Introduction
The candidate promised to abolish the central bank. He now leads a government running the most actively managed monetary regime in the hemisphere. Monthly inflation ran at 25.5% in December 2023 and printed 2.1% in May 2026, and the parallel-market premium that averaged above 100% across 2022 and 2023 has averaged 1.4% under the regime in force since January 2026. Real and large, the disinflation arrived all the same. What abolition was meant to install never came, and four exchange-rate regimes arrived in its place.
Argentina · Deep Dive · 2022–2026
Monthly Inflation and the Exchange-Rate Rule
Headline and core m/m % against the announced crawl or band-edge rate
That gap between promise and delivery resolves into one finding. A monetary revolution's result was delivered by the opposite means. Milei campaigned to abolish monetary discretion; the program kept the central bank, ran it hard through four successive peso regimes, lifted capital controls only halfway, and never dollarized. Disinflation came anyway. What carried it was a fiscal surplus guaranteed by no statute and ratified at the ballot box, financed in part by borrowed reserves, with the zero-deficit rule still a benchmark rather than a law. That achievement is genuine. The institution that would lock it in was not built.
A residue sits underneath the headline. It shows in nine consecutive consumer-price prints at or above 2% per month, and in real transactional balances still below where they stood at the start of the program. In force behind those prints, the band indexes its ceiling to the inflation it was built to contain, and in May it permitted a monthly depreciation of 3.4%. Disinflation that stalls at a 2% floor while the nominal anchor moves with lagged prices is a different object from disinflation that converges, and the difference is the regime's open problem.
This piece grades the program against its own promises. Three questions order the grade: whether confidence in the peso returned, what anchored it, and whether the arrangement now in force can keep holding it. Full empirical treatment, methodology, and references sit in the working paper; what follows is the verdict and the evidence that carries it.
The promises
Milei campaigned on three monetary commitments. Two were institutional, the dollarization of the economy and the closure of the Banco Central de la República Argentina, and the third was fiscal, a deficit of zero. Each was stated without qualification. Days after the November 2023 runoff the office of the president-elect reaffirmed that the closure of the central bank "no es un asunto negociable," not a negotiable matter. During the campaign Milei had called the institution itself "la peor basura que existe en esta Tierra," the worst filth that exists on this Earth. His fiscal pledge carried into the inaugural address in three words, "no hay plata," there is no money.
An economic argument sat beneath the rhetoric. Argentina removed thirteen zeros from its currency across four reforms between 1969 and 1992, has defaulted on sovereign debt nine times since independence, and entered 2023 with annual inflation above 200% and a central bank financing the Treasury directly. In that setting the promise amounted to credibility imported wholesale. Dollarization is the limiting case of anchor outsourcing: the country stops asking markets to trust its monetary institutions and adopts a unit those institutions cannot print. A serious academic case, developed by Ocampo and Cachanosky before the election, rests on that premise, that after seven decades of failed domestic anchors the cheapest remaining source of monetary credibility was somebody else's.
Abolishing the central bank was the institutional complement, because a dollarized economy that keeps its central bank leaves the machinery of re-pesification on the shelf for the next government. That zero-deficit commitment completed the design, since every prior Argentine reform had eventually been undone by a fiscal authority returning for financing.
Those promises form one list. Arrangements that arrived form another: a crawling peg run by the central bank, a band negotiated with the IMF, a band indexed to the inflation it was meant to contain, and a program under which the institution slated for closure buys up to USD 10bn of reserves in a single year. Every entry on the second list became policy between December 2023 and January 2026. Those promises set the standard of evaluation rather than the benchmark of success, and the question worth asking is what the substitutions did to the anchor.
What was delivered
Two days after the inauguration, the monetary program began. On 12 December 2023 the economy ministry under Luis Caputo moved the official rate from 366 to 800 pesos per dollar, a step of 118%, committed the central bank to a crawling peg of 2% per month, and pledged that the bank would extend no new financing to the Treasury. Both commitments held. That crawl ran unbroken through January 2025, and the zero-financing pledge shows in the monetary decomposition below.
