Political earthquakes are hard to predict. Fortunately, they rarely have lasting consequences.
That, at least seems to be two key lessons from last year’s tumultuous political events that many market participants in the developed world have taken to heart. A heavy electoral calendar looms in Europe.
Source: Bloomberg Finance L.P., Deutsche Asset Management Investment GmbH; as of 2/28/17
In this Special introducing our European election watch, we argue that both lessons can be misleading at best and misguided at worst. In our view, last year’s events suggest that polling can actually be quite useful, if used wisely. If and when the political status quo gets uprooted, however, it takes time for the eventual impact to become clear. And the consequences of this year’s elections in Europe – negative and positive – may prove more lasting than many think.
Our starting point is the strange mixture of nervousness and complacency that seems to have taken hold in European markets in the face of the French elections in particular (in Section 1). Section 2 revisits last year’s electoral surprises, while Section 3 takes a broader look at how and why politics might matter to markets, highlighting the need for a closer look at the political and economic institutions of the country in question. Sections 4 and 5 examine the upcoming votes in the Netherlands and France. The Special concludes with perspective on the real lessons learned from 2016’s events for this year’s European elections in general and France in particular.
Remember the run-up to the Brexit referendum and the U.S. presidential elections? At the time, brokers and analysts were going through the motions of assessing the implications of all potential outcomes. Among market participants, however, there was pretty limited interest even in the final months before each vote. After all, it was widely thought that a Clinton win and a vote of the UK to remain in the European Union (EU) were foregone conclusions.
The mood could not be more different ahead of the French presidential elections. Especially in bond markets, investors are increasingly willing to consider risk scenarios. In the case of France, this can be readily seen by comparing the spread of French and German government bonds with the odds of Marine Le Pen, party leader of the far-right Front National, winning the presidency, according to betting markets.
Source: Oddschecker, Bloomberg Finance L.P., Deutsche Asset Management Investment GmbH; as of 2/28/17
Another indication of wariness can be found in gold and foreign-exchange markets. Lately, there has been a pattern of a positive correlation between gold and the U.S. dollar (USD) vs. euro exchange rate. The last time this unusual pattern appeared was during the Italian referendum on constitutional reform in December 2016. It is a clear indication of resurfacing concerns over Eurozone stability.
Source: Bloomberg Finance L.P., Deutsche Asset Management Investment GmbH; as of 2/28/17
* Correlation over 10 days
At the same time, however, equity markets in France and elsewhere have remained strong. This is particularly striking if you compare the CAC 40 index of French blue chips with spreads on government bonds.
Source: Bloomberg Finance L.P., Deutsche Asset Management Investment GmbH; as of 2/28/17
How to make sense of all this? There are two ways to answer this question. One is to revisit the lessons from Trump and Brexit (Section 2). Another, arguably even more useful one is to look at how, where and why elections have mattered to markets throughout recent decades (Section 3).
Memory is a flawed tool, particularly when it comes to analyzing your own predictions that failed to come true. Many investors have grown distrustful of opinion polls. This is based on a somewhat misguided view of just how surprising the Brexit vote and the U.S. election results were. The answer is surprising indeed to many participants in betting, not to mention financial, markets. Pollsters, by contrast, did not get it all that wrong and their raw data actually provided valuable pointers to watch.
In fact, polls on the eve of the Brexit vote showed a very tight race. In the final days leading up to the referendum, both traditional phone polls and several online polls hinted at “Leave” gaining ground. Based on the data, the eventual result of 51.89% voting against continuing EU membership was at least as plausible as the strong “Remain” majority many investors – and betting markets – had come to expect.
Of course, investors had plenty of reasons to take the British opinion polls with a soup spoon of salt. In recent years, opinion polls in the United Kingdom (UK) have had a patchy track record, getting, for example, the 2015 British general election spectacularly wrong. One issue, also relevant in other countries, is that traditional random polling by phone is getting harder, with fewer and fewer voters willing to participate in surveys and using fixed-line phones. Another source of concern ahead of the Brexit vote was that for much of the campaign, most online polls showed “Leave” – slightly – ahead. By contrast, many traditionally more reliable phone polls only tightened during the last month or so, having previously shown clear leads for “Remain.”
Such divergences between different polling methods point to systematic biases. Until the vote takes place, however, you cannot know for sure which sources of data will prove reliable this time around. At most, you might want to aggregate polls by assigning weightings based on past reliability, though that too would not have helped much in the case of Brexit. The alternative – particularly when recent events and systematic divergences between polling methods suggest the potential for large polling errors – is to apply a broader error-adjusted range.
The key is not getting into the business of trying to predict the direction of the suspected polling error. Instead, all you can say is that a broader range of outcomes is possible – in the case of Brexit a fairly decisive win for either “Leave” or “Remain.” Instead, many observers and betting markets predicted a surge in favor of the status quo on polling day. They did so mainly based on just one previous event, the referendum on Scottish independence. Yet, as any statistician will tell you, one or even a few data points is rarely enough to draw reliable inferences from. The belief in a belated status-quo bias was widely held despite, not because of, any particular pointers in that direction in the polling data.
Meanwhile, national polls in the U.S. presidential elections were actually pretty accurate this time around. After all, Hilary Clinton won the popular vote by 2.9m (or 2.1 %), even as she lost in the Electoral College. From the primaries onwards, polls picked up signs of Donald Trump’s strength well before most pundits began to take him seriously. They also suggested he had the potential – much more so than traditional Republican candidates – to appeal to white working-class voters in America’s industrial heartland. Just ahead of the November election, polls also pointed to an unusually large number of voters who appeared undecided between the two unpopular front-runners.
By far the biggest forecasting problem was that results from key swing states tend to be correlated (especially within particular regions of the U.S., such as the Midwest). This is tricky, not least because it can lead to surprising results in neighboring states not as widely polled as acknowledged swing states. In the event, Trump’s strength in the Midwest got him within spitting distance of winning Minnesota, usually a reliably Democraticstate.
Once again, the correct way to think about such sources of uncertainty is to apply a broader range – in both directions. As we argued at the time, uncertainty was unusually high and a Trump victory was at least as likely as a Clinton landslide and a Democratic sweep.
The lesson mentioned above – if uncertain, apply a broader range in both directions – also applies when you have no polling data at all in the last weeks of a campaign. In fact, last year’s biggest polling miss took place in Italy. The government lost the referendum, intended to make the country more governable, by a whopping margin of 18 %. The gap between “No” and “Si” was roughly twice as large as in the last set of polls 15 days before the referendum on December 4th suggested. Italian law prohibits the publication of new polling results during this 15-day blackout period. The last published polls showed “No” ahead by high single digits – but plenty of voters still undecided. With 15 days left, these voters had plenty of time to make up their mind. As a result, a wide range of results were plausible – from a narrow win for “Si” to the landslide victory “No” secured in the end.
Political earthquakes are indeed hard to predict. Polls are only a snapshot and even the ones published just before an election are usually a few days old by the time you read about them. A lot can still change in those final days. In much of the industrialized world, the share of voters making up their mind at the last minute has tended to rise in recent decades. But if you want to get a grasp of the potential seismic shifts, opinion polls shortly before a vote are actually a pretty good place to start. (Of course, you should also take into account how well polls and other tools have done in predicting a country’s electoral results.)
Arguably, last year’s biggest surprises did not take place on polling days, but in the weeks and months that followed. For example, the Brexit vote to leave the EU was supposedly going to trigger a recession in the UK via the loss of confidence among British businesses and consumers. So far, it has not materialized. More immediately, it was expected to lead to a sharp fall in UK asset prices, with repercussions in financial markets around the world.
In the event, the Brexit fall-out proved rather less devastating. The pound duly dropped – boosting pound-denominated foreign earnings of British companies and share prices in the process. For foreign holders of British equities, the end result six months on was pretty much of a wash rather than the feared bloodbath. In euro terms, the FTSE 100 Index has lagged far behind theS&P 500 Index or the DAX, but it is still up by a few percentage points. Meanwhile, the UK government has had no trouble at all finding willing buyers for its bonds – another source of worry before the referendum.
It has been a similar story with other recent political events. Time and time again, European equity markets have proven remarkably resilient. Just take the Greek elections in 2015, Brexit, and the constitutional referendum in Italy, as well as the election of Donald Trump. In all these instances, market turbulences proved both milder than many had feared and shorter lived than almost anyone had expected.
Source: Bloomberg Finance L.P., Deutsche Asset Management Investment GmbH; as of 2/28/17
* Euro Stoxx 50 Index, 25 days after each event
Meanwhile, the continuing rise in U.S. equity markets since the Republican sweep in November’s elections has served as reminder of how quickly markets can change their collective mind, after an event once seen as risky materializes.
Elections have consequences. However, figuring out what these consequences are is surprisingly tricky. The issue has been extensively studied in the United States, where a sizeable literature has emerged in recent decades on how politics influences business cycles. The key conclusions are worth outlining here, as they have also shaped the way such issues are viewed in Europe.
Broadly, there are two reasons why you might think that politics matters to financial markets.
First, governments might try to use the timing of economic-policy measures in order to improve their odds of getting reelected. For example, they might boost fiscal spending or monetary stimulus as the end of their term in office draws nearer. Of course, any resulting pre-election boom might eventually lead to rising inflation, requiring higher interest rates. If the government controls both monetary and fiscal policy, however, it will try to delay both monetary and fiscal tightening until after the vote is over. Pre-election booms would regularly be followed by post-election busts.
Second, political parties might systematically differ in their policy priorities. A left-wing party, such as the Democrats, might be more concerned about employment and growth than about inflation. If so, you might expect inflation to be lower under Republican rule and, perhaps, real economic growth to be weaker in the short term. (Naturally, growth would be influenced by plenty of other policy measures as well.)
Once you start to think about this, it rapidly becomes clear that statistically testing such hypotheses is by no means straightforward. You might expect that rational voters will eventually learn to see through political shenanigans, or that a party might adjust their stances on issues such as inflation in response to what the electorate seems to want. This would result in breaks in the statistical data. Changes in the monetary-policy regime are another cause of such breaks. Much of the early U.S. literature emerged before 1979, when the U.S. Federal Reserve (Fed) under Paul Volcker fully embraced an inflation-fighting agenda.
Such breaks complicate empirical tests. They also compound the problem that there are only a fairly limited number of data points to draw conclusions from. U.S. presidential elections take place only once every four years, while gross-domestic-product (GDP) numbers, inflation indices and employment figures come out once a quarter and once a month, respectively. These economic statistics also tend to be lagging indicators, subject to sizeable revisions.
For all these reasons, it has become increasingly common in recent decades to use stock- and bond-market data as near instantaneous assessments of economic prospects by market participants. Such studies have provided pretty solid evidence that U.S. administrations indeed appear to have timed economic-policy measures with the electoral cycle in mind.
Stock returns generally tend to be higher in the second half of a presidential term than in the first half. It is less clear whether stock-market returns systematically differ depending on the party in charge – though there is some evidence that small caps in particular tend to do better under Democrats.
To financial economists, all this is somewhat puzzling. In particular, why would markets not anticipate the likelihood of a pre-election boom well before it actually starts? If they habitually did so early in a presidential term that would smooth returns, and political stock-market cycles would disappear.
Part of the reason probably is that political decision making is more opaque in real time than it appears when you look at the data in retrospect. For example, markets may currently be underestimating the risks to the Trump administration being swiftly able to deliver on corporate-tax reform, as we argued in our recent report on the topic . We, and other market participants, will only know for sure, however, if and when Congressmoves – or fails to do so.
In the meantime, we are watching events ahead of the March 15th deadline for raising the U.S. debt ceiling and the April 28th deadline for passing a continuing resolution to keep the government running. Past experience suggests, however, that unless and until such looming deadlines become headline news, there is little market attention on congressional minutia. This makes it doubtful that markets would correctly anticipate policy measures far into the future.
Things get even more interesting – and complicated – once you look beyond U.S. shores.
At one extreme is the British example. In 1976, Lord Hailsham popularized the phrase “elective dictatorship” to describe the “concentration of powers (…) in an executive government” usually “formed out of one party which does not always fairly represent the popular will.” Thanks to the UK’s first-past-the-post electoral system, a plurality in the popular vote can translate into whopping majorities in the lower house (House of Commons). Ultimate legislative sovereignty resides in Parliament, but the upper house (House of Lords) can only delay but not prevent new legislation from passing.
Under the Westminster System as it emerged since 1949, whoever controls the House of Commons controls Parliament and effectively the country. As long as the governing party remains united in the House of Commons, the government of the day is not constrained by the sorts of checks and balances familiar in the U.S. Until 1998, when the Bank of England gained operational independence, the government controlled not just fiscal but also monetary policy. And to top it all, the Prime Minister had, until 2011, the de-facto power to call an early general election, whenever it suited him or her within a 5-year term.
Given all this, it should come as no surprise that for much of the post-war period, political business cycles have tended to be quite pronounced in the UK - “boom and bust” as the former Chancellor and Prime Minister Gordon Brown used to call it. Moreover, bond-market data also suggest that until 1998, interest rates tended to rise in sync with the likelihood of Labour winning, because of rising inflation expectations. The UK, in short, has historically been an excellent example of just how much elections matter under constitutional systems where the new government will face few restraints.
By contrast, in much of the rest of continental Europe, elections tell less than half the story. Constitutional constraints and economic institutions tend to mitigate the impact of any one electoral event.
Take Germany, with its longstanding tradition of a famously independent central bank, as well as fairly centralized labor-market institutions committed to wage moderation and price stability. In such a setting, any government would tend to struggle to engineer a boom-and-bust cycle, even if it wanted to. (Arguably, the closest post-war Germany ever got was after re-unification in the 1990s, a once-in-a-generation political event when mere economic concerns did not seem to matter as much.)
Because of proportional representation, moreover, it generally takes more than one party to form a German government. Clear-cut political swings have been fairly rare, and more often than not they were due to changing political coalitions rather than elections to the Bundestag. As a result of all this, elections have historically not mattered all that much to German financial markets. There is, for example, no clear evidence that stock-market returns vary depending on the stage of the election cycle, or even that markets have particularly cared about the relative strength of left- versus right-wing parties. Interestingly enough, the main causal relationships seem to point the other way – with strong stock markets (as a proxy for economic health) boosting the popularity of the government of the day.
Studies in other European democracies with proportional representation paint a similar picture. In countries such as Belgium, election results in and of themselves appear to have little stock-market impact. When governments are typically multi-party coalitions, too much depends on the haggling between party leaders after the voting is done. Such haggling appears all but certain in the aftermath of the upcoming elections in the Netherlands.
Elections to the Dutch House of Representatives (“Tweede Kamer”) would seem like an unlikely event to attract the attention of global financial markets. In part, this is because the Dutch system is about as far as you can get from rules designed to deliver a clear result. Representatives are elected under a system of proportional representation. Unlike in other European countries, there is no minimum threshold a party needs to reach in order to get into parliament. Effectively any party that captures more than 0,67% of the vote is entitled to at least one of the 150 seats in the Tweede Kamer.
As a result of this, no single political party has ever been able to reach an absolute majority in the Tweede Kamer in Dutch parliamentary history. In recent election cycles, parliament has been growing even more fragmented. At least a dozen parties appear on track towards gaining parliamentary representation in the Tweede Kamer. If past experience is any guide, several of these parties might well splinter before the next election.
According to Peilingwijzer, a Dutch Polling Indicator created by Tom Louwerse from Leiden University, which combines all Dutch national opinion polls into one estimate, the two leading parties are only likely to capture around 16% of the vote each. That alone points to prolonged coalition negotiations, resulting in plenty of compromises and limited substantive policy change.
The reason the world is nevertheless paying attention is that one of two frontrunning parties is the right-wing populist Party for Freedom (PVV), led by eurosceptic Geert Wilders. For much of the past year, the PVV was ahead in the polls. For markets, the biggest concern is his call for a referendum on the EU membership. Once a free-market liberal, he is now in favor of such hard-to-square positions as lowering the retirement age and increasing public spending, while at the same time lowering income and vehicle taxes. Last year, Mr. Wilders was convicted of insulting Moroccans and inciting racial discrimination. He has also called for closing down all Mosques and Muslim schools.
It looks unlikely that Mr. Wilders will be able to deliver on his promises. All major parties, including the liberal People’s Party for Freedom and Democracy (VVD) of current Prime Minister Mark Rutte, have ruled out entering a coalition with the PVV.
Recent polls, moreover, suggest the PVV losing and the VVD gaining momentum, though the two remain statistically tied for the top spot. Polls have historically been fairly accurate in the Netherlands, especially just before the elections. In recent elections, the PVV has struggled to meet, let alone exceed, its polling results. Mr. Wilders, moreover, is already one of the longest serving members of Tweede Kamer, having first been elected in 1998. Pollsters – and voters – have had plenty of time to come to grips with his rise. The left – and particular the Labour Part (PvdA), as the junior partner in the current coalition – looks set to see its share of the vote shrink.
Based on the latest trends, Mr. Rutte may be able to cobble together either a very broad coalition (including, once again, both the VVD and the PvDA) or a center-right coalition, encompassing the VVD, the Christian Democrats (CDA), the social-liberal Democraten 66 (D66) and three of the smaller parties. Plenty of other options also remain possible, however, depending, notably, on how well the leftwing Greens (GL), the radical left (SP) and the pensioners 50PLUS party end up doing.
Source: Peilingwijzer, Deutsche Asset Management Investment GmbH; as of 3/7/17
* Average results of 6 polling institutes
** Polling results with 95% confidence interval
What the resulting policy mix will be is anyone’s guess and unlikely to become all that much clearer on election night. One important reason for this is rarely discussed outside the Netherlands. The Senate (Eerste Kamer) is not up for grabs. Once a bill passes the House of Representatives, it still has to be approved by the Senate to become law. Constitutionally, the Senate can only accept or reject legislation, without offering amendments. In practice, however, it often plays an important role in debating both upcoming legislation and how laws will be implemented.
The result, as the Dutch Senate helpfully points out on its website, is that “The legislative process can be fairly protracted. Sometimes it takes five or six years before a bill becomes law.” Apart from participating in and potentially slowing the legislative process, the Senate can also scrutinize government actions directly. Senators are not formally bound by a government program, even if their respective parties participate in the governing coalition. Elections to the Senate are indirect – its members were last chosen by the Provincial Councils in 2015, with four-year terms until 2019.
Source: "https://www.eerstekamer.nl/begrip/english_2", Deutsche Asset Management Investment GmbH; as of 3/7/17
Even if the PVV under Mr. Wilders were to massively outperform its current polling and somehow manage to assemble a coalition in the Tweede Kamer, it would probably still struggle to gain much traction in the Senate (where the PVV only has 9 seats out of 75).
In sum, the Dutch election looks unlikely to cause big policy changes in the short term. The vote is nevertheless well worth watching. Already, the PVV has dragged other parties such as the VVD in its direction, notably on immigration. A strong showing may make it harder for the next government to play a constructive role in European negotiations. It could also reinforce market fears on the prospects of Le Pen winning in France. Conversely, a weak showing might reinforce market complacency. In our view, neither reaction would be justified.
The French political system, as it emerged since 1962, is frequently likened to the U.S. system of a strong president with a popular mandate. Such comparisons hide more than they illuminate.
French policy-making and governance differs in at least three critical respects from most other developed democracies.
First, the presidential system as pushed through by President Charles de Gaulle was a significant change to the 1958 constitution. Until the 1962 constitutional referendum, France had a parliamentary system and indirect presidential elections. As frequently noted, however, the change took place without a complete constitutional overhaul. It was intended to establish the directly elected president as the pre-eminent decision maker, but left plenty of related issues unsettled.
This has resulted in a somewhat ambiguous and changing relationship between the different branches of government. In particular, the judiciary, though strongly independent on matters of criminal, civil, administrative and fiscal law, initially played a much more limited role than in other Western democracies, such as the United States. Except in assessing the constitutionality of new laws, it has traditionally been less likely to challenge the government on overly political issues. As two veteran observers note, government by the judges, or “le gouvernement des juges is still a bogeyman in French political discourse.” The judiciary has become more assertive in recent decades. However, the extent to which it can and should interfere in politics, without itself becoming politicized, remains a hotly debated subject among French legal commentators and scholars.
Another important source of tensions, and the second unusual feature, concerns the relationship between the presidency and the legislature, specifically the National Assembly (Assemblée nationale). Should the president be seen as a consensual arbiter, searching compromises beyond party boundaries, as General de Gaulle initially intended? Or be more of a super prime minister, able to enact a partisan agenda, provided the president’s party commands a majority in the National Assembly? In recent decades, a series of constitutional changes have arguably pulled France towards the latter logic.
Notably, the five-year terms of the president and the National Assembly have been aligned in 2000, with legislative elections taking place one month after the presidential ones. The intention was to make periods of cohabitation less likely – previously a fairly regular occurrence when a president of one party would face a hostile majority in the National Assembly.
The new system has worked so far in that recent presidential winners have been able to secure majorities for their allies in the National Assembly. However, it may also have contributed to turn the two most recent presidents, Nicolas Sarkozy and François Hollande into more divisive, partisan figures. Both ended up as one-term presidents, far less popular than most of their predecessors.
It is also worth noting that the National Assembly, as the lower house, generally dominates the legislative process. On most matters, the upper house (Sénat) can only delay but not block legislation. Members to the Senate are chosen by an electoral college (of well over 100,000 grand electors, mostly locally elected officials such as mayors and delegates from municipal councilors). Because this over-represents small, mostly rural municipalities, the Senate has traditionally been dominated by center-right parties. In the history of the Fifth Republic, the Socialists and other left-of-the-center parties only held the majority once, between 2011 and 2014.
Beyond mere procedural matters, any description of French governance must consider the strong role the administrative state has traditionally played in French society. To get a sense of this, just consider the concerns the new Trump administration has caused by publicly targeting individual companies for shipping jobs abroad, and trying to cajole them into “putting America first” instead. In France – under both left-wing and right-wing governments – such targeted measures and economic nationalism more broadly would simply be considered good practice and barely cause a shrug. (The French, after all, invented mercantilism.)
In France, however, public criticism of individual companies may often not even be necessary. Remember how constitutional constraints and economic institutions tend to mitigate the impact of any one electoral event (Section 3)? There seems to be an added twist to this in the French context.
In the other examples of political business cycles considered above, the question was how governments might influence the private sector in order to maximize their chances of getting re-elected. A few years ago, a couple of economists wondered whether in France, the causality might also work the other way around. Specifically, Marianne Bertrand, Francis Kramarz, Antoinette Schoar and David Thesmar looked at the timing of the hiring and firing decisions of politically connected Chief Executives (CEOs). A CEO was deemed “politically connected” if he or she had previously worked in the public sector and served as an advisor to a government minister (“cabinet ministeriel”).
At the time of the study in the late 1990s and early 2000s, 11% of publicly listed companies accounting for some 63% of the market capitalization of the French stock markets were headed by a former civil servant. More than half of these CEOs had also served as advisors to a government minister.
Next, the study looked at how many jobs each firm created in any given municipality, as well as how many plants were started and shut down annually in a given city, and ran the data against the 1995 and 2001 municipal elections. It duly found that politically connected CEOs seemed to be highly sensitive to the electoral calendar. In election years, the firms in the sample were generally more likely to create jobs and less likely to shut down plants. This seemed to be particularly pronounced in municipalities were the election outcome was uncertain. The results broadly held, both for companies that had at some point been nationalized and ones that had never been state-owned, provided they were headed by a politically connected CEO.
These findings illustrate a few interesting features worth keeping in mind when thinking about this year’s elections. In France, large chunks of both the public and the private sector have long been run by fairly cohesive elites. Most CEOs in above study, for example, shared common educational backgrounds, having studied at the same, very limited set of “Grandes Ecoles.” Switching from the civil service to the private sector and back again remains common.
Superficially, this might look like crony capitalism. The reality is arguably more complicated. Admission to the “Grandes Ecoles” such as Polytechnique for engineering, the École nationale d'administration (ENA) for future civil servants or the business school École des Hautes Etudes Commerciales (HEC-Paris) is highly selective, but basically based on individual merits, usually through a series of competitive exams. The French civil service generally takes a similar approach to identifying potential high-flyers.
Of course, that will be of little comfort to the many millions of people either looking for a job or unhappy with the one they have. Employment in general and “youth” unemployment in particular has been a major source of frustration – for decades. This is key to understanding both how this year’s presidential candidates might fare – and what the eventual winner may be able to accomplish.
To start with the issue of what the next president might be able to do, remember the 1980s? If you happen to be British or American, you might think of tax cuts, deregulation and other supply-side reforms. The French experience was very different.
François Mitterrand first came into office in 1981. It was the third time Mitterrand had run for the presidency, becoming the first left-wing President of France to be elected by universal suffrage. Mitterrand embarked on a far-ranging program of radical economic policies, with new taxes (including on wealth), increased workers’ rights, raising the minimum wage and generous new welfare entitlements. Nationalizations of major industrial groups as well as banks further reinforced the role of the state in the economy, already a dominant actor for much of the post-war period.
The result of all this were a rapidly deteriorating trade balance, rising unemployment and stubbornly high inflation, even as key trading partners such as Germany, the UK and the U.S. saw inflation fall. Markets panicked, with the franc coming under attack. By mid-1983, the experiment was abandoned, but not before the left lost its credibility on economic policy – and soon thereafter control of the National Assembly.
The episode neatly illustrates just how radical a new French administration, with a popular mandate and little previous governing experience, can be. Provided that it commands popular support and as long as it controls the National Assembly, there are few alternative centers of power that can slow, let alone stop, changes. The checks and balances in this particular instance did not come from within the French political system, but from the outside, via market pressures.
Alas, for those favoring market-friendly reforms, radicalism in the opposite direction is rare. Once again, checks and balances within the political system play a secondary role. Instead, mass protests, marches and national strikes have time and time again slowed, stopped and watered down reforms over the past twenty years, including under the two most recent presidencies.
Since the beginning of the year, Marine Le Pen has been leading in most polls for the first round of voting on April 23rd, and is currently polling at around 25-27%. She looks all but certain to qualify for the run-off on May 7th, which will be required if no candidate wins 50% or more. Having succeeded her father, Jean-Marie Le Pen, as leader of the Front National (FN) in 2011, she has aimed to soften her party’s image, while continuing to call for a crackdown on illegal immigration and a moratorium on legal immigration. In economic-policy terms, Ms. Le Pen has been sharply critical of free trade, free markets, globalization, the European Union and the euro. Financial markets fear that she may lead the country out of the Eurozone.
The current runner-up is political newcomer Emmanuel Macron. Aged 39 years, Mr. Macron is a former investment banker with Rothschild, who has never held elected office. A former advisor to outgoing President Hollande, Mr. Macron pushed for business-friendly reforms as Minister of the Economy between 2014 and 2016. His economic-policy priorities include – relatively modest – cuts in taxes and public spending, partly by not replacing some 120,000 civil servants due to retire over the next few years. Most promisingly, he plans to increase labor-market flexibility, increase spending on training the unemployed and tighten the criteria for claiming unemployment benefits. He is also the most europhile leading candidate. Most polls currently give him 23-25% in the first round. In a run-off with Ms. Le Pen, current polling suggests Macron would win, securing somewhere between 59-62%, with Le Pen at 38-40%.
Next in line is former Prime Minister François Fillon with 19-21% in the first-round polls. Mr. Fillon is the standard-bearer of France’s largest center-right party (renamed Les Républicains in 2015). He was the surprise winner of his party’s first-ever presidential primary and had been planning to run on a far-reaching economic-reform program of cutting taxes, spending and red tape. In recent weeks, he has instead been embroiled in a nepotism scandal over payments to his wife and two of his children. Putting family members on the parliamentary pay-roll is not unusual in France, but it remains unclear whether Mr. Fillon’s actually did any work. With a criminal indictment for fraud looming, Mr. Fillon has vowed to nevertheless continue his campaign. Polls before the latest revelations still put Mr. Fillon ahead of Le Pen in a second round match-up. In recent weeks, Mr. Fillon has been at 56-59%, compared to 40-43% for Le Pen.
None of the other candidates currently looks likely to get into the second round. However, there has been some speculation that Benoît Hamon, the socialist candidate and former Minister of National Education under Hollande, might team up with far-left firebrand Jean-Luc Mélenchon. Mr. Mélenchon has consistently been polling at around 10-12%, having won 11% at the 2012 presidential elections. Mr. Hamon has seen his polls fluctuate between 13-16%. It is unclear, however, how voters would react to any endorsement. The deadline for formally filing candidacies is March 17th, and the final list of candidates will be published on March 21st.
For markets, the worst-case scenario would be such an alliance allowing a far-left candidate to enter a second-round match-up against Ms. Le Pen; given Mr. Mélenchon’s eurosceptic stances, this could create massive uncertainty on the future of both the Eurozone and French economic-policy making.
Given the polling situation, it is tempting to discount the chances of Ms. Le Pen winning. This would be premature in our view, however. Admittedly, French polling has historically had a pretty good track record in correctly projecting results within a few percentage points well in advance of both first-round presidential elections and potential second-round match-ups. However, this has been less true for other contests, where big polling misses have been quite common.
A key stumbling block in the French context seems to be turn-out. This has generally been high for presidential contests, but less so, for example, in regional elections. It is tricky to know in advance whose voters will turn out. However, current polling suggests that Ms. Le Pen’s voters are most committed to their candidate. Those favoring Mr. Macron, her most likely opponent, are far less sure about their choice.
Moreover, the usual concerns about drawing conclusions from a handful of examples apply, only more so. There have only been 10 presidential elections so far, and none quite like this one. In particular, passed front-runners have tended to be veteran politicians. By the time François Mitterrand, Jacques Chirac, Georges Pompidou, Valéry Giscard d'Estaing or indeed Nicolas Sarkozy and François Hollande got elected, they had been in electoral politics for decades. Typically, they served in government, and more often than not they had either already run for the presidency or at least considered it. Voters had had plenty of time to form a view well in advance of the actual campaign.
By contrast, Mr. Macron is a political newcomer, with a limited presence and untested appeal outside the main urban areas. It remains to be seen, meanwhile, whether Mr. Fillon will be able to recover from his scandal. Ms. Le Pen is the subject of several legal investigations as well, notably into party funding. These may not harm her prospects among core supporters, but could make it even harder to win over new voters.
Plainly, plenty can still happen both before the first round and between the two rounds. Ms. Le Pen is a skilled and experienced campaigner and her party has been building up its party infrastructure throughout France for decades. Already, there are signs of foreign meddling, with Russia and its state-sponsored media stepping up efforts to damage Mr. Macron.
Of course, uncertainty cuts both ways. For example, polls significantly overstated the electoral strength of FN candidates, including both Ms. Le Pen and her niece Marion Maréchal-Le Pen in the run-off to the 2015 regional elections. Having come out ahead nationally in the first round, the FN was unable to win a single regional presidency in the second round. The reason the polls missed this was an unusual increase in turn-out. In 2015, many voters appeared to come out specifically to block FN candidates. It was a similar story in 2002, when Jean-Marie Le Pen shocked the nation by getting into second round, only to be resoundingly beaten by Jacques Chirac.
Market nervousness has mainly centered on concerns about France potentially leaving the Eurozone. Even if Le Pen won the presidency, this would hardly be a forgone conclusion. A more immediate question, however, is whether markets are missing the real story.
When following the French elections, we are keeping a keen eye on political betting markets.
Source: Oddschecker, Bloomberg Finance L.P., Deutsche Asset Management Investment GmbH; as of 3/8/17
It is doubtful whether betting markets can give reliable signals for this particular contest. However, they seem to provide an excellent proxy of what financial-market participants are currently thinking about each candidate’s prospects. Already, this has had a big impact on sovereign-bond spreads.
The link looks much less clear between the odds of various candidates and the French stock market. This seems a touch odd. After all, an alternative way to read the current situation is that two of the three leading candidates, Mr. Macron and Mr. Fillon, are running on a platform of economic reform. Their combined share of the vote for the first round has been hovering in the low- to mid-40s. By contrast, traditional left-wing candidates have been left far behind.
Clearly, change is in the air. And clearly, French voters, like their British and American counterparts are fed up with the status quo. But what if you live in a country like France, where for decades, the status quo has already meant being ruled by economic nationalists of various sorts?
As the American futurologist Roy Amara observed, humans tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run. The same is probably true of seismic political shifts.
We would continue to argue that Brexit and the election of Donald Trump illustrate new fault-lines, not captured by the traditional left and right political spectrum. Following these events, it seems sensible enough to look for signs of popular anger in other parts of the world.
However, be prepared that any revolt against the status quo might play out very differently, depending on the country in question. As we have seen, forces such as the FN in France, the PVV in the Netherlands and their counterparts in many other European countries, have been around for decades. In many European countries, they have participated in regional, local and even (directly or indirectly) in national governments.
For better or for worse, European politics remains a mish-mash of national politics. This is why this Special has focused on the two next contests, though we will, of course, have more to say about other upcoming elections in Germany, and, potentially, Italy. In the meantime, we would caution against reading too much into any one result.
France, in particular, has long followed its own political cycles, often proudly at odds with the Anglo-Saxon world. As other countries turn inwards, France might just surprise the world by embracing such “foreign” concepts as competition, free markets, innovation and – dare one say – “Laissez-faire.” Either way, it will be worth watching, particularly as we learn more on who the two final candidates will be and what they might do.
1 . See, in particular, the ground-breaking analyses by Downs, Anthony (1957) “An Economic Theory of Democracy”, Harper and Brothers; and by Nordhaus; William (1975), “The Political Business Cycle”, Review of Economic Studies, Vol. 42 (2), pp. 169-190
2 . Lindsey, David; Orphanides, Athanasios; and Rasche, Robert (2005), “The Reform of October 1979: How It Happened and Why”, Staff working papers in the Finance and Economics Discussion Series (FEDS)
3 . See, for example, Johnson, Robert; Chittenden, William; and Jensen, Gerald (1999), “Presidential Politics, Stocks, Bonds, Bills, and Inflation”, The Journal of Portfolio Management, Vol. 26 (1), pp. 27-31
5 . See, for example, Hays, Jude; Stix, Helmut; and Freeman, John (2000) “The Electoral Information Hypothesis Revisited”, Prepared for presentation at the Conference on “Globalization and Democracy”, University of Minnesota, May 26-27, 2000
6 . Pierdzioch, Christian; and Döpke, Jörg (2006), “Politics and the Stock Market: Evidence from Germany”, European Journal of Political Economy, Vol. 22(4), pp. 925–943; also see Hays et al. quoted above on the evidence on German bond markets.
8 . Note, however, that there have been comparatively few studies of political business cycles in developed markets outside the U.S., the UK and Germany. Partly, this is because financial markets (in particular for sovereign bonds) are less liquid in smaller European countries. Many of them also saw a change in the monetary-policy regime after the adaption of the euro, which creates a break in the data. For a solid overview in German of both the U.S. literature and the international evidence, see Bechtel, Michael (2009), „Tanzt auch die Politik auf dem Börsenparkett? Wahlen, Parteipolitik und die Entwicklung von Aktienmärkten“, Österreichische Zeitschrift für Politikwissenschaft, Vol. 38 (2), pp. 133 – 162
9 . For a basic overview on the Dutch political system, see “The Dutch Political System in a Nutshell”, published in 2008 by the Netherlands Institute for Multiparty Democracy and the Instituut voor Publiek en Politiek and available online at http://nimd.org/wp-content/uploads/2015/02/Dutch-Political-System.pdf
11 . The Economist, 12/11/16, “The Netherlands has found Geert Wilders guilty of hate speech”; both Mr. Wilders and public prosecution service (whose request of a 5000 euro fine the court had rejected) vowed to appeal the verdict
13 . For an overview of the other parties in contention and their prospects at the start of the campaign, see Louwerse, Tom, “The 2017 Netherlands election: Polls suggest mid-sized parties are now the new norm in Dutch politics” at the European Politics and Policy blog of the London School of Economics, available online at http://blogs.lse.ac.uk/europpblog/2017/01/24/2017-netherlands-election-mid-sized-parties-the-new-norm/
18 . Grossman, Emiliano and Sauger, Nicolas (2010), “The End of Ambiguity? Presidents versus Parties or the Four Phases of the Fifth Republic”, in: “France’s Political Institutions at 50”, Routledge, pp. 181-194
19 . Bertrand, Marianne, Kramarz, Francis, Schoar, Antoinette and Thesmar, David (2006) “Politicians, Firms and the Political Business Cycle: Evidence from France” available online at http://www.crest.fr/ckfinder/userfiles/files/pageperso/kramarz/politics_060207_v4.pdf. Another interesting finding was that politically connected firms did not, in fact, derive net benefits financially. This suggests that in France, unlike “in other countries political connections might have a large cost” (p. 28) to the shareholders of such companies.
20 . For an overview, see Sachs, Jeffrey and Wyplosz, Charles (1986), “The Economic Consequences of President Mitterrand”, Economic Policy, Vol. 1 (2), pp. 261-322, who convincingly argue that the French economy was already in bad shape when Mitterrand took over and, less convincingly, that he should receive some credit for changing course so quickly