- During the lockdown, some 20% of the Dutch economy came to a standstill.
- Economic activity rebounded after the reopening in June …
- …. but high-frequency data indicate that ongoing measures and negative sentiment continue to inhibit the revival.
- Total consumer spending in June was still 10% lower than last year.
- Dutch economy will not recover to year-end 2019 level until end of 2022.
The Dutch economy largely reopened for business on 1 June
Part of the lockdown measures were relaxed on 11 May. Although some spending categories benefited from this, we already concluded that the revival was not spectacular (see: Consumenten gaven 19% minder uit tijdens de lockdown.). The measures were eased further on 1 June amidst hopeful talk of a ‘reopening’ of the economy (see list of the relaxed measures below). But not everything was given the green light; night clubs, for instance, remain closed and large events that require a licence are not yet permitted. That said, the Dutch economy has on the whole returned to the pre-coronavirus normal. But is this reflected in the economic data? In this publication we analyse high-frequency data to determine how successful the reopening of the Dutch economy has been.
Economy rebounded in June, but not to pre-coronavirus level
ABN AMRO transaction data indicate that, despite the reopening, total average consumer spending in June was still below the level in the same month in 2019. We calculated on 20 May that consumers spent about 19% less during the lockdown. This ‘nowcast’ (fast calculation) was confirmed by Statistics Netherlands (CBS), which reported on 22 June that consumer spending was some 17% lower in April than in April 2019. DNB data show that households saved almost EUR 10 billion in May, the biggest increase in savings since the series started in 1998. Based on our transaction data, we have calculated that average consumer spending in June was still 10% below the level of June 2019. Sectors such as entertainment and travel are still contending with substantial contraction due to the ongoing partial restrictions. But even sectors that largely reopened for business at the start of June are struggling to lure back customers. Spending on clothing, footwear and wellness have not rebounded to pre-coronavirus levels, and the tale is similar for hospitality venues. Only retail remained buoyant, with nominal revenue up 8.2% and volume up about 6% in May 2019 (CBS). Our transaction data indicate that especially supermarkets and local speciality shops generated more revenue than normal, both online and offline. This trend was sustained in June.
Across the board, consumption in June remained below the level of June 2019, partly due to the 1.5 metre rule, continuing fears of infection and increased income insecurity. There is no sign of ‘catch-up growth’, suggesting that a V-shaped recovery with a fast recovery to pre-coronavirus economic growth can be ruled out. Worse still: the initial contours of a 95% economy are slowly becoming visible.
Mobility patterns recovering slowly, but not to pre-coronavirus levels
Remarkably, in contrast with other high-frequency data that we see, the mobility data barely show a trend break since the lifting of the lockdown on 1 June. After hitting a low point at the end of March, mobility soon started to rise steadily again (see Apple Maps data) and this increase continued in the month of June. Driving and walking returned to pre-coronacrisis levels in mid-June. Public transport capacity was still limited in June, as is reflected by the data showing that public transport usage is some 40% lower than normal. The Google data also confirm that the mobility pattern is gradually returning to normal. Movements to and from supermarkets returned to pre-coronavirus levels even before the lockdown ended. Commuter mobility remained stuck at some ‑20%. Movement patterns to shopping streets and recreational facilities are largely recovering, but paused at ‑8% at the end of June. Our transaction data show that this was largely due to the sharp decline in recreational spending and, by extension, recreation-related mobility. These data suggest that continuing caution is preventing a full return to pre-coronavirus mobility patterns.
Nor is business staging a full recovery
Much of the focus in recent times has been on consumer spending, partly because of the forced closure of hospitality venues. But the coronavirus is obviously also taking its toll on activity higher in the chain, i.e. in the business to business sectors. As well as hospitality, retail and leisure, sectors such as transport and industry are also being hit.
The transport sector provides services to all other sectors, both directly and indirectly, which makes it extremely sensitive to the changing fortunes of the economy. Aviation and shipping were already confronted with the coronavirus crisis at the end of January, when China largely shut down. Most passenger flights to and from China were cancelled because of travel restrictions. The shipping sector contended with diminishing freight volumes as Chinese industry closed down as well as with logistical problems due to the closure of port transhipment services and local container shortages. In recent months shipping has had to deal with severe supply disruptions across the globe, lower industrial output and a sharp decrease in international trade. According to CBS figures, the number of movements of seagoing vessels in Dutch ports was no less than 17% lower than in the same month last year.
