Understanding long and short margin mechanics
Dec 7, 2025
Most investors do not fully understand margin mechanics — in particular, what happens once long and short positions share an account. This post teaches the basics, corrects a few common misconceptions, reconciles the theory against two real custodial statements — a market-neutral ~45/45 book and a ~130/30 extension — and closes with what changes under prime brokerage. Two of the identities hold to the penny. It is not advice about what to trade or how much risk to take.
General disclaimers apply. I am not your accountant, broker, or financial advisor, and this is not investment advice. The information here is for general educational purposes only, may not apply to your specific situation, and margin rules vary by broker and can change without notice. Always review your custodian’s margin agreement and consult qualified professionals before trading on margin.
What you should know
Reg T
Reg T Accounts
LONG SHORT
LMV | DR CR | SMV
|----------- |-----------
| EQ | EQ
-------------------- --------------------
Reg T Margin Reg T Margin
-------------------- --------------------
Finra Margin Finra Margin
-------------------- --------------------
Ex EQ Ex EQ
-------------------- --------------------
SMA SMA
The four legs. Read the ledger before the rules. LMV is the market value of what you own. DR is cash you owe — the margin loan. CR is cash you hold or are owed — deposits, sale proceeds. And SMV, the leg that trips people, is the market value of what you owe in shares: when you short, you borrow stock and sell it, so you hold the sale proceeds (a credit) and owe the shares back. SMV is not money anywhere — it is the live price of your debt, the cost of buying the shares back right now, and it marks against you: the stock rallies, SMV rises, your equity falls. (Custodial statements print it with a minus sign; the T-accounts in this post carry it on the right side — same debt, two sign conventions.) The ledger thus holds two assets and two debts, one of each in cash and in shares — LMV and CR against SMV and DR — and equity is what is left: EQ = LMV + CR - SMV - DR.
Anatomy of a short: sell 1,000 sh at 100
at sale stock -> 120 stock -> 80
CR (proceeds) 100,000 100,000 100,000 -- fixed at sale
SMV (owed) 100,000 120,000 80,000 -- marked: cost to buy back
EQ contribution 0 -20,000 +20,000 -- proceeds - SMV = short P&L
SMV is a debt denominated in shares; the market reprices it daily
Two legs move, two are set. A corollary worth carrying through the rest of the post: of the four legs, only the two valuations breathe day to day. LMV and SMV are remarked every night, so a pure market move lands entirely there. CR and DR are financing facts — fixed when you trade, deposit, or borrow, and untouched by price. The short above is the cleanest case: when the stock rallied, SMV rose because it is a valuation, while the proceeds you banked held still because they are financing — and the gap that opened between them was the loss.
What moves, and why
set by moves with
LMV buying / selling longs the market (remarked nightly)
SMV selling / covering shorts the market (remarked nightly)
DR borrowing cash only your transactions (+ interest)
CR proceeds, deposits only your transactions (+ the sweep)
a pure market day shows up in LMV - SMV; the financing legs hold still
Two small exceptions, both visible later on the statements. The first is easy: the loan creeps even when you do nothing, because DR accrues interest — the 130/30’s Pending Margin Interest is precisely that.
The second exception is the short-credit line, and it needs care because it seems to break the rule. On a real statement the short credit does move with price: your 130/30 shows it at 221,038 against an SMV of 216,757, and if the stock you are short rallies tonight, both numbers climb tomorrow. Here is why that is not financing changing its mind. Two different things hide under “short credit.” The proceeds — the dollars you actually collected when you sold short — are fixed financing and never move. The short-credit balance the statement prints is something else: it doubles as the cash the broker must keep posted with the stock lender, marked daily to about 102% of SMV. So when SMV rises by 5, the broker owes the lender 5 more, and it raises that 5 by sweeping it out of your free credit:
A short rallies overnight: SMV up 5
before after
free credit (type 2) X X - 5 -- 5 swept out to fund the call
short credit (type 3) S S + 5 -- topped up to ~102% of SMV
total CR X + S X + S -- unchanged: cash only changed pockets
SMV V V + 5 -- the real loss lives here
EQ = LMV + CR - SMV - DR falls by 5
Total CR is identical before and after — the two cash legs are equal and opposite — so the sweep adds nothing to equity. The only line that genuinely moved equity is SMV, and it fell by exactly the 5 of short loss, just as the clean model says. The lesson: the short-credit number jiggles because of where the cash sits, not how much there is, and that movement is invisible to equity. The slice by which the short-credit line sits above SMV — your 221,038 over 216,757 — is the excess short credit defined below.
Initial margin. Start with the strangest fact in the subject: how much you may borrow against a stock is set not by your broker but by the Federal Reserve. Congress decided, in the wreckage of 1929, that stock-market leverage is monetary policy — Section 7 of the Securities Exchange Act of 1934 handed the dial to the Fed, which adopted Reg T, 12 C.F.R. §220 effective October 1, 1934. The Board moved the rate around for four decades — between 40% and 100% — before parking it at 50% in 1974, where it has sat ever since.The Fed publishes the full history of initial requirements under Regulations T, U, and X. Under §220.12(a), the 50% applies directly to LMV (long market value): you put up half the purchase price. Shorts are stated differently: §220.12(c)(1) requires a credit of 150% of SMV (short market value), of which 100% is supplied by the sale proceeds themselves — so you deposit an additional 50%.
Maintenance margin. Initial margin is tested once, at the door; maintenance is the test that never stops asking. FINRA Rule 4210(c) sets the floor at 25% of LMV for longs. For shorts, the familiar “30%” is only part of the schedule: stocks at $5 or above require the greater of 30% of market value or $5.00 per share; below $5, the greater of 100% or $2.50 per share. The per-share minimums mean a real short book’s exchange requirement usually runs above a naive 30% — visibly so in the statements at the end of this post. Brokers then layer “house” requirements on top: higher, position-specific, and changeable without notice.Schwab’s schedule: base house maintenance of 30% for long equities at $3 or above, 100% below. Fidelity computes house requirements position-by-position via “rules-based requirements” keyed to volatility, liquidity, and concentration.
Three tests, run in parallel. Every margin account is continuously scored against three independent requirements: the Fed’s (Reg T initial — applied at trade time, and governing withdrawals through SMA), FINRA’s (the maintenance, or “exchange,” test), and the broker’s (house — the strictest). A custodial balance page is just these three tests printed as surpluses or calls: Fed surplus / SMA, exchange surplus, house surplus. Hold that trichotomy; the statements below are unreadable without it.
Restriction and calls. Ex EQ (excess equity above the Reg T requirement — the Fed surplus) is the buffer. If it goes negative, the account has no Reg-T buying power: new positions require fresh initial margin in full, and a purchase made anyway generates a Fed call, due within the Reg T payment period (standard settlement plus two business days). Separately, if equity falls below the maintenance requirement you receive a maintenance call — customarily a few business days to cure, though brokers may liquidate without notice and you have no right to an extension. The liquidation arithmetic is unforgiving: selling longs leaves equity unchanged (proceeds retire the debit, not the deficiency) while reducing the requirement by only the maintenance rate times the amount sold — selling shrinks the account around an unchanged hole. Curing a deficiency D by liquidation alone therefore takes a sale of D / rate — 4x the call at a 25% requirement, ~3.3x at 30%, and proportionally more at house rates.
