第15章将货币政策视为泰勒规则——一个从通货膨胀和产出缺口到利率的反馈函数。本章将深入探讨。人们为什么持有货币？什么决定了货币的最优数量？为什么中央银行持续制造过多的通货膨胀（时间不一致性）？财政政策如何通过政府预算约束与货币政策相互作用？
本章的高潮是价格水平的财政理论（FTPL）——在某些条件下，财政政策而非货币政策决定价格水平的激进主张。
This chapter connects to four of the book's Big Questions. It is the densest chapter for debates — monetary-fiscal interaction touches central banking, money's nature, inequality, and the role of government all at once.
CIA约束假设代理人必须持有货币才能购买消费品：
货币之所以有价值，是因为交易需要它。当名义利率 $i > 0$ 时，持有货币有机会成本（放弃的利息），创造了扭曲消费决策的楔子。
另一种方法：货币直接进入效用函数，捕捉它提供的流动性服务：
一阶条件将实际余额的边际效用等于持有货币的机会成本：
其中 $m = M/P$ 是实际余额，$i$ 是名义利率。
生产货币的边际成本基本上为零。效率要求每种商品的价格等于其边际成本。持有货币的"价格"——机会成本——是名义利率 $i$。由于货币的边际成本为零，有效价格为 $i = 0$。
由于费雪方程给出 $i = r + \pi$，而实际利率 $r$ 由基本面决定，弗里德曼规则意味着：
最优通胀率是实际利率的负值——中央银行应以时间偏好率通缩，使名义利率为零，消除持有货币的扭曲。
中央银行最小化损失函数：
其中 $y^*$ 是自然产出，$k > 0$ 反映中央银行将产出推至自然水平以上的愿望，$a$ 是通胀的权重。预期增强的菲利普斯曲线将产出和通胀联系起来：
在承诺下：中央银行宣布 $\pi = 0$ 并坚持执行。损失为 $k^2$。
在相机抉择下：在理性预期均衡中（$\pi = \pi^e$），通胀偏差出现：
相机抉择下的损失为 $L_{disc} = k^2(1 + b^2/a)$ ——严格劣于承诺。通胀偏差只有成本没有收益：两种体制下产出都保持在 $y^*$，但相机抉择增加了无谓的通胀。
相机抉择下的通胀偏差为 $\pi^* = bk/a$。调整中央银行的偏好和菲利普斯曲线斜率，观察偏差和损失如何变化。
图 16.1.承诺与相机抉择下的损失。差距是中央银行无法承诺的代价。更保守的银行家（更高的 $a$）缩小了通胀偏差。拖动滑块探索。
时间不一致性的解决方案：（1）中央银行独立性（Rogoff, 1985）：任命一位具有更高 $a$ 的"保守的央行行长"。（2）通胀目标制：明确的数值承诺。（3）声誉：在重复博弈中，长期信誉成本超过短期收益。（4）绩效合同（Walsh, 1995）：未达标的惩罚。
考虑效用函数 $u(c, m) = \ln c + \gamma\ln m$，具有预算约束和费雪方程 $i = r + \pi$。
第1步：实际余额的一阶条件：$\gamma/m = i \cdot (1/c)$，所以 $m/c = \gamma/i$。
第2步：货币的边际效用：$u_m = \gamma/m$。消费的边际效用：$u_c = 1/c$。最优性：$u_m/u_c = \gamma c/m = i$。
第3步：生产货币的社会成本为零。效率要求 $u_m/u_c = $ 边际成本 $= 0$。因此 $i^* = 0$。
第4步：由费雪方程：$1 = r + \pi^*$，所以 $\pi^* = -r$。当 $r = 4\%$ 时：最优通胀率为 $-4\%$/年（通缩）。中央银行应以时间偏好率缩减货币供应量。
参数：菲利普斯曲线斜率 $b = 0.5$，产出野心 $k = 0.02$，通胀权重 $a = 1.0$。
第1步：相机抉择下的通胀偏差：$\pi^* = bk/a = 0.5 \times 0.02 / 1.0 = 0.01$（每年1%）。
第2步：承诺下的损失（$\pi = 0$）：$L_c = k^2 = 0.0004$。
第3步：相机抉择下的损失：$L_d = k^2(1 + b^2/a) = 0.0004(1 + 0.25) = 0.0005$。
第4步：相机抉择的代价：$L_d - L_c = 0.0001$。社会承受了1%的无谓通胀，却没有任何产出收益。