Through 2024 the central bank dismantled the quasi-fiscal complex it had inherited. At the change of government the stock of remunerated liabilities, LELIQ and pases, stood at about two and a half times the monetary base, and the interest paid on them was itself a running source of money creation. Unwinding ran in sequence: BOPREAL bonds absorbed importer arrears, the exchange-rate puts sold to banks were repurchased, and in July 2024 the remunerated liabilities migrated into LEFI, liquidity letters issued by the Treasury in place of the central bank. That LEFI stock matured and the instrument was retired in July 2025, the migration completing in its extinction. Quasi-fiscal cost of the central bank moved from 1.6% of GDP in 2024 to −0.5% in 2025. A machine that had created money to pay the interest on its own debt was switched off by decision, with no statute to compel it.
That exchange-rate regime then revised twice more. From 1 February 2025 the crawl slowed to 1% per month. On 11 April 2025 the IMF approved a 48-month Extended Fund Facility of about USD 20bn, and the same day the central bank replaced the crawl with a band of 1,000 to 1,400 pesos per dollar, eliminated the exporter blend rate, and lifted the cepo for individuals. Reserve-accumulation targets were missed, recorded as such in the Fund's own review.
A stress test arrived in the austral spring of 2025. On 7 September the government lost the Buenos Aires provincial election by fourteen points, and sovereign risk repriced within days. On 9 October the United States Treasury Secretary announced, in a post on X in place of a press release, a USD 20bn swap framework financed through the Exchange Stabilization Fund, with USD 2.5bn drawn by month-end. By 26 October the government had won the midterm elections decisively. Inside seven weeks the program absorbed a provincial repudiation and a midterm ratification, with the most direct United States intervention in a Latin American currency market since the 1995 Mexican rescue falling between them.
On 1 January 2026 the current regime entered force. Its defining feature is indexation: the band's maximal monthly depreciation at month T equals headline CPI inflation at month T−2. Around it sits a remonetization program of reserve purchases, USD 10bn as a 2026 baseline, executed daily at no more than 5% of market volume, with a stated success criterion of a base-money corridor rising from 4.2% to 4.8% of GDP by December 2026.
| Promise | State | Delivered at writing |
|---|---|---|
| Dollarization | Reversed | Four successive peso regimes in twenty-six months |
| Abolition of the central bank | Reversed | The BCRA operating actively, buying up to USD 10bn of reserves in 2026 |
| Zero fiscal deficit | Delivered, not institutionalized | Financial surplus every month since January 2024; the statutory rule remains a pending IMF benchmark |
| Quasi-fiscal overhang | Delivered by discretion | Remunerated-liability complex eliminated by July 2025; cost moved from 1.6% to −0.5% of GDP |
| Capital controls | Partial | Lifted for individuals in April 2025; corporate restrictions remain |
That record is fixed. What it achieved is a question for the data, and the analysis freezes that data at the May 2026 consumer-price print. That freeze is itself a fact about the regime: the band indexes to a CPI whose methodology is under political dispute, still carrying expenditure weights from the 2004–05 household survey, its scheduled rebasing shelved after the resignation of INDEC's director in February 2026. A regime indexed to that number has to be read against a stated vintage of it.
Confidence returned
Among confidence statistics in the program, the parallel-market premium is the most legible. A currency whose blue-market price exceeded its official price by more than 100% on average across 2022 and 2023 now trades within about 1% of it, and across the four phases the gap means run 100.9%, 27.6%, 2.1%, and 1.4%.
Argentina · Deep Dive · 2022–2026
The Parallel Exchange-Rate Gap
Daily premium of parallel dollars over the official wholesale rate, %
That premium is the price of evading the currency, and its collapse from triple digits to near parity is the market's verdict that holding pesos no longer carries a confiscation discount. One qualification attaches: corporate capital controls were still partly in force at the vintage, so some of the narrowing is the tightening enforcement of a shrinking arbitrage window rather than a fall in the demand for dollars, and the published series cannot separate the two.
Argentina · Deep Dive · 2022–2026
Real Money Balances
Deflated by headline CPI, December 2023 = 100 · the confidence test in one picture
Money balances tell a more textured story. Indexed to 100 at December 2023, the real monetary base fell to a trough of 78.7 in March 2024 and climbed to 183.2 in August 2025, standing at 137.7 in May 2026. Part of that is mechanical, since the July 2024 LEFI migration shifted sterilized liquidity into the base and the raw index overstates the behavioral recovery. Real transactional M2 remains below its starting level, at 95.3 in May. Narrow base recovered and the broad aggregate did not, and the asymmetry is the hinge of the reading.
