Who Owns
the State?谁在占有
这个政权？
Brazil spends 47% of GDP through the public budget. Three locked categories — debt interest, pensions, and civil-service wages — consume 30.5 percentage points of that before a single road is built, a single clinic staffed, or a single child taught. This is not a spending problem. It is a capture problem.巴西通过公共预算支出占GDP的47%。三个锁定科目——公债利息、养老金和公务员薪资——在建一条公路、配置一家诊所或教育一个孩子之前，已消耗掉其中30.5个百分点。这不是支出问题。这是政权俘获问题。
A 47% GDP budget that serves less than its size suggests.一个占GDP 47%的预算，其服务效能远不及规模所示。
Brazil operates a welfare state sized like France but structured like a feudal rent-extraction system. The state is large enough to be consequential and captured enough to be redistributive in reverse: money flows in through regressive consumption taxes and flows out through bond interest to financial capital, through inflation-indexed pensions to the organized public sector, and through constitutionally protected wages to the federal bureaucracy. 巴西运营着一个规模堪比法国的福利国家，但其结构却像一套封建租金提取体系。国家庞大到足以举足轻重，又被俘获到足以实现逆向再分配：资金通过累退性消费税流入，通过债券利息流向金融资本，通过与通胀挂钩的养老金流向有组织的公共部门，通过宪法保护的薪资流向联邦官僚体制。
The standard critique of Brazil’s fiscal position focuses on its size: 47% of GDP is high for a middle-income country. The structural critique is more damning: of that 47%, 65% is locked by constitutional mandate or political untouchability before the democratic government can allocate a single real. 对巴西财政状况的标准批评聚焦于其规模：47%的GDP对于中等收入国家而言过高。而结构性批评则更为严苛：在这47%中，65%在民选政府支配任何一分钱之前，已被宪法授权或政治禁区锁定。
Revenue收入端
43% of GDP in taxes. ~50% from consumption (indirect) taxes: maximally regressive. Poorest decile pays 45% of income in taxes; richest decile pays 7%. Dividend income largely untaxed.税收占GDP约43%。约50%来自消费（间接）税：最大程度累退。最贫困十分位纳税占收入45%；最富有十分位仅为7%。股息收入基本免税。
Triple Lock三重锁定
30.5% of GDP locked: interest (6%), pensions (12%), civil wages (12.5%). These three categories are constitutionally mandated, indexed to inflation, and politically untouchable. Every elected government inherits them intact.30.5%的GDP被锁定：利息（6%）、养老金（12%）、公务员薪资（12.5%）。这三个科目由宪法授权，与通胀挂钩，在政治上不可触碰。每届民选政府原封不动地继承它们。
Residual剩余端
16.5% of GDP remains for everything else: health, education, welfare, infrastructure, security, justice. Public investment gets 1.5%: the lowest in the peer group, half the OECD average, one-fifth of China.剩余16.5%的GDP用于其他一切：卫生、教育、福利、基础设施、安全、司法。公共投资仅得1.5%：为同类국家中最低，是OECD均值的一半，是中国的五分之一。
Brazil's federal budget: where every 47 reais of GDP goes巴西联邦预算：每47雷亚尔GDP的去向
Brazil vs. 10 countries: expenditure by category, % of GDP巴西 vs. 10国：各科目支出占GDP比（2022/23年）
General government figures where available. Brazil: central government dominates (states have limited fiscal scope). All figures approximate from IMF, OECD, and World Bank sources. 🔴 = extreme high vs EM peers 🟢 = extreme low.尽可能采用一般政府口径。巴西以中央政府为主（各州财政空间有限）。所有数字来自IMF、OECD和世界银行，为近似值。🔴 = 相对新兴市场同行极端偏高 🟢 = 极端偏低。
