Base effect, real, nominal, points: a glossary of the arithmetic in tourism news

A base effect is the distortion a percentage change picks up from its starting value: the same absolute move looks large after a low base and small after a high one. Points measure the gap between two rates; percent measures a change relative to the starting number. A rate falling from 51.91% to 49.54% drops 2.37 points, about 4.6%.

A sentence that says occupancy fell 2.37% is using the wrong unit. The bulletin behind it says 51.91% became 49.54% (Source: Ministry of Culture and Tourism, accommodation bulletin 2025). That is a fall of 2.37 points, and about 4.6% in relative terms. The sentence picked the wrong one of the two, and sentences like it are everywhere in hotel reporting. After reading this glossary you will be able to name the arithmetic behind each change you quote, spot the wrong unit in someone else's sentence, and caption every number in your own reports so that it can be checked.

What does base effect mean in tourism data?#

A base effect is the distortion a percentage change picks up from its starting value. When the starting value is unusually low or high, the same absolute move produces a very different percentage.

The mechanism is one line of arithmetic: percent change is the new value minus the old value, divided by the old value. The old value is the base, and it does all the work. Change the base and the same absolute move changes size, with no change in the underlying business.

The Ministry's 2025 yearly accommodation bulletin shows how wide the base can swing within one year. March 2025 was the weakest month at 26.79% occupancy and August the strongest at 71.15%. A rise of 5 points from March's level is about 19%; the same 5 points from August's level is about 7%. Nothing about the business differs, only the denominator.

The reverse happens after an exceptional year. A high base makes an ordinary year read as a collapse even though the level is normal. Base effects also have causes worth naming: a one-off event, a facility closed for renovation, or a change in what the series covers. Coverage changes, such as the inclusion of new facility types, can break series continuity. Check the publisher's revision notes for such changes.

What to do: always print the level next to the change, name the base period, and compare like with like before you interpret anything. What not to do: call a base effect a recovery, a boom or a crash. It is the denominator moving, not the market. If the change looks alarming, run the six checks before you read a monthly drop before acting on it.

Percentage points or percent: which one is it?#

Points measure the gap between two rates; percent measures a change relative to a starting number. A rate that goes from 51.91% to 49.54% falls 2.37 points and about 4.6%.

One point is one unit of the percentage itself. When both numbers are already rates, you subtract, and the unit is points. When you divide, the unit is percent, and the sentence must name the base it was divided by, otherwise the reader cannot check it.

The error is common in owner reports and in press coverage: a rate gap written as a percent, which overstates or understates the move depending on the size of the starting rate. The same two-point move is a different percentage at every level: two points on a 10% rate is 20%, two points on a 50% rate is 4%. This is the base effect again, in its most common disguise.

A count is different from a rate. Visitors, nights and revenue are counts, so their changes are percents; occupancy, share and growth are rates, so their gaps are points. Test your sentence: if you replaced the word with the other one, would a reader still get the right size of change?

The first row is from the Ministry of Culture and Tourism's 2025 yearly accommodation bulletin (opens in a new tab) (December 2025); the second is arithmetic, not data.

Monthly, yearly or cumulative: which change is this?#

Monthly compares one month with the same month a year earlier, cumulative compares the year to date with the same span last year, and yearly compares whole years. They answer different questions and can point in different directions.

Year to date means 1 January to the last published month, measured against the identical span a year earlier. It is only valid when both spans are the same length. The partial-year trap works the other way too: comparing January to August 2026 with the whole of 2025 measures the calendar, not the business.

Quarters are not added. TÜİK publishes tourism revenue by quarter, and stacking four quarters mixes figures published at different times and revised on different schedules, so a cumulative revenue figure for 2026 does not exist in the official release. The same caution applies to any series the publisher does not itself total.

Calendar effects sit inside monthly comparisons: religious holidays, school breaks and Easter move between months, so a monthly change can be a scheduling artefact. The first half of 2026 is the current example of a flat period, with visitors at 25,758,545, down 2.39% (Ministry of Culture and Tourism, August 2026 bulletin). If you want the fuller reading of that period, see reading a flat year in the official data.

Cumulative figures inherit every revision inside the window, so a year-to-date number can move without any new month being published. Check the label before the number.

