CAIA® real estate indices: which price signal should you use?

On this page

An appraisal-based property index is a broad but delayed account of values assigned to holdings; a transaction-based index is a narrower account of prices at which selected properties actually sold. For a CAIA® Level II comparison, ask first whether the decision needs representative coverage or a timely signal of market-clearing prices. Neither series is a neutral window onto an identical pool of assets.

Start with the observation, not the return

A private building does not post a continuously observable price. An owner can estimate its value while continuing to hold it, or wait until a sale establishes a price for that particular asset. The two observations answer different questions. An appraisal says what a professional estimates a held property would be worth under a specified valuation process. A sale says what a particular buyer and seller agreed to exchange at that time. A property index turns many such observations into a time series, but its construction determines what the resulting movement can mean.

CAIA Association includes valuation methods for private real estate in its Level II curriculum outline. The distinction matters in an exam answer because a prompt may ask for a benchmark, a risk estimate or a turning-point signal. Those are not interchangeable uses. The BIS discussion of commercial property prices also notes that transactions are sparse and can be unrepresentative, while appraisals may be infrequent and subjective. That pair of limitations should shape the comparison before any return calculation begins.

What enters an appraisal series

An appraisal-based series can include buildings that remained in a portfolio throughout the measurement period. The valuation is refreshed according to the index provider's process rather than whenever a sale occurs. That gives coverage of held assets that a sales-only sample could miss. It is especially useful when the question concerns the reported experience of a portfolio whose properties rarely trade. The cost is that many current values are estimates, and estimates can absorb new market information slowly.

Appraisers may rely on comparable transactions, income assumptions and earlier values. Those inputs are reasonable ways to estimate a price where no trade exists, yet comparables can be old and a property's prior mark can anchor a new one. A newly adverse market may therefore appear as several smaller reported adjustments. This is a timing issue, not proof that the underlying buildings were insulated from the shock. The BIS report describes the possibility that both appraisal and transaction indices can lag or smooth developments, depending on the data available.

Suppose a candidate sees a calm appraisal return next to a jumpier listed-property return. The wrong immediate conclusion is that the private assets are inherently safer or independent of markets. Reported variation is partly a property of observation frequency and valuation method. A risk estimate built from stale marks may give a portfolio model too much confidence about diversification. Yet the appropriate remedy in an answer is to identify the measurement limitation and seek corroborating evidence; it is not to declare every appraisal useless.

What enters a transaction series

A transaction index starts from completed trades. A closed sale provides a price at which that asset cleared, which can make a change in buyer appetite visible before appraisal marks for unsold holdings are revised. The price can be compared with a prior value or incorporated into a method that adjusts for differences among buildings. It is still a price for the sold sample, not an automatic valuation for the entire commercial-property stock.

Consider which assets can appear during a downturn. Owners with the option to wait may avoid selling, while an owner facing a financing deadline may accept a lower price. Alternatively, a prime building may trade while weaker buildings remain unsold. Either pattern can make the observed sale sample different from the properties held by a fund. Changes in building quality, lease terms, refurbishment and location also complicate a simple average of sale prices. The BIS report expressly warns that transaction records may lack the characteristics needed to adjust for these differences.

This is why a transaction series is often valuable as a price-discovery signal but requires questions about sample size and mix. An apparent early fall could reflect genuine market repricing, a changed set of sellers, or both. A candidate should test the claim against the index methodology and other evidence. The value of a fresh price does not remove selection bias. In the same way, the breadth of appraisal coverage does not remove valuation lag.

An illustrative decision with round numbers

Imagine an invented fund of ten equal-sized office buildings. At the previous valuation each was marked at 100 units, giving an aggregate mark of 1,000 units. During a weak quarter, one building sells for 85 units. Its own previous appraisal was 100 units. A rough sale-to-prior-mark comparison is therefore 85 divided by 100, or 0.85: a 15% discount to that old mark. That is a useful warning that at least one buyer and seller cleared well below an earlier valuation. It is not evidence that every unsold building now deserves precisely the same 15% markdown.

Suppose the nine unsold buildings are later marked down by 5 units each, while the sold building leaves the portfolio at its sale price. The illustrative aggregate value becomes 85 plus nine times 95, or 940 units, a fall of 60 units from the original 1,000. The fund's reported change would also depend on income, expenses, cash flows and the index's treatment of sold assets; this simple calculation isolates the valuation question only. The 15% sale comparison and the 6% aggregate change are both arithmetically sound but describe different populations and measurement dates.

Now change only the identity of the seller. If the traded building had a large vacancy and the other nine had secure long leases, applying its discount across the whole pool would exaggerate the portfolio effect. If it was the strongest building, its price could instead understate trouble elsewhere. The examination skill is to say which information is missing before transferring one observation to the whole index. The numbers here are invented teaching values, not market data or a forecast.

Choose the series for the task

For a current market-turn question, a transaction series can be a useful early check, provided there are enough comparable sales to make its movement interpretable. For an account of the marked value of a broad held portfolio, an appraisal series may better match the asset population, provided its delayed adjustment is acknowledged. For a risk or correlation study, compare observation methods and dates before using either series as if it were a direct measure of economic volatility. A sales series can be noisy; an appraisal series can be smooth. Those descriptions concern the signal, not a universal ranking of index quality.

The candidate can apply a compact sequence: identify the property population, identify how each observation is produced, ask when it enters the series, then state the bias most relevant to the requested decision. In a benchmark question, also ask whether income is included and how properties enter and leave the index. A price-only series and a total-return series cannot be compared by treating their headline changes as the same measure. Methodology comes before interpretation.

An existing explanation of appraisal smoothing addresses the statistical correction after slow marks have been identified. This article's prior choice is which observation should inform a specific question in the first place. The CAIA programme page offers the broader study context. Keep those steps separate: first diagnose the index's data, then decide whether a modelling adjustment is warranted.

Common exam misreadings

The first mistake is calling the appraisal index accurate because it covers more assets. Coverage helps representation, but it does not guarantee a timely price. The second is calling the transaction index accurate because it uses actual sales. A trade is real, yet the sold sample may be thin and selective. The third is interpreting lower reported appraisal volatility as an economic promise of lower risk. Timing and valuation conventions can suppress measured swings even when the buildings' potential sale prices have moved.

Finally, do not compare numbers whose definitions differ. Ask whether the observations concern the same market segment, quarter, property quality, income treatment and index population. If a prompt gives none of that detail, the strongest answer is conditional: say what each series is suited to reveal, what it can hide, and what further information would resolve the decision. That is more useful than a blanket instruction to prefer one type of index.