By Alexander Apostolides on July 17, 2011

2nd estimate on "best case" cost estimate: 2.9 Bn euro

This is the second estimate of the Cost of the Naval Base explosion using several different sources and internalizing many things that where previously exogenous. I still come to within 5% of the original estimate. I want to thank Michalis Persianis at Kathimerini and Manos of LolGreece for all their comments and help.

Once again this is a preliminary estimate of a "best case" scenario and to begin a debate on how to tackle the calamity. That I came within 5% of the original with different sources makes me more confident that this is close to the cost. Most international commentators argue I am much to conservative and that the true cost of the "best case" scenario could be up to 30% higher.

It is important however to state that this shows the loss of wealth of Cyprus. It is by no way a full indication of the decline of the economy for 2011: that estimate is coming soon.


Our thoughts are to the families of the deceased at these moments, but it is important to quickly understand the effect of this apparent negligent homicide and get a public dialogue on the costs. I tried to make calculations based on energy intensity per sector, but being out of power today for more than six hours meant that I could not access vital data. Once again this is an effort to stimulate debate and not by any chance a final projection of costs, and should not be considered as the income for the year 2011 will fall by so much.

Total damages are close to 3 billion Euros, or 17% of the GDP of 2010. Even if we delay fixing 2/3rd of Vasiliko until after Christmas, the burden to the economy will still be high.

I am quite confident of this estimate, and have calculated a 30% confidence interval at a 95% level. I know this is rough, but if you notice I have at least internalized some issues such as costs of petrol etc. that where previously exogenous. Once again I have again biased all figures against what I expected (a high result), so this should be seen as a conservative estimate.

1) Vasiliko Power station
There is increasing hope in salvaging one turbine and perhaps recovering some of the investment, but until we hear a cost figure I stick to a low estimate of the recovery and repair. We need to ask original builders, Austrian Energy and Environment, which built the first phase, and Ansaldo which built the second.

2) Loss of Work Hours in Nicosia, Limmasol and Pafos
We argued that not all areas are affected by the power cuts. Thus we decided to try and capture that, by only having the power cuts affect some workers of the economy. We argue that based on the areas that power rationing exists at present, 80% of the Nicosia workforce, 50% of the Limmasol workforce and 60% of the Pafos workforce is affected.

The using the total workforce of the 2009 labour Statistics (376,300) we needed to calculate how much of this labour force is in each province (without having access to the internet due to another powercut). Thus we argued that the population per district of the latest Demographic report is representative of the labour force per district, as use the share of total population to find out the population share per district. This was used to find the total employment per district. Actually this is almost certainly an underestimate as the we actually know that the economic activity rate in Nicosia is about 9.9% higher than that in Limassol, according to the EU’s urban audit.

Thus the total employment per district was multiplied by the proportion of employed population affected per district to argue that 194,146, or 51% of the islands labour force will be affected.

We then used the labour survey of 2009 to estimate the average salary of the labour force per hour. We assume that salaries are constant since 2009 (effectively meaning negative salary growth) in order to bias the result downwards. The average salary of 2,130 euros per month is multiplied by 13 salaries to get an annual average salary figure. Then we estimated that with allowed time of work, the average worker works 40.2 hours a week for 50 weeks (taking out the large number of public holidays in Cyprus, to estimate the average labour hours worked. We then divide the average annual salary by the average hours worked per year to get a annual creation of value added of 16 Euros an hour. Again this is quite an underestimate as the labour cost per hour worked in Cyprus is considered on average as 17.43 Euros.

How much labour time is lost by the power cut? Based on research by Dr. Tim Leunig of the LSE on train delays, predictable cuts (delays) are better for people than the type of unpredictable power cuts we have been having, as one can plan to maximize his efficiency around the power outage. Since power cuts are unpredictable, we estimated that there will be power cuts at least until the last week of December, and that they will be on 3 hours a day for 5 days a week, leading to 360 hours of powerless working time for each affected labourer.

We then argue that the unpredictable nature of the cut does not allow you to be very effective during power outages, being able to produce only 80% of when you have electricity. Thus for every hour of lost time, on average 12.8 Euros of value added are lost. So for each worker that is affected the lost hours from the power outage from now until resolution before Christmas costs the economy 4608 Euros.

