Comparative Study
Brazilian and Turkish policy evidence highlights exchange-rate exposure, inflation expectations and backward indexation as distinct transmission channels. Controlled examples separate imported-price effects, persistent contract adjustment and administered-price changes without assigning unsupported national pass-through coefficients.
Disinflation is a decline in the rate at which prices rise, not necessarily a decline in the price level. Its progress can differ across goods and services, across contracts and across households even within one economy. Comparing countries therefore requires a common account of the price-setting mechanisms rather than a simple ordering of headline inflation rates.
Brazil and Türkiye offer two settings in which central-bank assessments explicitly discuss exchange rates, expectations and domestic inflation persistence. The comparison does not assume that either economy has only one source of inflation. Instead, it asks how external price changes and domestic adjustment rules interact with monetary conditions and fiscal decisions.
The common framework separates an initial imported-cost shock, subsequent price and wage adjustments, and the financing or demand consequences of policy. Official statements are used as dated evidence of the mechanisms identified by the respective authorities. The worked examples use invented parameters. They do not estimate national pass-through, prescribe an interest-rate path or reconstruct either country's inflation forecast.
In its December 2025 meeting minutes, Brazil's monetary policy committee attributed lower goods and food inflation partly to exchange-rate appreciation and more favorable commodity behavior. It described services inflation as more resilient, alongside a still-dynamic labor market and gradual moderation in activity. It also emphasized the importance of inflation expectations for the policy assessment. These are the committee's interpretation of conditions at that meeting, rather than a current forecast. Banco Central do Brasil, December 2025 Copom minutes.
The evidence identifies two channels that should remain distinct. A change in the domestic-currency cost of internationally traded inputs can affect goods prices, while domestic service pricing can respond to different costs, contracts and demand conditions. Improvement in the first channel does not establish that all underlying inflation pressures have disappeared.
The committee's January 2025 minutes also discussed the relationship among fiscal news, exchange-rate developments and inflation risks. That discussion is evidence that the authorities considered these channels jointly; it is not an independently estimated causal coefficient linking a particular fiscal announcement to prices. Banco Central do Brasil, January 2025 Copom minutes.
The historical sequence matters. A January risk assessment and a December assessment of observed developments are not competing measurements taken on the same date. Keeping those dates visible allows the analysis to discuss mechanisms without presenting an earlier warning or projection as a realized later outcome.
The Central Bank of the Republic of Türkiye's Inflation Report 2026-I identifies backward-indexed pricing, particularly in services, as a constraint on disinflation. It also discusses fiscal and incomes policies through production-cost, demand and expectations channels, including the treatment of administered prices and taxes. This is the February 2026 report's assessment and policy framework. CBRT, Inflation Report 2026-I, medium-term projections.
Backward indexation introduces a link between past inflation and the prices or wages set now. Current demand restraint may influence the adjustment, but a rule that gives substantial weight to the previous period can continue transmitting an earlier price increase. The relevant question is the contract or pricing practice, not merely the current headline reading.
The report's discussion does not establish that every service price is formally indexed or that indexation has one constant coefficient. An empirical comparison would need sector-specific evidence on adjustment frequency, reference periods and the influence of expectations. The mechanism can be present without every transaction following the same rule.
Nor is backward-looking adjustment exclusive to Türkiye. The Brazilian discussion of expectations and service-price resilience also points to domestic persistence. The comparative distinction concerns what each cited authority emphasizes and how the mechanisms operate; it should not become an unsupported claim that one country has only exchange-rate inflation and the other only indexation.
| Channel | Initial object to measure | Main comparative boundary |
|---|---|---|
| Exchange-rate change | Domestic-currency price of the relevant foreign currency | Quote direction and invoicing currency must match |
| Import-price transmission | Imported content and price adjustment over time | Direct purchases differ from embedded imported inputs |
| Backward indexation | Contract reference period and adjustment formula | Past inflation differs from current expectations |
| Domestic demand | Activity and pricing conditions in affected sectors | Services and goods need not respond equally |
| Fiscal and administered prices | Budget costs, price rules and timing | A temporary price restraint can shift rather than remove costs |
The channels interact. Exchange-rate depreciation can raise an imported input price, which then enters a later wage or service adjustment. Fiscal measures can change current measured prices and expectations about future financing. Monetary conditions can affect demand, financing and exchange rates, with different timing across those routes.
A comparative account should identify these links without counting one shock repeatedly as several independent contributions. Adding a direct imported-price effect to a later adjustment that already contains that imported effect can overstate the total unless the accounting defines each stage carefully.
Consider an invented consumer basket with a base price index of one hundred. Thirty percent represents a category exposed to imported costs; the remainder is unaffected in this exercise. Assume the local-currency price of the relevant foreign currency rises by 10 percent and that 60 percent of that change reaches the exposed category's consumer price during the measured interval.
The exposed category's index rises to 106, while the other category stays at one hundred. The basket index becomes 101.8: 0.30 × 106 + 0.70 × 100. The first-stage price-level effect is 1.8 percent under these assumptions. This is a finite arithmetic example, not an estimated relationship for Brazil or Türkiye.