Road haulage is less sensitive than other transport sectors to logistical problems and international restrictions (barring, of course, long tailbacks due to border checks). And it is also very flexible: transportation by lorry from A to B can usually be arranged in a few days. That’s more complicated for seagoing shipping, where cargo is sometimes en route for weeks from, for instance, China to Rotterdam. Besides being responsive to changing economic conditions, road haulage also caters to very diverse types of customers such as construction, agriculture, food, shops, e-commerce, hospitality and manufacturing. The number of road haulage movements is therefore a suitable indicator for measuring current economic activity in the Netherlands. The number of lorry movements clearly reflects when the lockdown occurred and when the restrictions were relaxed: from early January to mid-March, lorry movements were on average 4.5% higher than in the same period last year; from week 12 to week 22 (the lockdown period) the average number was 5.7% lower (year on year); and towards the end of the lockdown, lorry movements picked up again – presumably driven by the restocking of shops, restaurants and bars and the reopening of factories (e.g. car and lorry production lines).
The number of transport movements has not fully recovered. Since early June, the number of transport movements is still 1.4% lower on average than the same period last year (adjusted for national holidays and free days). The decrease is presumably due to the cancellation of large events, lower spending in shops (e.g. clothing), less hospitality consumption, and lower industrial output. As a result, the expansion at the start of the year, which was driven by the growing economy, has turned into contraction.
Still far fewer passenger flights
Aviation still has a long way to go before it is out of the coronavirus woods. The spike in infections grounded most flights to and from China in late January. And in March all international flights to and from Europe also ceased, while flying within Europe was discouraged as far as possible. Figures of Schiphol Airport show that the number of flights to and from that airport in April nose-dived almost 90%, plummeting from nearly 42000 in April 2019 to just over 4200 in the same month this year. The number of passenger flights is also still much lower than before. All sorts of travel restrictions remain in force. For several months the European Union (EU) refused to admit passengers from most other countries. On 1 July the EU reopened its borders to passengers from fifteen ‘safe countries’. Figures of FlightRadar24.com show that the number of flights to and from Schiphol in June was just over 200 per day – a far cry from the 43,000-plus flights (over 1400 per day on average) in June 2019. The number of flights in June was still
some 85% lower than in the same period last year. That is better than May, however, when the number of flights was down by some 90%, just as in April. Presumably, air passenger numbers are recovering slowly because air travel requires a certain amount of planning (such as asking for holiday leave). The number of flights is expected to rise in the coming weeks thanks to the relaxation of the travel restrictions and the start of the holiday season.
Investments and industrial output, too, have not yet recovered. According to the CBS, investments in tangible fixed assets were some 10.6% lower than in the same month last year. Industrial output in April was 11% down on April last year. The Nevi manufacturing purchasing managers index has progressively improved from 41.3 in April and 40.5 in May to 45.2 in June, but remains below 50 and thus still signals a decrease in business activity (see: Nederlandse industrie ondanks malaise voorzichtig optimistisch.).
Leading indicators for investments point to weak recovery in June
The CBS investment figures for June are not due until the end of August, which is why we will look here at two leading indicators: the MSCI Netherlands Financial Index, which is made up of share prices of Dutch financial institutions, and electricity prices in the options market. Both indicators have a strong correlation with investments (gross investments in tangible fixed assets) and economic activity (GDP growth, quarter on quarter). The MSCI gives an indication of the health of the financial industry, a key sector for the financing of investments. The second indicator approximates actual electricity usage. Demand for electricity is generally greater when companies are operating at full power. Consequently, electricity prices are an indicator of the level of business activity.
Although these high-frequency variables must be interpreted with caution, the general picture appears to be clear: economic activity visibly accelerated in June compared to May, but is still a long way short of the pre-coronavirus level. The MSCI index shows that the financial markets already anticipated a downturn due to the coronavirus in mid-February. Despite a significant recovery after the reopening, this index is still far below the level at the start of this year. The electricity options market also sustained a small decline in early 2020, which then accelerated due to the lockdown. Electricity prices appear to be largely recovering to the February level.