Again this is very optimistic and almost certainly and underestimate. The cost of planned outages is estimated at one third to one tenth of unplanned ones in Thailand, about one fifth to a quarter in Nepal, or about one half in Sri Lanka, etc. International commentators have argued that even these figures are very conservative as it omits, in a knowledge-based economy, the impact of data and service outages. LolGreece argues that the additional probability of an IT company going out of business following a major data loss incident can create additional significant costs.

By multiplying this by the amount of workers affected we get a staggering loss of the economy of 894,628,513 Euros of lost value added.

3) Additional loss due to Multiplier effect of no.2 (same for 5,7,9,13)
The loss of value added has cumulative effects the economy. Each 1 pound placed in the economy created more than 1 pound. Alternatively each 1 pound loss of the economy will reduce the income of the GDP by more than 1 pound. Using the input/output of the Cypriot economy we can then estimate the additional loss for the economy. Thus for 3, 5, 7, 9, 13 I calculate the knock on effect of the reduction of income when it is resolved through the economy of each cost.

However the Republic of Cyprus does not release its newest Input / Output tables, and as a result the only publicly available Input / Output table is shamefully from 1986. Thus I reduced the multiplier effect to lower numbers (biasing against a high figure) to take account of the increased import dependence of our economy since 1986.

4) Loss through increase electricity cost (50% rise) primary and secondary sector(excluding Utilities), and 6) Loss through increase in tertiary sector

The Cyprus Electricity Authority can in no way absorb all the cost of the smaller and hence expensive units that are in use today due to Vassiliko being knocked out. In addition many companies have to rely on alternative power sources such as generators to keep operating even basic systems going. Thus the cost of electricity for companies and households will increase, despite any suggestions to the contrary. A brief survey found that electricity is approximately 4% of cost in the Secondary sector and 2% in the Tertiary sector. An increase in the cost of electricity by 50% (partly increased unit costs and partly cost of rebuilding) will thus lead to a reduction of the value added, as cost will jump by 2% in Primary and Secondary, and by 1% in the Tertiary sector. Thus assuming all else is constant (i.e. value added of 2010) there will be a reduction to the value added share due to the increase cost (2% or 1%) of electricity Production.

Once again LolGreece is correct to argue that we need to add to the estimate some cost consequences of other power plants running at high capacity levels, as indeed we have already experienced system failures through the overload on Friday. This can lead to more costly more maintenance and the increase the frequency of outages.

8) Loss of Utilities value added

All major utilities are affected, with electricity generation being the first. We estimate the loss of value added as 40% for the last six months of this year. This is based on estimates of increased cost per unit of electricity produced (i.e. a reduction of the value added share of electricity, as more of the gross output is cost), and of a reduction of electricity capacity. Similar reductions are expected on all major utilities, water / sewerage, as they are also main clients of electricity and hence very much affected by this crisis where our electricity maximum capacity has fallen below our demand for electricity.

It must be clear to all that there is no doubt that the income created by the production of electricity has been reduced as even if we come to our previous capacity the power currently produced by Moni, Dekelia and auxiliary sources, cost much more per MW hour to make than Vasiliko, as power plants have very large economies of scale in production.

11) Diverted resources of police, army, fire-fighters etc.

The diversion of police, army and fire fighters due to traffic duties, helping trapped visitors and gassing peaceful protesters diverts police time. I arbitrarily picked a low number for the overtime and additional work load based on high overtime wages

12) A fall of 1% of hotel occupancy rate due to tourism cancelations

Using estimates I devised in another project that attempts to measure economic interdependence between the communities in Cyprus, I have an estimate of how a 1% in occupancy rate impact the hotel and restaurant sector. Notice I am only suggesting a very small decline in occupancy and not a mass exodus of tourists – just a reluctance of a small number to come to Cyprus after the negative publicity.

Things I have not estimated include: Increase in insurance premiums due to increased car crashes (traffic lights are out); cost of repairing the damage to the villages around the base; cost of rebuilding the state of the art base that was destroyed; cost of compensation to victims if negligent homicide is proven; additional deaths caused by heat during power blackouts during the expected August heat wave; additional cost of health provision for the injured of the accident and of the accidents due blackouts; Machine degradation and IT support due to system shutdown and failure.


  1. What about offsetting amounts, such as insurance payouts and EU solidarity funds?

  2. What I did here is sum up costs. There has been no attempt to say who pays what.