Pandemic increases risk of high inflation - Capital Economics
US Economics

Pandemic increases risk of high inflation

US Economics Focus
Written by Paul Ashworth
Cancel X

The pandemic has increased the odds that the US will eventually experience a period of high inflation, principally because we expect the Fed to be less committed to ensuring price stability in the future. The higher public debt burden, slower global labour force growth, and the possibility that globalisation will be partly reversed, are additional reasons to expect inflation to gradually rise over the longer-term to 3% or 4%. Any deflationary pressure from technology is likely to be muted, while several unique factors linked to the pandemic mean that the risk of a near-term slide into deflation is low.

  • The pandemic has increased the odds that the US will eventually experience a period of high inflation, principally because we expect the Fed to be less committed to ensuring price stability in the future. The higher public debt burden, slower global labour force growth, and the possibility that globalisation will be partly reversed, are additional reasons to expect inflation to gradually rise over the longer-term to 3% or 4%. Any deflationary pressure from technology is likely to be muted, while several unique factors linked to the pandemic mean that the risk of a near-term slide into deflation is low.
  • Inflation did fall in the early stages of the pandemic, but that was due to big declines in prices for the worst-affected sectors like air fares and hotels. As restrictions were eased, prices for those goods and services began to rebound and that reflation has been compounded by inflationary effects in other sectors.
  • Because the lockdowns triggered declines in both production and spending, with the latter recovering more quickly than the former, inventories are unusually lean for this stage of the economic cycle, which is putting upward pressure on goods prices. In addition, ongoing physical distancing restrictions have increased costs and reduced supply in many services sectors. As a result, we expect headline inflation to average 2.5% next year, before dropping back to 2.2% in 2022. Core inflation should average 2.3% next year and then edge back down to 2.2% in 2022.
  • Beyond the near-term forecast horizon, the pandemic has left the balance of longer-term risks to inflation more skewed to the upside. In particular, more activist monetary policy could eventually lead to a bout of higher inflation, not because central bankers don’t have the tools to rein in demand, but because they will be persuaded not to use them by pressure from politicians and society more broadly.
  • We are not too concerned by the Fed’s policy actions during that pandemic, which led to a massive acceleration in money growth. That won’t be inflationary when the economy is still operating well below its capacity and, besides, the Fed could rein in money demand by simply raising the interest rate it pays on excess reserves. But we are concerned by the potential long-run implications of the Fed’s recent changes to its policy framework. Those changes in isolation do not represent a full-blown return to the negligent monetary policy of the 1970s, but they are the first step down that path.
  • The pandemic has also generated another big increase in the Federal debt burden to more than 100% of GDP. There is not a direct causation between higher public debt and high inflation, but governments often end up resorting to various methods of financial repression, which encourage a rise in prices that reduces the real value of that debt.
  • The pandemic may further accelerate the deglobalisation trend that has been building, with the global shortages of some key goods illustrating the dangers of over-reliance on imports and complex global supply chains. We expect that US-China relations will continue to deteriorate even under a Biden administration. Just as globalisation and the dismantling of trade barriers put downward pressure on goods prices in recent decades, deglobalisation could have the opposite effect. The coming slowdown in global labour force growth and weaker global productivity growth are additional reasons to, at the very least, expect the downward pressure on goods prices to ease.
  • Finally, although we remain optimistic that new technologies will eventually trigger a revival in productivity growth, we expect any deflationary pressure to remain limited. The prices of digital hardware continue to fall rapidly, but digital services now account for a much bigger share of the economy and those prices are rising.

Pandemic increases risk of high inflation

In our US FocusInflation: the next 20 years” published in 2017, we concluded that:

Whether the biggest long-term risk is high inflation or deflation ultimately depends on policymakers. High inflation has followed periods of loose fiscal policy, whereas monetary policy errors were a key factor in the 1930s Great Depression, and arguably Japan’s current deflation too. As it stands, the Fed remains an ardent inflation fighter committed to its 2% target. The question is whether its commitment will remain that steadfast or whether, in response to academic fashion or political pressure, its focus changes?