Why a maintenance call liquidates at 4x (25% requirement)
before LMV 100 | DR 80 EQ 20 req 25.00 deficiency 5
sell X LMV 100-X | DR 80-X EQ 20 req 25-.25X (EQ unchanged)
cure 5 - .25X <= 0 => X >= 20 = 4 x the call
Buying power. SMA (Special Memorandum Account) confuses everyone, so start with what it is for: the Fed scores you at trade time, and the Board decided long ago that a surplus you once earned should not be confiscated by a later market decline — so the surplus is remembered. SMA is that memory: a Reg-T bookkeeping line, not cash — a high-water record of the account’s Fed surplus, which by construction never falls with market depreciation, only with use. It is credited by cash deposits, dividends, and 50% of the proceeds of long sales; it is debited by 50% of new purchases and 100% of cash withdrawals. You may withdraw SMA dollar-for-dollar (1:1), or commit it at 2:1 as buying power for fully marginable securities — subject, always, to the maintenance tests still passing afterward. SMA does not exist under portfolio margin.
SMA: a high-water mark, not a balance
time t0 t1 t2 t3
Fed surplus 10 25 8 -2
SMA 10 25 25 15
t1 gains ratchet SMA up; t2's decline cannot pull it down;
t3 withdraws 10 against SMA -- legal with a live surplus of only 8,
and the account emerges restricted, exactly as described above
Mixed-netting
Now put longs and shorts in the same account and ask the obvious question: does the surplus on one side cover the deficiency on the other? The regulation answers in two sentences:
12 CFR § 220.2 Equity means the total current market value of security positions held in the margin account plus any credit balance less the debit balance in the margin account.
12 CFR § 220.4(a)(1) All transactions not specifically authorized for inclusion in another account shall be recorded in the margin account.
T Accounts
LMV | SMV
CR | DR
|----------
| EQ
--------------------
Reg T Margin
--------------------
Finra Margin
--------------------
Ex EQ
--------------------
SMA
Aggregation is not a courtesy; it is the rule. §220.4(a)(1) establishes the single-account principle — all transactions not assigned elsewhere live in one indivisible margin account — and §220.2 defines equity over that whole account: longs plus credits, less shorts and debits. The surplus on one side therefore offsets the deficiency on the other mechanically, not by petition. The side-by-side LONG/SHORT panels in this post are presentation; only the aggregate column is law. This is also exactly what a custodial statement displays: one equity, three requirements.
Portfolio margin
PM Accounts
LMV | SMV
CR | DR
|----------
| EQ
--------------------
Port Margin
--------------------
Ex Eq
Portfolio margin replaces the position-by-position schedule with a stress test. The motivation is the one you would guess: 45 long against 45 short is far safer than either leg alone, yet the strategy-based schedule bills you for both. PM is the regulator agreeing to score the book the way a risk manager would — what is the worst this portfolio loses? — in exchange for higher minimums and a requirement that can change its mind intraday. The SEC approved customer portfolio margining on December 12, 2006; it went live April 2, 2007 and is carried in FINRA Rule 4210(g).
Risk-based. Brokers stress-test all long and short positions under simulated market moves and set margin equal to the worst projected loss, allowing hedges and offsets to reduce requirements while adding extra margin for concentration, volatility, or hard-to-borrow shorts. Because it is recalculated continuously, PM can change intraday and requires active risk management.OCC develops the TIMS methodology. FINRA incorporates SEC portfolio margin rules in Rule 4210(g). SEC oversees FINRA and reviews OCC rule filings.
15% rule. Under TIMS, the floor for any individual equity or narrow-based index — blue chip or not — is the worst loss across a ±15% stress of the underlying, so your equity must cover a 15% adverse move: 6.67x maximum buying power. Broad-based, high-capitalization index products are stressed only +6% / −8%. In practice you will see less than 6.67x: house add-ons for concentration, volatility, and borrow bind first.
§4210(g)(2)(F) The term “theoretical gains and losses” means the gain and loss in the value of individual eligible products and related instruments at ten equidistant intervals (valuation points) ranging from an assumed movement (both up and down) in the current market value of the underlying instrument. The magnitude of the valuation point range shall be as follows:
Portfolio Type Up / Down Market Move
(High & Low Valuation Points)High Capitalization, Broad-based Market Index2 +6% / -8% Non-High Capitalization, Broad-based Market Index3 +/- 10% Any other eligible product that is, or is based on, an equity security or a narrow-based index +/- 15%
Beta exposure
145/45. The notation is gross positioning relative to equity: 145% LMV against 45% SMV. The strategy is an active extension — the “long-short fund” of the marketing decks — built to keep 100% net market exposure while widening the manager’s room to express views: same beta as the index fund, more ways to be right (and wrong).
145/45 T-account (in millions)
LONG SHORT
LMV | DR CR | SMV
145 | 45 45 | 45
|----------- |-----------
| EQ 100 | EQ 0*
-------------------- --------------------
Reg T Margin 72.5 Reg T Margin 22.5
(50% of LMV) (50% of SMV)
-------------------- --------------------
Finra Margin 36.25 Finra Margin 13.5
(25% of LMV) (30% of SMV)
-------------------- --------------------
Ex EQ 27.5 Ex EQ (22.5)
(Eq - Reg T) (Eq - Reg T)
-------------------- --------------------
SMA 27.5 SMA (22.5)
Aggregating the account, the surplus on the LONG side absorbs the SHORT side’s deficiency, leaving a compliant but tight $5m of Fed surplus. (Against FINRA maintenance the same book carries $50.25m of excess; at inception, the Fed test binds.)
LMV 145 | SMV 45
CR 45* | DR 45*
|----------
| EQ 100
--------------------
Reg T Margin 95
(72.5 Long + 22.5 Short)
--------------------
Finra Margin 49.75
(36.25 Long + 13.5 Short)
--------------------
Ex EQ 5
(100 Eq - 95 Req)
--------------------
SMA 5
(Aggregated Surplus)
The asterisks mark the internal plumbing: the 45 debit is the loan financing the long extension; the 45 credit is the short-sale proceeds, held as collateral. They are equal in size but cannot cancel — proceeds are segregated in the short account and do not retire the debit. (The 100/0 split of equity between the panels is likewise presentational; only the aggregate is regulatory.)
The same 100 of equity, scored by all three rulers at once:
One equity, three rulers (the 145/45, per 100 of EQ)
0 25 50 75 100
+---------+---------+---------+---------+
EQ |=======================================| 100
Fed |=====================================| 95 -- Fed surplus 5: binds at inception
House |===========================| ~70 -- house surplus ~30 (37% of gross, illustrative)
FINRA |===================| 49.75 -- exchange surplus 50.25
45/45. Unlike the 145/45 investor — who wants to beat the market while riding it — the 45/45 investor does not care where the market goes. Fund the same 45-long, 45-short book two ways. Keep the full 100 of cash, and the credit balance is 100: 45 of it short-sale proceeds, 55 free, for an aggregate Fed surplus of 55.