| Country国家 | Debt Interest债务利息 | Pensions养老金 | Govt Wages政府薪资 | Social Asst社会救助 | Health (pub.)公共卫生 | Education教育 | Defense国防 | Investment公共投资 | Total总计 |
|---|---|---|---|---|---|---|---|---|---|
| Brazil 🇧🇷 | 6.0% 🔴 | 12.0% 🔴 | 12.5% 🔴 | 1.8% | 4.3% | 5.6% | 1.1% | 1.5% 🟢 | ~47% |
| USA 🇺🇸 | 2.8% | 5.1% | 6.0% | 4.5% | 9.0% | 5.0% | 3.5% | 3.5% | ~38% |
| China 🇨🇳 | 2.0% | 4.8% | 8.0% | 0.8% | 3.0% | 3.6% | 1.6% | 7.5% | ~32% |
| OECD avg | 2.5% | 8.1% | 10.0% | 2.5% | 6.5% | 5.0% | 2.0% | 3.0% | ~41% |
| South Africa 🇿🇦 | 4.6% | 3.0% | 10.4% | 3.8% | 4.2% | 6.2% | 0.7% | 3.5% | ~33% |
| Mexico 🇲🇽 | 2.8% | 2.8% | 6.5% | 1.8% | 2.8% | 4.3% | 0.7% | 3.0% | ~30% |
| Colombia 🇨🇴 | 2.8% | 3.5% | 7.5% | 0.5% | 5.7% | 5.3% | 3.2% | 4.0% | ~29% |
| Chile 🇨🇱 | 1.2% | 3.7% | 7.5% | 0.8% | 5.5% | 4.9% | 1.8% | 4.2% | ~28% |
| India 🇮🇳 | 3.1% | 3.3% | 6.5% | 0.4% | 1.3% | 4.1% | 2.3% | 5.0% | ~28% |
| Peru 🇵🇪 | 0.9% | 3.0% | 6.5% | 0.8% | 3.2% | 4.0% | 1.2% | 4.5% | ~23% |
| Indonesia 🇮🇩 | 1.7% | 1.5% | 5.5% | 0.6% | 1.7% | 1.3% | 0.7% | 3.5% | ~17% |
Sources: IMF Fiscal Monitor (Oct 2023, Apr 2024); IMF Art. IV Brazil 2024; OECD Economic Survey Brazil 2023; OECD Pensions at a Glance 2023; SIPRI 2023; World Bank GHE database; UNESCO education data. Figures are general government where available; federal/central otherwise.数据来源：IMF财政监测报告（2023年10月、2024年4月）；IMF巴西第四条款磋商2024年；OECD巴西经济调查2023年；OECD养老金概览2023年；SIPRI 2023年；世界银行全球卫生支出数据库；UNESCO教育数据。尽可能采用一般政府口径，否则采用联邦/中央政府数据。
Three categories alone consume what most countries spend on everything.仅三个科目的消耗，便等于大多数国家的全部支出。
Debt interest + pensions + government wages = 30.5% of Brazil's GDP. To put this in context: this “triple lock” of mandatory spending exceeds India’s entire government budget (28%), Indonesia’s entire budget (17%), and equals China’s, USA’s, and Colombia’s triple lock combined. The investment bar for Brazil — shown in red — is smaller than every peer’s investment bar. 公债利息+养老金+政府薪资=巴西GDP的30.5%。对比参照：这一"三重锁定"的强制性支出，超过了印度的全部政府预算（28%）、印尼的全部预算（17%），并等于中国、美国和哥伦比亚三重锁定之总和。巴西的投资条形——以红色标注——小于所有同行的投资条形。
6% of GDP extracted, then redistributed upward.6%的GDP被抽取，再向上再分配。
Brazil’s debt interest payments reached 6.6% of GDP in 2023 and 8.0% in 2024 — the highest among all major economies, developed or emerging. South Africa, in fiscal distress, pays 4.6%. Mexico and Chile pay 1.2–2.8%. The mechanism is structural: ~46% of Brazil’s debt stock is Selic-linked (floating rate), meaning every central bank rate decision immediately inflates or deflates the public sector deficit. At Selic = 13.75% (peak 2023), the annual interest bill for Selic-linked debt alone exceeded R$300 billion. 巴西公债利息支出2023年达到GDP的6.6%，2024年达8.0%——在所有主要经济体（无论发达或新兴）中居首。处于财政困境的南非支出4.6%，墨西哥和智利支出1.2–2.8%。机制上，巴西约46%的债务以Selic利率挂钩（浮动利率），意味着每次央行利率决定都会立即扩大或缩小公共部门赤字。当Selic利率达到13.75%（2023年峰值）时，仅Selic挂钩债务的年利息账单便超过3000亿雷亚尔。