第5步：如果"保守的银行家"有 $a = 4$：$\pi^* = 0.5 \times 0.02/4 = 0.0025$（0.25%）。偏差缩小了75%，证明了中央银行独立性的合理性。
政府的流量预算约束：
以实际值表示的跨期政府预算约束（IGBC）：
其中 $R_t = \prod_{j=0}^{t-1}(1+r_j)$ 是累积贴现因子，$s_t = T_t - G_t$ 是基本盈余。实际政府债务等于未来基本盈余的现值。
该定理需要强假设。关键失效条件：（1）有限期界/世代交叠：当代获益，后代买单。（2）流动性约束：信贷受限家庭花费意外减税。（3）扭曲性税收：所得税时间安排改变相对激励。（4）关于未来财政政策的不确定性。（5）行为偏差：现时偏好的代理人过度消费意外收入。
经验上，约20-40%的美国家庭似乎受流动性约束（Zeldes, 1989）。退税使支出增加约退税金额的20-40%——与完全李嘉图等价不一致。
多大比例的家庭受流动性约束？在0%时，完全李嘉图等价成立，减税对消费没有影响。在100%时，全部减税被消费（纯凯恩斯式）。现实介于两者之间。
图 16.2.消费对1000亿美元减税的响应，作为受约束家庭比例的函数。在0%受约束时，代理人完全内化未来税收并储蓄全部减税（李嘉图等价）。在100%时，全部减税被消费。经验估计（灰色带）表明20-40%的家庭受约束。拖动滑块探索。
政府减税1000亿美元，通过发行债券融资。假设 $r = 3\%$，税收将在明年增加1030亿美元。
在李嘉图等价下：家庭今天获得1000亿美元，但知道明年欠1030亿美元（现值=1000亿美元）。他们储蓄全部1000亿美元。消费不变：$\Delta C = 0$。债券市场吸收1000亿美元新债务，利率不变。
40%流动性约束家庭的情况：无约束家庭（60%）储蓄全部减税。有约束家庭（40%）全部消费。$\Delta C = 0.4 \times 1000亿 = 400亿$。财政乘数为0.4，而非零。
经验证据：Johnson、Parker和Souleles（2006）发现，美国家庭在第一个季度内花费了2001年退税金额的20-40%，这与李嘉图等价的部分失效一致。
由公式16.9，跨期政府预算约束必须始终成立。在李嘉图体制中，财政政策调整盈余以在中央银行确定的任何价格水平下满足IGBC。在非李嘉图体制中，盈余独立设定，价格水平调整：
如果政府增加债务（$B_0$）而不调整未来盈余，价格水平 $P_0$ 必须上升。通胀是财政现象，而非货币现象。
| 货币政策 | 财政政策 | 结果 |
|---|---|---|
| 主动（$\phi_\pi > 1$） | 被动（调整盈余） | 标准NK：货币政策决定 $\pi$ |
| 被动（$\phi_\pi < 1$） | 主动（固定盈余） | FTPL：财政政策决定 $P$ |
| 主动 | 主动 | 无均衡（过度确定） |
| 被动 | 被动 | 不确定（欠确定） |
在非李嘉图体制中，$P = B / PV(\text{盈余})$。观察价格水平如何响应名义债务或预期财政盈余的变化。
图 16.3.FTPL价格决定。价格水平调整以使实际政府债务等于盈余现值。在不增加盈余的情况下增加债务会导致通胀。在不减少债务的情况下降低预期盈余也会导致通胀。拖动滑块探索财政主导。
政府有名义债务 $B_0 = 100$ 并宣布了新的财政计划。
情景A（可信盈余）：每年基本盈余为5，永续，$r = 5\%$。$PV(s) = 5/0.05 = 100$。价格水平：$P_0 = 100/100 = 1.00$。无通胀。
情景B（较低盈余）：盈余降至每年4。$PV(s) = 4/0.05 = 80$。价格水平：$P_0 = 100/80 = 1.25$。通胀：25%。
情景C（战争或危机）：政府将债务翻倍至 $B_0 = 200$，盈余不变（$PV = 100$）。$P_0 = 200/100 = 2.00$。通胀：100%。
关键洞见：在FTPL下，通胀由政府负债与盈余现值之间的缺口决定——与货币供应增长无关。中央银行的通胀目标被财政主导所覆盖。
Time inconsistency reveals a fundamental weakness in discretionary monetary policy. The FTPL goes further: if fiscal policy is active, the central bank may not even determine the price level. Two direct challenges to central bank control.
Two challenges to central bank control emerge in this chapter. First, time inconsistency (Kydland-Prescott, Barro-Gordon): even a well-intentioned central bank has an incentive to inflate — announce low inflation, then surprise-inflate to boost output. Rational agents anticipate this, producing an inflation bias $\pi^* = bk/a$ with zero output gain. The solution: rules over discretion, central bank independence, inflation targeting contracts. Second, the FTPL: the price level satisfies $P = B/PV(\text{surpluses})$. If the fiscal authority sets surpluses independently, the price level is determined by fiscal policy, not monetary policy. In a fiscal-dominant regime, the central bank is along for the ride — it can adjust the nominal interest rate, but the price level moves to satisfy the government's intertemporal budget constraint regardless.
Against central bank independence: democratic accountability argues against unelected officials making decisions that redistribute wealth — inflation is a tax, and who pays that tax depends on monetary policy choices. The ECB's austerity-enforcing role during the Eurozone crisis is a cautionary tale of an independent central bank imposing enormous costs on peripheral countries. Against the FTPL: the theory requires a specific fiscal-monetary game structure. If the central bank credibly threatens to refuse monetization, the fiscal authority must adjust surpluses. Whether FTPL describes any actual economy is debated — Japan has run enormous deficits for decades without the fiscal dominance the FTPL would predict, suggesting institutional credibility can override the mechanical relationship.
Central bank independence is the mainstream consensus — but the 2020s tested it severely. Governments borrowed massively (COVID), central banks monetized debt (QE), and inflation arrived. The post-hoc debate is whether this was fiscal dominance (FTPL in action) or supply-side shocks that central banks eventually controlled. The Leeper taxonomy (active/passive monetary and fiscal) provides a framework for classifying regimes, but determining which regime a country is actually in requires judgment, not just data.
Central banks can control the economy — but only within institutional constraints. Their power depends on: (a) independence from fiscal pressure, (b) not being at the ZLB (Chapter 15), (c) understanding the transmission mechanism, and (d) the fiscal authority not undermining them through unsustainable deficits. All four conditions have been challenged in recent history. "Can central banks control the economy?" is best answered: "usually, approximately, under favorable conditions." That is not a dismissal of central banking — it is an honest assessment of a powerful but bounded institution.
How should monetary and fiscal policy coordinate? The strict separation (independent central bank, fiscal rules) may be too rigid for crises. The question of coordination connects to BQ01, which also reaches this chapter. And the international dimension adds another layer: for most countries, central bank power is further constrained by the exchange rate regime. Come back in Chapter 17 for the open-economy dimension, where the impossible trinity shows that exchange rate commitments limit monetary independence further.
The Barro-Gordon inflation bias and the FTPL both challenge the Fed's power. Independence helps, but fiscal dominance could override it.
高级MMT says the central bank is a servant of fiscal policy. The FTPL says the same thing, in equations. The mainstream says it depends on the regime.