Nominal and real: what is the difference?#

Nominal is the number as published, in the money of the day. Real is the same number after removing price change, so it shows whether volume actually moved.

The mechanism: divide the nominal figure by a price index and rebase it to a reference year. The result is expressed in the prices of that year. The index you choose changes the answer: consumer prices, producer prices and a tourism-specific deflator are three different denominators, so a real figure without its index named cannot be checked.

A nominal rise can be a real fall. For example, a hotel whose lira revenue rises 30% while its own cost base rises 40% has more nominal revenue and less real margin. Türkiye's tourism revenue for 2025, USD 65.23 billion, up 6.8% (TÜİK, Tourism Statistics 2025, 30 January 2026), is a nominal figure in a foreign currency, which matters for how it can be compared.

Per-visitor and per-night figures are averages, so deflating them needs the same care as deflating a total, and the visitor count in the denominator has its own revisions. Revenue in a foreign currency adds a second layer: the exchange rate moves the number before any price index is applied. Say which currency you are in. The rule for reports: never place two years of a nominal series side by side without saying that they are nominal.

Adjusted and unadjusted: what actually changes?#

Unadjusted figures include the season; adjusted figures have the repeating seasonal and calendar pattern removed so that neighbouring months can be compared.

Seasonal adjustment is a statistical estimate, not a correction of errors. Adjusted figures are revised as new months arrive, sometimes noticeably. Türkiye's accommodation data are strongly seasonal: in the Ministry's 2025 bulletin the monthly occupancy rate runs from 26.79% in March to 71.15% in August, so a July to August comparison is mostly season.

Use unadjusted figures for levels and for same-month or full-year comparisons; use adjusted figures for direction between adjacent months. Never mix the two in one sentence, and never subtract an adjusted figure from an unadjusted one. If an adjusted series is not published for the measure you need, say so rather than adjusting it yourself. The details of what adjustment does are covered in what seasonally adjusted means in official data.

Provisional, revised or final: whose label is it?#

The label belongs to the publisher, not to the reader. Use the publisher's own word, and check the page where the series is published for revisions.

Provisional means published before all returns are in; revised means a published figure has been changed; final means the publisher states that no further change is planned. The Ministry's revision and error policy (opens in a new tab) (read 7 November 2026) sets out two cases: major revisions from a change in definition, coverage, classification or calculation method, and any error found after publication. In both cases the public is informed on the web page where the official statistics are published.

That page gives its own example: under temporary Article 11 of Law No. 2634, as amended by Law No. 7334 of 28.07.2021, some facilities had to obtain a simple accommodation certificate, and those that did appear in the certified statistics from November 2022. A coverage change is a revision, and it breaks the series.

Labels differ by series. The Ministry's border bulletins are marked as provisional data, so check the label on the bulletin you are quoting, and do not add or remove one yourself. When a figure is revised, quote the new one, say that it was revised, and keep the date you fetched it. For any legal or fiscal consequence of a revised figure, confirm your own case with an accountant or legal adviser.

Average, median, and a rate that is already an average#

An average hides the spread and a median splits the group in half, and some published figures, such as an occupancy rate, are ratios that cannot be averaged again without care.

The mean adds the values and divides by their count; the median puts them in order and takes the middle. A few very large facilities pull the mean up and leave the median where it is. A rate is a ratio, not a count, so averaging rates without weights gives a small facility the same say as a large one.

The Ministry's 2025 yearly bulletin gives a worked example. The twelve monthly occupancy rates average 46.66% if each month counts equally, while the bulletin's yearly figure is 49.54%. The annual figure is not the mean of the months. The Ministry does not publish the formula behind its occupancy rate, including the treatment of extra beds or closed months, so do not reconstruct it and do not present the result as a room-based figure.

Say what the rate covers: the occupancy rate in Ministry-certified (İşletme and Basit) facilities is not the same thing as the occupancy of hotels, and certified facilities include types that are not hotels. When you quote an average, add the count behind it and, where it matters, the range. The counting side of that subject, what visitors and nights actually measure, is in the tourism statistics glossary.

What to do next#

Build a caption template and use it for every number you publish: measure, period, comparison period, unit (points or percent), whether the publisher labels it provisional, and the source with its date. The seven terms in this glossary are not separate problems; they are one habit, which is to state the base before you state the change. Start with the next report you write and apply the template to its first table.