5D Library › Shareholder Free-Riding Games, Dynamic Vesting Mechanisms and Defection Early-Warning Models · 中文

Chapter 6. When Clauses Deter

The bad-leaver clause

Vesting deters early departure by putting unvested shares at stake. As shares vest, that stake shrinks, and by the fourth year a founder who leaves to compete forfeits almost nothing. A bad-leaver clause restores the stake: a founder who leaves in breach, for example to compete, must sell even vested shares back at a low price. p11 compares the expected loss from competing with the gain.

Calculation

# Chapter 6. p11: A bad-leaver clause as deterrence. Founder B holds
# 30% of a firm worth 40 units, vesting monthly over 48 months after a
# 12-month cliff. B can leave to compete, gaining G = 6 units. Without
# a bad-leaver clause B keeps the vested shares and forfeits only the
# unvested ones. With the clause, a leaver who competes must also sell
# the vested shares back at 20% of fair value, if the clause is
# enforced (chance 0.7). B is deterred if the expected loss exceeds G.
VALUE, STAKE, G, BUYBACK, ENFORCE = 40.0, 0.30, 6.0, 0.20, 0.7

def vested(month):
    return 0.0 if month < 12 else min(1.0, month / 48)

def loss(month, clause):
    full = VALUE * STAKE
    unvested = full * (1 - vested(month))
    if not clause:
        return unvested
    return unvested + ENFORCE * full * vested(month) * (1 - BUYBACK)

if __name__ == "__main__":
    print("   month   vested   loss, no clause   loss, clause   deterred?")
    for m in (6, 12, 24, 36, 48):
        a, b = loss(m, False), loss(m, True)
        print("   {:5d}   {:6.3f}   {:15.2f}   {:12.2f}   {:>4} / {}"
              .format(m, vested(m), a, b, "yes" if a > G else "no",
                      "yes" if b > G else "no"))
    print("   gain from competing G = {:.1f}".format(G))
#    month   vested   loss, no clause   loss, clause   deterred?
#        6    0.000             12.00          12.00    yes / yes
#       12    0.250              9.00          10.68    yes / yes
#       24    0.500              6.00           9.36     no / yes
#       36    0.750              3.00           8.04     no / yes
#       48    1.000              0.00           6.72     no / yes
#    gain from competing G = 6.0

Without the clause, the loss from leaving to compete falls from 12.00 in month 6 to 9.00 in month 12, 6.00 in month 24, 3.00 in month 36 and 0.00 in month 48; against a gain of 6.0 it deters only in the first two years. With the clause, even if it is enforced only seven times in ten, the loss is 10.68 in month 12, 9.36 in month 24, 8.04 in month 36 and 6.72 in month 48, enough to deter throughout. The clause matters most precisely when vesting has done its work.

Detection and enforcement

A clause deters only if a breach is found in time and a court will enforce it. Broad non-compete covenants are unenforceable in some places, notably California, a fact that has been linked to the mobility of engineers in Silicon Valley (Gilson 1999); confidentiality, assignment of inventions and non-solicitation of clients are more widely enforced. p12 writes deterrence as dfP>Gd f P > G, where dd is the chance of detection, ff the chance of enforcement and PP the penalty.

Calculation

# Chapter 6. p12: Clauses work only if breaches are found and enforced.
# A defecting executive gains G = 6. A breach is detected before the
# damage is done with chance d (for example by the p09 test), and then
# a court enforces the relevant clause with chance f: a broad
# non-compete may be unenforceable where the law does not allow it
# (f near 0), while confidentiality, IP assignment and non-solicitation
# clauses are usually enforceable (f near 0.8). Penalty if enforced:
# P = 10 (damages plus forfeited equity). Deterred if d f P > G; the
# critical detection rate is G / (f P).
G, P = 6.0, 10.0

if __name__ == "__main__":
    print("   enforceable f   d = 0.3   d = 0.6   d = 0.9   critical d")
    for f in (0.0, 0.4, 0.8, 1.0):
        row = ["yes" if d * f * P > G else "no" for d in (0.3, 0.6, 0.9)]
        crit = "none" if f == 0 or G / (f * P) > 1 else \
            "{:.3f}".format(G / (f * P))
        print("   {:13.1f}   {:>7}   {:>7}   {:>7}   {:>10}".format(
            f, *row, crit))
#    enforceable f   d = 0.3   d = 0.6   d = 0.9   critical d
#              0.0        no        no        no         none
#              0.4        no        no        no         none
#              0.8        no        no       yes        0.750
#              1.0        no        no       yes        0.600

With no chance of enforcement, nothing deters at any detection rate. With enforcement four times in ten, even certain detection would not be enough. With enforcement eight times in ten, the breach must be detected at least 0.750 of the time; with certain enforcement, 0.600. A firm cannot buy safety by writing ever broader clauses; it needs clauses a court will uphold and a way of seeing breaches early, which is what Chapter 5 was about.

Hypothesis · not established

The author’s interpretation (not yet tested): firms that combine enforceable confidentiality and non-solicitation clauses with early detection lose fewer clients to departing executives than firms that rely on broad non-compete covenants alone. What would show it wrong: in a pre-registered comparison of firms in jurisdictions that do and do not enforce non-competes, client losses after executive departures do not depend on the combination of clauses and detection.

Summary

A bad-leaver clause keeps a stake at risk after vesting has run its course. But any clause deters only as much as the product of detection, enforcement and penalty; a broad clause that will not be enforced deters no one.

References

Gilson, Ronald J. 1999. “The Legal Infrastructure of High Technology Industrial Districts: Silicon Valley, Route 128, and Covenants Not to Compete.” New York University Law Review 74 (3): 575 to 629.

Fan, Y. (2026). Shareholder Free-Riding Games, Dynamic Vesting Mechanisms and Defection Early-Warning Models: Moral-Hazard Equity Dilution, Shotgun Buy-Sell Clauses and Behavioral Divergence Signals [股东搭便车博弈、动态股权成熟与背刺早期预警模型] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090243 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).