German Chancellor Angela Merkel prepares to welcome Georgia’s prime minister in September. (AFP/Getty Images)

Before Germany’s  Sept. 24 federal election, observers often commented that its economy was “firing on all cylinders.” GDP growth will likely exceed 2 percent in 2017, the strongest rate in half a decade. Unemployment levels are at their lowest since German unification in 1990.

Yet the strength of the German economy did not seem to help Chancellor Angela Merkel at the polls, as her Christian Democratic alliance (CDU/CSU) received only 33 percent of the vote — their worst result since 1949 and an 8.5 percentage point drop in support from the previous election in 2013.

This might come as quite a puzzling outcome, because the state of the economy and election results are supposed to be closely related. We expect voters to reward incumbents in good economic times and punish them when the economy performs poorly. This behavior has even its own name: economic voting.

At the same time, researchers have had a difficult time demonstrating consistent evidence for the electoral effects of the economy. A strong influence of the economy on election results seems to be at work in some elections but not in others. Is the 2017 German federal election one of those cases where economic voting doesn’t seem to fit the results?

It’s all relative — including economic performance

My research provides an answer for the sometimes-puzzling relationship between economic and electoral outcomes: researchers and pundits focus far too much on just the election-year performance of the national economy but pay little attention to the broader domestic and international economic context.

In a new article, I argue that we should focus on the relative performance of incumbents to understand how the state of the economy affects elections. In particular, we should take into account how economic outcomes on the incumbent’s watch compare to recent past outcomes in the country, as well as how the country is doing compared to others.

There’s a simple rationale at work. Think about an incumbent whose term delivers about 3 percent annual growth. Voters would have no reason to consider this a particularly strong performance if the country had already been growing by about 3 percent for some time, because there has not been a significant change in how the economy is doing.

In contrast, all else being equal, if the economy had displayed a considerably slower growth, say 1 percent, before the incumbent took office, the same outcome would suggest that the incumbent is highly competent. It’s the same 3 percent growth in the two scenarios — but voters probably will reward the incumbent at the polls only in the latter case, because of the improvement in the economy.

Moreover, national economies are not isolated from trends in the international economy. We expect our leaders to take advantage of global or regional booms and minimize the consequences of negative external shocks. Thus, any information about how the national economy performed relative to other countries that are likely to be similarly affected by external shocks would be useful for voters to judge an incumbent’s performance.

In basic terms, voters need reference points to evaluate an incumbent’s performance in managing the economy, as no economic outcome is inherently good or bad. How the economy had been doing in the period leading to the incumbent’s term and how other countries’ economies have been performing provide voters with two such reference points. Models of economic voting should incorporate the broader domestic and international economic contexts into their analyses.

I checked how this approach to economic voting might play out by analyzing 475 elections in 62 countries from 1965 to 2014. I found that incumbents who presided over relatively better (or worse) economic outcomes in domestic and international comparisons are consistently rewarded (or punished) in elections. Crucially, once we account for an incumbent’s relative performance during its term, election-year growth does not have a discernible effect on the incumbent vote share.

Does it matter if the country did better economically than other countries? My results showed that this was electorally consequential especially in countries with highly educated electorates. It is likely that in such countries there is more demand and supply for the kind of information that makes it easier for voters to have a sense of how the domestic economy fares in international comparisons. But the exact mechanism warrants further research.

What does this analysis say about the 2017 German election?

So to assess the impact of the economy on Merkel’s electoral fortunes in 2017, we should look at how the German economy during Merkel’s third term (2014-2017) performed relative to (i) her second term (2010-2013) and (ii) an international benchmark. Let’s use GDP growth as a measure of economic performance and the main export partners of Germany as the international benchmark group.

The cumulative GDP growth in Germany during Merkel’s third term was 7.1 percent. In contrast, Germany grew by 8.5 percent during her second term. Thus, Merkel’s government did relatively worse in her third term compared to her second term. And the German economy did worse relative to benchmark countries, too — Germany’s main export partners’ growth averaged 7.8 percent during Merkel’s last term, better than Germany.


How did the German economy do during Angela Merkel’s third term?Data: IMF, OECD; Figure: S. Erdem Aytaç

Here’s what this tells us:

1. German voters are probably not as impressed with how Merkel handled the economy as they used to be. Yes, the German economy might be sound, but it has been this way for quite some time anyway — and things actually were somewhat better a couple of years ago.

2. Some voters might have picked up news about the Netherlands or the U.K. growing faster than Germany.  This type of news was not a good signal about Merkel’s competence. Or the media might have been less sanguine when reporting the German economy than it used to be, due to its relatively poorer international standing.

3. The economy actually worked against Merkel’s electoral fortunes. In fact, my economic voting model pointed to about a 5 percentage-point drop in her vote share from the 2013 election. Note that this prediction takes into account that growth in 2017 was much higher than in 2013 (1.9 vs. 0.5 percent). Merkel ended up losing 8.5 percentage points — which likely reflects voter dissatisfaction on some other things as well, such as her stance on immigration.

In short, the economy matters to voters — but we need a broader domestic and international comparative framework to understand just how the economy guides voter choices. For incumbents, their relative performance matters.

Erdem Aytaç is assistant professor of political science at Koç University in Istanbul, Turkey.