules (through-the-cycle margin) target the same failure but trade protection for cost in normal times - a genuine, unresolved trade-off.
7. Liquidity spirals cross the risk-type boundary. A funding spiral is a liquidity event that becomes a credit event (counterparty default) that becomes a market event (fire-sale prices) - Pillar 4’s risk types are not separable in a crisis, which is why integrated stress testing (Stress Testing) is the only honest frame.

---

5. References

• Brunnermeier & Pedersen (2009)
• Acharya, V. & Pedersen, L.H. - Asset Pricing with Liquidity Risk, JFE 77(2):375–410 (2
• Pastor, L. & Stambaugh, R. - Liquidity Risk and Expected Stock Returns, JPE 111(3) (2003). Liquidity as a priced systematic factor. (Pillar 6 refs.)
• Bouchaud, J.-P., Farmer, J.D. & Lillo, F. - How Markets Slowly Digest Changes in Supply and Demand (2009)
• Duffie, Gârleanu & Pedersen - Over-the-Counter Markets, Econometrica 73(6) (2005)
• Hasbrouck & Seppi (2001) / Chordia, Roll & Subrahmanyam (2000)
• Bao, J., Pan, J. & Wang, J. - The Illiquidity of Corporate Bonds, JF 66(3) (2011)
• BCBS

---

6. Connected Graph Bridges

• Back: 05 · Failure Modes · Index Hub
• In-folder: 02 · Market vs Funding Liquidity · 04 · Margin & Funding Spirals
• Systemic / stress: Stress Testing & Scenario Analysis · Credit Risk & the Merton Model · Extreme Value Theory & Fat Tails
• Portfolio / pricing: Transaction Costs & Turnover Constraints · Optimal Execution & Almgren–Chriss
• Base: VaR & Expected Shortfall · Probability & Measure Theory