Who pays for this? The revenue that funds debt interest comes overwhelmingly from consumption taxes: ~50% of Brazil’s tax revenue is indirect, hitting low-income households hardest. The poorest income decile pays 45% of their income in taxes (mostly indirect). Million-dollar earners pay ~20.6% effective rate — dividend income is largely exempt. 谁在为此买单？资助债务利息的税收收入，主要来自消费税：巴西约50%的税收为间接税，对低收入家庭打击最重。最贫困十分位缴纳收入的45%作为税款（主要是间接税）。百万收入者有效税率约为20.6%——股息收入基本豁免。
THE RENTIER CIRCUIT食利者循环
↓
Source: Tesouro Nacional, Nov 2024. ~52% of debt held by banks + bank-linked investment funds.数据来源：国家财政部，2024年11月。约52%债务由银行及银行关联投资基金持有。
Tax structure: who pays how much, and how regressive is it?税收结构：谁支付多少，其累退性如何？
Money flows in via the most regressive tax structure in the G20. It flows out as 6–8% of GDP in interest to the financial sector. The Selic rate — set by the central bank, ostensibly for inflation control — is simultaneously the mechanism that determines the size of this extraction. Brazil’s poorest decile effectively subsidizes its financial sector. 资金通过G20中最具累退性的税收结构流入。又以GDP的6–8%的利息形式流向金融部门。Selic利率——由央行设定，名义上用于控制通胀——同时也是决定这种抽取规模的机制。巴西最贫困十分位实际上在补贴其金融部门。
Comparative note: In China, ~79% of government bonds are held by state commercial banks — interest payments circulate within the state system. In the USA, the holder base is diversified (30% foreign, 18% Federal Reserve, ~7% private banks); no Selic-equivalent floating-rate instrument exists. Brazil’s LFT structure is unique: floating-rate bonds held predominantly by private domestic financial capital represent a form of institutionalized rent that has no equivalent among peer economies. 比较说明：在中国，约79%的政府债券由国有商业银行持有——利息支付在国有体系内循环。在美国，持有人基础多元化（30%境外、18%美联储、约7%私人银行）；不存在类似Selic的浮动利率工具。巴西的LFT结构独树一帜：主要由私人国内金融资本持有的浮动利率债券，代表着一种制度化租金，在同类经济体中没有对应物。
Four anomalies. Two are visible only with a cross-country lens.四个异常，其中两个唯有横向视角才能看清。
0. Fiscal scale: Brazil's government is as large as Germany's — on one-third the income0. 财政规模：巴西政府规模堪比德国——但人均收入仅为德国的三分之一
Before examining who pays and who benefits, note the sheer scale. Brazil's total government expenditure (~47% of GDP) matches Germany and exceeds the OECD average — yet Brazil's GDP per capita ($18k PPP) is roughly one-third of Germany's ($62k). The four anomalies below are not the result of a small state; they are the result of a large state with deeply concentrated benefits.在考察谁缴税和谁受益之前，先注意纯粹体量。巴西的政府总支出（约47% GDP）与德国相当，超过OECD平均水平——但巴西人均GDP（PPP约18,000美元）约为德国（62,000美元）的三分之一。以下四项异常并非源于"小政府"；而是源于一个福利高度集中的"大政府"。
A. Tax incidence: who actually bears the burden? (effective rate by income decile)A. 税收归宿：谁实际承担税负？（按收入十分位的有效税率）
Effective tax incidence measures what each income group actually pays as a share of their income, accounting for all direct and indirect taxes. Most OECD countries are progressive (rich pay more). Brazil is the only G20 country where the effective rate across income deciles is inverted.有效税收归宿衡量各收入群体实际承担的税收占收入比例，同时计入所有直接税和间接税。大多数OECD国家呈累进性（富人缴纳更多）。巴西是唯一一个G20国家，其各收入十分位的有效税率是倒置的。