高级
现代货币理论在五年内从学术角落变成畅销书再到国会的谈话要点。它的核心主张——赤字不像我们所想的那样重要——要么是一代人中最重要的洞见，要么是最危险的。
高级铸币税——印钞的收入——是对货币持有者的通胀税。实际铸币税为：
其中 $\mu$ 是货币增长率，$m(\mu)$ 是实际货币需求（随 $\mu$ 递减）。在低通胀时，更高的 $\mu$ 增加收入。但在高通胀时，税基（$m$）侵蚀速度快于税率上升——形成铸币税拉弗曲线。
实际货币需求随通胀呈指数下降：$m(\mu) = m_0 \cdot e^{-\alpha \mu}$。铸币税收入 $S = \mu \cdot m(\mu)$ 呈倒U形。通胀推得太高会摧毁税基。
图 16.4.铸币税拉弗曲线。收入先随通胀上升，然后随实际货币基础被摧毁而下降。恶性通胀经济体（津巴布韦、委内瑞拉）运行在曲线的右侧——高通胀、低收入。拖动滑块探索。
You have now seen three formal models of why money has value (CIA, MIU, FTPL), plus the seigniorage Laffer curve that shows what happens when governments abuse money creation. The theoretical stakes are higher here than in Chapter 8.
Three approaches to modeling money, each with different implications. Cash-in-advance (CIA): you must have cash to buy goods. Money is a transaction technology — a physical constraint. The Friedman rule follows: deflate at the rate of time preference to make holding money costless. Money-in-utility (MIU): money enters the utility function directly — a reduced-form approach that skips the question of why money is useful and just assumes it is. FTPL: money's value depends on the government's fiscal backing. $P = B / PV(\text{surpluses})$. If the government credibly promises future surpluses, money has value. If it does not, the price level adjusts and money loses value. CIA and MIU tell you money is useful; the FTPL tells you what makes it valuable.
The credit theory of money (Graeber, Mehrling): money is not a commodity that evolved from barter — the textbook origin story is historically false (Graeber 2011). Money is a credit instrument — all money is debt. Bank deposits (most "money" in circulation) are bank IOUs. Central bank money is a government IOU. This matters because if money is credit, then the money supply is endogenous — banks create money by lending — not exogenous as CIA and MIU assume. MMT's version: money is a creature of the state. Taxes create demand for government currency. The government spends first, creating money, and taxes drain money to control inflation. This reverses the textbook causation entirely. The metallist view: historically, money that is not backed by a commodity eventually loses value. Bitcoin is an attempt to create a digital commodity — scarce, decentralized, not subject to government manipulation.
The mainstream uses CIA/MIU for tractability while acknowledging that these models do not resolve the "what is money?" question. The credit theory and MMT have gained influence post-2008 as the banking system's role in money creation became impossible to ignore — the Bank of England's 2014 paper "Money Creation in the Modern Economy" (McLeay et al.) was a turning point. The FTPL provides a formal framework where fiscal and monetary policy jointly determine the price level, offering a bridge between mainstream and heterodox thinking about money's nature.
Money is a social convention — it has value because people expect others to accept it. The different theories capture different aspects of this convention: CIA/MIU capture the transaction role, FTPL captures the fiscal backing, credit theory captures the banking mechanism, and chartalism captures the state's role in establishing the convention. No single theory is complete. The correct answer to "what is money?" is: it is a self-reinforcing equilibrium of mutual acceptance, maintained by institutions — the state, the banking system, the central bank. When those institutions fail, money fails, as every hyperinflation demonstrates.
Does money need a state? Bitcoin and other cryptocurrencies test this proposition — they attempt to sustain a monetary convention without government backing. The seigniorage analysis shows that money creation is a fiscal resource; if money is decentralized, who captures that resource? Come back in Chapter 17 where exchange rates, the dollar's reserve currency status, and digital currencies complicate the picture further.
Bitcoin satisfies some money functions but fails others. The CIA model says money needs transaction convenience; Bitcoin's volatility undermines that.
高级MMT says money is a creature of the state. The FTPL says its value depends on fiscal surpluses. They agree on more than you might expect.