Obstfeld regresses the log real monetary base on log expected inflation and reports a semi-elasticity near −0.12. Replication on his window lands at −0.125 and extension to the locked vintage at −0.113, while the same specification on transactional M2 returns an insignificant +0.050. Tested jointly, a stacked specification puts the difference in semi-elasticities at +0.163 with a Newey–West standard error of 0.064 (p = 0.013), so the elasticity is a property of the narrow base and not of the broad aggregate. What collapses in a confidence crisis and returns in a stabilization is the demand to hold the monetary unit itself, cash balances proper, while broad aggregates track payments activity and bank intermediation. Real narrow balances recovering ahead of real broad balances is confidence in the unit returning ahead of confidence in the financial system that lends it on. Private dollar deposits stand near USD 39bn in May 2026, and the deposit data carry the same reading from the other side, since a public that rebuilds peso cash while keeping dollar savings is holding two monies by choice.
Dating settles the recovery to the start of the program. A Bai–Perron procedure runs on two regressands, and the one free of the LEFI migration decides it: on real transactional M2 the single break dates to December 2023 and holds there across every trim and segment fraction tested. Discriminating evidence comes from the inflation semi-elasticity re-estimated with a dummy absorbing the December 2023 to March 2024 devaluation window: the slope shifts by +0.12 with a Newey–West standard error of 0.05 (p = 0.02), moving from −0.31 before the break to −0.16 after. Honest about its limits, the evidence shows the series integrated of order one and the levels money-demand relation failing a cointegration test, so the levels elasticity is a conditional correlation, and under a moving-block bootstrap the slope shift carries a two-sided p of 0.089. Demand for real balances returned with the regime change, before the band arrived, and whether the confidence is in the peso or in the fiscal regime standing behind it an elasticity cannot say. The decomposition can.
The anchor was fiscal and political
At the front end, the regimes and the fiscal turn moved together, so no estimate can separate the crawling peg from the budget in December 2023. A narrower claim defended here survives the collinearity: the durable anchor, the one still holding after the exchange-rate regime had been changed four times, is fiscal and political.
Argentina · Deep Dive · 2022–2026
Base-Money Creation by Source
Monthly factores de explicación, ARS bn · the anchor decomposition behind Test 2
Holding to an exact accounting identity, the decomposition shows where the money came from with no residual to estimate. Treasury financing goes to zero after December 2023. Sterilization and the interest complex wind down through the 2024 cleanup, and by 2025 creation comes from foreign-exchange purchases offset by sterilization. Seigniorage quantifies the heavy lifting. Measured as the monthly change in the base over annual-rate GDP and summed by year, it runs 1.13% of GDP in 2022, 1.73% in 2023, 2.98% in 2024, 1.61% in 2025, and 0.05% through May 2026. That 2024 peak reads as remonetization seigniorage, a one-off stock adjustment harvested while real balances were rebuilding during disinflation, and the 2026 program cannot repeat it. Surviving the denominator and the numerator alike, the gap holds: excluding the LEFI migration the 2024 figure is 2.41% of GDP, against 0.05% in 2026. That fiscal anchor did its work as the public rebuilt cash balances, and that work is largely done.
October 2025 offers a within-regime experiment, and that is what makes the sequence useful. With the IMF band held fixed across the window, any repricing inside it cannot be attributed to a change in the exchange-rate rule. The provincial loss in Buenos Aires on 7 September widened the EMBI by 454 basis points, and the midterm result on 26 October compressed it by 497, both surviving a placebo distribution built on every non-event trading day. Currency demand does not move at all, since every parallel-gap effect is indistinguishable from noise. A modest reading is appropriate by design: corrected for the family of tests, the election responses sit at suggestive rather than conventional significance, which matches the secondary role assigned to them. Markets priced the political outcome and left the exchange-rate rule unmarked.
Two facts the placebo distribution does not touch carry the case. Demand for real balances broke inside the December 2023 stabilization package, on the uncontaminated series. And the monetary contribution to the disinflation was a one-off seigniorage stock adjustment the 2026 program cannot repeat. What survived October 2025 was a fiscal surplus delivered each month by a government and ratified at an election, guaranteed by no statute. That anchor's fate when the politics turn is the durability question.
Durability is borrowed
The disinflation is delivered. Underneath it sits a borrowed balance sheet. Durability of the program is a financing question before it is a monetary one, and it turns on reserves the central bank does not net own and a maturity wall it has to roll.