Brazil's incidence is uniquely inverted. In every OECD country and most EM peers, the effective rate rises with income. In Brazil it falls. The poorest decile pays ~45% of income in taxes — dominated by ICMS and PIS-COFINS consumption levies. The richest decile pays ~7%: capital gains are largely exempt, dividend income is untaxed, and wealth transfers are avoided via legal structures. South Africa and the USA show the rich paying 3–7× the effective rate of the poor. Brazil's ratio is the inverse: the poor pay 6.4× more. Data: IPEA/Silveira et al. (2022) for Brazil; OECD Taxing Wages (2023); CEPAL.巴西的税收归宿是独一无二的倒置。在所有OECD国家和大多数新兴市场同行中，有效税率随收入上升（累进）。在巴西，它下降。最贫困十分位缴纳约45%的收入作为税款——主要是ICMS和PIS-COFINS消费税。最富裕十分位仅缴纳约7%：资本利得基本豁免，股息收入免税，财富转移通过法律结构得以规避。南非和美国均显示富人有效税率是穷人的3–7倍。巴西的比率恰好相反：穷人缴纳的比率是富人的6.4倍。数据：巴西IPEA/Silveira等（2022）；OECD工资税（2023）；CEPAL。
Brazil's inverted incidence is not a loophole or enforcement failure — it is the designed outcome of three interlocking structural features that have survived every reform attempt since the 1988 constitution.巴西倒置的税收归宿不是漏洞，也不是执法失败——而是自1988年宪法以来历经每次改革尝试而幸存下来的三项相互联锁的结构性特征所刻意产生的结果。
ICMS (state VAT, 25–30%) and PIS/COFINS (federal excise) are invisible in the price of food, electricity, clothing, and fuel — you never see a separate tax line, but roughly a third of what you pay is tax. A poor family spends 100% of income on such goods → effective tax rate ~45%. A wealthy household spending R$50k/month on consumption but earning R$500k/month pays the same absolute amount of consumption tax — which represents only ~5% of their income. The rate looks flat on goods; it is deeply regressive on income.ICMS（州流转税，25–30%）和PIS/COFINS（联邦增值税）以隐性方式嵌入食品、电力、服装和燃油的价格中——你看不到单独的税款行，但大约三分之一的支出是税。贫困家庭将100%的收入用于此类商品→有效税率约45%。一个月消费5万雷亚尔但月收入50万雷亚尔的富裕家庭，消费税缴纳的绝对金额相同——但仅占其收入约5%。商品层面税率看似统一；收入层面则是强烈的累退性。
Brazil is one of ~30 countries where shareholder dividends face zero personal income tax — and the only major economy with this structure. A company owner takes R$20k as salary (27.5% rate) and R$180k as dividends (0%). Combined effective rate: ~3.5%. This is the statutory design, not evasion. "Pejotização" amplifies it: high-income professionals incorporate as PJ entities (companies), pay lower corporate rates, then extract tax-free dividends — legally bypassing the income-tax system entirely. Recife lawyers, Rio de Janeiro surgeons, São Paulo consultants: none face the IRPF as employees do.巴西是全球约30个股东股息无需缴纳个人所得税的国家之一，也是唯一采用这种结构的主要经济体。一个公司老板月薪2万雷亚尔（27.5%税率），股息18万（0%税率）。综合有效税率：约3.5%。这是法定设计，而非逃税。"PJ化"放大了这一效应：高收入专业人士将自己注册为PJ实体（公司），缴纳较低企业税率，再以免税股息形式提取——从法律上完全绕开个人所得税体系。累西腓律师、里约外科医生、圣保罗顾问：没有人像普通雇员那样面对IRPF。
