La première coupure est la plus profonde

Le projet de loi de finances du gouvernement français pour 2025 prévoit un effort discrétionnaire de 1,4 % du PIB. Il s'agit d'un montant important, l'effort le plus important du plan budgétaire et structurel à moyen terme (PSFMT) pour 2025-2028. Il est toujours préférable d'anticiper les mesures plutôt que de promettre une vertu future, surtout lorsque les conditions politiques sont fragiles. Si l'on examine simplement la manière dont les ratios des dépenses et des impôts par rapport au PIB sont prévus dans le projet de loi du gouvernement, on constate que la majorité de l'effort provient d'une augmentation des impôts. Comme nous l'avons indiqué il y a deux semaines, compte tenu de la fragilité de la demande intérieure, ce choix est compréhensible à court terme. Le problème - souligné par exemple par l'agence Fitch lorsqu'elle a décidé d'assortir sa notation de la France d'une « perspective négative » - est que de nombreuses mesures du projet de loi 2025 sont temporaires. Maintenant que le gouvernement a présenté un « budget d'urgence », il devra esquisser un programme de mesures plus systémiques, en se concentrant probablement sur les dépenses, étant donné le niveau déjà élevé de la fiscalité.

À court terme, le marché se concentrera probablement sur la capacité politique du gouvernement à faire adopter le budget. Un processus de « non vote » via l'article 49.3 de la Constitution est probable, la vraie question étant alors de savoir si une motion de censure aboutirait. Nous partons du principe que le budget sera adopté, mais que le gouvernement devra peut-être accepter d'en atténuer les aspects les plus impopulaires. Nous examinons les précédents historiques pour évaluer le PSFM français. L'effort global pour parvenir à un déficit de 3 % en 2029 ne serait pas sans précédent - il se situe entre les trajectoires d'assainissement passées des années 1990 et 2010. Dans les deux cas, cependant, la France a bénéficié d'une réduction massive des taux d'intérêt qui, selon nous, ne peut pas être reproduite cette fois-ci, et la dette n'a diminué que marginalement à la fin de ces phases d'ajustement.

Nous jetons également un coup d'œil de l'autre côté de l'Atlantique. L'indice des prix à la consommation américain plus élevé que prévu pour le mois de septembre est un autre élément qui montre que la Fed a peut-être « sauté le pas » avec sa réduction de 50 points de base en septembre, mais nous prévoyons toujours des réductions de 25 points de base en novembre et en décembre.

2025 is not the hardest

Emergency action in France

Last Thursday, the French government transmitted to parliament the budget bill for 2025. There were no surprises relative to what had abundantly leaked in the press the previous week. This probably explains why the 10-year spread between OATs and Bunds did not react much on the release day. As of last Friday evening, it stood at 77 basis points (bps) – towards the upper end of the range observed since the dissolution of parliament in June, but still below a recent peak to 80 bps on 29 September – and still 3 bps above Spain (a more recent feature).

The government is communicating on an effort of EUR60bn, shared between tax hikes for one third and spending cuts for two thirds. This uses as a reference a “business as usual” scenario under which the deficit would have spontaneously increased from 6.1% of GDP in 2024 to 7.0% in 2025, 2 % of GDP higher than the government believes it can bring it thanks to the bill’s measures. As the High Council of Public Finance noted in its advice, this is however based on a very strong trend increase in public spending corrected for inflation, 2.8%, which would exceed by far trend GDP growth. This mechanically inflates the effort on the spending side. The council concluded that 70% of the adjustment came from higher tax, but also that the overall structural effort – i.e. the discretionary improvement in the deficit, net of any impact from the cycle – amounts to 1.2% of GDP for 2025, in line with the government’s claim, and 1.4% when taking on board the increase in interest payments of 0.2% of GDP. While the notion of primary structural adjustment is important from a macro point of view – this is equivalent to the “fiscal stance” and matters to gauge the impact on aggregate demand – it relies on estimates of potential growth and output gap which are always imprecise. When looking at ratios to actual, and not potential GDP, as we think many observers will, the story is simple: in 2025 tax in the government’s own forecast would reach 43.6% of GDP, up from 42.8% in 2024, while spending will only marginally fall from 56.8% of GDP to 56.5%.