高级政府应如何构建税收以最小化扭曲？拉姆齐规则（1927）：在商品中，对需求缺乏弹性的商品征更高的税（逆弹性规则）：
对缺乏弹性的商品征税造成的行为扭曲更少（更少的无谓损失，回忆第3章）。拉姆齐规则在给定收入要求下最小化总无谓损失。
两种具有不同需求弹性的商品。逆弹性规则要求对缺乏弹性的商品征更高的税。比较拉姆齐最优税率与统一税——相同收入，更少的无谓损失。
图 16.5.拉姆齐最优税率与统一税收的比较。拉姆齐规则将更高的税率分配给更缺乏弹性的商品，在筹集相同收入的同时减少总无谓损失。弹性差距越大，效率收益越大。拖动滑块改变弹性。
赛斯和扎克曼提议对超过\$5,000万的财富征收2%的年度税。沃伦将其作为竞选核心。经济学说这是可行的。政治说这是雷区。历史说欧洲已经试过了，大多放弃了。
高级You now have the Ramsey optimal taxation framework and the Atkinson-Stiglitz result. These are the profession's sharpest tools for thinking about the efficiency cost of redistribution.
Ramsey (1927): minimize total deadweight loss subject to raising a given revenue. The result: $\tau_i/\tau_j = \varepsilon_j/\varepsilon_i$ — tax inelastic goods more heavily. This is efficient but regressive, because necessities (food, housing) tend to be inelastic. Atkinson-Stiglitz (1976): if utility is separable between consumption and leisure, the optimal income tax alone is sufficient — no commodity taxes needed. Mirrlees (1971): the optimal marginal tax rate at the top depends on the Pareto tail of the income distribution and the elasticity of taxable income. Recent estimates (Diamond & Saez, 2011) suggest optimal top marginal rates of 50-70%.
Against high top rates: the elasticity of taxable income may be large when you account for tax planning, avoidance, and migration. The behavioral response to high rates may grow over time as people find new avoidance strategies. The supply-side view: tax cuts can increase revenue if we are on the wrong side of the Laffer curve — though empirical evidence suggests this is unlikely at current U.S. rates. Piketty's deeper challenge: the optimal tax framework takes the pre-tax distribution as given and asks how to redistribute optimally. But if $r > g$ drives wealth concentration, the pre-tax distribution itself is policy-dependent. The problem is not just redistributing a given pie — it is that the pie is being cut by market rules (inheritance, capital gains treatment, rent-seeking) that are themselves political choices.
Post-Piketty, the profession has paid more attention to wealth taxation, inheritance, and the political economy of pre-distribution — shaping market incomes through competition policy, labor market regulation, and education investment rather than redistributing after the fact. The HANK (Heterogeneous Agent New Keynesian) models integrate inequality directly into macroeconomic analysis, showing that the distribution of income and wealth affects aggregate demand, monetary policy transmission, and fiscal multipliers.
The efficiency-equity tradeoff is real but smaller than many assume. Moderate redistribution through progressive income taxation has modest efficiency costs. Very high marginal rates (above 70-80%) likely have larger costs, but current rates in most countries are well below the revenue-maximizing level. The bigger question may be what to redistribute (income vs. wealth vs. opportunity) rather than how much. Economics provides precise tools for the "how" of redistribution — Ramsey, Mirrlees, Atkinson-Stiglitz — but the "how much" is ultimately a normative question that economics can inform but not answer.
Global inequality dwarfs within-country inequality. The tools for addressing it — foreign aid, trade, migration — are completely different from domestic tax policy. Come back in Chapter 20 (Development Economics) where the inequality question scales to the planet. Within-country inequality is a problem optimal taxation can partially solve; between-country inequality requires fundamentally different approaches.
Ramsey says tax inelastic bases. Wealth is elastic. The European evidence confirms it. But does that settle the question?
高级Dan Riffle popularized the slogan in 2019. The claim: any billionaire proves the system is rigged. Optimal taxation theory asks a different question — what tax structure maximizes social welfare given behavioral responses?