IPTU (property tax) assessed values lag market prices by decades in most municipalities. ITCMD (inheritance & gift tax) is capped at 8% and managed at the state level — vs. 40–45% in France or the UK. The 1988 constitution explicitly mandated an IGF (wealth tax) in Article 153 — but Congress has never passed enabling legislation in 36 years. The result: wealth accumulated through asset appreciation and inheritance is structurally untaxed. All effective tax burden falls on consumption at the bottom. The rentier class collects; the salaried worker and the poor pay.全国大多数市的IPTU（物业税）评估价值落后市场价格数十年。ITCMD（遗产及赠与税）上限为8%，由各州管理——而法国或英国高达40–45%。1988年宪法第153条明确授权征收IGF（大额财富税）——但国会在36年内从未通过相应立法。结果：通过资产增值和遗产积累的财富在结构上未被征税，全部有效税负最终落在底层的消费上。食利阶级收取；薪资工人和穷人缴付。
B. Government bond holder structures: who collects Brazil's 6% GDP in interest?B. 政府债券持有结构：巴西6%GDP的利息收益流向何处？
The same government debt generates radically different political-economy dynamics depending on who holds it. When the state holds its own bonds (China, ~79%), interest circulates within the public sector. When private financial capital holds them (Brazil, ~52%), each interest payment is a fiscal transfer to a concentrated domestic rentier class — financed by the most regressive tax base in the G20.同样的政府债务，持有人不同，产生的政治经济学动态截然不同。当国家持有自身债券（中国，约79%），利息在公共部门内循环。当私人金融资本持有（巴西，约52%），每次利息支付都是向集中化国内食利阶层的财政转移——由G20中最具累退性的税基资助。
Bond anatomy: three instrument types and who holds each债券市场解剖：三类工具及各自持有人结构
Brazil's government debt uses three distinct instruments, each with different interest-rate mechanics and a different investor base. The distinction between them is central to understanding who actually benefits from monetary tightening.巴西政府债务采用三类截然不同的工具，各自具有不同的利率机制和不同的投资者基础。理解这些区别，是理解货币紧缩实际惠及谁的核心。
LFT — Selic-linked floating rate (~46% of debt)LFT — Selic浮动利率（约占债务46%）
Coupon = Selic overnight policy rate, reset daily. When BCB raises Selic, all existing LFT holders automatically earn more on bonds they already hold — no new issuance, no renegotiation. This is the mechanism: a rate decision by the central bank is simultaneously a fiscal decision. Unique to Brazil; no equivalent in China, USA, or EM peers.票息=Selic隔夜政策利率，每日重置。央行加息时，所有现有LFT持有人自动在已持有的债券上取得更高收益——无需新发行，无需重新谈判。这就是机制：央行的利率决定同时是财政决定。巴西独有；中国、美国及新兴市场同行无对等工具。
NTN-B — IPCA inflation-linked (~27% of debt)NTN-B — IPCA通胀挂钩（约占债务27%）
Pays a fixed real rate (e.g. 6%) plus IPCA inflation. Not sensitive to Selic changes directly — existing bonds do not re-price when BCB acts. Returns rise when inflation rises. Preferred by pension funds for liability matching: long-duration bonds (10–40 yr) that track the same inflation index as their benefit obligations.支付固定实际利率（如6%）加IPCA通胀。直接对Selic变化不敏感——央行行动时现有债券不会重新定价。通胀上升时收益增加。养老基金偏好，用于负债匹配：长久期债券（10–40年）跟踪与养老金支付义务相同的通胀指数。
LTN / NTN-F — Fixed nominal rate (~27% of debt)LTN/NTN-F — 固定名义利率（约占债务27%）
Fixed coupon regardless of Selic. Existing holders do not benefit from rate hikes — only new issuance prices higher. This is the instrument foreign investors use for carry trade: borrow in USD at 5%, buy BRL bonds at 13%, capture the 8-pt spread. When Selic rises further, they benefit only on rollover, not on existing positions.固定票息，与Selic无关。现有持有人不受加息直接惠益——仅新发行以更高利率定价。这是境外投资者进行利差交易的工具：以5%借入美元，以13%购入巴西雷亚尔债券，赚取8个百分点利差。Selic进一步上升时，仅在展期时受益，对现有持仓无利。
