PTWhat makes a real estate fund perform well?
With a real estate fund, anyone can buy a slice of office towers, warehouses and shopping malls on the stock exchange. The thesis measured the efficiency of 102 of these funds from 2012 to 2023 (how much return they deliver for the risk they take) and tested a popular rule: picking funds by their dividend yield. What mattered most was the fund itself, not the economy. Explained in 12 screens.
A real estate fund splits properties into units traded on the exchange
The fund pools the money of many unit holders, and a manager invests it in properties or in securities tied to property. Unit holders earn in two ways: from the income paid out and from the rise in the unit price.
Sections 1 and 3.1. IFIX, B3’s index of real estate funds, has existed since 2012 and brings together the most traded funds. Beside, the three broad types the thesis uses in its third article.
Almost everything known came from outside Brazil
The author searched two large databases of scientific articles (Web of Science and Scopus) for studies on the performance of real estate funds and read the 23 that remained in depth. None used data from Latin America, and almost all used the same kind of statistics.
Sections 2.3 and 2.4. This led to the thesis’s two bets: measuring efficiency with another method (DEA, next screen) and looking at the type of property and the momentum effect, both little studied.
An efficient fund earns more for the same risk
The thesis used Data Envelopment Analysis (DEA): the best funds in each year draw a frontier, and each fund gets a score from 0 to 1 depending on its distance to it. Risk enters in two ways: the total swing of the unit price and the swing that follows the market (the basis of the Sharpe and Treynor ratios).
Data from Economática and B3, 2012 to 2023. The version of DEA used (with the Range Directional Model) accepts negative returns. When a data point was missing, the author used the fund’s own average.
Our illustration Made-up funds and numbers, with a single risk measure. In the thesis, the frontier uses both risk measures at once.
The average score stayed near 0.46 and only fell clearly in 2019
From one year to the next, the average efficiency of the 102 funds barely changed. The exception was 2019: the average fell to 0.406 and scores became far more spread out. In 2020, the year the Selic policy rate hit its low of 2%, came the best average of the series.
After adjusting the risk measures (the thesis calls it “input-oriented”), the average rose a little in every year, and the statistical test (paired t) found a significant difference in each one.
The market index swung far more than the average fund
The author computed the same efficiency score for IFIX, the index that sums up the real estate fund market. The index’s score rose and fell from one year to the next, while the funds’ average stayed almost still.
Table 3.5 also lists 2024 (score 0), outside the study period, which we do not show. The thesis does not explain why the index scores so far below the funds; compare the shape of the lines, not their level.
Office-building funds had the most similar scores
B3 sorts funds into eight asset types. The typical score changes little from one type to another; what changes is how much funds of the same type differ. For the author, office buildings and shopping malls show steadier management, with long contracts and large amounts. Tap a type.
Approximate values (~), read from Chart 3.5. The box runs from the first to the third quartile (the middle half of the funds); the line is the median. Isolated funds far below the boxes (some scoring near zero, mainly in securities, hybrid and logistics) are not shown.
The six that improved most and the six that worsened most, as the thesis lists them. Most of the 102 had a slope near zero (Chart 3.4). This is not investment advice.
Only one factor had a clear effect on the score
The author compared the efficiency score with eleven indicators, eight of the fund itself and three of the economy, in a model that compares each fund with itself over the years. Before looking, place your bets:
The rule tested: buy the funds whose dividend yield rose most
The dividend yield (DY) is what the fund paid in 12 months divided by the unit price. The idea of momentum says that whoever is doing well keeps doing well for a while. The author ranked funds by the change in DY and built portfolios with those at the top (“winners”) and those at the bottom (“losers”).
Equal-weight portfolios, rebuilt every month and compared with IFIX and with the Selic rate. Three strategies: buy the winners (W), sell the losers (L) or do both at once (WL). Brick: physical properties; paper: real estate credit securities (CRI); FOF: units of other funds.
In brick funds, the “losers” earned more than the “winners”
If momentum held, the winner portfolios should earn more. In brick funds, the opposite happened in every window: the losers had a positive and growing average return (up to 1.47% in the 12-month window) and the winners, a negative one.
Our reading In brick funds, those whose DY rose least (or fell) earned more, which looks like a return to the average rather than momentum. The author reads the result as the “recovery capacity” of brick funds.
None of the 45 strategies passed the statistical test
Three strategies, three fund types and five windows give 45 combinations. The author tested whether each one’s return beat IFIX. In none was the difference large enough to rule out chance: the smallest p-value was 0.325, far from the usual 0.05.
The most promising was buying paper winners with a 12-month window (+1.2% on average), which the author sees as a “trend” without significance. For her, splitting funds by type may have diluted the effect, and the recent market may be pricing DY better.
Look at the fund and its property type before the dividend
For managers
Efficiency depended more on internal factors (the asset type and risk management) than on interest rates and inflation. Heavily traded funds had lower scores; the author suggests that a stable base of unit holders favors long-term strategies.
For investors
Compare funds by risk-adjusted return and within the same property type. The change in dividend yield, on its own, did not guarantee gains in the period studied. This page is not investment advice.
For researchers
Study how different forms of management affect efficiency, include property characteristics (location, size, profile of the region) and use longer series that cover full cycles.
What this research does not say
- Association, not cause. That more heavily traded funds have lower scores does not prove that trading lowers efficiency. · our observation
- Only the past. Historical data do not foresee crises or extraordinary events; management and long-term strategy are hard to measure. · according to the author
- The properties were left out. Location, size and the profile of the region were not part of the score. · according to the author
- Missing data filled with averages. Replacing missing values with the fund’s average smooths the series and may hide swings. · according to the author
- Numbers to check. The text places the drop in efficiency in 2016, but the tables show it in 2019; we follow the tables. In article 3, the thesis does not say whether portfolios use the 10%, 20% or 30% cut of funds. · our observation
Credits
CUNHA, Talita Ferreira. Performance financeira de fundos de investimentos imobiliários brasileiros: estratégias e modelos de avaliação. Doctoral thesis (Doctorate in Business Administration) – Universidade Presbiteriana Mackenzie, São Paulo, 2025. Advisor: Prof. Dr. Leonardo Fernando Cruz Basso.
Research supported by a scholarship from Caixa Econômica Federal, a Brazilian federal bank. Data from Economática, Comdinheiro and B3. This version simplifies the work for a general audience; the full tables are in the original text.