45/45 with 100% Cash
LONG SHORT
LMV 45 | DR 0 CR 100 | SMV 45
|---------- |----------
| EQ 45 | EQ 55
-------------------- --------------------
Reg T Margin 22.50 Reg T Margin 22.50
(50% of LMV) (50% of SMV)
-------------------- --------------------
Finra Margin 11.25 Finra Margin 13.50
(25% of LMV) (30% of SMV)
-------------------- --------------------
Ex EQ 22.50 Ex EQ 32.50
(Eq - Reg T) (Eq - Reg T)
-------------------- --------------------
SMA 22.50 SMA 32.50
LMV 45 | SMV 45
CR 100 | DR 0
|----------
| EQ 100
--------------------
Reg T Margin 45
(22.5 Long + 22.5 Short)
--------------------
Finra Margin 24.75
(11.25 Long + 13.5 Short)
--------------------
Ex EQ 55
(100 Eq - 45 Req)
--------------------
SMA 55
Now withdraw 50. The positions are unchanged, but they stand on 50 of equity: the Fed surplus collapses to 5, and what read as “cash on the ledger” reveals itself as collateral. (In exposure terms the book is now 90/90 — leverage is a ratio, and you just halved the denominator.)
45/45 with 0% cash
LONG SHORT
LMV 45 | DR 0 CR 50 | SMV 45
|---------- |----------
| EQ 45 | EQ 5
-------------------- --------------------
Reg T Margin 22.50 Reg T Margin 22.50
(50% of LMV) (50% of SMV)
-------------------- --------------------
Finra Margin 11.25 Finra Margin 13.50
(25% of LMV) (30% of SMV)
-------------------- --------------------
Ex EQ 22.50 Ex EQ (17.50)
(Eq - Reg T) (Eq - Reg T)
-------------------- --------------------
SMA 22.50 SMA (17.50)
LMV 45 | SMV 45
CR 50 | DR 0
|----------
| EQ 50
--------------------
Reg T Margin 45
(22.5 Long + 22.5 Short)
--------------------
Finra Margin 24.75
(11.25 Long + 13.5 Short)
--------------------
Ex EQ 5
(50 Eq - 45 Req)
--------------------
SMA 5
What you might want to know
Theory should survive contact with a custodian. Below are two actual Fidelity margin statements — the first a market-neutral book, the second a long extension — mapped onto the ledger above, panel by panel. Fidelity’s nouns differ from Reg T’s, but the page becomes readable once you see that it prints two families of rules, not one. The three rulers of the first half are the family of adequacy: how much equity must stand behind the positions. Everything else on the page belongs to the family of custody: whose property each balance is, who may touch it, and for how much. The remainder is settlement plumbing.
First, the type system. Fidelity (like most custodians) tags every balance with the regulatory account it lives in: cash (1), margin (2), short (3) — the §220 segregation made visible. The core sweep is type-1 cash; longs sit in type 2; short positions and their sale proceeds sit in type 3, with the “Short Credit” kept close to the position by mark-to-market.
The custody rulebook. Margin looks like arithmetic, but underneath it is a deal about property: you want to borrow, the broker wants collateral, so you hand the broker rights over your stuff — and a stack of laws then spends all its energy on three ordinary questions.
Whose stuff is it? Reg T’s answer is to split your account in two. In the cash account (§220.8), everything must be paid in full and no credit exists, so the property there is simply yours — hands off. In the margin account (§220.4), all the borrowing lives, so the property there is yours-but-pledged. And the broker may not squint across the wall: under §220.3(b), property helps the margin side only by actually moving there — which is the journal you will meet shortly.
What may the broker do with the pledged stuff? The agreement you signed says, roughly, anything; federal law trims it, and every trim is a memory. In the 1920s, brokers pooled customers’ securities under blanket liens and pledged the pool for whatever the house needed — its own speculation included; when the brokers failed in 1929, the banks simply kept the collateral, and customers in perfect standing learned their stock had answered for strangers’ debts and the firm’s own bets. The hypothecation rules (8c-1 and 15c2-1, New Deal vintage) are that disaster written as three prohibitions: the broker may not mix your collateral with other customers’ under one lien without your written consent, may never mix it with the firm’s own property, and may not re-pledge more than everything its customers collectively owe it. The customer-protection rule answers a stupider catastrophe — the 1968–70 paperwork crisis, when Wall Street’s back offices drowned in their own volume, firms like Goodbody and du Pont were carried out the door, and customer securities simply could not be located, never mind returned. Congress passed SIPA in 1970, the SEC wrote 15c3-3(b), (e) in 1972, and its bright line is that your stuff must verifiably exist: every security you have fully paid for, and everything above 140% of your loan — a number derived below — must sit untouched in the broker’s possession or control, and the cash the broker owes you must be backed by real deposits in a special reserve bank account.
And if the broker goes broke? SIPA liquidation hands the segregated property back to you in kind, splits whatever customer property remains pro rata, and SIPC advances up to $500,000 per customer, $250,000 of it for cash. (Lehman’s London clients ran the experiment of living without these rules; the prime-brokerage section returns to them.) Adequacy says whether you may lever; custody says what the leverage is secured by — and how far the security reaches.
The two rulebooks on one page
adequacy: how much equity custody: whose property, who may touch it
Fed initial §220.12 security interest margin agreement; UCC 8 & 9
FINRA maintenance 4210(c) account walls §220.8 / §220.4 / §220.3(b)
house the agreement consent & commingling 8c-1, 15c2-1
re-use cap aggregate indebtedness; 140% of DR
segregation 15c3-3(b): possession or control
cash lockup 15c3-3(e): reserve formula
backstop SIPA / SIPC
The point of all of it is one sentence: the customer side of a broker-dealer must be a closed loop. Your property may fund your credit — dollar for dollar, traceable, returnable — and may never become the broker’s working capital. Notice, though, that the rules are also a subsidy, which is the part people miss: they permit re-use up to the aggregate cap and the 140% line, and if brokers could not touch customer collateral at all, every margin loan would be funded off the broker’s own balance sheet and priced like unsecured credit. The regime is a deal — your loan funded by your own assets, cheaply, inside a loop the broker cannot raid. The rule is not “the broker can’t touch your stuff”; the rule is “the broker can touch exactly as much of your stuff as it takes to fund your loan, and not one share more.”
The closed loop
pledge collateral re-pledge, <= 1.4 x DR
--------------------------> -------------------------->
you broker bank
<-------------------------- <--------------------------
margin loan, DR funding, ~ DR
above 1.4 x DR .............. the box (possession or control)
customer cash ................ reserve bank account, 15c3-3(e)
the loop funds your loan; nothing leaks into the broker's own book
The words, precisely. Start with the uncomfortable fact underneath all of it: you do not, strictly speaking, own any shares at your broker. What you own is a security entitlement (UCC 8-501) — a legal claim against your broker to shares like yours, held for you somewhere in a pool.
