Does Deregulation Work? Comparison Between Spanish and Italian Average Taxi Tariffs
Author
Mario Ignazio Tateo
Abstract
This study analyzes the structural differences between average taxi tariffs in Italy and Spain to verify the existence of statistically and economically significant discrepancies. Using data extracted from the Numbeo platform for a sample of 22 Italian cities and 13 Spanish cities, nominal tariffs were adjusted for the local cost of living to obtain Purchasing Power Parity (PPP) values. Descriptive and inferential analysis, conducted using a Student’s t-test for independent and homoscedastic populations, reveals a strong and significant difference between the two countries (p-value ~ 0.001). The findings show that Italian tariffs are systematically and markedly higher than Spanish ones, suggesting that differing regulatory choices and their respective national normative frameworks may exert a decisive impact on the final price for the consumer.
Methodological Note
The objective of this analysis is to demonstrate a potential significant difference between average taxi tariffs in Spain and Italy. To do so, taxi tariff data were extracted from the Numbeo platform for 22 Italian cities and 13 Spanish cities. These data are subsequently used to calculate tariffs adjusted for the cost of living in each city (100 = cost of living in New York City). These adjusted values represent the ultimate object of the analysis, which consists of a Student’s t-test evaluated on two normal, independent, and homoscedastic populations.
Tariffs and Cost of Living Index in Italian Cities
City
Average Taxi Tariff
Cost of Living Index
Bari
5.00
57.4
Bergamo
9.20
58.1
Bologna
4.10
68.8
Brescia
8.50
71.4
Cagliari
10.00
63.2
Catania
6.50
55.4
Florence
3.30
72.3
Genoa
6.00
63.4
Milan
6.90
75.8
Modena
5.90
66.2
Naples
4.00
60.2
Padova
10.00
65.9
Palermo
3.00
56.0
Pisa
5.00
71.8
Reggio Nell’emilia
7.50
61.4
Rimini
4.80
59.5
Rome
6.00
61.2
Trento
7.00
65.6
Treviso
8.00
65.2
Trieste
7.75
65.7
Turin
3.50
63.7
Verona
7.50
67.4
Tariffs and Cost of Living Index in Spanish Cities
City
Average Taxi Tariff
Cost of Living Index
Alicante
2.10
52.9
Barcelona
3.00
61.1
Bilbao
4.90
55.9
Granada
5.00
48.9
Las Palmas de Gran Canaria
3.00
51.8
Madrid
3.35
60.3
Malaga
1.68
52.6
Oviedo
4.71
57.3
Palma de Mallorca
4.30
59.9
Seville
1.52
49.5
Tarragona
3.50
53.4
Valencia
4.00
52.8
Valladolid
4.00
61.8
Results and Discussion
The first step to undertake is adjusting the taxi tariffs to the cost of living of each individual city. To this end, it is necessary to compute the ratio between the average tariff of a single city and its cost of living index, multiplying the result by 100:
Once the adjusted tariffs are obtained, we can proceed with a preliminary analysis:
Summary Statistics - Italy:
Min. 1st Qu. Median Mean 3rd Qu. Max.
4.564 7.240 9.634 9.891 11.878 15.835
Mean and Standard Deviation - Italy:
Mean SD
9.890614 3.323919
Summary Statistics - Spain:
Min. 1st Qu. Median Mean 3rd Qu. Max.
3.071 4.910 6.472 6.268 7.576 10.225
Mean and Standard Deviation - Spain:
Mean SD
6.267931 2.158362
As evident from the exploratory data analysis, there is a marked difference between the Spanish and Italian average and median tariffs.
The side-by-side box-plot shows a net difference of 3.162 PPP Euros between the Spanish median tariff and the Italian one. This reinforces the hypothesis that the regulatory framework in the Iberian nation favors more affordable tariffs for final consumers.
For a formal analysis, we must perform a Student’s t-test, first ensuring that the variables satisfy the assumptions of normality and homoscedasticity:
Shapiro-Wilk normality test
data: tariffe_ppp_ita
W = 0.95613, p-value = 0.4153
Shapiro-Wilk normality test
data: tariffe_ppp_spa
W = 0.97583, p-value = 0.9528
We can accept the null hypothesis of normality for both variables, with a p-value for Italy and Spain of 0.4153 and 0.9528 respectively.
F test to compare two variances
data: tariffe_ppp_ita and tariffe_ppp_spa
F = 2.3717, num df = 21, denom df = 12, p-value = 0.1254
alternative hypothesis: true ratio of variances is not equal to 1
95 percent confidence interval:
0.7756958 6.2534994
sample estimates:
ratio of variances
2.371659
With a p-value of 0.1254, we accept the null hypothesis of homoscedasticity.
Two Sample t-test
data: tariffe_ppp_ita and tariffe_ppp_spa
t = 3.5059, df = 33, p-value = 0.001334
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
1.520403 5.724962
sample estimates:
mean of x mean of y
9.890614 6.267931
We strongly reject the null hypothesis H0 of equal means, with a p-value of ~0.001.
This paper compared taxi tariffs in Italy and Spain, isolating the effect of local price levels by calculating purchasing power parity (PPP) adjusted tariffs.
The initial exploratory analysis highlighted a clear visual and numerical discrepancy between the two markets, which was subsequently confirmed rigorously through inferential statistics. After verifying the assumptions of normality (via the Shapiro-Wilk test) and homoscedasticity (via Fisher’s F-test), the Student’s t-test indicated a strong rejection of the null hypothesis of equal means with a highly significant p-value ~ 0.001. The Italian median tariff stands well above the Spanish one, registering a net gap of 3.162 PPP points.
These empirical findings support the initial hypothesis that the regulatory framework and local governance models play a crucial role. The Spanish market, historically characterized by greater flexibility or structural reforms oriented toward competitiveness in non-scheduled public transport, results in significantly more affordable and accessible tariffs for the end user compared to the Italian system.
In conclusion, this study demonstrates that the variability of taxi prices between Italy and Spain is not merely a reflection of the overall cost of living differences across individual cities, but rather the effect of regulatory asymmetries and internal competitive dynamics within the two countries. Future research developments could include explanatory variables in the model regarding license density per capita and the impact of alternative mobility services (e.g., Uber/Cabify) to map the determinants of this gap in even greater detail.
