TARIN
The cedi and prices · to September 2026

When the cedi falls, how fast do prices follow?

When a dollar costs 10% more cedis, imported goods cost about 3% more within three months and about 5% more within a year. Local goods follow almost as much, about 4% within a year, so a weaker cedi raises prices well beyond imports. World fuel and food prices explain only a small part of the 2022 surge. GSS's split between imported and local goods starts in 2021, so this runs from February 2021 to September 2026, and it is recomputed with each release.

32%of a fall in the cedi has reached imported-goods prices after three months
51%after a year · 90% range 39–63%
44%for local goods after a year
2 monthsbetween the cedi's moves and import inflation, at the closest fit (1–3 months fit almost as well)
How many months later

Import inflation follows the cedi, about two months behind

Over a year, import inflation moves most closely with the cedi's change over the year ending two months earlier (a correlation of 0.86; one to three months fit almost as well). Each 10% rise in cedis per dollar goes with about 4.5 more points of import inflation two months later. The cedi's fall peaked in November 2022, when a dollar cost 121% more cedis than a year before. Import inflation peaked two months later, at 62.4% in January 2023. After the cedi's strongest month in the 2025 recovery, June 2025, import inflation kept falling for nine months, to −0.6% in March 2026: the only month in this record when imported goods cost less than a year before.

In September 2026 a dollar cost 5.3% fewer cedis than a year earlier. Imported goods cost 2.4% more than a year earlier, and local goods 6.4% more.

Change on a year earlier

Imported goods Local goods Cedis per US dollar

Change on a year earlier, %, January 2022 to September 2026. Cedis per US dollar: the Bank of Ghana's interbank mid-rate, averaged over each month; a rise means the cedi fell. Imported and local goods: GSS's consumer price index by origin. Hover or tap for any month; Cedi moved two months later slides the cedi's line along to show the lag.

Which lag fits best

Imported goods Local goods

Correlation between prices' change over a year and the cedi's change over the year ending 0 to 15 months earlier, January 2022 to September 2026. Highest at two months for imported goods and one month for local goods.

How much, how fast

About half of a fall reaches import prices within a year

Set each month's change in import prices against the cedi's change in that month and in each of the twelve before, and the share passed on builds up like this: 7% in the same month, 20% after one month, 32% after three. Then it slows: 36% after six months and 51% after a year. Local goods follow a similar path, to 44%, and all items to 46%. With five and a half years of data the range is wide: 39–63% after a year for imported goods.

Imported goods 90% range Local goods

Share of a change in cedis per US dollar that has reached prices, by months since the change, %. Shaded: the 90% range for imported goods. Monthly changes, February 2021 to September 2026 (68 months), with each calendar month's usual change taken out.

Episodes

The cedi's three big moves since 2022

In the 2022 crash a dollar went from 6.01 to 13.07 cedis, and over the year from January 2022 imported goods rose 62.4%, more than the 40.3% rise the model above predicts. World diesel had risen 66% in the year before, but the section below finds world prices explain only a small part of 2022. In the 2024 slide (11.65 to 15.99 cedis), imported goods rose 18.0%, less than the 27.8% rise the model predicts. In the 2025 recovery (15.53 back down to 10.28 cedis), they fell 0.6%, close to the 2.6% fall the model predicts.

EpisodeCedis per dollarImported goodsPredictedLocal goodsAll items
2022 crashyear from Jan 20226.01 → 13.07+118% by Nov 2022+62.4%+40.3%+51.3%+53.6%
2024 slideyear from Dec 202311.65 → 15.99+37% by Oct 2024+18.0%+27.8%+26.4%+23.8%
2025 recoveryyear from Mar 202515.53 → 10.28−34% by Jun 2025−0.6%−2.6%+4.9%+3.2%

Prices: change over the twelve months from the start month. Cedis per dollar: monthly averages, at the start and at the cedi's turning point. Predicted: what the model behind the chart above gives for imported goods over the same twelve months, from the cedi's moves plus the usual rise in prices that has nothing to do with the cedi.

The cedi or the world?

World fuel and food prices explain only a small part of the 2022 surge

GSS's split between imported and local goods starts in 2021: too short to separate the two (see below). Prices as a whole go back to 1998, so the question is asked there. Each month's change in all prices is set against the cedi's change and against world prices of diesel (the US Energy Information Administration's New York Harbor price) and of food (the World Bank's index), in dollars, in the same month and each of the twelve before. Over these 28 years, a 10% rise in world diesel adds about 0.9% to Ghana's prices within a year, and a 10% rise in world food about 0.4%, a figure too uncertain to tell from zero. The cedi's own share hardly moves: 43% within a year, against 43% without world prices. Diesel rather than crude oil, because Ghana imports refined fuel: over the 2022 crash year world diesel rose 25%, while Brent crude fell 3%.

In the 2022 crash year all prices rose 53.6%. By this measure the cedi accounts for about 21 points of that, and world fuel and food together for about 6 (4 to 8, whichever world prices are used), though world diesel had risen 66% in the year before. About 11 points is the usual rise in prices, in a year when neither moves. That leaves about 15 points that neither the cedi nor world prices explain; the next section looks for them. In the 2024 slide world prices played almost no part (under a point). In the 2025 recovery the cedi's strengthening would have cut prices by about 13 points; they still rose 3.2%, with about 8 points explained by neither.