Where your "shares" actually live
issuer's register: Cede & Co. -- DTC's nominee: the owner of record
DTC's books: your broker nnn,nnn sh -- the broker's securities account
broker's books: you 100 sh -- your security entitlement (UCC 8-501)
two layers of bookkeeping stand between you and the company
The rest of the vocabulary describes what happens to that claim when you borrow. A security interest is the law’s name for “this property answers for that debt” (UCC 1-201(b)(35)) — the personal-property cousin of the bank’s mortgage on your house. Its practical meaning is twofold: on default the secured party may take the property and sell it without first suing you, and it stands ahead of every unsecured creditor in line. Collateral is the property doing the answering (UCC 9-102). A lien is the older, broader word for any claim that ties property to a debt — law French for “bond,” from ligare, to bind — and the Bankruptcy Code still defines a security interest as just a “lien created by an agreement” (11 U.S.C. §101). Liens come in two kinds: some arise by statute whether you like it or not — the mechanic who fixed your roof has one — and some arise by signature. The broker’s is the second kind: you granted it in the margin agreement, so read what you granted. Fidelity’s provides that property in a margin account “may be pledged or repledged, hypothecated (loaned) or rehypothecated,” separately or together with other property — three permissions in one clause: your securities may leave the building, they may be mixed into a pool with other people’s, and the pool may secure a sum larger than what you personally owe.The clause also waives any duty to retain like securities on hand — breadth the federal rules then trim to the aggregate-indebtedness ceiling and the 140% line.
Hypothecation deserves its own sentence, because the word is doing real work. Roman law distinguished the pignus, a pawn — collateral handed over, the pawnbroker holds your watch — from the hypotheca, collateral that stays with the debtor while the creditor holds only a claim. Margin is hypotheca: you keep the dividends, the votes, the upside, the position on your screen; the broker keeps a right against all of it. Rehypothecation is the broker doing the same thing one level up — pledging your collateral onward to fund itself. Walk the chain once: you pledge 140 of stock against a loan of 100; the broker carries that same stock to a bank and borrows roughly 100 against it; now one block of collateral secures two debts, yours to the broker and the broker’s to the bank — the closed-loop diagram above, in motion. This sounds like a scandal and is in fact the design: UCC 9-207(c)(3) says flatly that a secured party with possession or control “may create a security interest in the collateral.” The risk it creates is equally specific: while your stock is out on re-pledge, it answers for the broker’s debt too, and the bank’s claim is senior — the seed of every Lehman story, and the precise thing the 140% cap exists to confine.
And the Code is generous to your broker in ways worth sitting up for. First, the word perfect, which here means complete, not flawless. A security interest attaches when the agreement binds you (UCC 9-203); it is perfected when it binds the world — other lenders, buyers of the collateral, and above all your bankruptcy trustee, who avoids any interest left unperfected and demotes its holder to unsecured (UCC 9-317; 11 U.S.C. §544). Perfection normally costs publicity: the lender files a financing statement in a public registry so that anyone can search and see the property is spoken for — the whole apparatus exists to abolish secret liens. Your broker perfects by existing. For investment property the Code accepts control in place of filing (UCC 9-314) — control being dematerialized possession, the pignus again with the watch replaced by a book entry, and possession has always been self-perfecting because holding the collateral is the notice. Since your entitlement lives on the broker’s own books, a grant to your own securities intermediary counts as control automatically (UCC 8-106(e)): the moment you sign, the interest is attached, perfected, and invisible — no filing anywhere, nothing for the world to search. It also outranks every other secured creditor, even one who obtained control (UCC 9-328(3)) — which is why a bank lending against your brokerage account will make you move the assets to its custody, or extract a subordination from the broker, rather than rely on its own paperwork. A system built to abolish secret liens thus blesses exactly one — automatic, invisible, senior — on the theory that custody itself is the notice. None of this is accident: the 1994 rewrite of Article 8 chose settlement certainty for intermediaries over debtor protections, deliberately, after the back-office failures of the prior decades. State law, in short, hands the broker everything; the federal ladder above is the give-back. The asymmetry shows even at the edges: for the broker to borrow the shares you have fully paid for, it needs a separate signed agreement and full collateral (15c3-3(b)(3)) — at Fidelity the fully-paid lending MSLA, under which your shares leave type 1 and are re-tagged type 6, you collect a slice of the borrow fee, and the fine print does the pricing: dividends arrive as cash-in-lieu taxed at ordinary rates, the vote travels with the shares, and SIPC does not cover what is out on loan — the collateral is your protection.
Now the type system makes sense. The wall between the types is collateral law, not bookkeeping taste: a type-1 dollar and a type-2 dollar are the same dollar under two different legal regimes, and compliance is scored per account. The lien is the agreement’s; federal law does not create it, it caps it.The stack: 12 CFR §220.8(a)(1) — purchases in a cash account only with sufficient funds or a good-faith agreement to pay in full promptly; §220.4(a)(1) and §220.2 — the single-account principle, and equity defined over “the margin account”; 17 CFR 240.8c-1 and 240.15c2-1 — no commingling of customers’ collateral under one lien without written consent, none with the firm’s own, no hypothecation beyond customers’ aggregate indebtedness; 17 CFR 240.15c3-3(b)(1) — possession or control of all fully-paid and excess-margin securities, with (a)(5) drawing the excess-margin line at 140% of the debit and (b)(3) requiring full collateral even to borrow fully-paids. Assets in types 2 and 3 are the pool that secures the debit and the stock borrow. Assets in type 1 are the customer’s unencumbered property — segregated, unable to collateralize the loan. Margin Equity, and therefore the house and exchange surpluses, is accordingly computed over types 2 and 3 alone.The 45/45 statement below is the proof: Margin Equity of 334,136.39 against a Total Account Value of 743,688.88, the 410k core invisible to every margin test.
Within the pledged pool stands a second wall, and here is the part that surprises people. The lien covers everything in types 2 and 3 — but the broker’s permitted use of the collateral covers only a slice, and the slice is measured by the strange-looking number 140. Under the customer-protection rule (SEA 15c3-3), the broker may treat as working collateral only securities worth up to 140% of your debit; everything above that line is “excess margin” and must be held in possession or control — in the broker’s vault or at a control location, free of any lien, not loaned out, not re-pledged — operationally just like the fully-paids in a cash account. Two things to fix about the number. First, its direction: this is not collateral you must post, like the short rule’s 150% or stock loan’s 102% — it is a ceiling on what the broker may take out. Second, it is not arbitrary; it is the funding chain solved for collateral. The broker lent you DR and is allowed to raise that money by re-pledging your securities to a bank — but banks lend against stock at a haircut, historically about 70 cents per dollar of collateral, so raising DR takes DR / 0.7 ≈ 1.4 × DR of stock. The ceiling is sized so that your loan can fund itself, and nothing else: the surplus of your property cannot be conscripted to fund the broker’s inventory or its other customers. The broker designates which names fill the 140% bucket — the most loanable ones, with the rest going to the box. And strictly this is a long-side instrument, a rule about securities measured against the debit: the short proceeds in type 3 are cash, governed instead by the reserve formula and the 102% collateral chain to the stock lender.
