How is the economic confidence index calculated?
The economic confidence index is a composite indicator that summarises how consumers and businesses assess current conditions and the near future rather than measuring one price, production quantity or income level. The series published by TurkStat combines consumer confidence with seasonally adjusted confidence components for services, retail trade and construction, together with manufacturing-related real sector confidence components published by the Central Bank of the Republic of Türkiye. The published number is not a simple average of five headline indices. Selected question balances from the surveys are first normalised against their own historical distributions. Sector weights are then distributed across the component questions belonging to each sector, and the combined result is rescaled so that its long-term mean equals 100. Consequently, 100 is not a fixed threshold for growth, contraction or recession. It locates current sentiment relative to the average confidence level in the reference sample. A monthly movement should be read together with the components that changed, the survey collection period, seasonal adjustment and any methodological update. A modest rise in service confidence, for example, may not offset a fall in a more heavily weighted manufacturing component. Consumer expectations may improve while construction order assessments weaken. The index is not an early release of realised GDP and cannot replace output, employment, sales or price statistics, because it records perceptions and expectations. The calculation below is deliberately simplified. TurkStat’s actual computation uses numerous normalised question balances and the official coefficients applicable to the published series. Before comparing two periods, users should confirm the methodology note, revision status, seasonal-adjustment treatment and reference sample. Confidence indicators are valuable because they are timely and can signal changing direction, but a single composite number can conceal substantial disagreement among households and sectors. It should therefore support, rather than substitute for, a broader evidence-based economic assessment.
From survey components to a common sentiment measure
1. Constructing survey balances The raw material consists of qualitative answers supplied by households and enterprises about current conditions and expectations. A question balance is commonly formed by subtracting the percentage of negative replies from the percentage of positive replies, while neutral replies do not directly lift the balance. Questions may concern recent business conditions in services, sales in retail trade, or orders in manufacturing. Their natural ranges and volatility differ, so adding the raw balances would give unstable components excessive influence. Official methods also apply the survey design, response weights and collection window. When interpreting a month, analysts should ask whether a major policy announcement or market event occurred before or after respondents submitted their assessments. 2. Normalisation and sector weights Each selected balance is standardised using its historical mean and variation, which places questions with different volatility on a more comparable scale. In the European Commission method, the Economic Sentiment Indicator assigns sector weights of 40 percent to industry, 30 percent to services, 20 percent to consumers, 5 percent to retail trade and 5 percent to construction. TurkStat follows the harmonised business and consumer survey approach. A sector’s weight is not simply multiplied by its headline confidence index; it is divided among the selected normalised questions used for that sector. The methodology note accompanying the current TurkStat release remains the authority for the precise national components, sample and coefficients. 3. A simplified weighted example Suppose the normalised values for the five sectors are 0.20, 0.10, -0.15, 0.05 and -0.10. Using weights of 40, 30, 20, 5 and 5 percent produces 0.20×0.40 + 0.10×0.30 - 0.15×0.20 + 0.05×0.05 - 0.10×0.05 = 0.0775. In the official procedure, each sector contains several question balances, so its sector weight is allocated across those components before the available weighted values are summed. The composite is then rescaled to an index with a long-term mean of 100 and a specified standard deviation. This illustration cannot reproduce an official release; it only demonstrates why component weights and normalisation matter.
Interpretation, comparison and methodological limits
4. Reading the 100 benchmark correctly A value above 100 indicates that sentiment is above its average in the historical normalisation sample, while a value below 100 indicates a below-average position. It does not prove that the economy is expanding or contracting, and it must be separated from the direction of monthly change. An increase from 98 to 99 still leaves the level below its historical mean, although the direction has improved. A small one-month movement may reflect sampling variation, rounding or a temporary news effect. Analysts should distinguish the level, trend and component contributions and avoid deriving a certain investment, employment or GDP conclusion from the headline alone. 5. Seasonal effects, calendar and revisions Confidence series can repeat seasonal patterns related to holidays, tourism, weather or regular ordering cycles. Seasonal adjustment attempts to remove those recurring movements, yet recent adjustment factors may change as new observations enter the model. A revision can also follow changes in survey questions, classifications, samples or the historical period used for normalisation. Comparisons should therefore use the same series type and methodology vintage. Mixing an old rounded figure from a press release with a newer revised database observation may create an apparent movement that is purely methodological. Users should retain the publication date, adjustment status and revision note when documenting an analysis. 6. Combining sentiment with realised indicators The index is timely and may provide an early indication of turning points, but survey answers are assessments and expectations rather than realised production, income or spending. A robust reading compares its direction with industrial production, retail sales volume, employment, inflation, foreign trade and national accounts. The composite can hide divergence when sector components move in opposite directions, so consumer, real sector, services, retail and construction measures should also be inspected. Economic confidence alone is insufficient for a portfolio decision, credit assessment or company sales forecast. Current releases, methodology and revision calendars should be verified from official sources, and the result should be treated as one piece of economic evidence rather than a deterministic forecast.