中级正常时期（$\phi_\pi > 1$）：财政乘数 $\approx 0.5$–\$1.0$。政府支出增加总需求，但中央银行提高利率，挤出投资。
零利率下限（$i = 0$）：财政乘数 $> 1$，可能为 \$1.5$–\$1.0$。中央银行无法提高利率，因此没有挤出效应。财政政策在最需要时更加有效（Christiano, Eichenbaum & Rebelo, 2011; Woodford, 2011）。
两种商品的弹性分别为 $|\varepsilon_1| = 0.5$（缺乏弹性，如食品）和 $|\varepsilon_2| = 2.0$（富有弹性，如电子产品）。收入目标：$R = 400$。
第1步：逆弹性规则：$\tau_1/\tau_2 = \varepsilon_2/\varepsilon_1 = 2.0/0.5 = 4$。缺乏弹性的商品应被征收4倍的税率。
第2步：收入约束：$\tau_1 Q_1 P_1 + \tau_2 Q_2 P_2 = 400$。以基期 $Q_0 = 100$、$P_0 = 10$ 和需求 $Q_i \approx Q_0(1 - \varepsilon_i\tau_i)$ 为例：
令 $\tau_1 = 4\tau_2$：数值求解得 $\tau_2 \approx 8.3\%$，$\tau_1 \approx 33.2\%$。
第3步：无谓损失比较。拉姆齐：$DWL = 0.5 \times 0.5 \times 0.332^2 \times 1000 + 0.5 \times 2.0 \times 0.083^2 \times 1000 = 27.6 + 6.9 = 34.5$。
统一税率（$\tau_1 = \tau_2 = 0.20$）：$DWL = 0.5 \times 0.5 \times 0.04 \times 1000 + 0.5 \times 2.0 \times 0.04 \times 1000 = 10 + 40 = 50$。
结果：拉姆齐方法相比统一税收减少了31%的无谓损失。效率收益来自将税收负担集中在反应较低的商品上。
津巴布韦恶性通胀与日本失去的几十年：货币-财政互动的两个极端。
津巴布韦（2007-2008）：2008年11月峰值通胀率达到约每月796亿%。政府通过印钞为巨额财政赤字（土地改革、军事支出）融资。随着通胀加速，实际货币基础崩溃——经济体滑向铸币税拉弗曲线的错误一侧。津巴布韦元变得一文不值；交易转向美元和南非兰特。这是财政主导的教科书案例：中央银行从属于财政需要，FTPL方程 $P = B/PV(s)$ 在 $PV(s) \to 0$ 时得到体现。
日本（1990年代至今）：相反的极端。政府债务超过GDP的250%，但通胀几十年来保持在接近零或负值。日本银行在1999年将利率降至零，并实施了大规模量化宽松。财政和货币扩张都未产生通胀。可能的解释：（1）预期日本财政盈余最终会调整（尽管债务高企仍为李嘉图体制）。（2）通缩均衡是自我实现的——代理人预期零通胀，这在零利率下限处自我验证。（3）人口下降使自然利率永久低于零。
教训：津巴布韦和日本框定了货币-财政体制的光谱。津巴布韦展示了当财政政策主导且盈余崩溃时会发生什么。日本表明，如果保持财政信誉，即使巨额债务也不一定产生通胀——但也表明摆脱通缩均衡极其困难。
凯拉尼政府的债务为GDP的85%。中央银行遵循 $\phi_\pi = 1.5$ 的泰勒规则（主动货币政策），政府宣布了15年间每年GDP 2%的基本盈余。
如果政府兑现：李嘉图体制。如果盈余不足：$P_0 = B_0 / PV(盈余)$。如果盈余从85亿KD降至60亿KD的现值，价格必须上涨 \$1.5/6 = 42\%$ ——财政主导覆盖了通胀目标。
凯拉尼约40%的家庭受流动性约束，因此减税对总需求有正面（但部分）影响——李嘉图等价对他们不成立。
You now have the full toolkit: the intertemporal government budget constraint, Ricardian equivalence and its failures, the FTPL, and the state-dependent multiplier. This is the final stop.
The government's intertemporal budget constraint ties debt, taxes, and spending together: real debt equals the present value of future surpluses. Ricardian equivalence says that if this constraint holds and consumers are rational, the timing of taxes does not matter — only the present value of spending matters. A tax cut financed by borrowing is fully offset by increased private saving. The FTPL goes further: if fiscal surpluses are set independently of the price level, then the price level must adjust to satisfy the government budget constraint. Fiscal policy determines inflation, not monetary policy. The empirical multiplier evidence is state-dependent: approximately 0.5-1.0 in normal times (when the central bank offsets fiscal expansion by raising rates), but 1.5-2.0+ at the zero lower bound (when monetary policy cannot offset, so crowding out disappears). Government spending helps the economy most precisely when the economy needs it most — in deep recessions at the ZLB.