Now, as we argued two weeks ago, these choices are understandable in the current configuration. Indeed, domestic demand has been weak so far in 2024 in France – consumer spending has been flat over the first half of the year and both corporate and household investment has been contracting – and drastic action on government spending could have precipitated the country into recession. The government chose to act on individuals at the upper end of the income distribution, whose propensity to consumer is lowest (EUR2bn from a “minimum tax” levied on those making more than 250K a year), and focused much of the effort on the corporate sector to the largest enterprises (EUR 8bn), with arguably the best capacity to “take this on the chin” without deeply revising their investment and hiring plans.

Still, even the smaller businesses will not completely escape the austerity drive, with potentially some adverse effect on employment. Since the mid-1990s, successive governments in France, across the political spectrum, reduced the social contribution rate on lower wages to boost employment. While this has probably contributed to the decline in unemployment in France on trend, these various measures entailed a sizeable cost to the government (c. EUR75bn per annum, i.e. 2.5% of GDP), while dis-incentivising business to promote workers: the upward slope of the social contribution rate means employers must consent to a disproportionate rise in their overall cost of labour when they want to raise their employees’ direct salary. That the system needs to be reformed makes little doubt, but the consensus view was that the overall financial envelope should be better distributed to avoid the worst threshold effects, not that it should necessarily cut, especially if this means raising labour costs close to the minimum wage, where the price elasticity of labour demand is the highest. This is however the choice made here, with a gross effect of EUR 5bn (4bn net when taking into account the mitigation from the corporate tax). In the same vein, social contributions are going to rise significantly for apprentices. The system had become very costly, and in some cases triggering windfall effects, but the emergence of apprenticeships has undoubtedly contributed significantly to the overall rise in employment in France these last few years (1/3 of the total job creation between 2019 and 2022).

Now, on top of trying to minimize the impact on demand, the government had little time to work on the budget and the recourse to “low hanging fruits” to send a signal was understandable. The thorniest issue in our view is not so much the much-discussed tax versus spending breakdown of the 2025 effort, but rather the possible trajectory beyond next year to ensure a proper stabilisation of French debt.

Looking beyond 2025

The government pledges to bring the deficit back to 3% of GDP in 2029 from 5.0% in 2025. In its own medium-term planning, this entails the continuation of a decent size structural adjustment (0.6% of GDP every year on average) after the steep effort of 2025 (1.4%). This “front loading” is commendable, and on the whole the trajectory looks do-able, as it would sit right in the middle of the two previous episodes of French fiscal consolidation, 1994-1999 and 2011-2015, as illustrated in Exhibit 1 (in the graph, “year zero” is the lowest point of the primary structural deficit before the improvement starts, and we stop the “consolidation phase” the last year the primary structural deficit improves). There is thus nothing unprecedented in what the government is proposing to do. Of course, the level of debt is much higher than at the beginning of these two consolidation phases (see Exhibit 2), and the government is not expecting to see the debt ratio fall before 2027.

A key problem – which has been mentioned by Fitch as one the factors motivating its decision to put the French rating on negative watch last Friday – is that it is very difficult to extrapolate from the measures announced in the budget bill for 2025 what could constitute the elements of a multi-year, maintained austerity drive. Indeed, most of the tax measures are designed to be temporary. In 2026 already the surcharge on corporate tax for the largest companies will fall, before – in principle – disappearing altogether in 2027. The same will apply on the income tax surcharge for the wealthiest individuals. The postponement of the inflation indexation of pension from January to July 2025 will have a permanent effect on the level of the pension payments in France (6 months of inflation will be lost “forever”) but in “delta” – i.e. triggering another braking effect on spending – it is a one-off. Beyond 2025, France will need to produce more systemic changes, and focus will need to be on spending. In 2025, some savings are expected by reducing the day-to-day spending of some ministerial departments (Defence, Police and Justice will be spared), while in Education the decline in the number of pupils – a consequence of population ageing – will allow some limited staff cuts. But if the modes of action, organisation and perimeters of public administration stay untouched, parametric approaches will quickly touch their limits.

