Contribution
It replaced lists of “domestic influences” on foreign policy, and the unitary state of state-centric theory, with one framework in which a negotiator plays a domestic and an international game at once. “It is fruitless to debate whether domestic politics really determine international relations, or the reverse. The answer to that question is clearly ‘Both, sometimes.’ The more interesting questions are ‘When?’ and ‘How?’” (p. 427). Its answer is the win-set: the set of agreements that would be ratified at home, whose size and overlap decide whether a deal is possible and who gains from it.
Key insight
What a government can concede abroad is bounded by what it can get ratified at home, and the bound is a resource as well as a constraint. A negotiator with a narrow win-set can extract more and risks more; a divided government can sometimes agree to more than a united one, because some of its factions want what foreigners demand. Domestic and international politics are therefore not modelled in sequence but together: when foreign pressure reverberates at home or issues are linked across the table, “domestic outcomes are not exogenous, and the two levels cannot be modeled independently” (p. 456).
Assumptions
- A chief negotiator is the only formal link between the levels, at first with no policy views of his own; the assumption is relaxed in the last section (pp. 435–436, 456).
- Ratification is generic — “any decision-process at Level II that is required to endorse or implement a Level I agreement, whether formally or informally” (p. 436) — and need not be democratic; forms of political power are assumed reducible to “some common denominator” for counting votes (p. 437).
- Ratification is up or down: a Level I agreement cannot be amended at Level II without reopening Level I (p. 437).
- No-agreement is the reversion point against which ratification judges a deal (p. 442).
- Constituents are myopic about the other side’s Level II, an assumption Putnam calls “not unrealistic empirically” (p. 453 n. 73).
- Metaphor, not theory. “Metaphors are not theories”; formal solutions for two-level games are left as “a difficult challenge” (p. 435).
Key results
- The Bonn summit of 1978 (pp. 427–430). Germany’s fiscal stimulus, Japan’s growth pledge and Carter’s commitment to decontrol oil prices were possible “only because a powerful minority within each government actually favored on domestic grounds the policy being demanded internationally” (p. 428). International pressure was “a necessary condition”, and without “domestic resonance” it would not have sufficed (p. 430).
- The state as a plural noun (p. 432). State-centric work runs “a unitary-actor model run amok”; on nearly all important issues central decision-makers disagree. “If the term ‘state’ is to be used to mean ‘central decision-makers,’ we should treat it as a plural noun: not ‘the state, it …’ but ‘the state, they …’” Because state structures “vary little from issue to issue or from year to year, such explanations are ill-suited for explaining differences across issues or across time (unless ‘time’ is measured in decades or centuries)”.
- The two tables (p. 434). National leaders sit at both; “any leader who fails to satisfy his fellow players at the domestic table risks being evicted from his seat”.
- Win-sets (pp. 437–440). Larger win-sets make agreement more likely; the relative size of the win-sets shapes the division of gains, so a small domestic win-set is a bargaining advantage (Schelling), up to the point of deadlock (Fig. 1).
- Voluntary and involuntary defection (pp. 438–439). Reneging by an egoist against failing to deliver because ratification failed. The Bonn accord “contained no provisions for temporal balance, sequencing, or partial conditionality”, yet Germany and Japan enacted their parts “more than six months before” Carter’s decision on decontrol, and the promise held “because reneging would have had high political and diplomatic costs” (p. 438).
- Three determinants of win-set size (pp. 442–450). (1) Level II preferences and coalitions: the cost of no-agreement, homogeneous (“boundary”) against heterogeneous (“factional”) conflicts, abstention and politicization, and multi-issue packages, including “synergistic linkage” that creates options beyond domestic control (pp. 446–448). (2) Level II institutions: ratification rules (a two-thirds Senate), party discipline, and autonomy. “The greater the autonomy of central decision-makers from their Level II constituents, the larger their win-set”, as with central-bank insulation, yet “the stronger a state is in terms of autonomy from domestic pressures, the weaker its relative bargaining position internationally” (p. 449). Several levels of ratification, as in the European Community, multiply the analysis. (3) Level I strategies: side-payments targeted at swing constituents, and “generic goodwill” (pp. 450–452).
- Uncertainty (pp. 452–453). Negotiators understate their own win-sets and doubt the other’s; the useful claim is a “kinky” win-set.
- Restructuring and reverberation (pp. 454–456). Foreign pressure can change domestic minds, positively or as backlash.
- The chief negotiator (pp. 456–459). His motives include shifting “the balance of power at Level II in favor of domestic policies that he prefers for exogenous reasons”; he will “normally give primacy to his domestic calculus … not least because his own incumbency often depends on his standing at Level II” (p. 457); he has, in effect, a veto (Wilson and Versailles); and his “fixed investment” in a supporting coalition constrains which deals he will back (p. 458).