Argentina · Deep Dive · 2022–2026
Gross vs Net International Reserves
USD bn, end-year · gross reserves recover while NIR stays deeply negative — the buffer is borrowed and encumbered (swap lines, FX-deposit reserve requirements, Fund credit)
Gross international reserves ended 2025 at USD 27.6bn, and the Fund projects USD 36.1bn for end-2026. At end-2025 the net international reserve position was −USD 11.8bn. From the gross figure the Fund's definition nets out the activated swap lines, the foreign-currency reserve requirements on dollar deposits, and frozen Fund credit, arriving below zero, against an encumbrance near USD 39bn. Whatever buffer defends the band is in large part claims that other parties hold against it. How durable that encumbrance is even when a headline reports its removal shows in the United States swap. A Treasury line drawn in October 2025 was repaid in December by drawing an equal BIS line that still sits inside the reserve wedge, so the borrowing changed counterparty without leaving the balance sheet.
Argentina · Deep Dive · 2022–2026
Can the Wall Be Rolled?
Required market rollover rate on external amortizations, by NIR-accumulation scenario · above 100% means private rollover alone cannot cover the maturity wall
External amortizations falling due run USD 8.6bn in 2026, near USD 11.9bn in 2027, and a peak USD 14.6bn in 2028. Fund credit itself peaks at 1,329% of Argentina's quota in 2026, the largest exposure in the institution's history. Whether the wall can be rolled is a question the maturity table alone does not answer. Netting the financing need against the reserve accumulation the program targets yields the rollover rate the program implicitly requires. On the full reserve path the wall is rollable, the required rate rising to 82% by 2028. Halve the accumulation, and the 2027 and 2028 walls demand rollover above 100%, more than the entire stock of maturing debt. A stochastic version sharpens the conditionality. Drawing the reserve path around the program target with export-price variability, the probability that the wall requires rollover above the private ceiling is near zero in 2026, about 22% in 2027, and about 47% in 2028. In the deterministic central case, the wall is rollable on the program's own reserve path, that path is the one the missed targets show is hard to hold, and the tail in which it cannot be rolled is material by 2028.
Real appreciation is the durability risk the exchange-rate rule itself generates. Falling from a phase mean near 92 under the IMF band to 87.7 under the indexed band and 84.1 in May, the ITCRM real multilateral index leaves the peso dear and presses it onto the weak side of its own ceiling, where defending it costs reserves the net position does not have. On the program's own terms, the corridor closes the section. Base money over GDP stands at 4.40% in May 2026, inside the BCRA's 4.2% to 4.8% corridor for December 2026. That accounting is self-certified, since the corridor is the bank's own forecast of money demand, produced by the institution doing the emitting, so the absorption is a claim the emitter validates against its own projection.
The institution that was not built
Disinflation is real and its durable anchor is fiscal and political. That monetary framework borrowed its credibility from the anchor and did not manufacture its own. What that means turns on a single property of the arrangement now in force.
A monetary or exchange-rate rule provides a nominal anchor when two conditions hold together. The first is exogeneity of the reference series: the quantity the rule fixes one period ahead is measurable against information that excludes the authority's own realized prices, so the authority cannot move the anchor by moving its own outcomes. The second is costly reversal: abandoning the rule requires an act whose cost exceeds the value of reneging. A fixed or foreign anchor pins the price level for any pass-through and independent of the fiscal regime. This indexed band fails the first condition outright. Its ceiling follows $$e_t = e_{t-1} + \pi_{t-2}$$, the authority's own lagged inflation, so the price level inherits a stochastic trend and the exchange-rate rule supplies no nominal anchor at any admissible pass-through. At no admissible pass-through does the indexed band constitute a commitment whose breach is costly. It is no commitment.
This makes the condition a complement to the fiscal theory of the price level, not a rival to it. When the band supplies no monetary anchor, determinacy requires the fiscal block to pin the level through the government's budget constraint. That is the paper's thesis stated as theory: the exchange-rate rule anchors nothing, so the anchor that remains is the fiscal one. That condition also predicts a procyclical loosening the inertial-inflation reduced form does not. Real permitted depreciation under the band equals $$\pi_{t-2} - \pi_t$$, positive whenever inflation is falling, so the band widens precisely during disinflation. Data carry it: the permitted monthly depreciation widened from 2.5% in January 2026 to 3.4% in May while headline inflation fell from 2.9% to 2.1%.
