Merger (Risk) Arbitrage

Merger arbitrage captures the small spread between a target's market price and the announced acquisition price, in exchange for bearing the risk that the deal breaks. It is not arbitrage in the riskless sense — it is selling deal-completion insurance. The payoff profile is the defining feature: a high probability of a small, capped gain and a low probability of a large loss. That is a short volatility signature dressed up as event-driven equity, and traders who do not see it that way blow up the same way short-vol traders do. The discipline lives entirely in estimating deal probability and sizing for the break, not in spotting the spread, which anyone can read off a screen.

The basic spread

When an acquirer announces it will buy a target, the target's price jumps toward but stays below the offer price. That residual gap is the deal spread — your gross return if the deal closes. It persists because the deal is not certain and not instant: regulators, shareholders, financing, and material-adverse-change clauses can all kill it, and your capital is tied up until close.

gross_spread = offer_price - target_price
gross_return = gross_spread / target_price
annualized   = gross_return * (252 / expected_days_to_close)

The spread is the market's compensation for two things bundled together: the probability of a break and the time to close. A 2% spread on a deal expected to close in two months is ~12% annualized — attractive if the break probability is low. The entire game is decomposing that spread into its probability and severity components, because the screen only shows you the bundled number.

Cash deals: the binary payoff

In an all-cash deal the arithmetic is clean. You buy the target; if the deal closes you receive cash at the offer; if it breaks the target falls back toward its unaffected price (where it traded before the announcement, adjusted for whatever the market now thinks of standalone value). Model it as a binary outcome:

E[PnL] = p * (offer - entry)  -  (1 - p) * (entry - break_price)

where p is the probability of completion and break_price is the estimated post-break price (almost always well below entry — the announcement pop unwinds). Set E[PnL] = 0 to get the market-implied break probability, which is the single most useful number in the trade:

def implied_completion_prob(entry, offer, break_price):
    # Price-implied probability that completes; below this you're being underpaid
    upside = offer - entry
    downside = entry - break_price
    # p*upside = (1-p)*downside  ->  p = downside / (upside + downside)
    return downside / (upside + downside)

def expected_return(p, entry, offer, break_price, days, financing=0.0):
    win = p * (offer - entry)
    loss = (1 - p) * (entry - break_price)
    carry = -financing * entry * days / 252
    return (win - loss + carry) / entry

Your edge is having a better estimate of p than the market's implied one: legal read on antitrust, financing certainty, board and shareholder dynamics, precedent for the regulator and jurisdiction. The asymmetry is brutal and worth internalizing: a typical winner makes the 2-4% spread; a break can lose 15-30% as the target collapses to or below its unaffected price. You need a high hit rate just to break even, and one mis-assessed deal erases many wins. This is why the same downside-focused risk measures used in tail trading belong on a merger book.

Stock deals and the hedge ratio

In a stock-for-stock deal the acquirer pays in its own shares — say r acquirer shares per target share (the exchange ratio). Now the target's value is tied to the acquirer's price, so an unhedged long-target position is implicitly long the acquirer. The arb is to long the target and short the acquirer at the exchange ratio, isolating the spread from the acquirer's price moves:

spread_per_target = r * acquirer_price - target_price
hedge: short r acquirer shares for each 1 target share long

def stock_deal_pnl(target_entry, acq_entry, ratio, target_close, acq_close, completes):
    long_leg = target_close - target_entry
    short_leg = -ratio * (acq_close - acq_entry)   # short acquirer
    if completes:
        # at close, 1 target -> ratio acquirer shares, hedge unwinds to the spread
        return (ratio * acq_close - target_entry) + short_leg - (target_close - target_entry) + long_leg
    return long_leg + short_leg

Subtleties that decide whether the hedge actually works:

• Fixed vs floating exchange ratio. A fixed ratio means you hedge a constant r

acquirer shares. A floating ratio (the acquirer pays a fixed dollar value, so the share count floats) inverts the hedge and may include collars that cap or floor the ratio — your hedge ratio becomes state-dependent and you may need options to hedge the collar's optionality.

• Borrow on the acquirer. The short leg requires borrowing the acquirer's stock;

if it is hard to borrow, the trade is expensive or impossible, and crowded deals get expensive to borrow exactly when everyone piles in. This is pure short-selling mechanics risk.

• Cash-and-stock and contingent value rights add legs and optionality; treat each

component separately and hedge what is hedgeable.

