Real Estate Investment Indexes: Construction and Potential Biases

Real estate indexes do not all measure the same thing. An appraisal-based index, a transaction-based index, and a public-market index can report different returns and different volatility for exposure to the same underlying asset class. CFA Level II tests whether you can identify what each index actually measures before comparing their risk or return figures.

Key Takeaways

• Appraisal-based indexes rely on periodic professional valuations, which lag transaction prices and smooth reported returns.

• Smoothing lowers measured volatility and reduces measured correlation with other asset classes, which can overstate the apparent diversification benefit of real estate.

• Transaction-based indexes use actual sale prices, which reduces smoothing but relies on a non-random sample of properties that sold during the period.

• Public-market (securities) indexes, such as listed REIT indexes, price continuously but reflect leverage and stock market sentiment, not only the underlying property.

• Survivorship bias occurs when failed properties or funds drop out of the sample, which overstates historical index performance.

• No single index type is free of bias; each trades one measurement problem for another.

What You Need to Know for CFA Level II

• The LOS requires you to discuss construction and bias, not to calculate an unsmoothing adjustment.

• Know the three index types by pricing source: appraisal, transaction, and public securities.

• Be able to explain why low reported volatility in appraisal-based indexes does not mean low true economic risk.

• Understand that transaction-based indexes reduce appraisal anchoring but can retain reporting delays and sample-selection bias.

• Recognize that comparing leveraged public real estate returns with unlevered private property returns requires an adjustment for leverage before any risk comparison is valid.

• Treat survivorship bias as a separate issue from smoothing and lag.

Compare index measurement methods

Real estate index construction starts with a basic question: what price is the index actually tracking? There are three common answers.

Appraisal-Based Indexes

Appraisal-based indexes use periodic professional valuations of a pool of properties, typically those held by institutional investors such as pension fund real estate accounts.

Valuations occur quarterly or annually, not continuously. Appraisers often reference recent comparable sales and the prior appraisal value, which anchors new valuations to old ones.

Transaction-Based Indexes

Transaction-based indexes use actual closing prices from properties that sold during the measurement period.

The index only includes a property when a sale occurs, so coverage depends on market activity.

Public-Market Indexes

Public-market indexes track listed securities, such as real estate investment trusts (REITs), that trade on an exchange.

Prices update continuously during market hours, driven by the same investors and sentiment that price other equities.

Two distinctions matter most for Level II.

First, appraisal-based and transaction-based indexes track the direct property market. Public-market indexes track a listed security that is a claim on property plus corporate leverage and management decisions.

Second, listed-security indexes are generally easier to track using tradable securities or funds. An index itself is a benchmark; accessing private-property index exposure can require specialized vehicles.

Candidates should avoid treating smoothed reported risk from an appraisal-based index as the true economic risk of holding real estate. The smoothing process itself, not the underlying asset, produces the low reported variance.

Explain smoothing, lag, and sample bias

Each index type solves one measurement problem and creates another.

Smoothing and Lag in Appraisal-Based Indexes

Appraisers tend to anchor new valuations to the prior reported value, adjusting gradually toward new market information rather than repricing immediately.

This behavior, common when comparable transaction evidence is limited, produces a series that moves more slowly than the true market. The statistical result is smoothing: reported returns show lower period-to-period variance than actual economic returns, and reported correlation with other asset classes, such as public equity, appears lower than it truly is.

A portfolio that allocates to real estate based on this smoothed correlation may underestimate its true exposure to broad market risk.

Sample-Selection Bias in Transaction-Based Indexes

Using actual sale prices removes appraiser anchoring, but a transaction-based index only reflects properties that sold. Sellers are not a random sample of the market.

In a downturn, owners of distressed or lower-quality properties may be forced to sell, while owners of stable properties hold and avoid transacting. In a strong market, premium properties may trade more actively as buyers compete for them.

Either pattern biases the measured index return away from the return on the full property universe, not just the subset that transacted.

Survivorship Bias

When an index drops funds or properties that fail, liquidate, or stop reporting, the remaining sample looks stronger than the original population.

This can inflate historical average returns and understate historical volatility, independent of smoothing or selection effects.

Public-Market Index Limitations

Listed real estate securities trade with the leverage, governance, and liquidity characteristics of the issuing entity. Measured volatility reflects corporate capital structure and equity market sentiment, not only property-level fundamentals.

Comparing an unadjusted leveraged REIT return with an unlevered appraisal-based return combines differences in leverage, valuation timing, and investment structure.

Those differences must be separated before attributing the gap to property-market risk alone.

Candidates should avoid assuming that a transaction-based index eliminates selection bias simply because it uses real transaction prices instead of appraisals. It reduces appraisal anchoring, while reporting delays and transaction-selection issues can remain.

Working Example

A pension fund compares an appraisal-based property index with a listed real estate index. To isolate recognition timing, assume the underlying property portfolio suffers a 15% economic value decline at the start of Quarter I and then remains unchanged.

For this illustration only, assume the listed index also falls 15% immediately and remains unchanged thereafter. Matching that decline is an explicit assumption, not a general prediction for a leveraged REIT index.

Step 1: Compare immediate recognition.

The listed series reflects the assumed shock in Quarter I. The appraisal series initially reports only a 3% decline because valuations incorporate information gradually.

Step 2: Compare cumulative results.

Both series finish at 85 from a starting value of 100, a cumulative decline of 15%. The appraisal index spreads recognition across several observations. Its quarterly percentage returns must be compounded; index-point changes are not interchangeable with percentage returns.

Step 3: Interpret measured risk.

The smoother appraisal path can show lower contemporaneous volatility and correlation even though the assumed economic loss occurred at the start. That measurement effect does not remove the pension fund's exposure.

The example isolates timing. A real listed-versus-private comparison also needs adjustments for leverage, property mix, financing, fees, and other corporate exposures. A listed index decline alone does not establish the exact simultaneous change in every private property's value.

Common Exam Traps

Treating smoothed reported risk as true economic risk

Appraisal-based indexes understate volatility because of valuation lag, not because the underlying property market is actually calm. Always ask whether the index price source updates continuously or periodically before comparing risk statistics across index types.

Assuming transaction-based indexes eliminate selection bias

A transaction-based index reduces appraiser anchoring but still depends on which properties happen to sell. Properties that transact are not a random sample of the full property population, so the index return can diverge from the broader market return.

Comparing leveraged listed returns with unlevered property returns without adjustment

Public-market real estate securities carry entity-level leverage. Comparing their volatility directly with an unlevered, property-level appraisal-based or transaction-based index does not isolate the difference in underlying property risk.

Practice Question

A real estate analyst observes that after an unexpected rise in interest rates, a listed REIT index fell 12 percent within the same week. An appraisal-based index tracking similar properties reported a decline of only 2 percent for the full quarter in which the rate increase occurred. Most properties in the appraisal-based sample were not revalued after the rate increase until the following quarter.

Which of the following is the most likely explanation for the difference in reported results?

1. The appraisal-based index proves that the underlying properties carry little price risk from interest rate changes.

2. The appraisal-based index's transaction-based construction eliminates sample-selection bias relative to the REIT index.

3. The appraisal-based index understates contemporaneous volatility because valuation lag delays recognition of the price change.

FAQs About Real Estate Investment Indexes: Construction and Potential Biases