Economies, governments and elections: much talk

It is often said that elections disrupt the macroeconomy. The recent elections in Kenya have prompted us to apply a beginner’s test to the theory. Economists generally subscribe to it. When asked to provide a forecast of core variables, the temptation is to build in a slowdown in growth, a boost to public spending and subdued private investment. This is the result of either laziness or thorough analysis.

Nigerian planning officials are divided on the point. The Economic Recovery & Growth Plan 2017-2020 has growth slowing from 4.8 per cent in 2018 to 4.5 per cent in 2019 (election year), and then picking up to 7.0 per cent in 2020. We understand from conversations with senior officials that the elections were built into the forecasts. It is the work of the federal ministry of budget and national planning, and is dated February 2017. The more recent 2018-2020 Medium-Term Expenditure Frameworkfrom the same ministry has growth rising steadily from 3.5 per cent next year to 4.5 per cent in 2019 and 7.0 per cent (again) in 2020. These projections we have taken from local media reports.

Since the framework has been approved by the Federal Executive Council, the official line seems to be that the elections will not be disruptive. Our next step therefore is to examine the revised quarterly national accounts at constant prices for any trends around the presidential elections of April 2011 and March 2015, making allowances for the fact that the data series is not seasonally adjusted.

Q1 2015 saw double-digit contraction of the economy on a quarter-on-quarter (q/q) basis. The first-quarter phenomenon (reaction after the holiday period) would have been a factor, and perhaps the elections too. Predictably, for private consumption, the q/q contraction was still greater (-26 per cent). When we look at the accounts on a year-on-year (y/y) basis, we find that that economy slowed from Q2 2014 through to Q3 2016 (with one small blip in Q3 2015). This coincides exactly with the slide in the oil price, which has exposed Nigeria’s vulnerability to the Dutch disease. One symptom of the disease is paucity of government revenue so we are not surprised to find that government consumption contracted y/y throughout 2015 and 2016.

Any impact of the elections of March 2015 was dwarfed by the decline in the oil price from above US$100/b and its negative pass through to the economy. Q1 2011 also saw q/q contraction of the economy (of 9 per cent), and again we identify the first-quarter phenomenon. When we look at q/q consumption trends in Q1 2011, the find the reverse of what we might have expected: private consumption increased (after the holiday period) by 9 per cent q/q while the government element shrank (when the authorities might have been distributing the largesse) by -28 per cent q/q.

One type of cynic would claim that the government had already spent big for electoral purposes. Another would seek to diminish the value of the data. We are neither and find the national accounts inconclusive. However, the FGN’s public finances are more revealing. Spending picked up strongly throughout 2010, notably for recurrent items, which we can attribute to the near-doubling of government salaries. (The public finances have not since recovered.)  When we come to the run-up to the 2015 elections, we see that spending was fairly constant. As with the national accounts, we conclude that the slide in the oil price was the principal determinant. Entering the political space, we might add that the authorities chose not to ramp up borrowing for an electoral agenda.

So over two elections in Nigeria, we have one clear case of the impact on the broader economy (FGN spending in 2010). Otherwise, the evidence is inconclusive. The exposure of Nigeria’s Achilles heel since mid-2014 has determined the direction of the economy, and other considerations have been marginal.

More compelling is the reaction of stock markets to elections. The NSEASI opened 2015 very badly, and was down 20 per cent ytd by mid-February. There followed a bounce before voting and then a period without any clear direction before the index resumed its downward direction in the third quarter. In Kenya the index (NSE 20) also enjoyed a bounce directly before and after the recent elections on 08 August despite the negative newsflow warning of a close result and serious violence in response to a result not accepted by the loser (as happened in 2007). Investors apparently decided that the constitutional referendum in 2010 in Kenya had removed the sting from elections.

A stronger and longer lasting bounce on Wall Street followed the US presidential elections in November. It was said that household spending would benefit from tax cuts, defense stocks from the president’s agenda for the armed forces and bank stocks from an easing of regulation of the industry. The first two benefits have not materialised because of the impasse in the US Congress although, worryingly, the third may come to pass in some form.

There are other non-numerical influences of elections to mention. One such is the distraction of government officials, who have less time and appetite to sign documents, grant permits and authorize waivers. The loser is deal flow.

Our search for correlation between elections in Nigeria and economic activity has led us only to the steep rise in salaries/public spending in 2010. In other respects, there is much talk and theorizing. One of the most popular clichés is that markets do not like uncertainty. In extreme cases this is probably true but our three examples tell us that those same markets can also give new governments the benefit of the doubt before they see their policies in action.

Gregory Kronsten

Head, Macroeconomic & Fixed Income Research

FBNQuest