In this new Focus, we look at how the pandemic and the resulting recession have affected the balance of risks between deflation and higher inflation. Our conclusion is that balance has been skewed more toward the possibility of excessively high inflation developing, particularly because of the Fed’s more relaxed attitude to inflation and the renewed surge in public debt, with the near-term risks of a slide into deflation much lower now.

Inflation rebounding more quickly than normal

With inflation persistently undershooting the 2% target over the past decade, even as the unemployment rate fell to almost unprecedented lows, we have long argued that deflation could be only one recession away. But, because of several factors that made this year’s recession unique, it won’t be the one that tips the economy into deflation. Although the recent collapse was the most severe on record in terms of lost output and the peak in the unemployment rate, it will probably end up being less disinflationary than a run-of-the-mill recession, let alone the downturn caused by the financial crisis just over a decade ago.

There are three factors that made the pandemic-related recession unique: First, although the magnitude of the decline in GDP was unprecedented, as large swathes of the country went into a near-complete lockdown, the recovery was also unusually rapid after those restrictions were lifted. The strong responses from monetary and fiscal policy played key roles in driving the early stages of that recovery.

Second, in a normal recession, aggregate demand falls off quickly, with supply reacting much more slowly, leading to a glut of unwanted inventory, which puts downward pressure on prices until it is cleared. But the pandemic lockdowns triggered simultaneous declines in both output and spending. Furthermore, because of problems with getting supply chains restarted, output has rebounded more slowly than spending. The upshot is that, uniquely for this early stage of a recovery, inventories are unusually lean and that is already putting upward pressure on goods prices.

Finally, ongoing physical distancing restrictions have increased costs for most businesses and, for the worst-affected services sectors like restaurants, air travel and hotels, resulted in a dramatic drop off in supply too. As a result, even though the elevated unemployment rate should be putting downward pressure on prices in those labour-intensive services sectors, we are instead seeing inflation rise.

The initial impact of the lockdowns triggered by the pandemic was deflationary, with core CPI and PCE inflation falling sharply. But the decline was mainly due to very big declines in prices for the worst-affected sectors – including air fares, hotels, motor vehicle insurance and clothing. (See Chart 1.) More recently, however, as restrictions have been eased, prices for those goods and services have either stabilized or rebounded and the recovery has been compounded by inflationary effects in other sectors.

Chart 1: Selected CPI (Feb = 100)

Source: Refinitiv

Admittedly, core inflation remains well below the Fed’s 2% target and lower than it was before the pandemic struck. (See Chart 2.) But the reversal has occurred within a few months, whereas it took several years before core inflation bottomed out a decade ago.

Chart 2: Core Inflation (%)

Source: Refinitiv

In addition, the lack of inventory is putting upward pressure on a broader range of goods prices. As Chart 3 shows, the business inventory-to-sales ratio was at a six-year low in September, because the post-lockdown rebound in output has lagged the rebound in sales. (See Chart 3.) During the last recovery a decade ago, it took two years for firms to work down their excess inventory. But in this cycle that adjustment was completed in only two months.

Chart 3: Business Inventory-To-Sales Ratio (%)

Source: Refinitiv

The inventory shortages are most acute in the motor vehicle sector, which has had a knock-on impact on used vehicle prices too. But, as Chart 4 shows, it is a broader issue with the inflation rate for household furnishings and supplies quickly rebounding, driven by a massive surge in appliances prices, which have also been in short supply. Appliances price inflation is running as hot now as it was in late 2018; when the economy was enjoying the tail wind of a massive tax cut and the Trump administration had just imposed significant tariffs on imported laundry appliances. Contrast that with the aftermath of the financial crisis, when prices fell sharply for several years, before finally beginning to rebound in 2011, when the excess inventory had been worked off.