Beyond the need to agree on some systemic reforms, the French government will have to be lucky with its growth trajectory. The Treasury has revised down its estimate of potential GDP growth in France to 1.2%, but still expects actual GDP growth to exceed this trend every year after 2025 (from 1.1% next year to 1.4% in 2026 and 1.5% in 2027 and 2028). In clear, this means that there is no space for any “accident” on the way. A source of difficulty there lies in the “feedback loop” from the fiscal adjustment itself: if the Treasury sees GDP growth at 1.1% in 2025 despite a primary structural effort of 1.4% of GDP the same year, it means that the underlying growth rate – i.e. how the economy would fare with a neutral fiscal stance – would have to be well above 1.5%, even when positing a very low multiplier. The government is counting on a drop in the savings rate of households and businesses in 2025 which would completely offset the rise in the government’s saving to support aggregate demand. As we have discussed two weeks ago, these are strong assumptions: while the households’ savings ratio is historically very high, which could “justify” a mean reversion next year, the general sense of uncertainty over France’s economic and political prospects may induce more precautionary saving, while expecting a rise in business spending when taxation is rising and profit margins are already moving down is brave.

True, the French economy will benefit from the ECB’s new stance. Such a lucky configuration happened before. GDP growth was very strong during the 1990s adjustment (see Exhibit 3). There was an “accident” in 1996, precisely when the fiscal stance was at its most restrictive – an interesting observation when forecasting 2025 - but GDP grew by more than 2% almost every year. This was made possible, largely, by the massive decline in French interest rates as they were converging towards the German level as part of the European monetary convergence process. Yet, public debt did not start falling before 1999, 5 years into the fiscal adjustment process. ​

A decline in interest rates is however not a source of certain success. During the 2011-2015 adjustment the ECB provided extraordinary support, bringing interest rates in negative territory, and ultimately resorting to Quantitative Easing. This allowed to reduce the interest payments of the French government significantly. When Paris submitted its 2011-2014 Stability Programme to the European Commission in 2011, the Treasury was positing a gradual increase in the interest payments to 3% of GDP by 2014. The outcome was a slight fall to 2.1% (see Exhibit 4). Yet, public debt started declining in 2018 only, and in 2014 stood nearly 10% of GDP higher than targeted by the Stability Programme.

The mid-2010s fiscal efforts were thwarted by the poor performance of the French economy at the time. The government was counting on a replication of the 1990s trajectory, with GDP growth exceeding 2% (see Exhibit 5), which failed to materialise.

Interestingly, the government this time is forecasting an increase in the average interest rate on debt over the entirety of its forecasting horizon (see Exhibit 6). This may be surprising given the ECB’s new stance – we expect it to bring its policy rate down to 2% in June 2025. Yet, the transmission of this monetary easing through the yield curve may not be massive, since (i) the ECB is reducing the size of its balance sheet, lifting the net supply of government bonds and (ii) a return to the next to zero inflation regime of the pre-Covid regime is unlikely, which will put a floor on the term premium. While the share of public debt which was issued at the trough in interest rates is declining (it takes about 7 years to roll the French debt), even with an accommodative ECB the average cost of the stock of debt is likely to rise on trend.

“Politicking ahead”

So, in a nutshell, the medium-term plan looks achievable by historical standards, but the government will need luck on the growth trajectory, and of course, crucially, will need to come up with systemic reforms which will have the power to prolong the adjustment beyond the “one offs” of 2025. This is of course daunting in the currently fragmented political landscape. Still, of course we cannot predict the political set-up which will prevail over the next 5 years, but our impression is that at the current stage the market’s focus really is on whether the government will be able to pass the 2025 budget bill through parliament. The key dates are the vote in parliament on the “receipts” part of the budget for the central government, on 29 October, before a vote of the overall package (receipts and spending) on 19 November. Then normally the bill goes to the Senate, and the two chambers are given until 21 December to converge (if they do not, the National Assembly has the last word), offering a few days for the Constitutional Court to opine. The budget bill must be promulgated by 31 December. This calendar can however be disrupted if the government triggers a “no vote approval” via article 49.3 of the Constitution. The bill will then pass through without a formal vote, unless a motion of no confidence is tabled and is supported by the absolute majority of the National Assembly.

