| Assumption (M-M 1961) | Real-world friction | What it generates |
|---|---|---|
| No information asymmetry — managers and investors have identical information | Managers know more about future cash flows than the market does | Dividend changes are credible signals (Lintner 1956; Bhattacharya 1979); initiations trade up ~3-5%, cuts trade down ~6-9% |
| No taxes — dividend income and capital gains are taxed identically | Pre-2003: dividends taxed at ordinary income (up to ~39.6%); capital gains at 20%. Post-JGTRRA 2003: qualified dividends at 15-20% | Tax-clientele effect (Miller-Modigliani 1961; Allen + Michaely 2003) — high-bracket investors hold low-yield stocks; tax-exempts hold high-yield stocks |
| No agency costs — managers act perfectly on behalf of shareholders | Managers may retain excess cash for empire-building, prestige acquisitions, or perks | Jensen's (1986) free cash flow theory — dividends as a commitment device that forces managers to disgorge excess cash; reduces the agency cost of free cash flow |
| Frictionless transactions — investors can costlessly buy or sell to synthesize their preferred dividend | Transaction costs, taxes on realized gains, indivisible shares, time + cognitive friction | Homemade-dividend arbitrage is approximate, not exact; some investors genuinely prefer cash dividends for reasons M-M ignored (retirees, mental accounting, income mandates) |
Total wealth = Shares held x Ex-div price + Cash dividend + Net cash from synthetic trades
The defaults above (100 shares, $50 cum-div, $2 dividend) recreate the canonical M-M worked example. Initial wealth: 100 x $50 = $5,000. After ex-dividend: stock drops $2 to $48 (the cash leaves the firm, so per-share value falls by the dividend amount). With preference=1 (no dividend): reinvest the $200 dividend at $48 buys 4.17 shares, ending with 104.17 shares x $48 = $5,000. With preference=2 (as-paid): 100 shares x $48 + $200 = $5,000. With preference=3 (more dividend, $400 target): sell 4.17 shares at $48 = $200, plus the $200 dividend = $400 cash, holding 95.83 shares x $48 = $4,600 + $400 = $5,000. All three preferences end at $5,000 — total wealth is invariant. The synthetic-dividend mechanic is the entire proof. The real-world frictions that break it (transaction costs, taxes on the share sale in preference=3, indivisible shares, tax brackets) are exactly the wedges that signaling, tax-clientele, and agency theories fill in.
Going Deeper — Jensen's free cash flow theory + the agency-cost lens. Michael Jensen (1986) added a third major friction to the M-M dividend baseline: managers and shareholders have CONFLICTING preferences over excess cash. Managers prefer to retain cash (it funds prestige acquisitions, empire-building, perks, and reduces the need to access disciplined capital markets); shareholders prefer to receive it (so they can redeploy capital to higher-return alternatives). Dividends are a COMMITMENT DEVICE — once a firm raises its dividend, the political cost of cutting it later is so severe that management effectively commits to disgorging that cash flow stream forever. This is why mature firms with weak investment opportunities create value by raising payouts even when M-M dividend-irrelevance says they shouldn't — the value creation comes not from the dividend itself but from REDUCING the agency cost of free cash flow. Empirically: Lang + Litzenberger (1989) showed firms with Tobin's Q below 1 (i.e. firms whose assets are worth less than book — strong indicator of negative-NPV investment opportunities) have the largest positive announcement effects on dividend increases — exactly the prediction of Jensen's theory. AI prompt: 'For this ticker, walk through the M-M, signaling, tax, and agency lenses on its most recent dividend announcement. Which friction is doing the work? What does the empirical announcement-effect literature predict for a firm of this profile?'
Sit with the ideas.
MatureCo trades at $40 per share, generates $4 of EPS, and has paid a $1 quarterly dividend (16% payout ratio of TTM EPS, before special dividends) for 30 consecutive years. Management unexpectedly raises the quarterly dividend 25% to $1.25 — a level the firm could have funded for the last decade but never did. Modigliani-Miller dividend-irrelevance (1961) says payout policy does not change firm value in a frictionless market, yet the stock jumps 8% on the announcement. What is the strongest interpretation under the post-MM signaling literature?