Chart 4: Household Furnishings & Appliances Inf. (%)

Source: Refinitiv

The upshot is that core goods inflation has already rebounded sharply. (See Chart 5.)

Chart 5: Core CPI Inflation (%)

Source: Refinitiv

The costs of complying with ongoing physical distancing requirements mean that core services inflation is also holding up better than might have been expected.

Housing inflation has begun to fall, but the slump is smaller than we saw in the aftermath of the collapse of the housing bubble a decade ago, not least because low interest rates have also supported home sales and prices. (See Chart 6.) But for the Fed’s preferred PCE measure – food and health care services inflation are rising, as supply is restricted and businesses, like restaurants and dentists, incur substantial costs to ensure physical distancing is maintained.

As a very labour-intensive sector, we would normally expect the high unemployment rate, via slower wage growth, to be putting downward pressure on inflation in food services, which is exactly what happened in the aftermath of the last recession. But the cost-push upward pressure on prices, from the impact of supply constraints linked to complying with physical distancing restrictions, has dominated.

Chart 6: Selected PCE Services Inflation (%)

Source: Refinitiv

Health care services inflation is a little more complicated. The CPI measure was already elevated before the pandemic began because the BLS introduced a new methodology for tracking health insurance costs in 2017. That new methodology, which wasn’t applied to the historical data, shows a much more rapid pace of appreciation. (See Chart 7.) this year, however, medical insurance inflation has dropped back sharply, which explains that fall in the medical care services CPI inflation rate.

Chart 7: Health Care Services Inflation (%)

Source: Refinitiv

Whereas the CPI measure only covers out-of-pocket expenses by households, the PCE measure is much broader – incorporating health care spending by private insurers and the government on behalf of households. After bottoming out in 2015, PCE health care services inflation has been gradually rising since then. Moreover, the acceleration has increased since the pandemic began, presumably because treating coronavirus victims has reduced the ability of hospitals to treat other patients.

This is crucial because health care services accounts for almost 18% of the core PCE, making it roughly the same size as the housing component. The upshot is that the surge in health care services inflation has offset the moderation in housing inflation during this recession.

This unusual upward pressure on services inflation generated by the pandemic restrictions is also observable in the survey evidence. In the aftermath of the financial crisis, the ISM services prices paid index remained at depressed levels for several years. In contrast, within a few months of this recession ending it had already climbed above the pre-pandemic level. (See Chart 8.) It is currently consistent with a further rebound in the core PCE inflation rate toward the 2% target.

Chart 8: ISM Prices Paid & Core PCE Inflation

Source: Refinitiv

Finally, even the downward pressure on wage growth and price inflation may not be as severe as one would have expected given the surge in the unemployment rate to a peak of nearly 15%.

First, the unemployment rate has fallen much more quickly than is normal for the early stages of a recovery, dropping to slightly below 7% in October. In contrast, although the unemployment rate only peaked at 10% coming out of the last recession, it took several years before it fell below 7%.

Chart 9: Unemployment Rate By Reason (%)

Source: Refinitiv

Second, the spike in the unemployment rate during the pandemic was, unusually, almost exclusively people reporting that they were only on a temporary layoff. Admittedly, the rate of permeant job losers has picked up in recent months, but it remains below the peaks reached in the aftermath of the last two recessions. (See Chart 9.)

As a result, the decline in the proportion of small businesses reporting that jobs are hard to fill has been much smaller than the headline unemployment rate would have suggested. (See Chart 10.) There is a corresponding mismatch between the unemployment rate and the net proportion of households reporting that jobs have become hard to get, with the latter also suggesting that there is not as much labour market slack.

Chart 10: Jobs Hard to Fill Index & Unemployment Rate

Source: Refinitiv

As a result, although the survey evidence on future compensation plans does point to a decline in the growth rate of average hourly earnings, that decline should be smaller than the drop in the early stages of the post-financial crisis recovery. (See Chart 11.)

