Estimating the economic cost of electricity shortages is an exercise in enormous uncertainty. To reduce the uncertainty, or, rather, to get a sense of its range, I relied on a variety of methods.
Firstly, several economists have attempted various ways to calculate the direct costs. Typically, they assumed the loss by each industry would be equivalent to 3 hours (the time span of undelivered electricity) of production in a normal working day. But, the length of a working day differs across industries; some work 8-hour days and others the full 24-hours over three shifts. They then weighted the loss in each industry by the contribution each one makes to the whole economy before accounting for the number of working days. The estimated cost over the three months of load shedding ranged from R50 billion to R119 billion at the time (early 2008). These estimates would suggest a potential drop in nominal GDP growth of 2.5-6%, if the unmet electricity needs were sustained at a rate of every fourth day for a year. In general, these direct costs are considered to be underestimates of the total costs because they do not account for the effect on confidence.
Secondly, I looked at estimates of economic losses due to power outages in other countries – most notably California in the USA, and India. From estimates of the amount of energy shortfall and the proportion of GDP lost as a result, I was able to produce an equation relating expected percentage of GDP lost to the percentage of electricity demand unmet.
%GDP lost = 0.4153 x % electricity demand unmet – 0.00097.
The intercept is essentially zero as one would expect, i.e., if no electricity is lost then there should not be any GDP loss. Using the roughly 5000 GWh lost over three months of the electricity crisis (about 8.5% demand unmet), or the 8% electricity saving Eskom requested, as estimated unmet demand, this equation implies a 3.3-3.5 per cent lower GDP if that degree of unmet electricity needs were sustained in South Africa.
Thirdly, I looked at differences in the trend lines in quarterly GDP figures before and after the electricity crisis but before the onset of the global financial crisis. The two trend lines followed similar paths, i.e., similar growth rates, but the intercept of the trend line after the crisis was lower. This is consistent with a once off setback during the crisis and normal growth after the crisis. The difference in intercepts implied that a sustained electricity shortfall of the same degree for a full year would result in a 5 per cent drop in GDP growth. This is certainly an over-estimate, given the assumption that the decline was entirely caused by the electricity shortages.
It is not unreasonable to say that electricity shortages of the sort South Africa experienced in late 2007 and early 2008 would result in a slowdown of annualised nominal economic growth of somewhere between 2.5 and 5 per cent. Those values should be considered extreme estimates, or lower and upper bounds respectively.
It is virtually impossible to be precise about likely losses in the future. Quite apart from the lack of precision in estimating the losses in the past, there is no reason to expect electricity shortages to consistently hit crisis levels for a full year. On the other hand, electricity usage is below trend because of the financial crisis and global economic conditions. As these conditions ease, demand for electricity will surge and South Africa again may be caught short. What can be said with certainty is that any degree of unmet demand for electricity will result in fairly large economic losses.
A single 3-hour shutdown would lose the country around 2.25-3.55 billion Rand in today’s money. What would that loss mean? The lower estimate is equivalent to losing 27,108 to 40,900 low cost houses, or wasting one seventh of the annual low cost housing budget. It is equivalent to losing around 10,000 thirty-seat busses, or a whole train (locomotive plus thirty coaches) with ten kilomtres of rail from public transportation. At the minimum wage level, it would provide a year’s employment to about one thousand five hundred people. Or, it could provide 9.2 days’ worth of the entire social grant; one hour of lost electricity is worth seventy four hours of social welfare.
One could go on citing examples but this ought to be sufficient to impress upon people that any further loss of electricity supply should be avoided and that the development of sufficient future capacity is essential. Economic losses cost SA fifteen to one hundred and thirty three times more than to produce the ‘missing’ electricity. Ironically, the 2007-2008 electricity crises were equivalent to forgoing a new power station. Failure to produce the electricity this country requires would be highly detrimental to our economy.
FMF Feature Article / 14 August 2012