How Accurate are the Cost Estimates of Numbeo and Expatistan?
Whenever we look at places to stay for long-term destinations, one of the items we look at is cost. While we’d love to be the high-rolling ballers who rock up to Saint Tropez in our yachts, that’s not really how we roll. We’re fans of geographic arbitrage, where we can get the same or a better quality of living than we can when we’re back home in Texas (for example, when I first published this, the high in our hometown was 104 degrees, whereas it was 73 degrees in Istanbul…quality of life WIN!).
I’m a big fan of using the website Numbeo for getting an estimate of how much a place will cost compared to our baseline at home. For the purposes of this evaluation, I am also going to compare the cost of living as provided by Expatistan.
Methodology
If Numbeo gives out a number for a cost of living, it is very difficult to use that as a comparison. How you travel may be much different than how we travel. For example, our friends at Geekstreamers will spend more on lodging and cook in more, whereas we’d rather save a little on lodging and spend more on eating out. However, we do know what our cost of living is in Fort Worth and can create a baseline from the Numbeo estimate of cost of living and then apply that to the places that we have traveled.
So, let’s say that our cost of living, net of housing in Fort Worth is $5,000 per month. According to Numbeo, the cost of living in Fort Worth, Texas is $3,520.18 per month for a family of four. Yes, I know, we don’t have any kids, but the numbers for a family of four are much more in line with our numbers than for a single person. Therefore, our cost of living, in this example, is 142% of Numbeo’s estimate. In Istanbul, where we are now, Numbeo says that the cost of living for a family of four is $1,391.60, so, applying our 142% factor, we would expect to spend $1,976.60 net of housing.
I do not include housing or travel in the calculations. Travel is because there is no incorporation for travel in Numbeo. You live somewhere, so you don’t have to travel there. I do not include housing because we use AirBnB most of the time for our travel, and rent on AirBnB is going to be higher than what you would pay if you could find a local rental and rent for a standard lease term of 12 months. For example, rent in an apartment we’d consider living in back in Fort Worth is about $3,000 per month. According to Numbeo, rent is 74.63% cheaper in Istanbul than in Fort Worth, so we should pay $761.10 to get the same quality of apartment in Istanbul. We are paying more than that for our apartment here, and, while it’s nice, it’s not the same as what we would rent in Fort Worth. Therefore, due to the impermanence of our stays, I am not including lodging in these comparisons.
The Results
| City | Our spend | Numbeo Estimate | Expected spend | Our Results | Expatistan Estimate | Expected spend | Our Results |
|---|---|---|---|---|---|---|---|
| Cologne, Germany | $3,243.31 | 79.2% | $4,375.22 | 74.1% | 88.4% | $3,927.16 | 82.6% |
| Manchester, England | $3,099.24 | 80.7% | $4,454.88 | 69.6% | 116.7% | $5,182.12 | 59.8% |
| Istanbul, Turkey | $1,434.45 | 39.5% | $2,179.07 | 65.8% | 59.6% | $2,646.96 | 54.2% |
| Izmir, Turkey | $2,177.00 | 36.3% | $2,006.35 | 108.5% | 57.1% | $2,536.97 | 85.8% |
| Split, Croatia | $2,846.20 | 57.0% | $3,146.47 | 90.5% | 75.9% | $3,371.50 | 84.4% |
| Puerto Vallarta, Mexico | $2,005.82 | 46.5% | $2,571.01 | 78.0% | 49.3% | $2,187.95 | 91.7% |
| Medellín, Colombia | $2,812.01 | 40.6% | $2,241.56 | 125.4% | 50.7% | $2,249.48 | 125.0% |
| Playa del Carmen, Mexico | $3,464.65 | 43.6% | $2,405.43 | 144.0% | 47.3% | $2,099.27 | 165.0% |
There’s starting to be a pattern.
For us, the average of our spend compared to what Numbeo would expect is 94.5% and the average compared to Expatistan is 93.6%. The standard deviation for Numbeo is also lower: 0.29 compared to Expatistan’s 0.36. When I just use it for Europe (sample size alert!), the standard deviations drop again: Numbeo’s is 0.18 and Expatistan’s is 0.24.
Resultantly, I’m a little more inclined to trust Numbeo’s numbers than Expatistan’s, but it’s too early to make any definitive statement about that.
So, for now, my conclusion is that neither is more accurate than the other, and that they are mostly accurate, particularly for European travel. We’re not going to go wrong by orders of magnitude using either as a projection for spending, but would probably be wise to budget the top end, so perhaps 125% of what is projected for spending, given the results we’ve seen. It’s better to have a happy surprise and come in under budget than a negative surprise and come in way over budget.
What have you seen regarding these websites and their accuracy in your travels? Let’s talk about it in the comments below!