A seasonally adjusted annual rate (SAAR) is a way to remove predictable seasonal patterns from periodic data (monthly or quarterly) and then convert that de‑seasonalized value into an annualized rate. The goal is to see the underlying trend or to compare activity “apples to apples” across different months or quarters without seasonal swings distorting the comparison.
Key takeaways
– SAAR removes recurring seasonal effects (holidays, weather, school schedules, etc.) and annualizes the de‑seasonalized value.
– Common formula (monthly): SAAR = (unadjusted monthly value ÷ seasonality factor for that month) × 12.
– For quarters, multiply the de‑seasonalized quarterly value by 4 instead of 12.
– SAAR is widely used in auto sales, housing, employment, GDP reporting, and other business/economic metrics.
– Limitations: relies on a stable seasonal pattern; can be misleading during structural breaks or one‑off shocks. Agencies often revise seasonal factors as new data arrive.
Understanding seasonality and SAAR
– Seasonality: regular, predictable variation that repeats within a year (e.g., higher retail sales in December, more ice cream sold in summer).
– Seasonality factor: a multiplier that describes how a specific month or quarter typically compares to the average period (factor > 1 means above average, factor < 1 below average).
– De‑seasonalizing: divide the raw (unadjusted) value by the seasonality factor to remove the seasonal effect.
– Annualizing: multiply the de‑seasonalized period value by 12 (monthly) or 4 (quarterly) to express it as an annual rate (SAAR).
How to calculate SAAR — practical step‑by‑step (monthly data)
1. Choose a baseline period for seasonal factors
• Use several years of data (commonly 3–5+ years) so monthly patterns average out; longer series can improve stability.
2. Compute the overall average for the baseline
• Overall average = (sum of all monthly values in the baseline) ÷ (number of months).
3. Compute each month’s average
• For each calendar month (Jan, Feb, …, Dec) compute the average of that month across baseline years.
4. Calculate the seasonality factor for each month
• Seasonality factor for month m = (average value for month m) ÷ (overall average).
• Interpretation: factor = 1.25 → that month is typically 25% above the average month.
5. De‑seasonalize the observed month
• De‑seasonalized value = observed unadjusted month ÷ seasonality factor for that month.
6. Annualize the de‑seasonalized value
• SAAR = de‑seasonalized value × 12.
Worked example (monthly)
– Baseline: business had total annual revenue = $144,000 last year → average monthly = $12,000.
– June last year = $20,000 → June seasonality factor = 20,000 ÷ 12,000 = 1.6667.
– This year June actual = $30,000 → de‑seasonalized = 30,000 ÷ 1.6667 = 18,000.
– SAAR = 18,000 × 12 = $216,000 → indicates a year‑over‑year increase versus $144,000.
Quarterly SAAR
– Same idea; seasonality factors by quarter and annualize by multiplying de‑seasonalized quarterly value by 4:
SAAR_quarter = (unadjusted quarter ÷ quarter seasonality factor) × 4.
Practical tips for computing seasonality factors
– Use multiplicative model (division) when seasonal variation scales with level (most common for sales, employment).
– Use additive model (subtract seasonal component) when seasonal swings are roughly constant in absolute terms.
– Apply smoothing or moving averages to reduce outlier influence.
– Recompute seasonality factors periodically (e.g., yearly) because seasonal patterns can shift.
– For official statistics, agencies use statistical methods such as X-13ARIMA-SEATS to model and adjust seasonality and to handle trading‑day and holiday effects.
How to do this in Excel or Python (short examples)
– Excel sketch:
1. Column A: Date; Column B: Value.
2. Compute month number: =MONTH(A2).
3. Use AVERAGEIFS to get average per month across baseline.
4. Compute overall average with AVERAGE.
5. Seasonality factor = month average ÷ overall average.
6. SAAR = (Value ÷ SeasonalityFactor) * 12.
– Python (pandas) sketch:
• groupby month to get monthly means, compute overall mean, derive factors, then apply to each row and multiply by 12.
SAAR vs non‑seasonally adjusted (NSA) data
– NSA values = raw observed values (no adjustment). Use NSA to see actual counts and cash flow in a period.
– SA or SAAR values = adjusted for seasonality (and annualized). Use SAAR to compare trend strength across periods, to compute growth rates that are not confounded by seasonal timing.
– Example: comparing home prices in July vs January without adjustment could be misleading; SAAR lets you compare underlying price movement.
When to use SAAR
– To compare performance across different months/quarters or across years without seasonal bias.
– To spot underlying trends and cycles more clearly (e.g., whether sales momentum is rising or falling).
– To calculate annualized figures for reporting or forecasting (note: SAAR is not a forecast—it's an annualized view of a de‑seasonalized period).
Limitations and cautions
– SAAR assumes the seasonal pattern is reasonably stable; structural changes, major shocks (pandemic, policy changes), or changing consumer behavior can invalidate factors.
– Revisions: agencies and analysts may revise seasonal factors and therefore SA or SAAR figures as more data come in.
– Annualizing a single month/quarter can exaggerate short‑term swings—use SAAR alongside multi‑period averages and context.
– SAAR shows rate (annualized level) not necessarily the actual expected total for the year; it projects the de‑seasonalized period forward as if the same rate persisted all year.
Common use cases
– Auto industry: monthly car sales reported at SAAR to compare seasonally varying monthly sales.
– Housing: convert monthly median prices/transactions to SAAR to compare across months/years.
– Employment and GDP: agencies publish seasonally adjusted and annualized series to indicate trend growth.
Where agencies get more advanced
– Statistical offices (U.S. Census Bureau, Bureau of Labor Statistics) use established software (X-13ARIMA-SEATS) to model and remove seasonality, account for trading‑day and holiday effects, and produce more robust seasonal factors.
– If you need high‑quality adjustments for policy or formal analysis, use official seasonally adjusted series or proven seasonal‑adjustment software rather than simple averages.
Sources and further reading
– Investopedia — Seasonally Adjusted Annual Rate (SAAR):
– U.S. Census Bureau — X-13ARIMA-SEATS: /
– Bureau of Labor Statistics — Seasonal Adjustment
Editor’s note: The following topics are reserved for upcoming updates and will be expanded with detailed examples and datasets.