Why 140, not 100
the broker must raise your DR, say 100, to fund your loan
a bank lends ~70 against 100 of stock -- the haircut
collateral needed: 0.70 x C = 100 => C = 100 / 0.70 ≈ 140
the ceiling lets your loan fund itself at a haircut, and no more;
everything above it is "excess margin": yours, in the box
The 140% line on the 130/30 below
DR 213,560.78 -> line = 1.4 x DR = 298,985.09
LMV 935,653.60
re-usable "margin securities" 298,985.09 32% -- may be re-pledged, loaned
excess margin, locked 636,668.51 68% -- possession or control, as if fully paid
Even in a levered extension, two-thirds of the longs sit beyond the broker’s reach for re-use — and the 45/45 is the limiting case: with DR = 0 the line sits at zero, the lien exists on paper and reaches nothing, every long in type 2 is excess-margin and segregated. This is also the provision that makes a margin account survivable in a broker failure: in a SIPA liquidation, securities in possession or control are customer property returned in kind; the exposure is confined to the slice inside the 140% actually out on re-pledge. The type codes themselves appear in none of this law. 1/2/3 is back-office convention from the NYSE-era account coding; the law speaks of accounts, the wire format of types.
Equity scored over types 2 and 3 alone is why a margin call can coexist with ample cash. The call fires on margin-segment equity; type-1 cash counts in NAV but not in collateral. And cash drifts into type 1 by default — deposits and ACH land in core, dividends and sweep interest accrue there, sales of fully-paid securities pay there. Retail configurations often draw the core against a call automatically; managed and institutional configurations typically wall it off, so the cure is an explicit journal — a type transfer, 1 → 2:
journal 100,000: type 1 -> type 2 (TAV / NAV unchanged)
before after
type-1 cash 100,000 -> 0
CR (type 2) x -> x + 100,000
Margin Equity ME -> ME + 100,000
house surplus (60,000) -> 40,000 -- call cured
exchange surplus s -> s + 100,000
SMA sma -> sma + 100,000
The journal is the cheapest cure on the menu: NAV-neutral — no trade, no settlement, no tax event — yet it raises margin equity, all three surpluses, and SMA dollar-for-dollar, instantly; if a debit exists, the cash pays it down first and the interest stops. Compare liquidation, which costs 1/rate times the call plus market impact and realized gains. The only price is the one the type system exists to mark: the cash stops being unencumbered property and joins the pledged pool. The reverse move, 2 → 1, is correspondingly not free — it is a withdrawal for Reg T purposes, bounded by SMA and the surpluses, exactly the “Available to Withdraw” logic on the statement. And one more place the wall bites: you cannot short against type-1 assets at all, so the 1 → 2 journal is usually the first instruction in onboarding an extension account, not just the cure for a call.
Second, the vocabulary, since the industry and the regulation do not share a dictionary. NAV (net asset value) is assets less liabilities — at the account level, exactly §220.2 equity: NAV = EQ = LMV + CR - SMV - DR, Fidelity’s “Total Account Value,” and the denominator of every strategy ratio in this post. GMV (gross market value) is LMV + SMV: the total securities footprint regardless of direction, the denominator of “Margin Equity %,” and the quantity that financing, borrow, and margin requirements actually scale with. NMV (net market value) is LMV - SMV: the statement’s “Securities Market Value” line — not your wealth, your directional exposure. Gross measures the plumbing; net measures the beta.
Cash is not one number. A margin ledger holds credit and debit balances by type — free credit in type 1, margin credit or debit in type 2, short credit in type 3 — and what enters NAV is their sum, net cash = CR - DR. Colloquial “cash” is only the free credit — a regulatory term of art, not a loose one. Rule 15c3-3(a)(8) defines free credit balances as the broker’s liabilities to you that are payable on demand: cash with no strings on it — neither pledged as collateral nor waiting on a settlement. On this ledger that means net type-1 cash plus any type-2 credit (on the 45/45 below: 410,325.99 − 773.50 + 9.36 = 409,561.85), and it is exactly the quantity the reserve formula locks into the special bank account — because a free credit balance is, economically, an unsecured loan from you to your broker. Free credit is spendable now and withdrawable subject to the surpluses; everything else called “cash” fails one of the two tests. The short proceeds inside CR fail the first — pledged: the margin agreement lets the broker re-use that cash to collateralize the stock borrow, institutionally at 102% of marked value — so CR enters equity in full but “Available to Withdraw” only in part. Settled cash is free credit whose generating trades have completed settlement (T+1 for US equities since May 2024). An unsettled credit fails the second — sale proceeds in the limbo between execution and settlement: contractually yours, redeployable at once in a margin account, withdrawable only on settlement — and in a cash account, buying with it and then selling the new position before the original sale settles is a good-faith violation under §220.8. An unsettled debit is the mirror, a purchase awaiting payment; at settlement it consumes free credit or rolls into DR — which is precisely how a debit is born. In the cash account the same receivable has a deadline instead of a destination: §220.8 bans credit, not debits, so a type-1 debit exists only in transit — a charge or unsettled purchase awaiting the sweep or prompt payment — and one left unpaid ends in liquidation and the 90-day freeze, never in DR. The 45/45’s −773.50 is exactly this: the day’s charges waiting for the overnight core redemption, already netted on the statement’s Settled Cash line.
"Cash" on the 45/45 ledger below
CR 733,708
|
+-- free credit 409,562 spendable now (type 1 core, type 2 credit)
|
+-- short proceeds 324,146 pledged: collateralizes the stock borrow (type 3)
|
+-- marked to SMV 319,988 cancels against SMV inside EQ
|
+-- excess 4,158 unswept short P&L; counts in NAV
Life of a sale under T+1
T (trade) T+1 (settle)
sell ---> unsettled credit ---> settled cash
redeployable on withdrawable,
margin; not no strings
withdrawable
Excess short credit is defined by one subtraction: the type-3 credit minus current SMV — the wedge between the collateral actually posted against the borrow and the marked size of the debt it secures. The wedge has three sources: short-side gains the sweep has not yet harvested (SMV fell; the credit still reflects the last mark), the collateral cushion where the posting target exceeds the mark (the institutional 102% builds in a structural two points), and plain timing between marks. It can run negative too — a rallying short opens a deficit that the next sweep tops up out of type 2 or new borrowing. The asymmetry to hold onto: NAV sees the wedge instantly, because every credit counts in equity the moment it exists; spendability waits for the sweep to release it into type 2. So CR decomposes exactly — free credit, plus marked proceeds, plus this excess: three claims with three different legal lives.