Signs of growing confidence
The CBS confidence indicators paint a more positive picture. Entrepreneurs are less negative. Producer confidence rose from ‑25.1 in May to ‑15.1 in June. The outlook on expected business activity showed a particularly strong improvement. The deterioration in this outlook which occurred in April has thus been largely cancelled out. Consumers, too, are less sombre according to the CBS. In the second half of June consumer confidence stood at ‑26 versus ‑29 in the second half of May. The willingness to buy and consumers’ estimation of their own economic situation in the coming twelve months both produced less negative readings. Even leaving aside the methodological issues, drawing conclusions based on confidence indicators is a tricky exercise. However, our own sentiment-indicator – which distils the prevailing sentiment from Dutch news based on Natural Language models – provides further evidence of a strong revival (see also: Consument gegrepen door negatief sentiment.). The rapidly falling infection rate combined with the accelerated relaxation of the measures has injected a dose of optimism into the popular mood.
Is a revival in confidence good news for the future?
The high-frequency indicators for the month of June show that the Dutch economy is up and running, but not yet at full power. Economic activity is rebounding quickly, but is still inhibited by ongoing measures to contain the virus, cautious consumers and businesses and a contraction in foreign demand for Dutch products and services. The capacity of the Dutch economy cannot be fully utilised. We expect this ‘95% economy’ to remain the new reality until a vaccine becomes available and a large proportion of the Dutch populated is vaccinated. This, we assume, will be achieved in mid-2021. But until then, the Dutch economy (GDP) will not return to its pre-coronavirus level. All in all, ABN AMRO expects the Dutch economy to contract by 5.6% in 2020, followed by a modest recovery of 2.7% in 2021 – which is by no means sufficient to make up for the loss sustained in 2020. We do not expect GDP to return to its pre-crisis level before the end of 2022.
To sum up: a strong recovery in Q3, followed by moderate growth
Q3 2020 will see a revival estimated at about 5.5% to 6% q-o-q, as the coronavirus crisis peaked in Q2 when the economy ground to a virtual standstill during the lockdown (our expectation for Q2: -10% q-o-q). Q4 growth will be meagre (our expectation: between 0.5% and 1% q-o-q). On top of the coronavirus-related restrictions and continuing caution among consumers and businesses, second round effects will also have a stronger impact in Q4 2020 and Q1 2021. Higher unemployment, tighter financial conditions, corporate bankruptcies and supply disruptions will dampen the revival.
The coronavirus is a new phenomenon for economists. As most official consumption data appear with a time lag of weeks and sometimes months, the direct impact of the coronavirus on the Dutch economy is hard to estimate. Our payment systems process millions of transactions every day. Based on these transactions, we can calculate consumption in aggregate terms and broken down by service and product groups. These data are not shared externally, only the analyses based on these data are made public on our website (without commercial motive).
These data have not been adjusted for inflation or other fluctuations.
April 2020 also had one working day less than April 2019
 We ran all MSCI variables that are available via Bloomberg for the Dutch economy (22 variables) through a ‘best subset regression’. Based on three selection criteria (R2, SME and AIC) we found that the MSCI Financial Index has the greatest predictive value for GDP growth.
 Both variables are also influenced by other factors. Bank share prices correlate in the short term with interest rates, regulatory changes and credit default expectations. Electricity prices respond strongly to variances in solar and wind supply, capacity storage restrictions and oil and gas prices.
 Interpreting consumer confidence is not straightforward. This confidence indicator is based on the balance of the positive and negative answers. Consumers who indicate that their financial or economic situation has not changed or say they don’t know are not included. This method makes it difficult to determine whether consumers have really become more or less positive. If their underlying financial or economic situation initially deteriorated but subsequently remained the same, respondents will change their answer from ‘worse’ to ‘unchanged’, leading to a more positive balance and therefore misleadingly suggesting a revival of consumer confidence.
 We use a combination of Natural Language Processing models (DistilBERT, Vader en TextBlob) to analyse news articles. These models quantify the ‘positive or negative feeling’ that people get when they read a news article based on the wording in the article. We standardise these results between the values of +0.25 and -0.25, which is the margin within which sentiment moves (+ is positive; – is negative). News articles about culture, art, food and sports are filtered out. We look at a selection of articles about politics, climate, economics, international relations, opinion and social themes, We apply this method to articles about the Netherlands and analyse some 250 articles per day. The data go back to 2005.