Within each instrument: who holds it? (% of that instrument's outstanding)各类工具内部：谁持有？（占该工具未偿额比例）
Brazil's 46% LFT stock is the structural centrepiece. When BCB raises Selic, R$7.3 trillion in LFT bonds re-price overnight. 67% of that stock (≈R$4.9T) is held by domestic private banks and investment funds. Foreigners own just 5% of LFT (≈R$370B). The 2021–23 tightening cycle — Selic from 2% to 13.75% in 18 months — added an estimated R$200–250 billion per year in additional interest transfers to this group relative to the near-zero-rate baseline. This transfer is funded by Brazil's tax system, which is the most regressive in the G20. 巴西46%的LFT存量是结构性核心。央行加息时，7.3万亿雷亚尔的LFT债券隔夜重新定价。其中67%（约4.9万亿雷亚尔）由国内私人银行和投资基金持有。境外投资者仅持有LFT的5%（约3700亿雷亚尔）。2021–23年紧缩周期——Selic在18个月内从2%升至13.75%——每年估计为这一群体增加了约2000–2500亿雷亚尔的额外利息转移，相较于接近零利率的基准水平。这一转移由巴西的税收体系资助，而该体系是G20中最具累退性的。
46% LFT (floating, re-prices daily) + 27% NTN-B (inflation-linked, pension-held) + 27% fixed LTN (foreigners 22% via carry). Where foreign investors concentrate (fixed LTN), existing holdings do not benefit from rate hikes. Where rate hikes benefit holders automatically (LFT), foreigners hold only 5%. The political economy is entirely domestic: BCB → domestic private financial capital → funded by regressive consumption taxes from the poor.46%LFT（浮动，每日重新定价）+27%NTN-B（通胀挂钩，养老基金持有）+27%固定LTN（境外22%通过利差交易）。境外投资者集中的地方（固定LTN），现有持仓对加息不自动受益。加息自动惠及持有人的地方（LFT），境外仅持有5%。政治经济学完全是国内的：央行→国内私人金融资本→由穷人的累退消费税资助。
79% state commercial banks. No LFT equivalent: bonds are fixed-rate or policy-rate-linked, but not Selic-style daily repricing. When PBoC raises rates, it raises the cost of state banks, not transfers to private rentiers. Interest income flows to state banks → state dividends → budget. Rate decisions cost the state; they do not enrich private domestic financial capital.79%国有商业银行。无LFT对等工具：债券为固定利率或政策利率挂钩，但不是Selic式每日重新定价。人民银行加息时，提高国有银行的成本，而非向私人食利者转移。利息收入流向国有银行→国家股息→预算。利率决定使国家承担成本；不会富裕国内私人金融资本。
No LFT equivalent. Treasuries are fixed-rate (notes/bonds) or inflation-linked (TIPS). Rate hikes raise the cost of new issuance, not existing portfolios. Private banks hold 7%: tiny. 30% foreign + 17% Federal Reserve dominate. When Fed raises rates, the primary transmission is through mortgage markets, bank lending, and exchange rates — not automatic daily repricing of existing government bonds held by domestic private capital.无LFT对等工具。国债为固定利率（票据/债券）或通胀挂钩（TIPS）。加息提高新发行成本，而非现有投资组合成本。私人银行持有7%：极小。30%境外+17%美联储占主导。美联储加息时，主要传导渠道是抵押贷款市场、银行贷款和汇率——而非对国内私人资本持有的现有政府债券进行自动每日重新定价。
A 12% GDP pension bill for a country with a median age of 35.一个中位年龄35岁的国家，却背负12%GDP的养老金账单。