Usual rise The cedi World fuel and food Neither
EpisodeAll pricesUsual riseThe cediWorld fuelWorld foodNeither
2022 crashyear from Jan 2022+53.6%+10.5+21.5+6.5−0.4+15.5
2024 slideyear from Dec 2023+23.8%+9.4+12.0−0.80.0+3.2
2025 recoveryyear from Mar 2025+3.2%+8.6−13.3+0.2+0.3+7.5

All items: the change in prices over the twelve months from each episode's start, split into parts that add up to it, in percentage points (rounding can leave them a tenth apart). Usual rise: what prices do in a year when neither the cedi nor world prices move. The cedi, world fuel, world food: their moves in that year and the twelve months before, at their average effect from 1998 to 2026. Neither: what's left. World prices, in US dollars: diesel at New York Harbor ($ a gallon; ultra-low-sulfur since 2006, No. 2 heating oil before; EIA, via FRED) and the World Bank's food price index.

For imported goods alone, adding world prices raises the cedi's share to 54–59% within a year, against 51% without them. But five and a half years can't pin down world prices' own effect on them: depending on how they're entered, world diesel's comes out between 6% and 15%, and food's between −11% and 4%, and a negative effect can't be right. So the world is measured on all items only.

Checks on the answer

Does it hold up?

The same measure, done four other ways. Imported goods after a year, unless a card says otherwise.

Other currencies50–58%euro, yuan and a basket, after a year

Measured against the euro, 50% of a fall reaches import prices within a year; against the Chinese yuan, 58%; against a basket of the three (half dollar, a quarter each euro and yuan), 54%. The answer barely depends on the currency.

Without the 2022 crash34%after a year, from March 2023

From March 2023, once the crash had passed through (43 months), the share after a year is 34%, and it builds more slowly: 16% after three months. The 2022 crash pulls the average up.

Falls and rises of the cedi38% / 17%of a fall / of a rise, after three months

After three months, 38% of a fall in the cedi has reached import prices, but only 17% of a rise. After a year the two are close: 46% and 44%. Prices go up faster than they come down. Only 17 of the 68 months saw a stronger cedi, so this is the least certain result here.

28 years of all items43%all items after a year, 1998 to 2026

All-items prices go back to 1998. Over 1998 to 2026 (329 months), 43% of a fall in the cedi reaches them within a year; over 1998 to 2020, 42%; since 2021, 46%. The recent years aren't unusual.

Method

How it's measured, and what it leaves out

The cedi. The Bank of Ghana's interbank mid-rate for the US dollar, averaged over each month, in today's cedis. Months with fewer than 10 published rates, and the month in progress, are left out. Prices. GSS's consumer price index for imported goods, local goods and all items: StatsBank before January 2022 for imported and local goods (January 2021 for all items), then the monthly CPI annex, which agrees with StatsBank to its rounding where both exist. Only values that pass TARIN's checks are used. World prices. Diesel: the US Energy Information Administration's daily spot prices at New York Harbor, as FRED (the Federal Reserve Bank of St. Louis) averages them by month, in US dollars a gallon: ultra-low-sulfur diesel since 2006, chained onto No. 2 heating oil before, to September 2026. Food and Brent: the World Bank's Commodity Price Data (the Pink Sheet), monthly, in US dollars: its food price index (2010 = 100) and Brent crude, from the workbook of 2 October 2026, with data to September 2026.

How much, how fast. Each month's change in prices (in logs) is regressed on the change in cedis per dollar in the same month and in each of the twelve before, with a dummy for each calendar month. The running sum of the cedi's coefficients is the share passed on after that many months. Standard errors are Newey–West with four lags; the shaded range is 1.645 of them either side, a 90% range. How many months later. The correlation between prices' change over a year and the cedi's change over the year ending k months earlier, for k from 0 to 15. The cedi or the world. All items, February 1998 to September 2026: the same regression with world diesel and the food index added, each also in the same month and the twelve before. An episode's parts are each one's coefficients applied to its moves, scaled so the parts add up to the change in prices. The range quoted for world prices comes from six other ways of measuring them, Brent instead of diesel among them. What else in 2022. The same regression for each of GSS's 13 groups of the basket (StatsBank, from 1998), each group's parts weighted by its share of the 2021 basket in January 2022.

What it leaves out. Freight costs and policy (the Bank of Ghana's rate, fuel taxes and levies, utility tariffs, transport fares) aren't in the model, so part of any episode can be them; for 2022, the section on what else happened lists them without measuring their size. World fuel and food prices are measured for all items only: GSS's import split is too short. GSS's imported goods are a basket of consumer goods priced in shops, not prices at the border. And every share here is an average over the years measured, not a forecast.

Downloads. cedi.csv (the monthly series), cedi-passthrough.csv (the shares with their standard errors), cedi-world.csv (each episode split into its parts), cedi-2022-groups.csv (2022 by group of the basket) and cedi.json (everything on this page). The series behind it: the cedi's daily rate and world diesel, food and Brent prices in the catalogue, and imported and local goods in the data explorer.