AfDB approves US$68.44m second-phase reform loan for Eswatini

The budget-support operation targets public-finance management, business conditions and domestic energy capacity.
A national index and a family’s bills measure different things. A worked example explains weights, comparison periods and why slower inflation need not mean lower prices.
You can read that inflation has eased and still find that the money set aside for ordinary shopping no longer stretches far enough. Those observations do not automatically contradict each other. A national price measure describes a broad pattern; a household has its own mix of purchases, commitments and changes in circumstances. The useful question is what each number measures before deciding whether they disagree.
This guide explains how to read the figures rather than announcing the latest inflation rate. Its numerical examples are deliberately hypothetical, not current Eswatini prices, official forecasts or a survey of household spending. They are small enough to check with a calculator. The aim is to make a statistical release more useful when looking at an actual receipt or bill.
Eswatini’s Central Statistical Office describes the Consumer Price Index, or CPI, as a measure of price changes for household goods and services. Its January 2026 methodology says the expenditure weights were based on the 2016/17 Household Income and Expenditure Survey, with June 2020 as the index reference period. It describes price collection across all four regions. These are features of that published methodology, not a claim that a particular family follows the national basket. [1]
An index reference of 100 is a measuring starting point, not an amount of money that every household once spent. A later index of 120 would mean a 20% increase from that reference for the measured basket. It would not mean the annual inflation rate was 20%, unless the comparison period happened to be that same reference period. Always identify the two dates being compared.
Imagine an index of 100 a year ago, 103 last month and 104 now. The year-on-year increase is 4%: divide 104 by 100, subtract one, then multiply by 100. The month-on-month increase is about 0.97%: use 103, rather than 100, as the denominator. Both statements can describe the same latest observation.
Neither result should be multiplied casually into a forecast. Repeating a one-month increase for twelve months assumes a future path that has not happened; compounding also differs from simple multiplication. For a news reader, the first safeguard is more basic: do not compare a monthly figure with an annual one as if they were equivalent. The CSO release presents those measures separately. [1]
Consider a fictional household whose selected monthly purchases originally cost E2,000: E800 for food, E400 for transport, E300 for electricity and E500 for everything else in this example. Suppose the price of the same quantities rises by 6%, 12%, 10% and 0% respectively. These are invented changes, not tariff or inflation claims about Eswatini.
The new amounts would be E848, E448, E330 and E500. Together they total E2,126, an increase of E126, or 6.3%. The transport increase is larger in percentage terms, but food also matters because the household spends more on it. A percentage becomes meaningful for this budget when it is connected to an original amount.
Now imagine another household allocating 25% of this selected budget to food, 5% to transport, 10% to electricity and 60% to the unchanged category. Applying the same hypothetical changes gives 3.1%, rather than 6.3%. Different spending shares can produce different experiences without requiring either household to be mistaken.
In the first example, transport represents 20% of the original selected budget. A 12% transport price increase therefore adds 2.4 percentage points to the example’s overall 6.3% increase: 20% multiplied by 12%. It does not add twelve percentage points to the whole budget. Food contributes another 2.4 points, electricity 1.5 points and the unchanged category zero.
That simplified arithmetic is a teaching example, not a reconstruction of the official CPI formula. The distinction is useful when reading official tables: the price movement of one category and its contribution to a broader movement are different questions. Eswatini’s CSO report includes a separate presentation of group contributions. [1]
Suppose a fictional basket rises from E1,000 to E1,100 in one year, then to E1,155 the following year. Inflation has slowed from 10% to 5%. The basket still costs more in the second year, and it costs 15.5% more than at the beginning. A smaller positive rate means slower growth, not a return to the old price.
This is why a statement about easing inflation does not by itself establish that living has become cheaper. Falling prices would be a different movement. The IMF’s introductory explanation distinguishes the rate of price increase from the price level and explains why income that fails to keep pace can lose purchasing power. [3]
Price comparison works best when the quantity and service are comparable. Spending more on electricity after using more units is different from paying more for the same usage. A smaller package at an unchanged shelf price also needs a unit comparison. An annual school payment appearing this month can make cash flow harder without representing a new monthly price rise.
These distinctions are reasons to annotate a household record, not dismiss a difficult month. Mark whether a change comes from the unit price, the amount used, a new purchase or the timing of payment. A short list of comparable repeat purchases can be more informative than comparing two shopping totals that contain different items.
The Central Bank of Eswatini’s January 2026 Economic Review and Inflation Report separates recorded inflation from an outlook discussing assumptions and forecasts. That document is a historical example of the distinction, not a statement of the latest outlook. A forecast remains conditional; a subsequent statistical observation answers what was measured over a completed period. [2]
When sharing a number, carry its label with it: source, observation month, comparison period and publication date. A report issued in one month may describe the previous month. A prediction for a full year is not the same thing as a single month’s year-on-year reading. Even a correct number becomes misleading when its label disappears.
A practical reading routine is to find the original release, check the dates and units, examine relevant category movements, then compare like-for-like items in your own records. Keep the national measure for understanding the wider economy; use the household record for identifying where personal spending changed. Neither should pretend to do the other’s entire job.
The payoff is a clearer conversation. Instead of saying only that prices feel higher or that inflation has fallen, it becomes possible to name the comparison: this item, this quantity, this period and this change. That is useful evidence for a household discussion, a question to a service provider or a better-informed reading of the next statistical release.
Media: International Monetary Fund · What is Inflation? | Ask an Economist — IMF (2022)
Prepared with AI assistance; sources linked below.
Published 18 Nov 2026
Suggest a correction
The budget-support operation targets public-finance management, business conditions and domestic energy capacity.

Zulu D quoted E100 for admission, and organisers are inviting families to bring picnic blankets.

The ICT ministry aimed to restore services by Wednesday evening. The reports reviewed did not confirm a full recovery.