France spends 14% of GDP on pensions — with a median age of 42. Italy spends 16% — with a median age of 46. Brazil spends 12% — with a median age of 35. There is no peer country in the world that spends at this proportion relative to its demographic position. The demographic excuse — that Brazil is aging — only partially explains the current bill; the structural excess — civil servant replacement rates of 90%+, military pensions with 15% contribution coverage, untouchable since 1988 — explains the rest. 法国养老金支出占GDP 14%——中位年龄42岁。意大利占16%——中位年龄46岁。巴西占12%——中位年龄35岁。世界上没有任何同类国家在与其人口结构位置相匹配的比例上有如此高的支出。人口老龄化的借口——巴西正在老龄化——只能部分解释当前账单；结构性超额——公务员替代率90%以上、军事养老金仅15%缴费覆盖率、自1988年以来不可触碰——解释了其余部分。
Pension spending vs. demographic position: Brazil as a global outlier养老金支出 vs. 人口结构：巴西作为全球异常值
MILITARY PENSIONS军事养老金
Contribution coverage (only 15% of military pension expenses are covered by contributions). Taxpayers fund the other 85%. The per-beneficiary deficit is R$159,000/year vs. R$9,400 for INSS workers. The 2019 pension reform explicitly excluded the military.缴费覆盖率（仅15%的军事养老金支出由缴费覆盖），纳税人承担其余85%。人均赤字为R$159,000/年，而INSS工人仅为R$9,400。2019年养老金改革明确将军队豁免。
CIVIL SERVANT PENSIONS (RPPS)公务员养老金（RPPS）
Replacement rate: federal civil servants receive ~90% of final salary as pension. OECD average is ~60%. The per-beneficiary deficit is R$69,000/year for a group covering only ~5% of total retirees.替代率：联邦公务员领取约90%的最终工资作为养老金。OECD均值约为60%。该群体人均赤字为R$69,000/年，仅覆盖约5%的退休人员群体。
INSS (PRIVATE SECTOR)INSS（私人部门）
Contribution coverage for INSS — the system that covers 35M+ retirees and is relatively well structured. Per-beneficiary deficit: R$9,400/year. This is the only regime subjected to the 2019 reform. The others were exempted.INSS缴费覆盖率——该体系覆盖3500万以上退休人员，结构相对合理。人均赤字：R$9,400/年。这是唯一接受2019年改革的制度，其他均获豁免。
1.5% of GDP: half of OECD, one-fifth of China, one-third of its Latin neighbors.1.5%的GDP：OECD的一半，中国的五分之一，拉美邻国的三分之一。
The triple lock leaves 16.5% of GDP for everything else. After health (4.3%), education (5.6%), social assistance (1.8%), and defense (1.1%), public investment receives 1.5% of GDP — the lowest in this peer group. India is pushing 5%+ as part of a deliberate infrastructure push. China has maintained 7–8% for decades, building the capital stock that sustains its manufacturing competitiveness. Brazil’s public capital stock stood at 35% of GDP in 2015, versus 87–92% for Latin American and EM peers. 三重锁定之后，仅剩16.5%的GDP用于其他一切。在卫生（4.3%）、教育（5.6%）、社会救助（1.8%）和国防（1.1%）之后，公共投资获得1.5%的GDP——为同类组别中最低。印度作为蓄意推进基础设施的政策，正在突破5%。中国数十年来维持7–8%，建立起支撑其制造业竞争力的资本存量。2015年巴西公共资本存量仅为GDP的35%，而拉美和新兴市场同行为87–92%。
This is not an accident. It is the arithmetic residual of a constitution that mandated pension and wage growth faster than GDP, while providing no equivalent constitutional mandate for investment. Every fiscal adjustment since 1988 has been accomplished by compressing capital expenditure, because mandatory spending cannot be compressed. 这不是偶然。这是宪法运算的残差——该宪法规定养老金和薪资增长快于GDP，却没有对投资提供对等的宪法保障。1988年以来的每次财政调整，都是通过压缩资本支出来实现的，因为强制性支出无法被压缩。