That condition leaves a signature in the comparative record. Brazil's Real Plan erased indexation memory before the currency switch, through a transitional unit of account later frozen, so indexation served as the bridge to its own removal. Argentina wrote indexation into the exchange-rate rule instead. Post-stabilization inflation persistence, the summed autoregressive coefficient of monthly inflation, is 0.26 for de-indexed Israel against 0.60 for Argentina under the indexed band, the higher persistence the condition predicts when the anchor is fed its own lagged prices.
Argentina · Deep Dive · 2022–2026
Stabilization Paths Compared
Monthly inflation, aligned at t = 0 (program start) · Brazil broke indexation; Argentina 2026 institutionalized it
A second reading runs through the central bank's own balance sheet. For years the bank created base money to service the interest on its remunerated liabilities, so the stock of money grew by the arithmetic of its own debt rather than by any decision about the stance of policy. A monetary aggregate produced this way cannot be read from its published total alone, because its growth carries the deliberate emission of a policy choice and the mechanical compounding of a balance sheet at once. Mises's calculation argument names the defect: an authority unable to distinguish its own action from the automatic consequences of its past commitments has lost the ability to reason from the aggregate it controls. That illegibility was contingent and was in fact removed, by discretion, when the sterilization complex wound down and the LEFI stock matured, and base-money growth once again reports what the central bank decides. Nothing in statute prevents a later central bank from rebuilding the stock and reintroducing the defect.
A third reading comes from the deficit-bias argument of Virginia public choice. Public spending concentrates its benefits and disperses its costs, so the political return to spending exceeds the political return to restraint at nearly every margin, and the standing result is a bias toward deficit in any government that finances itself by discretion. Argentina has the surplus without the rule. That bias did not vanish when the surplus arrived; it was overridden, and an override is not a constraint.
On the diagnosis these readings converge with the mainstream; on the prescription they divide. Mainstream completion of the program is an independent central bank with an inflation target and a float once reserves allow. Removing the discretion the calculation reading identifies points instead to dollarization, the credibility-import limb Ocampo and Cachanosky advance, on the narrow ground that it most directly satisfies both components of the condition. An obvious objection is 2001, when Argentina tried maximal irreversibility and it broke. Yet the objection cuts the other way once the two components are separated. Convertibility was a currency board held by statute, exogenous to current prices but repealable, so reversal was cheap once overvaluation and fiscal strain made breaking it attractive, and the 2002 pesification exercised exactly that option. Dollarization differs in the second component: removing the domestic unit raises the cost of reversal from a statute to a re-monetization. What 2001 teaches is that a reversible anchor dressed as a hard one fails in the way the condition's second component says it must.
The verdict
This program's record reads in four states. Delivered: the disinflation from 25.5% to 2.1% per month, the parallel premium from triple digits to near parity, a fiscal surplus held every month since January 2024, and the legibility of the monetary aggregate restored. Reversed: dollarization and the closure of the central bank, replaced by an actively managed peso regime. Partial: the cepo, lifted for individuals and still binding on corporates. Held by discretion: the surplus, whose deficit rule remains a benchmark, and the reserve position, negative on net and financed by borrowed buffers. That achievement is real, and its weakest link is that nothing locks it in.
Falsifiable by design, this reading may already be meeting its falsifier. Headline inflation has printed at or above 2% in every month since September 2025, and part of that plateau is administered-price catch-up and part may be indexation inertia, which five post-regime observations cannot separate. A corridor that held through a genuine political shock, with the budget loosening and the disinflation surviving, would weaken the claim that the anchor is borrowed. Core inflation breaking decisively below 2% under the band would weaken the inertia mechanism, and the May core print of 1.9%, the first below that line, is the paper's own falsifier beginning to show. Pre-registered for the December 2026 print, the persistence-break test is the scheduled adjudication of both.
A central bank that abandoned its rule and a country trying to build a rule out of its own inflation are the same study read from opposite ends, and what anchors money is the question both are asking. Until those prints exist, the verdict stands as the evidence leaves it. Disinflation is real, and the institution that would make it last is still the one promise the program has not kept.
Read the full paper (PDF, opens in new tab)
Get new essays and research by email — subscribe on Substack (opens in new tab)