How we measure the Venezuelan housing market

Every figure is computed from the property listings published on Propiedash that are active at the moment of calculation. We do not buy data, we do not merge in third-party feeds, and we do not model. If a listing is not on Propiedash, it is not in these numbers.

That is the strength and the limit of this record at once: the sample is our own inventory, which we can describe exactly, and which is not the whole country.

As of the current edition we hold 19,879 distinct active properties. Of those, 10,497 are homes for sale that pass every check below and form the sale aggregates. A property offered both for sale and for rent is counted on both sides, which is why the sale and rent counts add to more than the total.

A city, or a property type, needs at least 3 valid distinct properties before it gets its own published figure. Below that we publish the real count and the words "sample too small" instead of a median.

Venezuelan housing is priced and advertised in United States dollars, and that is how it is stored and reported here. No bolivar conversion is applied, so no exchange rate enters these figures and none of them move because the rate moved.

The median is the headline metric everywhere, because a handful of very large or very cheap listings would drag an average away from what a typical home costs. Averages appear only as secondary context, clearly labelled.

One vote per property. The same home is often published several times, by several agents of one agency or by different agencies. Every figure counts it once: its listings are grouped under the identity our property registry gives them (or their cadastral record, when there is one), the group contributes the median of its own asking prices, and only then is the city, property type or national median taken. The counts printed beside each figure are distinct properties, not listings.

Price per square metre is each listing's price divided by its declared area, and the reported figure is the median of those values, not the national price divided by the national area. Only listings that declare an area contribute, so the square-metre sample is smaller than the price sample and is printed separately.

Gross rental yield is the median annualised rent divided by the median sale price of the same city. It is gross: before maintenance, taxes, agency fees and empty months.

Only homes. Commercial locals, offices, warehouses, hangars, whole buildings, land, farms and parking spaces are excluded from every figure that carries the word housing. The types that count as a dwelling are: apartamento, casa, townhouse, penthouse, anexo, habitacion.

This matters more than it sounds. Before we applied it, 18 of the 63 cities we published held fewer than three residential listings, so their "housing price" was computed from commercial stock. One coastal city published a median of 7.450.000 dollars off two 11.000 square metre commercial units, a hotel and a plot of land.

Sanity bands. A sale price outside 1,000 to 50,000,000 dollars, or a price per square metre outside 50 to 15,000 dollars, is treated as a data-entry error and dropped rather than silently corrected.

These are asking prices, not closed transactions. Venezuela has no public register of housing transactions that anyone can read, so every price index in this market measures what sellers ask. Asking prices lead transaction prices and usually sit above them.

The sample is concentrated. Caracas is the largest share of it by a wide margin, and a national figure therefore leans on one metropolitan area. The share is printed beside the national figure on every edition.

We do not yet publish change over time. Building a defensible series takes consecutive editions computed the same way, and we would rather publish no trend than one we cannot stand behind.

Coverage is uneven. Some states hold very few listings. Where a place has too little to say, the record says so.