Deal probability is the whole job

The spread is observable; p is not. A disciplined book estimates p from a mix of:

Factor	Raises break risk	Tooling
Antitrust / regulatory	overlapping markets, hostile regulator	legal review, precedent base rates
Financing	leveraged buyer, tight credit	credit spreads, commitment letters
Shareholder vote	thin premium, activist opposition	ownership, proxy advisory stance
MAC / business shock	cyclical target, macro stress	sector and macro monitoring
Jurisdiction	multi-country approvals	timeline by regulator

Base rates matter more than any single deal's story: across history the large majority of announced deals complete, but completion rates vary enormously by deal type and regime, and they cluster. When financing dries up or regulators turn hostile, breaks arrive together — your "diversified" book of 30 deals is really one bet on the deal-completion regime. Calibrate p against historical base rates for the specific deal archetype, then adjust for the idiosyncratics; do not start from the optimistic press release. This regime-clustering is why merger arb correlates with credit and equity beta in crises despite looking market-neutral — much like statistical arbitrage spreads that look independent until they all break at once.

The short-volatility payoff, made explicit

Plot the P&L of a single merger position against the underlying market or against deal outcome and it is the payoff of a short put: you collect a premium (the spread) and you are fine across most outcomes, but in the bad tail you take a large, convex-down loss. Concretely:

• In calm markets, deals close, spreads grind in, you collect the carry — steady,

bond-like returns. This lulls people into leverage.

• In a risk-off shock, multiple deals break or widen at once, financing evaporates,

and the book takes a cluster of large losses simultaneously. The "uncorrelated alpha" reveals its hidden short-market, short-credit, short-vol exposure.

This means merger arb should be sized like a short-vol strategy: position sizing and risk management must budget for the joint break scenario, not the average deal. Cap single-deal exposure (no deal large enough that its break is a portfolio event), cap sector/regulator concentration (a single antitrust regime shift can hit many deals), and stress the book against a correlated-break scenario rather than relying on the calm-period volatility, which radically understates the risk. Hedging the residual market beta with index puts or a short index position turns the implicit short-vol into something closer to pure deal-risk capture, at the cost of some carry.

Building the book: annualized return and capital efficiency

A single spread is a binary lottery; a merger-arb book is a portfolio of them, and the right way to compare deals is on annualized, probability-weighted return per unit of risk — not on the raw spread. Two deals with the same 3% spread are wildly different if one closes in three weeks and the other in eight months, and different again if one has a 99% completion base rate and the other 90%.

def deal_score(entry, offer, break_price, p, days, borrow_cost=0.0):
    exp_ret = (p*(offer-entry) - (1-p)*(entry-break_price)) / entry
    ann = exp_ret * (252 / max(days, 1))
    # crude risk proxy: probability-weighted loss given break
    downside = (1 - p) * (entry - break_price) / entry
    risk_adj = ann / max(downside, 1e-4)
    return dict(ann_return=ann - borrow_cost, downside=downside, score=risk_adj)

The capital-efficiency subtlety: deals that close fast let you recycle capital into the next deal, so a 1.5% spread closing in a month can beat a 4% spread closing in a year on annualized terms — provided you have a pipeline to redeploy into. This is why deal flow is a strategy input, not a market given: in a thin M&A year the binding constraint is not finding edge but finding enough deals to keep capital working without forcing it into marginal, low-p situations. The temptation to reach for yield by adding deals with thinner spreads and worse break odds is the most common way a book quietly raises its tail risk while the headline carry looks unchanged.

The downside leg and corporate-action traps

The breakprice input deserves more respect than it usually gets. It is not the unaffected price frozen at announcement — by the time a deal breaks, the standalone business and the whole sector may have moved, and a deal often breaks because the target's fundamentals deteriorated, so the post-break price can be well below the old unaffected level. Estimate breakprice from current comparable valuations conditional on the break reason, not from a stale pre-announcement print. The other trap is corporate actions during the deal: special dividends, record dates, and rights issues change the effective offer and the hedge ratio mid-trade, and an exchange ratio that looked fixed can be adjusted by the merger agreement's anti-dilution provisions. Read the actual deal terms; the screen spread assumes a clean structure that the legal document frequently contradicts.

Honest limits and decay

The merger-arb premium is real and economically grounded — it is payment for warehousing deal and liquidity risk that natural holders want off their books — but it is thin and crowded. Spreads have compressed structurally as dedicated capital and multi-strategy funds piled in, so gross returns are lower than the historical record and a larger share of the spread is now pure risk compensation rather than inefficiency. Capacity is bounded by deal flow: in a quiet M&A year there simply are not enough attractive spreads, and forcing capital into marginal deals lowers p for the book as a whole. The edge that persists is underwriting — superior estimation of deal probability and disciplined sizing for the break — not the mechanical act of buying the spread, which is commoditized. Treat every position as a short put, price the tail, diversify across uncorrelated deal risks where you genuinely can, and accept that some years the right size is small.