As we have been documenting daily (most recently in "More Container Ships Score "Astronomical" $100,000/Day Rates"), even as some signs of price normalization emerge, most notably the recent drop in commodity prices (lumber, copper)...
... widespread supply-chain disruptions remain and have become a serious challenge in many parts of the US economy, with large effects on both output and inflation.
The Goldman chart below shows that businesses report the worst supplier delivery delays in decades and expect delivery times to remain longer than normal through year-end and potentially well into 2022. In special questions included in regional business surveys, a substantial majority of firms in the manufacturing sector have consistently reported experiencing supply chain disruptions over the last several months, although as the Fed's Vice Chair for supervision Randy Quarles just said, he hopes that supply-chain bottlenecks are "likely temporary" as the economy reopens.
But what if they are not? It wouldn't be the first time the Fed is dead wrong about something being temporary only to become permanent. For an example of that look not further than the Fed's ballooning balance sheet which once upon a time was also a "temporary" fixture of monetary policy only to gradually become an immutable part of the Fed's response.
To answer this and other questions, over the weekend Goldman's economists looked at the underlying causes of the supply-demand imbalances in many areas of the economy, especially the goods sector, and discuss the timeline for normalization. The bank also discussed when and to what extent the price spikes generated by these imbalances should begin to reverse and what that means for the broader inflation outlook this year and next.
The bottom line for those pressed for time is that according to Goldman the current one-off inflationary boost from supply chain bottlenecks will eventually become a one-off disinflationary drag, with the bank expecting "fading fiscal support for households and a shift in spending to services over the next few months to ease demand pressures." And on the supply side, look for the semiconductor shortage to improve later this year, especially for the autos sector. Overall, the contribution to year-on-year core PCE inflation from supply-constrained categories has risen from -20bp before the pandemic to +105bp today, but we expect it to decline to +35bp by end-2021 and to -55bp by end-2022
Below we pull the most notable highlights from the report:
Goldman first looks at the demand side, where consumer spending on durable goods has risen to levels 15-25% above trend (Exhibit 2, left) due to the boost to disposable income from fiscal support, the unavailability of many services, and pandemic preference shifts. The further surge in demand for durables this year likely reflects the impact of the stimulus checks (Exhibit 2, right), which past experience has shown are disproportionately spent on durables and used for down payments on autos. Even under better circumstances, suppliers would likely have struggled to keep up with such a rapid surge in demand
On the supply side, there are three main contributing factors.
First, company anecdotes reveal that some businesses initially cut back because they expected demand to fall in a recession, and others had to stop production because of the virus. For example, automakers closed plants to improve safety at the start of the pandemic, resulting in a large drop in car production (Exhibit 3, left). They also reacted to the decline in demand at the start of the pandemic by ordering fewer microchips, which then led chipmakers to reduce production. Similarly, cash-strapped rental car companies that initially faced a sharp drop in demand sold roughly a third of their fleet (Exhibit 3, right).
Second, production has been hampered by supply chain disruptions, in particular the negative supply shocks that caused the global semiconductor shortage. The shortage of microchips has affected many consumer goods including electronics, appliances, and automobiles. The chip shortage caused auto production to fall for a second time over the last four months, bringing the shortfall of domestic auto production since the start of the pandemic to about 2.5mn (Exhibit 3, left). Combined with strong demand, this has resulted in a dramatic depletion of new and used car inventory (Exhibit 3, center), making it hard for consumers to find new cars and for rental companies to rebuild their fleets.
Third, the global shipping industry has experienced a number of problems—a shortage of shipping containers, the Suez blockage, clogged ports, and covid outbreaks at key ports—that have slowed deliveries and substantially increased shipping costs (Exhibit 4).
The effect of these supply-demand imbalances is evident in the large drawdown in retail inventories and reports from businesses that they have struggled to keep inventory at acceptable levels (Exhibit 5). These shortages have caused the price increases discussed below.
* * *
Which brings us to the $64 trillion question: what is the Timeline for Normalization, i.e., when will these supply-demand imbalances start to improve?
Here Goldman is - as usual - optimistic that things will normalize more or less on their own, as elevated demand for durable goods should moderate over the next few months as fiscal support for household incomes diminishes and spending shifts to services.
Meanwhile, on the supply side, Goldman argues that the semiconductor shortage should begin to improve later this year. Here's why:
In the short run, chipmakers are prioritizing producing chips for autos, partly at the expense of chips for consumer electronics. In the medium run, supply will increase as existing semiconductor plants that ordered new machinery to boost production at the end of last year bring their new capacity online. It takes about 6-9 months for new tools to be delivered, a month for installation and testing, and two months for the new capacity to start producing chips, meaning that global supply should increase around year-end. In the longer run, new plants being built in the US and abroad should be running in about two years, which will further boost global supply to meet rapidly rising demand. The net effect, our sector analysts believe, is that the large recent price increases will end this year, but the global semiconductor market will remain tight until 2023 and price trends will remain firmer than the deflationary pre-pandemic norm.
Next, Goldman notes that auto production schedules and guidance from major US automakers indicate that the chip shortage likely had its peak impact in Q2 (Exhibit 6). If true, auto production should jump in July as many plants skip the usual summer shutdowns, and should remain strong in coming months as automakers complete cars that they have produced with everything but the components requiring chips. The bank assumed that "the chip shortage should be resolved across the industry by early next year, and the flow of new production should normalize between now and then." However, even here Goldman concedes that "it will take until at least mid-2022 to fully restore auto inventories."