The mark-to-market sweep
short gains (SMV falls): excess released
type 3 ------------------------------------------> type 2
short credit free credit,
tracks SMV or pays down DR
<------------------------------------------
short loses (SMV rises): collateral topped up
Since marked proceeds and SMV cancel inside equity, a short position contributes to NAV only through this excess:
NAV = NMV + net cash
NAV = LMV + free credit + excess short credit - DR
A freshly marked short is NAV-neutral; only its P&L moves equity. Both statements below close on both identities to the penny.
The 45/45.
Both statements below are live. Color key: LMV · SMV · CR · DR · EQ — hover any colored item to trace it through both ledgers, T-accounts and statement lines alike, in either direction.
~45/45, market-neutral LMV 329,969 | SMV 319,988 (whole dollars) CR 733,708 | DR 0 |------------- | EQ 743,689 ---------------------------------------------------- Total Account Value 743,688.88 -- EQ = NAV Securities Market Value 9,980.83 -- NMV, 1.3% of EQ: beta ~ 0 Non-Core Money Market (1) 0.00 Core Sweep/Fund (1) 410,325.99 -- type-1 cash Cash (1) Debit -773.50 -- type-1 debit: charges awaiting the sweep Margin Credit 9.36 -- CR, type 2 Short (3) Credit 324,146.20 -- CR, type 3: short proceeds Committed to Open Orders 0.00 -- pending activities ---------------------------------------------------- Available to Trade Settled Cash 409,552.49 -- 410,325.99 - 773.50 Unsettled Cash Credit 0.00 -- nothing in flight Unsettled Cash Debit 0.00 Buying Power Cash Only 409,552.49 -- settled cash Cash & Marg (marginable) 1,071,055.48 -- 2 x (409,552.49 + 125,975.25 SMA) Cash & Margin (non-margin) 463,553.36 -- ~ cash + house surplus, at 100% req Additional Buying Power -- per-asset-class requirements Day Trade/Min Equity Call 0.00 Corporates 1,785,092.46 -- (cash + SMA) / 30%, exact Municipals 2,317,766.80 Governments 4,635,533.60 -- = 2 x municipals here ---------------------------------------------------- Securities Market Value (MV) Securities Market Value 9,980.83 -- NMV = LMV - SMV Non-Core Money Market (1) 0.00 Cash (1) 0.00 Margin (2) 329,969.30 -- LMV, 44% of EQ Short (3) -319,988.47 -- SMV, 43% of EQ ---------------------------------------------------- Margin Balances Margin Equity 334,136.39 -- EQ of types 2+3 only Margin Equity % 51.41% -- 334,136.39 / 649,957.77 GMV Pending Margin Interest 0.00 -- no debit, no interest Margin Interest Rate 0.00% House Surplus 57,242.18 -- 334,136.39 - 276,894.21 Exchange Surplus 134,719.84 -- req 199,416.55; 25/30 floor 178,488.87 SMA 125,975.25 -- Fed high-water ledger Day Trade Call 0.00 House Security Requirements 276,894.21 -- 42.6% of GMV House Option Requirements 0.00 ---------------------------------------------------- free credit 409,561.85 -- 410,325.99 - 773.50 + 9.36 excess short credit 4,157.73 -- 324,146.20 - 319,988.47 NAV 743,688.88 -- 329,969.30 + 409,561.85 + 4,157.73 - 0
Three readings. The exposure arithmetic is the strategy itself: 330k long against 320k short, netting a three-quarter-million-dollar account down to ten thousand dollars of market — β ≈ 0 by construction. The exchange requirement does not reconcile to the textbook 25/30 floor, instructively: 30.7% of gross, the gap being the per-share minimums from the maintenance section — published rates are floors, not formulas. And buying power is the cleanest identity on the page: twice settled cash plus SMA, exactly the 2:1 of the SMA section extended by free cash — and the Additional Buying Power ladder is the same arithmetic with the requirement swapped, corporates reconciling to exactly a 30% requirement on the same base. Note also what Margin Equity excludes: the type-1 core never enters it until journaled — cash standing beside the margin account is not cash standing in it.
The 130/30.
Here the four-legged equity definition of §220.2 reconciles exactly:
EQ = LMV + CR − SMV − DR = 935,653.60 + 221,037.57 − 216,756.87 − 213,560.78 = 726,373.52.
~130/30, beta ~ 1 LMV 935,654 | SMV 216,757 (whole dollars) CR 221,038 | DR 213,561 |------------- | EQ 726,374 ---------------------------------------------------- Total Account Value 726,373.52 -- EQ = NAV Securities Market Value 718,896.73 -- NMV, 99% of EQ: beta ~ 1 Non-Core Money Market (1) 0.00 Core Sweep/Fund (1) 0.00 -- no type-1 cash Cash (1) Credit/Debit 0.00 Margin Debit -213,560.78 -- DR: finances the 30-pt extension Short (3) Credit 221,037.57 -- CR, type 3: ~102% of SMV Committed to Open Orders 0.00 -- pending activities ---------------------------------------------------- Available to Trade Settled Cash 0.00 -- fully deployed Unsettled Cash Credit 0.00 Unsettled Cash Debit 0.00 Buying Power Cash Only 0.00 Cash & Margin (marginable) 427,440.88 -- 2 x SMA, to the penny Cash & Margin (non-margin) 213,720.44 -- 1 x SMA, to the penny Additional Buying Power -- per-asset-class requirements Day Trade/Min Equity Call 0.00 Corporates 712,401.46 -- SMA / 30%, exact Municipals 1,481,050.60 Governments 2,137,204.40 -- = 10 x SMA ---------------------------------------------------- Available to Withdraw Cash Only 0.00 Cash & Margin 216,469.74 -- ~ SMA ---------------------------------------------------- Securities Market Value (MV) Securities Market Value 718,896.73 -- NMV = LMV - SMV Non-Core Money Market (1) 0.00 Cash (1) 0.00 Margin (2) 935,653.60 -- LMV, 129% of EQ Short (3) -216,756.87 -- SMV, 30% of EQ ---------------------------------------------------- Margin Balances Margin Equity b 726,373.52 -- = NAV here (no type 1) Margin Equity % 63.03% -- 726,373.52 / 1,152,410.47 GMV Pending Margin Interest 492.13 -- ~ $24/day, ~20 days accrued Margin Interest Rate 4.07% -- on the full DR; CR does not offset House Surplus 294,821.65 -- 726,373.52 - 431,551.87 Exchange Surplus 404,624.04 -- req 321,749.48; 25/30 floor 298,940.46 SMA 213,720.44 -- Fed high-water ledger Day Trade Call 0.00 House Security Requirements 431,551.87 -- 37.4% of GMV: the binding test House Option Requirements 0.00 ---------------------------------------------------- free credit 0.00 excess short credit 4,280.70 -- 221,037.57 - 216,756.87 NAV 726,373.52 -- 935,653.60 + 0 + 4,280.70 - 213,560.78 ---------------------------------------------------- statement note: "Unpriced position(s) exist in this account."
