Fundhouse Insights Latest Articles

How rental property compares as an asset class

Published: December 31, 2019 by Peter Foster, Fundhouse

Buy-to-let or rental property plays an important part in many investors’ portfolios, but it is inherently complicated to evaluate the returns and costs relative to other asset classes. In addition, while listed instruments are accessed through unit trusts and controlled by various intermediaries[1], rental property is typically controlled by investors themselves. As a result, investors often lack clarity into the relative performance of their investments and whether it is the best investment vehicle for achieving their goals.

The appeal of rental property is easy to understand. Rental property is tangible and seems more controllable to investors than a share listed on an exchange. Property values and rentals have escalated dramatically over the last 20 years, anchoring investor expectations on future performance. Listed growth assets like shares and REITs[2] have delivered poor returns recently which entices investors to look elsewhere for growth and income. Direct property investments can be financed through a mortgage, while listed investments cannot. At the same time, lending rates are at favourable levels compared to history. Finally, certain rental expenses (like maintenance or the interest portion on the mortgage) are tax deductible, which makes rental property relatively tax efficient under certain conditions.

The total costs and returns of rental property are opaque and complicated to calculate, which makes it difficult to compare accurately to listed asset classes. This often means that investors compare “apples to oranges” when building return expectations from rental property relative to other investments. So, comparing “apples with apples”, what reasonable expectations can we develop around rental property to fairly compare it to other asset classes? We are primarily focussing on residential rental property here, which would include flats, townhouses and freestanding houses. The house price data used throughout the commentary is primarily sourced from the Absa House Price Indices, which recorded house price growth on loans they extended to consumers from 1967-2016 across small, medium and large residential properties. The indices were discontinued in 2016, after which we used annualised data from the Lightstone Property Report. The Lightstone indices are calculated a bit differently but tell the same story directionally.

We can start by gaining some perspective on historical growth rates relative to other asset classes. Property prices have closely kept up with inflation over the last 5 years, as the table below demonstrates. The data shows that despite recent weakness in local house prices, the asset class has managed to outperform the price returns on other local risk assets over the last five years. This is an important point to highlight because as investors we often interpret our recent experiences as being repeatable going forward[3]. As the table shows, this is not a reasonable expectation over long periods of time, where house prices have underperformed share prices by around 1.5% per year for 50 years, and by over 2% per year for the last 40 years.

Table 1: Inflation adjusted[4] price returns over last 50 years (per year)

  SA Equities* SA Listed Property* House Prices**
1yr 5.7% -9.6% -0.5%
3yr -1.1% -13.6% -0.8%
5yr -2.9% -9.9% -0.3%
30yr 3.6% N/a*** 2.8%
40yr 3.8% N/a*** 1.7%
50yr 2.5% N/a*** 0.9%

*Source: FE Analytics
**Source: Thomson Reuters Datastream (1967-2016), Lightstone Residential Property Indices (2016+)
***Asset class did not exist

Property prices are often seen as being an inflation hedge, which means that it keeps up with inflation over time. While this has been the case as Table 1 shows, where house prices have generally grown in excess of inflation by 0.9% per year over 50 years, house prices don’t necessarily move in line with inflation consistently. The grey areas on Chart 1 below demonstrate that property prices can underperform inflation for long periods of time (i.e. is not necessarily an inflation hedge), as it had for 13 consecutive years between 1985 and 1998. The chart also shows that the 2000-2008 period of rapid growth in property prices is anomalous by historical standards and is certainly not the norm.

Chart 1: Inflation adjusted 5yr annualised house price growth

Similarly, investors would expect that rental income is protected and that inflation can be passed on to renters through rental increases, but this is also not possible all the time. The data shows that only landlords in the Western Cape, the Free State and owners of flats in general were able to pass inflation on to renters over the last 10 years. As a result, average rental property investors have seen their inflation adjusted incomes decline for more than a decade. Granted, this time period is short and is confined to a difficult economic climate. Longer term data from the US shows that rentals do generally keep up with inflation over time. The point we are making is that, as with property prices, rentals can deviate from inflation for extended periods of time.

The important points to take away from the discussion thus far are that property prices tend to lag share prices over long periods; property prices and rental income don’t necessarily move in line with inflation consistently; and that we can’t treat the 2000-2008 property bull market as a normal environment.

Given these learnings, what are reasonable expectations for rental property over the long term, relative to alternative asset classes? To answer this, we evaluated the returns, costs, fees and taxes on a total return[5] basis and built reasonable and fair assumptions for each of these components. We did this for rental property, local equities, local listed property, local bonds, local cash and for a typical balanced fund. The most important assumptions relating to rental property are shown in Table 2 below:

Table 2: Investment assumptions

Assumption Reasoning
Investment Type Discretionary (i.e. not retirement money) that is therefore fully taxable.
Mortgage % of Home Price We assumed that the property will be mortgaged at the maximum level that ensures the investment still breaks even on a cash flow basis, i.e. where rental income exceeds mortgage costs (41% in this case). To make the investment comparable to other asset classes, we must assume that the investor cannot allocate additional capital over the period.
Rental Growth We have assumed that rental growth will tend towards inflation over long periods of time.
House Price Growth Here we have used the average house price growth over the last 50 years, or 0.9% p.a. over inflation. Simplistically, it would make sense for house prices to increase in line with inflation instead. However, there are factors like urbanisation, scarcity of land, etc, which suggest that above inflation increases in house prices over time are not unrealistic.
Reinvestment Rate Rental income cannot be reinvested in the property like interest on a fixed deposit or dividends on a share can be reinvested. To compensate, we have assumed that any profits be reinvested into a typical balanced fund.
Tax Rate The assumed marginal tax rate used is 45%
Investment Horizon 20 years (or typical mortgage period)

Based on the primary assumptions above (and others not discussed here), we have assessed the reasonable after cost, fee and tax expected total returns of the various asset classes as shown in the charts below. For a fair comparison between the asset classes, we have broken down the total return for each asset class into its various return and cost components.