Goldman then posits that outside of sectors affected by the semiconductor shortage, supply should also gradually catch up to demand. The bank cites a report by the Council of Economic Advisors which notes that the supply-demand imbalance for toilet paper at the start of the pandemic was resolved not through investment in new capacity—which would have been very costly and unnecessary once the demand surge passed—but rather though raising the utilization rate on existing productive capacity. The same pattern appears to be happening now in many areas. Industries with below-normal inventory-to-sales ratios are operating at above-normal utilization rates on average, and many appear to have the ability to boost utilization rates further if necessary (Exhibit 7).
But what about the biggest supply shortage of all - that of labor? Isn't it critical to have the available labor before assuming anything about production?
Here Goldman agrees that whether there will be enough labor available in the US to address the current supply-demand imbalances as immediate disruptions are overcome is a very "legitimate question." After all, the bank's economists note, "the labor market is already facing widespread worker shortages, and reopening services industries will need to hire more workers too."
Of course, goalseeking its cheerful outcome, the bank has determined that "these worker shortages will diminish in coming months, primarily because we expect the expiration of federal unemployment benefits to substantially increase labor supply." As a result, by early September, Goldman expects over 5 million workers to lose their extended duration benefits entirely, over 3.5 million to lose their $300 weekly federal top-up, and nearly 6 million to lose benefits from the Pandemic Unemployment Assistance program, though many of these are likely already working part-time.
Goldman is not alone in its cheerful assessment: business surveys - the same ones that scream there is a record labor shortage - also indicate that companies anticipate being able to raise shipments substantially through year-end (Exhibit 9), suggesting that they expect to at least partially overcome current supply constraints. Here, too, companies expect that the labor market will magically renormalize as soon as the emergency unemployment benefits expire. This will prove to be a very myopic assumption but we'll cross that bridge in a few months...
* * *
Assuming Goldman is correct, and both demand and supply-chain bottlenecks are resolved (on their own) sometime in late 2021 and early 2022, what does this imply for inflation? After all, by now even Joe Biden knows that supply-demand imbalances have raised prices in a number of categories to levels well above pre-pandemic trends.
The most visible examples are new, used, and rental cars; the microchip-affected consumer electronics category of video, audio, photographic, and information processing equipment; and categories that have seen very large increases in demand during the pandemic such as sports and recreational vehicles, sporting equipment, furniture, and appliances.
As Goldman notes, "the boost from these categories, combined with large base effects in reopening travel services categories where prices are normalizing from depressed year-ago levels, have pushed year-on-year core PCE inflation to 3.39%, the highest rate since 1992."
Looking ahead, the squid economists predict that these supply-constrained categories are likely to remain firm for at least a couple more months. While used car auction prices seem to have finally stabilized, consumer prices usually lag auction prices by a month or two.
Consumer electronics prices are also likely to rise somewhat further as the full impact of the microchip shortage materializes. More broadly, both domestic and foreign upstream input prices have risen sharply over the last year, and while bottlenecks have likely now peaked globally, changes in consumer goods prices tend to lag changes in input costs by at least a couple of months.
So while Goldman does not anticipate a normalization in the coming months, looking further ahead, the bank expects most of these temporary increases in price levels to at least partially revert to pre-pandemic trends: "this means that the current one-off inflationary boost will eventually become a one-off disinflationary drag" Goldman cheerfully predicts, adding that "this drag will be larger and arrive sooner where prices adjust downward or retailers restore regular discounts, and it will occur more gradually where prices are downwardly rigid and reversion to trend instead occurs through slower future price increases."
Of course, Goldman is right. The question is when will all this happen. Here is the bank's answer:
"we base our assumptions about price reversion by category on the timelines discussed above for supply-demand imbalances to be resolved. For new cars, we expect only slight reversion to trend this year and one-third reversion by end-2022 because some of the price increases so far appear to reflect persistent cost increases. For used cars and rental cars, we assume one-quarter reversion by end-2021 and two-thirds reversion by end-2022, reflecting immediate downward pressure from a more normal flow of new cars this year, but continued upward pressure from inventory shortages that will last well into next year. We make a similar assumption for sports and recreational vehicles. For consumer electronics, we assume prices will rise further above trend this year due to the chip shortage, but revert one-third of the way to trend by end-2022. Finally, in the furniture, appliances, and sporting equipment categories, we assume only slight reversion to trend this year because of order backlogs for furniture and further effects of chip shortages on appliances, and then roughly one-third reversion by end-2022."
These very optimistic assumptions also serve as the basis for category-level forecasts that Goldman uses in its bottom-up core PCE inflation model, which the bank updated over the weekend in the June edition of its Monthly Inflation Monitor. And only here does Goldman concede that its entire forecast may be dead wrong (again), admitting that its "assumptions about trend reversion in price levels are deliberately somewhat conservative because statistical analysis suggests that past category-level departures from trend have a mixed record of reverting."
In other words, in a perfectly normal, non-centrally planned world things should normalize by 2022. But in this one, where everything is now mandated by the Fed's central planning committee, all bets are off. But of course, Goldman can't say that, so instead the bank concludes with a solid dump of GIGO saying that based on its model output, the total contribution to year-on-year core PCE inflation from these supply-constrained categories has risen from a fairly normal -20bp before the pandemic to +105bp today. But under the bank's forecast the total contribution will decline to +35bp by December 2021 and then to -55bp by December 2022 as currently-elevated prices partially revert to pre-pandemic trends. In other words, today's inflation will become tomorrow's deflation. That's why Goldman concludes that "the drag from these categories and the fading base effects in reopening travel services categories are the two main reasons that we expect core PCE inflation to decelerate to 3% by December 2021 and 2% by December 2022, even as highly cyclical categories such as shelter accelerate as the economy continues to recover."
In other words, we now know what Goldman's economic model has spewed out. What happens in the real world, however, is another matter entirely.