The plumbing is the lesson. The debit and the short credit sit one line apart, nearly equal in size, and cannot cancel: proceeds are segregated collateral in type 3, so the account pays 4.07% on the full loan while also paying borrow fees on the shorts — the running cost of the extension. (Institutional prime arrangements claw some of this back through short rebates; retail margin does not.) And the house test binds first: 37% of gross against an exchange requirement near 28% — nine points of “house” between this account and the regulatory floor.
For the custodians’ own vocabulary, see Fidelity’s margin FAQ and balance-field definitions, and Schwab’s margin requirements schedule. Their numbers, not their prose, are the authority — which is why the reconciliation above is done in arithmetic.The “Day Trade Call” row is a relic in transition: FINRA’s new intraday margin standards (Regulatory Notice 26-10) replace the pattern-day-trader framework beginning June 2026.
Other topics
Multiple accounts
One customer rarely means one account: the taxable and the IRA, the trust, the second strategy. Reg T’s default for all of them is separation, and every instrument that overcomes it has its price written into its conditions.
Separation is the default. §220.3(b)(1): the requirements of one account may not be met by considering items in any other — and when cash or securities are used to meet requirements in another account, written entries shall be made. That clause is the statutory name for the journal above: the broker may not look across accounts; assets must move. The carve-outs are two and narrow — good-faith-account assets serving in lieu of margin, and transfers between the margin account and the SMA.
One margin account per customer, with listed exceptions. Reg T contemplates a single margin account per customer — subdivisions are bookkeeping, scored as one unit — and §220.4(a)(2) permits separate margin accounts for the same person only for clearing arrangements and, in (a)(2)(iii), accounts over which the creditor or a third-party investment adviser holds investment discretion. That clause is the legal chassis of the managed-account industry: every program stands as its own margin account at the custodian, and each passes its three tests alone. The two statements above may well belong to one customer; the law would still score them separately.
Discretion sets the requirement, not the collateral. Discretion can run down a chain. A financial advisor holding the client’s mandate sub-delegates a strategy to a subadvisor; the subadvisor’s firm is the “third-party investment adviser” of (a)(2)(iii), and its sleeve stands as its own margin account.
What the subadvisor controls is the sleeve’s positions — and positions are what set a margin requirement, so the subadvisor sets its own. What it does not control is the collateral that meets that requirement when the collateral sits in a sibling account it does not manage. The reason is separation: because the sleeve is a separate margin account, assets elsewhere in the client’s relationship reach it only by the instruments that lift separation anywhere — a journal, a guarantee, a cross-collateralization consent — and each is the client’s to grant, because each pledges the client’s other property. The subadvisor is a party to none of them.
So the party that creates the requirement is, by design, not the party that controls what meets it, and the sleeve runs in one of two regimes. With no instrument in place, it is self-contained: scored on its own equity, with a call cured into it by whoever holds its cash — the client or the lead advisor — never by the subadvisor reaching across. With a cross-collateralization agreement already in place, the custodian scores the combined position, so a sibling account’s excess absorbs the sleeve’s deficit automatically — convenient, until you notice the liability runs both ways and the sleeve’s own trading is drawing down a pool it does not see.
Tri-party sleeve: discretion vs. collateral control
client owns both accounts A and B
A the sleeve discretion held by the subadvisor [(a)(2)(iii)]
B sibling account discretion held by the lead advisor / client
requirement on A : the subadvisor's positions set it
collateral for A : the client's property -- reaches A from B only by an
instrument (journal, guarantee, 4210(f)(6)); two-way liability
Make it concrete with our own book. Sleeve A is a 45/45 — 45 long, 45 short, the long side financed by a 45 debit, the short proceeds held as 45 of credit — and worked on its own its equity is zero: a market-neutral overlay contributes none. What it stands on is the collateral in B: a single long equity, the client’s funding stock, which is not market-neutral at all. So A’s requirement is set by A’s gross, which scarcely moves, while A’s survival is set by a price A neither chose nor can trade.
A 45/45 sleeve standing on B's funding stock
A 45/45, market-neutral: LMV 45 | SMV 45 | DR 45 | CR 45 equity 0
B the collateral: one long stock worth V equity V
combined under (f)(6): LMV 45+V | SMV 45 | DR 45 | CR 45 equity = V
V=100 V=80 V=33
combined equity 100 80 33
maintenance req 49.75 44.75 33.00
surplus 50.25 35.25 0.00
equity = V (A adds none); req = 24.75 + 0.25 V; surplus hits 0 at V = 33
A's own P&L stays ~0 -- every one of these calls is B's stock, not A's book
A spent all its effort cancelling market beta and inherited single-name beta through the back door. When the funding stock slides, the consolidated surplus slides with it — by about three-quarters of the decline, since the stock is a long and its fall eases the long requirement by the other quarter — and a deep enough drawdown calls an account whose own book never moved. The subadvisor cannot add equity and cannot trade B; his only self-help is to shrink his own gross, forcibly de-risking a market-neutral strategy because someone else’s collateral fell. That is the price of standing on collateral you do not manage.
Consolidation is a document, not a default. Under FINRA 4210(f)(6), maintenance margin for a customer’s several accounts may be computed on the net position — provided the customer consents that the assets in each may be used to carry or pay any deficit in all.The interpretation under 4210(f)(6): maintenance margin “may be determined on the net position of said accounts,” given the customer’s consent that money and securities in each account may carry any deficit in all of them. Read the condition: netting is purchased with cross-liability, never granted alongside separateness. And it reaches only the maintenance tests — each margin account still passes Reg T alone, the separation rule again. Across customers the instrument is the guarantee, 4210(f)(4): an account guaranteed by another may be consolidated with it for margin, on a written guarantee with the firm holding access to both pools.
Two accounts, one customer
default (§220.3(b)) consolidated (4210(f)(6), by consent)
A: EQ 100 | req 60 surplus 40 A+B: EQ 150 | req 140 surplus 10
B: EQ 50 | req 80 call (30) maintenance only; Reg T still per account
price: A's assets answer for B's deficit
Margin nets exactly as far as assets answer for each other’s debts. Netting and liability are one fact seen from two sides, and the document is where the choice is made.
Prime brokerage
Everything above describes the default regime: a customer, a custodial broker, three public rulers. Institutional long-short runs on a negotiated one.
The architecture is a no-action letter, not a rule. US prime brokerage rests on the SEC’s letter of January 25, 1994 to the Prime Broker Committee: execute anywhere, give the trades up, settle everything into one margin account at the prime. The letter’s function is to bless the give-up against Reg T’s payment provisions, and it does so by letting each executing broker treat the prime-broker account as a broker-dealer credit account — citing “§220.11,” a section that has been §220.7 since 1998.Two cross-references are fossils of the 1998 restructuring of Reg T: the 1994 letter points at §220.11 for the broker-dealer credit account (now §220.7), and Rule 8c-1(c) still cites “12 CFR 220.4(c)” for the special cash account (now §220.8). It also sets the floors: minimum net equity of $500,000 — $100,000 for an account managed by a registered investment adviser — restored within five business days if breached, against a prime carrying $1.5 million of net capital.