Maintaining discipline within the parliamentary groups supporting the minority government is not proving easy. The tax hikes in the budget bill have been publicly criticised by some heavyweights in President Macron’s party. Former Interior Minister Gerald Darmanin signalled he would vote against the receipts part of the budget. There is however little risk in our view that deputies of the central parties would go as far as voting a motion of confidence if the 49.3 process is triggered, which should now be the baseline.

So far, the left has managed to remain united in its opposition to Barnier, and while the Rassemblement National (RN) has so far pledged to “give Barnier a chance”, supporting a budget which is likely to be unpopular is likely to be a step too far for them. There is however a difference between actively voting for a budget and abstaining in a motion of no confidence (which the left would likely table anyway).

RN is thus likely to be the deciding force as to the fate of such a motion of no confidence. Since the institutional solutions in case of a budget rejection – and government resignation - are scarce (no new general elections can be organised before July 2025), and RN has endeavoured to appear as a stabilizing force, it is plausible that they would choose not to join a motion of no confidence. It may however be necessary for them that they obtain concessions from the government to justify such approach to their electorate. One of the spending-brake measures – the 6-month delay in the indexation of pensions to inflation – has been immediately criticised by the RN, but the government has expressed a readiness to “discuss” it with parliament. While we are bracing ourselves for a lot of political noise in the weeks ahead, it is our baseline assumption that this budget bill will pass– possibly with some amendments – and shape the French economy in 2025. ​ ​

Unlucky Fed

The US dataflow has not been kind to the Fed since it inaugurated its change a stance with a 50-bps cut and a quite dovish dot plot. We explored in some details last week how the employment report for September continued to paint the picture of a resilient US economy, casting a doubt on the urgency of proceeding by “jumbo cuts”. The consumer price index for September, released last week, did not provide much respite. Core CPI rebounded slightly in year-on-year terms to 3.3% from 3.2%, while the market was expecting an unchanged reading. Beyond the monthly volatility, what is emerging is a general stabilisation – or in a more pessimistic fashion the emergence of a “line of resistance” - of core inflation above 3% since late spring (see Exhibit 7), a still uncomfortable pace relative to the Fed’s target (given the usual gap between the CPI and the PCE, the Fed’s favourite measure of inflation, 2.5% on core CPI would be consistent with target). The short-term momentum is not very reassuring, with the second acceleration in a row on the 3-month annualised rate (see Exhibit 8). The Fed had communicated on its tolerance of still robust gains in rents, which weight in core CPI exceeds 40%, given the favourable trend in real-time indicators of new leases, but the acceleration in “services less shelter” has been steep in September (from 0.7% on a 3-month annualised basis in August to 3.0%). “Explaining away” the September print is not straightforward.

The minutes of the FOMC meeting in September released last week indicate that the debate on the magnitude of the first cut was much fiercer than what the lone dissent of Michelle Bowman suggested at the time of the announcement, and of course the recent dataflow is helping those who support a prudent pace of “restriction removal”. Yet, we do not think there is enough to seriously alter our baseline course of easing by 25-bps increments in November and December, in line with the market’s current expectations (45 bps worth of cuts priced in as of last Friday). Among the Fed speakers who took to the wires after the CPI print, only Bostic has expressed a readiness to “skip” the November meeting – and he has been on the very hawkish spectrum of the FOMC in the current cycle so far – while Barkin, Goolsbee and Williams downplayed the significance of the September inflation number. For 2025, a continuation of cuts, stopping above the Fed’s estimate of the neutral rate as proxied by its forecast of the long-term level of the Fed Funds rate (2.9%) is our baseline, but with a high level of uncertainty, as we have been arguing for months, given the impact the outcome of US elections could have on fiscal and trade policy.