Now change only the imported-exposure weight to 20 percent and the interval-specific pass-through to 40 percent. The exposed category rises to 104 and the basket to 100.8. The difference between the two examples comes entirely from those two assumed parameters, even though the exchange-rate movement is the same.
| Hypothetical case | Exposed basket weight | Interval pass-through | Basket index after shock |
|---|---|---|---|
| A | 30% | 60% | 101.8 |
| B | 20% | 40% | 100.8 |
Neither label is assigned to a country. Actual comparison would require consistent estimates of imported content, pricing behavior and the time horizon. A pass-through estimate over one month cannot be compared directly with another covering a year. Differences in the size or type of shock can also limit the interpretation of a coefficient estimated in another period.
The example excludes indirect production effects and subsequent wage or demand responses. Import content embedded in domestic products would need an input account rather than being assumed equal to the share of finished imported goods in the consumer basket. It also excludes changes in foreign-currency export earnings, which can alter the income effect of the same currency movement.
A separate invented pricing rule assigns 70 percent weight to previous-year inflation and 30 percent to a current reference rate. If those rates are 20 and 5 percent respectively, the contractual adjustment is 15.5 percent: 0.70 × 20 + 0.30 × 5. A rule using only the current 5 percent reference would generate a smaller adjustment.
This calculation demonstrates persistence under a specified formula. It does not estimate an actual wage or rent contract in either country. Nor does it show that all of the difference is an avoidable policy cost; contracts can embody distributional concerns, negotiations and other conditions absent from the example.
Timing is central. An annual adjustment can keep incorporating an old inflation rate after more recent monthly price changes have slowed. Comparing an annual contract reset with a current monthly inflation reading can therefore exaggerate the apparent contradiction. The observation dates and reference periods must be aligned.
Expectations are a separate input. A firm may use expected future costs rather than a mechanical historical formula, or it may combine the two. An analyst should not label all persistent inflation as formal indexation without evidence about pricing behavior. The CBRT report's emphasis identifies a relevant channel, while the actual distribution of rules remains an empirical question.
Consider a hypothetical energy service with a resource cost of 120 per unit and an administered consumer price of one hundred. If ten units are supplied and the government covers the difference, the budget cost is two hundred. The lower consumer price has not removed the resource requirement; it has placed part of the financing elsewhere.
If the consumer price subsequently rises to 120, that category's index increases 20 percent. With a 10 percent basket weight and all other prices unchanged, the immediate aggregate price-level effect is 2 percent. This is a one-time level calculation. It is not 20 percent inflation recurring indefinitely, and it is not an estimate of an actual Brazilian or Turkish measure.
A full assessment would include who pays taxes or receives transfers, how the measure is financed and whether consumption responds. A subsidy can affect distribution, demand and public liabilities; removing it can also have distinct distributional effects. The current CPI contribution alone does not settle the welfare comparison.
The financing method can affect future conditions, but that connection requires evidence rather than an automatic assertion that every subsidy creates the same inflation. Temporary support, permanent commitments and contingent guarantees have different fiscal implications. The CBRT's discussion of policy coordination and the Brazilian committee's fiscal-risk discussion provide reasons to examine these distinctions, not a universal numerical multiplier.
External exposure is not confined to consumer prices. Suppose a borrower owes one hundred units of foreign currency, while the domestic-currency exchange rate moves from five to 5.5 per foreign unit. The local-currency value of the liability rises from five hundred to five hundred fifty, before any repayment or hedge.
That increase can occur even as the rate of domestic consumer-price inflation declines. Disinflation therefore does not by itself demonstrate that a foreign-currency borrower has become less exposed. Income currency, export receipts, liquid assets and hedges determine the net position.
The same currency movement can help an exporter receiving foreign currency while raising the domestic cost of imported equipment or debt service. National net effects require the distribution of those positions across firms, households and the public sector. A single exchange-rate change cannot establish the sign of every balance-sheet effect.
For Brazil and Türkiye, the appropriate comparative question is the documented exposure of the relevant borrowers and sectors. No national foreign-currency debt share is inferred here from a general description of either economy. The hypothetical balance sheet identifies the measurement needed before that channel is assigned an aggregate magnitude.
A credible comparison needs a matched inflation concept, observation window and source vintage. Headline consumer prices, underlying measures and service categories answer different questions. Year-over-year inflation can also decline because an earlier increase leaves the comparison period, even if the current price level continues rising.
The official evidence used here spans meetings in 2025 and a report in early 2026. It establishes actual assessments of transmission channels at those dates. It is not used to claim which country had achieved a better outcome by the information cutoff or to reproduce a current forecast from older projections.
The substantive comparison is that exchange-rate relief can coexist with domestic persistence, while backward-looking adjustments can postpone the effect of changing current conditions. Fiscal and administered-price decisions influence both the observed timing of price changes and the distribution of their costs. Balance-sheet exposure adds another reason why falling inflation need not translate into uniform financial relief.
The controlled examples keep these mechanisms separate enough to be tested: imported weights and pass-through determine the initial basket effect, contract rules determine persistence, and administered prices determine when consumers or budgets bear a cost. Applying those distinctions to actual country data offers a stronger basis for evaluating disinflation than assigning either economy one cause or importing another country's estimated transmission coefficient without checking its horizon, definitions and economic setting.