Market Analysis

Why Every Addis Ababa Price Estimate Online Disagrees — and How to Read Them

September 26, 2026
41 views
addis ababa property price estimate, ethiopia property market data, how to value property addis ababa, addis ababa price per sqm analysis
Why Every Addis Ababa Price Estimate Online Disagrees — and How to Read Them

Four sources, four different medians

Search for the median property price in Addis Ababa and you will find, from credible operators:

SourceMedian home priceStated basis
TheAfricanVestor~ETB 14 millionAll residential, median; notes average ~ETB 20m pulled up by villas
EthiopiaRealty~ETB 16 millionResidential, with 80% between ETB 10m and 35m
LivingEthio~ETB 24 millionBroader property set, median
EthiopiaPropertyCentre~ETB 36.3 millionHouses only, not apartments; +32% YoY

That is a spread of more than 2.5×. It is not because some of these sources are wrong. It is because they are measuring different things — and almost nobody tells you which.

Reason 1: Houses versus apartments

This single distinction explains the largest gap. In the same market, standalone houses average around ETB 36 million while apartments average ETB 13.5 million and older condominiums ETB 6.5 million.

Any source reporting "houses for sale" is measuring a completely different market from one reporting "apartments for sale". EthiopiaPropertyCentre's ETB 36.3m is a house median. Comparing it to an apartment median is a category error that is easy to make and very easy to publish.

Reason 2: Mean versus median

The Addis market has a long right tail — villas at ETB 65m, penthouses around ETB 95m, diplomatic-grade units well above. A small number of very expensive properties drags an average far above the typical transaction.

For almost any buyer, the median is the number that describes their experience. If a source leads with an average, ask for the median. The gap between the two is the tail of the market, not the market itself.

Reason 3: Asking prices versus what homes actually sell for

Published figures are usually listing prices, because closed-sale data in Ethiopia is not public. Ethiopia does not publish a housing price index, so every estimate is built from asking prices.

The gap is measurable. Sources put real sale prices 6–10% below asking, with 8% a reasonable central estimate, and 88–95% of homes selling below their original asking price. The gap is smallest for scarce prime apartments in Bole or Kazanchis, and largest for expensive villas and stale luxury stock.

So a source reporting median asking price of ETB 14 million is describing a market where the typical completed sale is closer to ETB 12.9 million. Neither number is dishonest. Only one of them is what you will pay.

Reason 4: Sample coverage and what gets listed

Every portal's figures are bounded by what its users chose to list. Owner-direct listings differ from agent-listed stock, which differs from developer inventory. Formal, documented stock is systematically under-represented relative to informal and cash transactions — which biases upward.

This is also why the most rigorous sources publish their sample size and label their own confidence. A figure based on 300 listings and a figure based on 3,000 are not equally reliable, and most pages do not tell you which you are reading.

Reason 5: Gross versus net yield

The same trap applies to rental yields. You will see 8–12% quoted as "rental yield in Addis Ababa" and 5–7% quoted as "rental yield in Addis Ababa". Both are correct. One is before vacancy, management, maintenance, service charges and tax. The other is after.

When a source quotes a yield, establish which one it is before comparing it to anything.

Reason 6: A statistically honest estimator will refuse to answer

The most useful thing a valuation tool can do is say "I don't have enough data for this".

When we built our own apartment price estimator, we backtested it by holding out each apartment in turn and asking the model to price it without seeing it. The result: median error 10.8%, mean 14.9%, and 77% of real asking prices falling inside the published range. The model also refuses roughly 1% of cases as unsupportable.

That refusal rate matters more than the headline accuracy. Look at the actual distribution in Addis Ababa: Bole alone holds about 40% of apartment stock, Megenagna has around 24 apartments, Kazanchise around 22, CMC around 17, Sarbet around 17, and roughly fifteen other areas have between one and seven apartments each.

You cannot produce a defensible price per square metre for an area with three listings. Any site that confidently quotes one is interpolating, not estimating. A tool that declines is more trustworthy than a tool that guesses.

How to read any Addis property estimate

  1. Check the property type. Houses-only figures tell you nothing about an apartment.
  2. Prefer the median over the average, always.
  3. Subtract about 8% from any asking-price figure to approximate a completed sale.
  4. Ask for the sample size and the date. A figure with no date is not an estimate, it is an impression.
  5. Check whether the area is one where the data exists. If nobody publishes a per-m² figure for your sub-city, treat any number you are given as an opinion.
  6. Distinguish gross from net on any yield figure.
  7. Compare price per m², never total price. It is the only figure that is actually comparable across listings.
  8. Treat the zonal valuation as a separate number from the market price — that is the figure your transfer tax is assessed on.

What good data looks like

At its best, an Addis property estimate should tell you: the property type it covers, whether it is asking or closed prices, the sample size, the date, the neighbourhood, and how often it declines to answer. If it gives you a single confident number with none of that, you are being sold a figure, not given an estimate.

For current benchmarks by neighbourhood, see Addis Ababa apartment prices in 2026, and for income-focused analysis, rental yields by unit size.

Sources

  • TheAfricanVestor, EthiopiaRealty, LivingEthio, EthiopiaPropertyCentre — the differing published figures analysed above
  • IMF World Economic Outlook, National Bank of Ethiopia — inflation and currency context
  • Leave-one-out backtest, 298 apartments — estimator accuracy, coverage and refusal rates

This article exists because the numbers on this page genuinely conflict. Where sources disagree, we have said so rather than picking the one that suits us.

Related Articles

Discover more insights and stories

Sell Your Property