The liberating effect of killing off a state-owned rival

When the final flight of the Hungarian national flag carrier, Malev landed in Budapest there were widespread concerns about the European Commission’s decision to declare Government payments to the company as illegal state aid, leading inevitably to the company’s shut down. Four and a half years on, evidence appears to show that shutting down the incumbent state-owned fat cat was probably the best thing that could happen to Hungarian passenger air transport. 

For most of its history Malev was a state-owned company. After many years of unsuccessful privatisation attempts the company was sold to Russian owned AirBridge (later AirUnion) in 2007. The years of private ownership were not easy for Malev. It never managed to become profitable enough to survive without continued life-support by the Government. In 2010 the airline was re-nationalised, and at around the same time the European Commission started an investigation into allegations of illegal state-aid paid to Malev between 2007 and 2010. As a New Year surprise, on 9 January 2012 the European Commission declared Government payments to Malev as state aid and ordered Hungary to recover around EUR 130 million from the airlines. This was too much for the financially ailing company that decided to drop the curtain on 3 February 2012.

At the time of the decision there were wide-spread concerns in Hungary that tourism might suffer as a result of eliminating the Hungarian flag carrier, and that other airports in the region will be able to benefit, diminishing the regional significance of Budapest Airport – and Budapest itself. On the other hand, competition lawyers and economists had often expressed concerns about the level playing field in markets where state-owned enterprises are present, and about the efficiency of state-owned enterprises. It is fairly common that state-owned enterprises enjoy benefits that their private competitors don’t. It wasn’t different in the case of Malev. Until 2007 the airline was under the same ownership as Budapest Airport, which, even in the absence of explicit preferential treatment would have been enough to keep competitors out of the market. The other problem was that with state-ownership Malev dragged on a corporate structure, which proved to be an ungainly fossil in the highly competitive passenger air transport market. The elimination of Malev – resulting in the elimination of these two factors – was good news for anyone caring about the welfare airline competition can deliver to passengers.

4 years on, it is possible to look back and try to assess the effects of shutting down a state-owned flag carrier of Hungary. I take a first step in this direction. This blog post is nothing but a first look at the data and conclusions drawn from therein. For confirmation one would need to spend more time looking closer at the data and the industry itself.

Eurostat offers detailed, airport-level monthly data on passenger numbers. In 2015 there were 10,228,352 passengers flying from/to BUD. This was a 15% increase (1,343,511 passengers, or roughly an average 112,000 passengers per month) from the final year before shutdown (2011: 8,884,837 passengers). So it would appear that eliminating Malev gave a boost to passenger air traffic to/from Budapest. It is possible that passenger numbers increased in other airports as well so the large increase at BUD is a sign of an overall EU trend rather than the effect of shutting down Malev. I take a cursory look at the data to see if this is likely to be the case.

The Eurostat data provides data on passenger numbers for 936 airports. The figure below shows the number of passengers by airport, distinguishing between Budapest Airport, all EU airports and all airports within 400kms from Budapest. The vertical red line is when Malev was shut down. The first figure displays the annual number of passengers for the three group of airports. BUD shows an increasing trend after Malev was shut down, which the other two lines don’t seem to follow. From this picture, it does seem to appear that the post-2011 increase was not an overarching EU trend.

To confirm these visual findings one would need to do some formal testing. Is the increase in passenger numbers in BUD really larger than the increase at other airports? Isn’t this difference due to some characteristic of BUD other than the shutting down of Malev? To find out would require rigorous econometric analysis. Below I only provide a quick ‘sniff test’ (using a simple difference-in-difference model) to check if this is something worth pursuing.

Ideally one would need to find out how passenger numbers would have evolved at BUD had Malev not been shut down. In practice, researchers would try and find an airport that is sufficiently similar to BUD but for the fact of the closedown of a major incumbent airline. As this is just a cursory look at the data I only take the first three steps in this direction.

First of all, I assume that changes at all other 935 airports (averaged out) are a good indicator of that changes that would have happened at BUD without the Malev shutdown. I did not have any other airport or time specific data so I estimated the following simple model: Y = b0 + b1*TREATxPOST+ airport fixed effects. The parameter of interest b1 tells us the difference between the change in passenger numbers at BUD and the change in passenger numbers at other airports. When compared to all other EU airports the number of passengers at BUD increased – on average – by 96,000/month (statistically strongly significant)

But the sample of all EU airports contains a large number of very small airports. A better comparison would only look at airports that are similar – in size – to BUD. BUD has a sample average 658,166 passengers per month. For rough approximation I removed airports that are smaller than 100,000, and larger than 1 million passengers per month (averaging over the 13 years of the sample). This left me with a sample of 132 airports. These are shown on the figure below.


A formal comparison to this more similar sample of airports reveals that the increase at BUD is still around 55,000 passengers per month (again, strongly statistically significant).

It goes without saying that this is only a very preliminary analysis. Ideally one would need to collect information on airport characteristics to ensure that the control group is sufficiently similar. One of the most important of these is that ideally in the comparator airport there would also be a failing state-owned airline. Prague airport seems to be good for this purpose at least on two characteristics: it is in the same region as Budapest and it is the hub for Czech Airlines, another state-owned airline with a similar fleet size. The other (partial) similarity is that Czech Airlines and Prague Airport are in the same ownership (just as Malev and BUD until 2007). The figure below compares passenger numbers for the two airports. The increase still appears to be higher for BUD.


The monthly increase in the number of passengers (averaged over the periods before and after Feb 2012) show around 100,000 more at BUD than in Prague. So assuming that Prague is a good depicter of what would have happened had Malev not been shut down, this would be evidence that the shutdown of Malev significantly contributed to more passenger journeys to/from BUD.

It is always nice to find a story where theory (the inefficiency of state owned enterprise) and data go so nicely hand in hand. As stated repeatedly, the above exercise was nothing more than a preliminary look at the data. It suggests that the shutdown of state-owned Malev did indeed eliminate some market inefficiencies and improved our lives in the shape of increased flight numbers. More formal confirmation would require more data, maybe a longer than 4 year perspective, and a more elaborately constructed control (for example through a synthetic control). The first step has been done, the ball is now rolling…

Here’s the data and Stata code I used.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: