Une semaine à la fois

L'extrême droite est arrivée largement en tête, et l'alliance de Macron en troisième position, du premier tour des élections françaises, mais le RN a obtenu des résultats inférieurs de 2 à 3 points à ce que certains des derniers sondages suggéraient. Pourtant, les projections de sièges placent toujours le RN près du seuil de la majorité. Ces extrapolations étaient cependant "mortes à l'arrivée" puisque l'alliance de gauche et les centristes retireront du second tour leurs candidats arrivés en troisième position pour tenter d'empêcher le RN de franchir le seuil, même si la portée exacte de ces "retraits mutuels" reste incertaine. Ce qui est en revanche assez clair, c'est que si les macronistes ont perdu tout espoir de s'approcher d'une majorité, une victoire de l'alliance de gauche serait désormais arithmétiquement très difficile compte tenu du nombre substantiel de circonscriptions dans lesquelles ils sont arrivés en troisième position. L'absence de majorité solide pour l'un ou l'autre des principaux blocs reste un résultat très plausible.

Un article paru dans le FT la semaine dernière a relancé l'intérêt pour les conditions dans lesquelles l'instrument de protection de la transmission de la BCE pourrait être déclenché pour aider la France. Selon notre interprétation, le fait que la France - et l'Italie - soient de nouveau sous procédure de déficit excessif ne serait pas nécessairement un obstacle tant que le gouvernement se montre prêt à se conformer aux recommandations du Conseil de l'UE. Cependant, compte tenu des doutes qui existent dans certains pays clés, la BCE devrait s'assurer que l'action est "proportionnée" au risque, ce qui suggère qu'un élargissement des spreads devrait sérieusement affecter l'économie pour que le TPI soit déclenché. Cet outil est là pour atténuer une crise, pas pour l'étouffer dans l'œuf. Quelle que soit la décision du gouvernement français, il ne doit pas compter sur le soutien immédiat de la BCE.

Enfin, si l'on considère l'actualité américaine de la semaine dernière, nous sommes paradoxalement encore plus confiants dans le fait que la Fed réduira ses taux en septembre, dans un contexte de meilleures nouvelles sur l'inflation et de signes supplémentaires d'un léger ralentissement de l'économie, mais aussi plus préoccupés par la possibilité qu'elle ne puisse pas réduire davantage ses taux en 2025 - ce qui aurait des répercussions sur les taux à long terme - alors que la probabilité d'une victoire de Trump et de la mise en œuvre de son programme inflationniste est de plus en plus élevée.

Vous pouvez lire l'analyse de Gilles Moëc ci-dessous.

French elections: round 1

The results of the first round of the French elections were broadly in line with the polls’ main messages, with Rassemblement National (including the component of the centre-right with which it forged an agreement) coming out first at 33% of the votes (slightly lower than some of the latest polls which put the party often in the 35-36% range), followed by the left-wing alliance (“Nouveau Front Populaire”) with 28%, the Macronist alliance at 21% and the “maintained” centre right at 10%. Across the seat projections for the second round made available immediately after the results, two pointed to RN coming out short of an absolute majority – albeit by a small margin if one takes the upper end of the range (230-280 for IPSOS, and 240-270 for IFOP, the threshold is at 289), while Elabe has RN passing the majority threshold within its range (255-295).

These projections are however extremely fragile since the very high turnout – even higher than what the polls were suggesting – means that the outcome of the second round will revolve around 307 three-way races, and 6 four-way ones, out of a total of 577 constituencies (note that 81 were adjudicated in the first round already, we don’t have results yet in 11). ​ However, the leaders of the left alliance announced they would withdraw their candidates where they came out third if RN came out first, and the Prime Minister called for similar withdrawals from Macronist candidates. It was unclear as of late Sunday night what is the exact scope of these withdrawals since some leading members of the Macronist alliance rejected the idea of withdrawing candidates when the opponent of RN is a member of the far-left component of Front Populaire. We will have to wait until Tuesday night – the deadline to register candidacies in the second round - to get a full picture.

Despite all these limits, some scenarios can already be ruled out. Even if this was already a very low probability before the first round, hitting the absolute majority threshold is now arithmetically out of reach for the Macronists. They got only 4 deputies elected in the first round (against 40 for RN). They qualified 312 candidates for the second round, which means that they would have to win 91% of their races to reach 289 seats, while some of these candidates will withdraw. But it would also be a tall order for the left alliance. They got 32 deputies elected in the first round already. In principle, their 407 candidates qualifying for the second round would need a “pass rate” of 63% for the alliance to hit 289 seats, but we counted 115 constituencies in which their candidates came in third position, i.e., with a near zero chance to win them. This means they would need a “success rate” of 88% in the races in which they came first or second, a very high threshold, especially since vote transfers from moderates to the hard-left candidates are likely to be difficult.

In a nutshell we are left with the two main scenarios on which we have been focusing so far: one in which RN wins an absolute majority by a narrow margin, and one – probably more likely on the basis of the results and the “mutual withdrawals” against RN - where none of the three blocks commands a majority. Note that in the latter configuration, forging a wide alliance from the moderates of the left to the centre-right, with the Macronists as the backbone, would likely be very fragile as the probability of seeing the RN and the hard-left command more than 289 seats together looks high. This means that any government based on such “federation of moderates” would constantly be at risk of being stopped by a motion of no confidence supported by the two extreme groups, for instance to reject a budget. We covered these issues in some details two weeks ago. There are institutional tools to deal with this sort of setup, but steering the economy would still be cumbersome.

Exploring the Transmission Protection Instrument

As we wait for the final outcome of the French elections – which may take weeks if not months after 7 July, to get a better sense of the kind of policies any new French government would seek to implement – a debate has emerged on the ECB’s capacity to provide help in case of market disruption. Focus is of course on the Transmission Protection Instrument (TPI), under which “the Eurosystem will be able to make secondary market purchases of securities issued in jurisdictions experiencing a deterioration in financing conditions not warranted by country-specific fundamentals”. When the TPI came out in July 2022 we were struck by how imprecise its “rules of engagement” were. The entire official documentation on the TPI spans a meagre 735 words. This vagueness was partly by design: to stop massive market pushes against a signature, creating uncertainty on how quickly and under which conditions central bank action could come can be asset. But we also suspect that to produce more granularity, in-depth conversations across the Governing Council would have been needed, and it is not obvious to us that a strong consensus existed – or exists today – on exactly how and when TPI can be activated.

The notion of “country-specific fundamentals” is itself subject to interpretation. In our understanding, economic policy should be part of those fundamentals. In clear, a government could hardly expect to benefit from ECB intervention if the market is rationally responding to concrete policy decisions which are likely to deteriorate its public debt trajectory. This point is supported by the reference in the documentation to assessing whether jurisdictions which may receive TPI support “pursue sound and sustainable fiscal and macroeconomic policies”. ​

The document then lists a series of criteria towards such positive assessment. Not being under an Excessive Debt Procedure is mentioned – which at first glance would exclude France and Italy – but so is “not being assessed as having failed to take effective action in response to an EU Council recommendation under Article 126(7) of the Treaty”. Since Article 126(7) itself refers to Excessive Deficit Procedure – the recommendation lists the measures a government should take precisely to exit EDP - this suggests that “being under EDP” is not necessarily enough to deprive a government from the benefit of TPI, as long as it is taking steps – in accordance with the guidance by the EU Council– to correct its fiscal trajectory.

In any case, the ECB would have extensive leeway to trigger TPI or not. The assessment is presented as “an input” into the decision-making, the word “condition” has been on purpose avoided. The spirit of TPI, in our understanding, is that a country in fiscal difficulties could benefit from ECB support but would at least need to show some readiness to accept instructions from Brussels. If a stubborn government in France were to ignore them, the benefit of TPI could still be extended to the peripheral countries to mitigate any contagion. Note in any case that the instrument’s documentation makes it plain that triggering TPI would need to be “proportionate” – a nod to the German Constitutional Court. In other words, it would need to be plain that the spread widening is having significant adverse macroeconomic effects. German Finance Minister Lindner has already expressed his reservations – to say the least – on using the tool. In our view, TPI would be there to mitigate a crisis, not to nip it in the bud altogether.

Fed cuts getting more likely, but so is a Trump victory

While the French – and peripherals’ - spread has risen since the announcement of snap elections, financial conditions did not tighten much as global long-term interest rates were kept in check by growing signs of materialisation of a soft landing in the US which would allow the Federal Reserve (Fed) to cut rates in September. Last week’s dataflow has strengthened this view, but at the same time the first presidential debate in the US forces us to raise the probability that Donald Trump returns to the White House with a mandate to implement his inflationary platform from next year onward. So, we are somewhat paradoxically at the same time more confident the Fed will cut this year and more confident long-term interest rates will not fall much.

Let’s start with the dataflow. The print for the Personal Consumption Expenditure deflator (PCE) for May confirmed the message from the Consumer Price Index (CPI): core inflation is falling again. Core PCE fell in line with market expectations to 2.6% year-on-year from 2.8% in April, with both major components (goods and services) contributing to the deceleration (see Exhibit 1). The changes are easier to see in the short-term momentum. In Q1, the Fed had indeed reasons to get nervous as, after nicely coming very close to 2% in the second half of 2023, the 3-month annualised change of core prices rose above 3% and then to 4% between January and March (see Exhibit 2). This was what triggered the shift in the Fed’s tone in the early spring: while one month of re-acceleration can easily be dismissed as “noise”, two could not be ignored and three called indeed for a message change. This should be symmetric though. ​ The reassuring April print could have been attributed to mere data randomness, but these favourable dynamics continued into May, and was broad-based, even if the relapse in goods’ prices was larger than the deceleration in core services (see Exhibit 2).

The San Francisco Federal Reserve Bank produces a breakdown of core PCE between a cyclical and an acyclical component. This is based on submitting each line of the PCE index to a simple statistical test: does it follow a Phillips curve, in clear, is possible to substantiate that it responds to a change in the unemployment rate. A bit less than 40% of the PCE components (measured by weights) pass that test. The rest follows idiosyncratic, industry-specific patterns. An example of the latter is the price of car insurance, which has played a big role in inflation gyrations in the US lately. Exhibit 3 illustrates clearly that the resilience of core PCE in late 2023/early 2024 was the product of a re-acceleration of the acyclical component, while the contribution from the cyclical one was by then only slightly falling, much less abruptly than at the beginning of 2023 (see Exhibit 3).

Prudence is needed when interpreting this breakdown. Indeed, that some price components usually respond to cyclical conditions does not mean that all, or even most of their recent gyrations can always be ascribed to changes in the state of economy. Yet, the absence of re-acceleration of the cyclical component in the first few months of 2024 would pour some cold water on the idea that the monetary policy stance has not been restrictive enough recently.

Federal Open Market Committee (FOMC) member Mary Daly – who has been increasingly vocal on the dovish side for a few weeks – was the first to react and took the May print as proof that “monetary policy was working”. She was clear though that the Fed would need more confirmation in the data releases ahead before changing the stance, concluding with “it’s really too early to tell”. ​

Yet, beyond the information brought by the inflation prints, the Fed cannot ignore the signals coming from the real economy. The latest revision to Q1 GDP – which at 1.4% annualised came short of potential – was significant for consumer spending, which is now estimated at 1.5%, down from a first shot at 2.0%. 1.5% is “fine” but by no means roaring. Personal consumption rebounded in May, gaining 1.0% annualised after a 0.5% decline in April, but the carry-over for Q2 currently stands at a modest 1.2%. Consumption seems to be operating below trend now. What is also striking is that some aspects of the US surprising resilience in the face of the monetary tightening are disappearing. This would be consistent with a resumption of the decline in the cyclical component of core inflation at a faster pace. ​

We have often commented in Macrocast how residential investment was surprisingly strong despite the rise in mortgage rates above 7%. It contributed a still solid 0.6 percentage points to Q1 GDP. This is changing. After months in healthy territory above 1.5mn, building permits have been correcting in April and May. Housing starts, which had re-accelerated at the end of last year, are also down (see Exhibit 4). Non-residential investment has also supported US growth nicely, thanks in particular to the federal programs CHIPS and IRA. While the series is quite volatile and we would not want to draw definitive conclusions from one print, factory orders excluding defence and air transportation – normally a good gauge of investment dynamics – fell in May. This could be a signal that momentum is getting lost there as well.

There is nothing alarming in such dataflow, but even under the assumption that “only” a soft landing is starting to materialise, we think the Fed could take the risk of starting to remove some of the current restriction in its stance: even if the neutral rate is higher, bringing the Fed Funds rate to 4.75%-5.00% from the current level would still leave the central bank with a lot of assurance against another revival of inflation.

We note however that, counter-intuitively, the US 10-year yield rose on Friday despite the good news on the inflation front. We are tempted to link this to Joe Biden’s counter-performance in the first presidential debate against Donald Trump the night before. Indeed, the probability that the Republican candidate wins in November is today higher – the polls surveyed by 538 suggest that ordinary voters and not just pundits saw Trump as the clear winner of the debate and the readiness to vote for Biden fell even among those who lean Democrat - and so is the probability that on top of winning back the White House, the Republicans also win both the Senate and the House of Representatives, allowing for a high “conversion rate” of their economic platform into actual policy. The latter is probably key in the current thinking among Democrats. Replacing Joe Biden would not necessarily help them retain the presidency – setting up a campaign on a new name in a few months could be difficult, especially if it comes after weeks of public wrangling across the various factions of the party to choose the replacement – but they could still move to this solution if polls start indicating that sticking with Biden would not only cost them the Presidency but also Congress and possibly lots of state-level positions.

At the top of Donald Trump’s agenda, three items – curbing immigration, prolonging expiring tax cuts and raising trade tariffs – would all be inflationary. Their implementation could make it difficult for the Fed to continue easing into 2025. This is likely to put a floor on how low long-term interest rates can go, even if the Fed cuts in 2H 2024. ​