MMT's challenge cuts deeper than it first appears. A sovereign currency issuer does not face a budget constraint in the same way a household does — the government can always create money to pay debts. The real constraint is inflation, not solvency. This is not as radical as it sounds: the FTPL actually shares more with MMT than either side typically acknowledges. Both agree fiscal policy affects the price level. Both reject the crude "bond-financed vs. money-financed" distinction. Where they diverge is on whether the inflation constraint is manageable in real time (MMT says yes, through taxation and job guarantees) or is an equilibrium outcome governments must respect (FTPL says inflation adjusts whether the government wants it to or not). The empirical multiplier literature is still contested — identification is hard because government spending is endogenous to economic conditions. And the frontier is moving toward heterogeneous-agent models (HANK) where the distributional effects of fiscal policy matter as much as the aggregate effects.
The FTPL (Leeper, Sims, Cochrane) formalized the intuition that fiscal policy matters for inflation. The mainstream now recognizes two regimes: Ricardian (monetary dominant, where fiscal policy adjusts surpluses to stabilize debt) and non-Ricardian (fiscal dominant, where the price level adjusts). The question "does government spending help?" became inseparable from "what regime are we in?" Post-2020, the profession saw both frameworks tested simultaneously — governments borrowed massively for COVID relief, central banks monetized debt through QE, and inflation arrived. The post-hoc debate is whether this was fiscal dominance (FTPL in action) or supply-side shocks that central banks eventually controlled. The answer is probably both, in different proportions across countries.
Government spending can help the economy — the evidence supports positive multipliers, especially in recessions and at the ZLB. But the effect is genuinely state-dependent: the multiplier is not a fixed number but a function of monetary policy stance, the fraction of liquidity-constrained households, the fiscal regime, and the state of the business cycle. MMT is right that sovereign currency issuers do not face a solvency constraint, but wrong to dismiss the inflation constraint as something that can be managed through ad hoc policy tools. The mainstream has a more nuanced view than either MMT advocates or fiscal hawks typically present. The honest answer to "does government spending help?" is: yes, under the right conditions (recession, ZLB, credible future fiscal adjustment), with diminishing and eventually negative returns as those conditions weaken. This is not a cop-out — it is the actual result of rigorous analysis, refined across four stops from the Keynesian cross to DSGE to the FTPL.
The empirical multiplier literature continues to evolve. HANK models suggest that the distribution of fiscal transfers matters as much as their size — sending checks to liquidity-constrained households has a larger multiplier than across-the-board tax cuts. The post-COVID inflation episode will be studied for decades, and its lessons for monetary-fiscal coordination are still being drawn. The deepest unresolved question is institutional: how should monetary and fiscal policy coordinate? The strict separation (independent central bank, fiscal rules) may be too rigid for crises, but the alternative (coordinated monetary-fiscal expansion) risks fiscal dominance and inflation. The BQ01 path is complete — but the question itself will outlast any model we build to answer it.
The FTPL and MMT agree on more than either admits. The question is whether the inflation constraint is a policy choice or an equilibrium outcome.
高级With the full multiplier framework, revisit the 2009 debate: the ZLB says the multiplier was large, but how large?
中级| 标签 | 公式 | 描述 |
|---|---|---|
| 公式 16.1 | $P_tc_t \leq M_t$ | CIA约束 |
| 公式 16.4 | $\pi^* = -r$ | 弗里德曼规则 |
| 公式 16.7 | $\pi^* = bk/a$ | 相机抉择下的通胀偏差 |
| 公式 16.9 | $B_0/P_0 = \sum R_t^{-1}s_t$ | 跨期政府预算约束 |
| 公式 16.10 | $P_0 = B_0 / \sum R_t^{-1}s_t$ | FTPL价格决定 |
| 公式 16.11 | $\tau_i/\tau_j = \varepsilon_j/\varepsilon_i$ | 拉姆齐逆弹性规则 |