Chart 3: Components of expected annual returns per asset class[6]

From a net total return perspective, we would expect equities to deliver the highest net returns over time, returning in excess of 1% more per year than rental property. Nonetheless, rental property does present a compelling investment opportunity to investors where we would expect it to outperform listed property, bonds, cash and a typical balanced fund on a normalised basis. The notion of a normalised environment is important because, as we have seen, the market can be driven to extremities which can persist for longer than investors care to remain invested.

As demonstrated in the charts, gross returns generated by rental property are expected to be significantly higher than for equities (primarily due to leverage, which is discussed later), and this is usually the figure that investors hold on to. However, the high overall costs of rental property are a material detractor from gross returns, and the full scope of the costs is often overlooked when calculating and comparing returns to the other asset classes. A proper assessment of costs goes much further than just rates and taxes on the property, and should include the initial purchasing costs like transfer, legal and realtor fees, ongoing maintenance, periodic renovations, recurring vacancies, the mortgage repayment, etc. In aggregate, a rental property investor will pay away over half of their gross returns to costs, fees and taxes over time, which is in line with the tax inefficient income generating asset classes, like bonds and cash.

Table 3: Overall cost ratio[7]

  Rental Property SA Equity SA Listed Property SA Bonds SA Cash SA Balanced
Overall Cost % 54% 35% 39% 56% 52% 38%

As touched on before, a net benefit of rental property is that the investment is typically financed with borrowed capital (i.e. the property is leveraged), which is not readily available to other asset classes[8]. The extent to which a rental property can be leveraged, however, is limited to the ability of the investor to cover the mortgage payments. In this analysis we assumed that the investment must pay for itself, and that the property can only be leveraged to the extent that the net rental covers the mortgage payment (41% in this case). Leveraging enhances the returns on rental property, but it also enhances the costs, resulting in an overall net benefit of around 0.5% per year (8.18% net total return on the leveraged property versus 7.67% on the unleveraged property). There is no real benefit in investing in rental property over equities or listed property if the rental property is unleveraged, and only a marginal benefit over investing in a typical balanced fund.

Having developed a return expectation for rental property, we can evaluate the risks that investors need to consider that are specific to rental property. The first such risk is related to the leveraging of the property. Interest rates typically change quicker than landlords can adjust rents, which means that financing costs can exceed rental income if interest rates rise. Ultimately, rental property can cost investors hard cash if market conditions change, which is not something all investors are comfortable with.

Secondly, rental property is acutely exposed to liquidity risk. The underlying instruments in unit trusts can have liquidity events which can have an adverse impact on performance, but a unit trust is liquid by design. If an investor decides to sell a unit trust, they are largely guaranteed the day’s closing price and should receive the proceeds from the sale three days later. This is not the case with rental property, where settlement averages around 14 to 22 weeks. As is true with all asset classes, liquidity typically dries up when investors need it most. The situation then becomes more precarious for rental property investors because it takes much longer to unwind an unprofitable or unsustainable investment, while at the same time having to cover a mortgage liability.

A third specific risk is that physical property tends to have longer market cycles than other asset classes, which means that the entry point is incredibly important to avoid locking in long term underperformance. As an example, inflation adjusted house prices only exceeded their 1984 values in 2004 – after a period of 20 years. House prices are currently still about 12.5% below the 2007 peak in real terms. Chart 4 below shows the drawdown[9] in inflation adjusted property prices relative to equities. While equities are much more volatile, they tend to recover quicker and have shorter cycles. The point is that a property investment at the wrong time can be wrong for a long time.

Chart 4: Inflation adjusted price drawdowns[8]

We have established that rental property can be a compelling investment opportunity relative to other classes, but only if leveraged, otherwise there is little benefit in owning rental property over a typical balanced fund. Leveraging introduces additional risks which, in combination with low levels of liquidity and long market cycles makes the importance of getting the entry point right immensely important.

Rental property is often seen as a relatively simple way to make money: Buy a property using debt and have renters pay off the loan. But as we have seen, developing an understanding of what to buy and when to buy it requires a significant amount of skill, which investors will have to demonstrate themselves as opposed to the comfort of relying on intermediaries, as is the custom in listed investments. Given the risks and time commitments to manage rental property for a marginal return benefit over other asset classes means that rental property is probably not suitable to all investors. In any event, exposure to the asset class should always be considered as a component of a well-diversified portfolio.



[1] Intermediaries include fund managers and financial advisers.

[2] REITs, or Real Estate Investment Trusts, are listed property holding companies, like Growthpoint or Redefine.

[3] This is known as “recency bias”

[4] Normal returns minus inflation

[5] Total return means we considered capital growth and the income generated

[6] “Costs” for rental property includes the mortgage payment. “Fees” include fund management, advice and platform fees where applicable.

[7] Costs, fees and taxes as a % of gross total return.

[8] There are ways to leverage the returns on listed instruments, using a variety of tools, but these are not typically used by most investors.

[9] Drawdown is the % decline from the previous peak.