- Conclusion (p. 460). Eight features the approach highlights, ending with “the divergences of interest between a national leader and those on whose behalf he is negotiating”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | International agreement is possible only where the parties’ Level II win-sets overlap, and larger win-sets make it likelier | strong as definition; moderate as explanation | the definitions of pp. 437–438; cases (Falklands, Bonn) |
| C2 | A smaller domestic win-set strengthens a negotiator’s bargaining position but raises the risk of deadlock and involuntary defection | moderate | Schelling; Fig. 1; Panama Canal and debt cases |
| C3 | Domestic division can make international agreement easier, through transnational alignments and synergistic linkage | moderate | the 1978 Bonn summit, from Putnam and Henning’s study; the stylized preference example of p. 444 |
| C4 | The autonomy of central decision-makers enlarges their win-set and weakens their bargaining position | moderate as deduction | from C1–C2; central-bank example |
| C5 | The chief negotiator’s own preferences, incumbency and coalition investments narrow the feasible set below the overlap of win-sets | moderate | Versailles, the neutron bomb, Vietnam |
| C6 | Reverberation and synergistic linkage make domestic outcomes endogenous to international bargaining | moderate | Bonn, IMF–Italy; argued on pp. 455–456 |
Concepts
- Level I and Level II — bargaining between negotiators toward a tentative agreement, and each side’s separate process of ratifying it (p. 436).
- Win-set — “the set of all possible Level I agreements that would ‘win’ … when simply voted up or down” at Level II (p. 437).
- Involuntary defection — failure to deliver on a promise because ratification fails, as against reneging (p. 438).
- Synergistic linkage — a Level I deal that creates a policy option previously beyond domestic control, and so alters feasible outcomes at Level II (p. 447).
- Reverberation — international pressure that shifts domestic opinion, positively or negatively (pp. 454–456).
- Kinky win-set — a claim that the proposed deal will be ratified but one slightly better for the other side will not (p. 453).
Connections
Putnam builds on Walton and McKersie’s behavioral theory of labor negotiation, Schelling’s power of a bound negotiator, Allison’s bureaucratic politics (“intra-national games, the overlap of which constitutes international relations”, p. 431), Snyder and Diesing on divided crisis governments, and Shepsle and Weingast on win-sets and the ex post veto. He lists Tsebelis’s “nested games” among related models of linked games (p. 441 n. 43). In the record it is one of the four accounts A28 named beside CLAIM-423, with Pierson’s path dependence (LIT-1372), Allison and Tsebelis.
Bearing on the record
-
CLAIM-423: met, fully, for Putnam’s half of the defeat condition. The claim falls if Putnam already states that components’ time horizons decide which component holds authority. He does not. Read in full, the article locates authority over a deal in three things: the distribution of power, preferences and coalitions at Level II; the ratification institutions, including how insulated decision-makers are; and the negotiators’ strategies, with the chief negotiator’s own veto and coalition investments. Time horizons appear only in passing, and never as what allocates authority among a polity’s components:
- State structures are slow, so they cannot explain variation across issues or years “unless ‘time’ is measured in decades or centuries” (p. 432). That is a point about which variables can explain fast change, not about which component governs which.
- Electoral calendars shape timing: German expansionists pointed to the 1980 elections (p. 428), Carter’s advisers postponed decontrol until after the 1980 election (p. 429), Italy played “we can lose the election” in Trieste (p. 459). The leader’s incumbency makes him put the domestic game first (p. 457).
- Repeated play lowers the temptation to defect among players “who expect to meet again” (p. 438), and offending foreigners “may be costly in the long run” (p. 455).
- The Bonn deal had no “temporal balance, sequencing, or partial conditionality” (p. 438).
The nearest Putnam comes to the claim’s mechanism is the autonomy of central decision-makers, such as an insulated central bank (p. 449). He treats it as an institutional fact that enlarges a win-set, not as following from a horizon. So the claim is not a reading of Putnam. Two of its pieces are already his, and the claim should credit him: the state as a plural of contending decision-makers (“the state, they”, p. 432), and authority in a deal shifting between factions as foreign pressure tips “the balance of power at Level II” (p. 457). Its other half, the state’s continuity as an actor while the organization behind its positions changes, Putnam does not take up: his game is one negotiation, and nothing in it runs across successive coalitions.
-
CLAIM-423’s other open items are untouched. Pierson’s 2000 article is still unread, and no case shows a horizon, rather than a formal rule, deciding authority.
-
CLAIM-331. Putnam’s one remark on slowness (p. 432) fits the premise that slower variables sit at a higher level of organization (state structures against issue politics), and is no argument for it.
-
LIT-1375’s summary (“domestic coalitions set the win-set within which an international deal is possible”) is accurate, though narrow: the win-set also depends on ratification institutions and on the negotiators’ strategies.
-
No instruction for machine-learning practice; nothing for the anthology.
Limitations
- A framework, not a model. No equilibrium is derived, and Putnam says formal solutions are left for future work (p. 435). The propositions hold ceteris paribus and are illustrated, not tested; the principal case, Bonn 1978, is from his own earlier study.
- One negotiation at a time. Level II actors and preferences are taken as given within a game, so how coalitions, careers and institutions change between negotiations is outside it.
- Strategic voting at Level II, and direct ties between Level II players across borders, are named and set aside (p. 453 n. 73, p. 459).
Open questions
- Can two-level games be solved formally, and which propositions survive?
- How do repeated two-level games across successive governments differ from one game, and does the slower side gain from being repeated?
- How should cross-table alliances among Level II players be modelled (p. 459)?