The 1994 architecture
EB A --give-up-->\
EB B --give-up--> prime broker: one margin account, one EQ
EB C --give-up-->/
|
+-- US broker-dealer ...... PM (4210(g)); 15c3-3; the 140% cap
+-- offshore affiliate .... arranged financing; the house ruler only
Reg T recedes in three steps. An institutional account at the US prime is portfolio-margined under 4210(g): no strategy-based initial, no SMA — the PM section above. The rules still bind the US broker-dealer regardless of who the customer is; entity status alone escapes nothing. The escape is the third step: arranged financing through the non-US affiliate. Reg T governs credit extended by US broker-dealers; the London entity is not one, so margin there is house risk appetite under the international prime brokerage agreement — which is how equity long-short carries leverage neither Reg T nor TIMS would sign. The US entity arranges; the affiliate lends; the three rulers collapse to one, the house’s.
Customer protection inverts with it. At the US entity, the customer-protection rule still applies to the fund as customer — possession-or-control, the 140% cap, the reserve formula — and cannot be contracted away; the agreement’s rehypothecation consent operates inside the cap. Move the assets to the offshore affiliate and the cap is gone: title-transfer collateral and contractual — historically unlimited — re-use, which is the Lehman lesson: LBIE clients whose assets had been rehypothecated woke up as unsecured creditors of the estate. The post-2008 settlement is documentary, not regulatory: negotiated rehypothecation caps, and prime custody — excess fully-paid longs swept to a bank custodian, outside the broker-dealer estate. Note the symmetry with the journal above: retail assets start unencumbered and are journaled into the pledged pool; prime assets start in the pool and are negotiated out.
Where the lien starts
retail / custodial prime brokerage
type 1: unencumbered by default PBA lien over the account by default
| |
| journal 1 -> 2: opt in | prime custody carve-out: opt out
v v
the pledged pool excess fully-paids at a bank custodian
The economics are repriced rather than regulated. Debit interest becomes a negotiated spread over SOFR; the short credit earns a rebate, so the asymmetry on the 130/30 statement — 4.07% on the full debit, nothing on the credit — dissolves into net pricing. Stock loan runs at the 102% convention already on the page. The call mechanics harden in exchange: same-day margin calls against morning wire deadlines and contractual default remedies, traded against margin lock-ups — methodology frozen, 30–90 days’ notice on requirement changes — which turn “changeable without notice” into a term of the documents.
The same 130/30, financed twice
retail: pay 4.07% on the full DR 213,561 | earn 0.00% on CR 221,038
prime: pay SOFR + spread on DR | earn a rebate on the short credit
gross, both directions net pricing: the legs offset
The fund wrapper prices the tail. In a margin account the customer is personally liable for a deficit; interpose an LLC — the fund-of-one — and the prime’s recourse stops at fund assets, so the investor’s worst case becomes NAV. That matters precisely because the short side’s losses are unbounded; the prime prices the non-recourse into its requirements, and occasionally asks for a guarantee. Series mechanics: Delaware’s series LLC (6 Del. C. §18-215) gives internal liability shields, and since August 2019 a registered series (§18-218) is a registered organization under the UCC — the prime can perfect against the specific series rather than the umbrella, which is what made credit committees willing to face series at all. The honest caveat remains: the cross-series shield is essentially untested in a major bankruptcy. Each series is a separate counterparty with its own account and its own margin, so the choice is per relationship: the shield, with no cross-collateralization and lower margin efficiency — or cross-margining, which is contractual piercing, re-linking what the statute separated. The prime’s minimums state that trade in dollars: under the 4210 interpretations, a cross-guarantee is a second route to the $100,000 floor — the guaranteed account may run at $100,000 net equity against a guarantor holding at least $500,000, plus $400,000 more for each additional account it guarantees.
Across primes, no instrument at all. Each prime margins only the book it sees; nothing nets requirements across broker-dealers. A market-neutral book split carelessly between two primes is margined as two directional books — the standing tax on post-Lehman counterparty diversification — and the remedy is allocation, pairs traveling together, not documentation. Within one dealer group the documents can still net the silos — the PB equities book against ISDA swaps and FCM futures under cross-product margin agreements, where the regulators permit — but never across groups.
What survives untouched is §220.2: LMV + CR - SMV - DR is double-entry, not regulation, and closes on a prime statement as exactly as on the two above. What changes is who sets the requirement — and what legal substance stands behind the credit balances when the music stops.
Exercises
- What is the maximum amount of cash you can withdraw from a “145/45” account without going into a Fed call, and what specific risks would you face immediately after doing so?
- What is the maximum withdrawable cash in a “45/45” account, and why are you unable to withdraw the full cash balance despite it appearing as a credit on the ledger?
- Construct the full balance sheet for a “275/275” strategy, including LMV, SMV, CR, DR, and Equity. Is it achievable under Reg T?
- If an account grants 6x Buying Power, what is the maximum effective leverage ratio available to the trader?
- If you hold 50% in a 2x Leveraged ETF as collateral, what is the maximum leverage you can achieve on that position? (Hint: FINRA scales maintenance requirements by the ETF’s leverage factor — Regulatory Notice 09-53.)
- Newly purchased mutual fund shares (excluding money market funds) carry no loan value for 30 days, under the new-issue credit prohibition of Exchange Act §11(d)(1). If you deposit a portfolio that is 50% Cash and 50% newly purchased Mutual Funds, what is your maximum achievable leverage ratio today versus after 30 days?
- From the 130/30 statement above: verify that fully-marginable buying power equals 2 × SMA, derive why buying power for non-marginable securities equals 1 × SMA, and explain why “Available to Withdraw” tracks SMA rather than the $221k short credit.
- Derive NAV = LMV + free credit + excess short credit − DR from the §220.2 definition of equity, and explain why a freshly opened, freshly marked short leaves NAV unchanged. Verify against both statements.
- An account holds ample type-1 cash yet receives a house call. Explain how this is possible, why journaling 1 → 2 cures it while leaving NAV unchanged, and what the account gives up in exchange.
- Account A holds equity of 100 against a maintenance requirement of 60; account B, same customer, holds 50 against 80. What happens to B’s call if the accounts are consolidated under FINRA 4210(f)(6), what does the customer concede to obtain the consolidation, and why does the Fed test not consolidate with it?
- Explain why the 130/30’s interest asymmetry — 4.07% on the full debit, nothing on the credit — largely disappears under prime brokerage, and what protection the fund gives up when its financing moves from the US prime to an offshore affiliate under arranged financing.
- A self-financed 45/45 holds its full cash in a 0.4% custodian sweep; a government money fund yields 3.8% and the account’s margin rate is 11%. Show that sweeping the free credit into the fund while financing the longs on margin is a negative-carry trade, identify the rate that is the cash’s true benchmark rather than the 0.4% sweep, and state the one financing arrangement under which the trade stops bleeding.