AXA IM: “The risk of a ‘double dip’ scenario is gaining ground”

Stockholm Syndrome

In the Euro area, the inflation figures for February vindicate the Governing Council’s decision to signal its intention to deliver another 50-bps hike on 16 March. The slowdown in economic activity is not yet having the expected moderating impact on price behaviour. We continue to think the monetary tightening has however started to work its way through the economy, even if it has not yet visibly impacted aggregate demand. The data for January has confirmed the steep decline in the credit impulse. The historical relationship with GDP with a lag of a quarter is relatively tight, even if base effects from the pandemic may have blurred the picture and ample liquidity buffers may delay the transmission to business and consumers decisions.

Cracks usually appear first in the most interest-rate sensitive sectors of the economy. The recent developments in Sweden are interesting from this point of view. Transmission should be swift there, given the dominance of variable rate mortgages and an extremely stretched housing market before the tightening began. GDP contracted significantly in Q4 there (-0.9%) amid sharp declines in house prices. The Riksbank is in a quandary. If it stops hiking rates, it risks fuelling the depreciation of the currency and hence imported inflation. Accepting a quite deep recession – and the collateral risks to financial stability – may be the only workable avenue there. The UK could be the “next shoe to drop”. The recent dataflow is not as bad as in Sweden, but the housing market correction has started. 

The ECB Governing Council is visibly tempted by additional tough moves beyond March, but we feel that the risk of a “double dip” scenario is also gaining ground. Transmission may be slower than in the UK or in Sweden because of structural features of the Euro area economy – with the dominance of conservative lending practices and fixed-rate mortgages in many member states – but it does not necessarily mean that down the road the impact is smaller. 

Stubborn inflation

In February, headline inflation failed to decelerate as much as expected in the Euro area, hitting 8.5% year-on-year, barely lower than in January (8.6%) and 20 basis points above expectations. Energy duly retreated but food prices continue to be a major issue at 15%. Food prices have significant ramifications beyond their share in the index (c.20% for the broader definition), as they trigger one of the most socially regressive effects – they t disproportionately hurt those of the bottom of the income distribution – which may call for further costly fiscal treatment, and as frequent expenditure, their impact on overall price perceptions and hence expectations can be salient. Even if there is nothing central banks can directly do to deal with this sort of shocks – often driven by weather conditions or the lagged effect of energy prices since agriculture is one of the most energy-intensive activities- it’s something they can’t ignore from a forward-looking point of view. 

This will compound the ECB’s frustration with core inflation, which has accelerated in February, hitting 5.6%, year-on-year up from an already concerning 5.3% in January and coming out 30 basis points above expectations. There is no solace to be found in base effects: on a 3-month annualized basis, the acceleration is also visible (see Exhibit 1). There is no solace either in the sectoral breakdown. Prices in both non-energy industrial goods and services. The former is a puzzle in its own right. Given the normalization in supply lines and the end of the global shift in consumption towards goods, one could expect some deceleration there, as it has been happening in the US. Two explanations can be brought forward: first, the lagged effect of the euro depreciation last year, which has affected import prices, and second the continuation of the pass-through from the energy price shock which has been stronger and more persistent than in the US. These effects are likely to gradually fade, but of course, the acceleration in services prices may reflect second-round effects from stronger wage growth, pointing to the beginning of a self-sustained process bringing further persistence. 

Exhibit 1 – Up, not matter how you look at it

Beware the macro-financial channel

In any case, the slowdown in economic activity is not yet having the expected moderating impact on price behaviour, although it now emerges that GDP probably fell in the Euro area in the last quarter of 2022. The 0.1 quarter-on-quarter gain reported in the first estimate had come as a positive surprise. Even if energy supply did not break in Q4, removing risks of a “sudden stop” in output in key sectors, business surveys were still consistent with a trip into shallow contraction (see Exhibit 2). Yet, the Euro area data had been pushed up by one of those regular bumper figures from Ireland, where GDP have first been estimated at 3.5 % qoq before being revised down by a factor 10 to 0.3%. In addition, the drop in German GDP has been brought to -0.4% qoq in the second estimate, from -0.2% initially. All components of final domestic demand declined there, and the acute drop in household consumption (-1.0%) was the biggest on record since the Great Recession of 2009 if one excludes the lockdowns. GDP did not fall into properly concerning territory in Germany at the end of last year only because inventories brought a positive contribution – which does not necessarily bode well for Q1 if a large share of this inventory build-up was involuntary – while imports crashed more than exports (-1.3% versus -0.9%), which is hardly reassuring either. The direct effect of these revisions in the two countries would send the Euro area print to -0.1%. It’s not a massive difference in the great scheme of things – even if the psychological effect of hitting negative territory matters – but at least it rebuilds some trust in the relationship between the surveys and quarterly national account. 

Of course, one could argue this is already “old history” and focus on the fact that technically, a recession – defined as two quarters of contraction in a row - is likely to be avoided anyway since the current dataflow is consistent with marginally positive growth in Q1 (0.2% qoq using our “adjusted PMI” relationship). Yet, we want to highlight that the “financial side” of cyclical analysis would point to severe difficulties ahead. We have updated with the January data our measure of the credit impulse – the year-on-year change in the flow of new loans to businesses and households. They confirm what we have already seen with the December batch: the decline in lending to corporations is very significant (see Exhibit 3). 

Exhibit 2 – Short trip in contraction in Q4 2022

Exhibit 3 – Deeply negative credit impulse

In general, the relationship between the credit impulse (summing the CI for households and businesses) and GDP is quite tight if one corrects for some volatility and exclude the pandemic period. In Exhibit 4 we compare the average over two-quarters of the credit impulse with GDP growth in the next quarter from Q1 2002 to Q4 2019. 

The fit is quite satisfying – we can explain 60% of the variance – and the elasticity is very close to 1. On the basis of this historical elasticity, GDP would fall by 2% year-on-year at the beginning of 2023 (the red “dot” on Exhibit 4). Given the carry-over it would take an unrealistic collapse in output in the coming months to get there.  As things stand today, the relationship is probably noisier. Even if the bulk of the pandemic impact on credit came in 2020 and early 2021 – with the impact of the government-guaranteed schemes - base effects may still reduce the information content of the most recent values of the credit impulse. Moreover, the credit impulse focuses only on the liability side of the balance sheet of non-financial agents. We have already suggested in Macrocast that the unusual level of liquid assets in the non-financial sector -itself a legacy of the pandemic – may blunt the effect of the tightening in financial conditions and credit contraction on spending decisions.

Yet, beyond the data noise, it is an important warning to keep in mind, and in any case what we think is clear from these figures is that it is only at a stretch that the central bank can argue that it has brought interest rates “barely” in restrictive territory.  To quote exactly from the latest ECB minutes: “any assessment of what level of rates could be seen as excessively restrictive was complex and uncertain, although it was generally felt that concerns of “overtightening” were premature at the present”. A risk, as always, is that “nothing happens” until the economy snaps brutally, and this data from the macro-financial channel could be a harbinger of painful developments ahead.