Three organized groups. One budget. One answer.三个有组织的集团。一个预算。一个答案。
The fiscal data answers the political economy question directly. Brazil’s government budget has been captured, in three distinct layers, by three organized interest groups who converted their political strength at the transition moment (late 1980s) into constitutionally protected fiscal entitlements. 财政数据直接回答了政治经济学问题。巴西政府预算已被三个有组织的利益集团以三个不同层次俘获，他们将在过渡时刻（1980年代末）的政治力量转化为宪法保护的财政权利。
BLOC 1: FINANCIAL CAPITAL集团一：金融资本
Banks and investment funds hold ~52% of Brazilian government bonds. The Selic rate — the mechanism of their enrichment — sits at the heart of Brazil’s fiscal system. Financed by the most regressive tax structure in the G20: the poor pay 45% of income to fund these transfers to financial capital. 银行和投资基金持有约52%的巴西政府债券。Selic利率——其致富机制——位于巴西财政体系的核心。由G20中最具累退性的税收结构资助：穷人缴纳收入的45%，以资助这种向金融资本的转移。
BLOC 2: ORGANISED PUBLIC SECTOR集团二：有组织的公共部门
Military (R$59B/yr pension, 85% taxpayer-funded), federal civil servants (90% replacement rate, R$69K/yr deficit per beneficiary), and general INSS (the only reformed system). The 1988 Constitution locked all three. 2019 reform touched only the third. 军方（每年养老金R$590亿，85%由纳税人资助）、联邦公务员（90%替代率，每位受益人赤字R$69,000/年），以及普通INSS（唯一接受改革的体系）。1988年宪法锁定全部三者。2019年改革只触及第三个。
BLOC 3: FEDERAL BUREAUCRACY集团三：联邦官僚机器
Federal civilian payroll at 12.5% of GDP versus 6–7% for Mexico, Colombia, Peru. Above the OECD average. High wages generate high pension contributions which generate the RPPS deficit. Self-reinforcing: wages set the pension base; the pension base mandates constitutional protection; constitutional protection blocks wage reform. 联邦文职薪资占GDP 12.5%，而墨西哥、哥伦比亚、秘鲁为6–7%，高于OECD均值。高薪资产生高养老金基数，进而产生RPPS赤字。自我强化：薪资设定养老金基数；养老金基数要求宪法保护；宪法保护阻止薪资改革。
The question “who owns the Brazilian state?” has a fiscal answer. Three organized blocs — financial capital, the organized public sector, and the federal bureaucracy — have collectively constitutionalized 30.5% of GDP as their pre-committed entitlement. The remaining 16.5% is formally democratic: subject to legislative appropriation, adjustment, and reform. In practice, it is also under pressure — from the same blocs, through Congress, through debt market signals, through the BCB. "谁占有巴西政权？"这个问题有一个财政答案。三个有组织的集团——金融资本、有组织的公共部门和联邦官僚机器——已集体将GDP的30.5%宪法化为其预承诺权利。剩余的16.5%在形式上是民主的：受立法拨款、调整和改革的约束。但在实践中，它同样处于压力之下——来自同样的集团，通过国会、通过债务市场信号、通过央行。
The democratic government can choose which hospital to build and which school to fund. It cannot choose whether to pay the interest, whether to honour the pension formula, or whether to reduce civil service compensation. Those choices were made in 1988 and have not been substantially revisited since. The constitution does not just organize the state — it distributes the state to its creditors. 民选政府可以选择建哪所医院、资助哪所学校。但它无法选择是否支付利息、是否兑现养老金公式，或是否降低公务员薪酬。这些选择在1988年已经做出，此后从未得到实质性重新审视。宪法不只是组织国家——它将国家分配给了它的债权人。