Appraisal-Based Property Indices and Smoothing

The most widely used commercial property performance indices, like the NCREIF Property Index in the US, aren't built from actual sale transactions the way a stock index is built from trade prices — most of the properties in the index don't sell in any given quarter. Instead they're built from periodic professional appraisals of properties that stay in the index unsold for years at a time. This single construction detail — valuing based on estimates rather than trades — produces a statistical artifact that anyone using these indices needs to understand before trusting the numbers: appraisal-based indices report real estate as far smoother and less volatile than it actually is.

Why appraisals smooth the truth

An appraiser valuing a building this quarter typically anchors partly on the building's own last appraised value and on whatever comparable sales happened to close recently, which are themselves often stale by the time they're used. This makes each new appraisal a partial, lagged blend of "what's really happening to values now" and "what we said it was worth last time" — mathematically similar to averaging a noisy signal with its own recent past. That averaging process, applied across an entire index of thousands of appraised properties, dampens quarter-to-quarter swings and smooths out genuine turning points, making the index understate true volatility and understate how quickly and how far property values actually move.

The same smoothing distorts how real estate appears to correlate with other asset classes. Because the appraisal-based index reacts to a market downturn with a lag — sometimes several quarters — it can show low correlation with equities or bonds purely as an artifact of that lag, even when the underlying economic forces hitting property values are the very same ones hitting stocks and bonds at the same time. Once the smoothing is corrected for (a process called "de-smoothing" or "unsmoothing," which tries to back out the true, un-lagged signal), real estate's actual volatility and its correlation with other risky assets both come out meaningfully higher than the raw appraisal-based index suggests.

For example, an appraisal-based property index might show a mild 3% decline during a year the broader stock market fell 25%, suggesting real estate is a strong diversifier. A transaction-based or de-smoothed version of the same market, correcting for appraisal lag, might show a true decline closer to 15–18% over the same period, simply arriving with a delay — a much less flattering, but more accurate, picture of how correlated the two markets actually are.