Chart 11: Plans to Raise Compensation Index & Average Hourly Earnings

Source: Refinitiv

Inflation expectations already building

The unexpected rise in inflation expectations is a final reason to believe that the recovery in actual inflation will be quicker during this recovery. Despite the decline in actual inflation, households’ medium-term inflation expectations have continued to trend higher this year. (See Chart 12.)

Chart 12: Households’ Inflation Expectations (%)

Sources: Refinitiv, NY Fed

Breakeven inflation compensation rates, calculated from nominal and inflation-protected Treasury securities, did fall during the early stages of the pandemic and are still at a relatively low level compared with the average for the past decade. But the rebound was nevertheless quite prompt, leaving the longer-term compensation rate above its pre-pandemic level in February. (See Chart 13.)

Chart 13: Breakeven Inflation Compensation – 5yr/5yr Forward (%)

Source: Refinitiv

Short-term inflation outlook

The upshot is that, despite the initial deflationary impact of the pandemic during the lockdowns, inflation is already rebounding more rapidly than is normal for this early stage of the economic recovery. Lean inventories and, at the margin, the weaker dollar are putting upward pressure on goods prices. The added costs of complying with physical distancing requirements and supply restrictions are boosting services prices.

To the extent that the Phillips curve is still relevant, there may be less slack in the labour market than the elevated unemployment rate suggests, which implies that the moderation in wage growth could be smaller than we saw following either of the last two recessions. Finally, because of the big, but temporary, declines in prices during the early stages of the pandemic, strong base effects will briefly push inflation well above the 2% target next spring.

Our forecasts for CPI inflation are shown in Chart 14. We expect headline inflation to average 2.5% next year, before dropping back to 2.2% in 2022. Core inflation should average 2.3% next year and then edge back down to 2.2% in 2022.

Chart 14: CPI Inflation (%)

Source: Refinitiv

Normally, we would expect core PCE inflation to be roughly 0.4% points lower than the core CPI rate but, with PCE medical services inflation accelerating, that gap is currently only 0.1% points. Since that acceleration started as far back as 2015, there is little reason to believe it will die out any time soon.

As a result, we also forecast that, driven by the base effects, core PCE inflation will temporarily climb above the 2% target next year before falling back in line with the target in 2022. (See Chart 15.) But, given the persistent undershooting of the target over the past decade, even just returning to the target on a sustained basis would be a remarkable shift. Whether inflation continues to rise further above the target, however, depends on how the longer-term structural forces play out.

Chart 15: PCE Inflation (%)

Source: Refinitiv

Long-term inflation outlook

In our 2017 Focus on the long-term inflation outlook we identified five factors that could trigger a return to higher inflation:-

  • The surge in global labour supply was a one-off step change that happened more than 20 years ago. Global labour force growth will slow over the next few decades.
  • Developing countries have less scope for catch-up productivity gains that reduce goods prices.
  • Globalisation was also a one-off step change and could go into reverse because of the backlash caused by rising income inequality in developed countries.
  • The commitment of central banks as inflation fighters could wane.
  • Government debt burdens are at generational highs.

In this next section we recap how those factors will affect the long-term outlook for inflation, with a particular focus on how the pandemic has altered the final three.

1. Global labour force close to peaking

Back in the 1970s, when capital and trade flows were more restricted and the world was split along ideological lines, US workers were sheltered from global competition. But the collapse of the Soviet Union and the Warsaw Pact communist bloc in Eastern Europe, symbolised by the pulling down of the Berlin Wall in 1989, brought 400 million people into the global labour force. Deng Xiaoping’s rise to power in China, after the death of Mao Zedong in 1976, gradually put American workers in competition with an additional 1.2 billion people, particularly after China’s accession into the WTO in late 2001.

As Charles Goodhart and others have noted, however, the integration of more than one billion people into the global labour force in the 1990s was essentially a one-off supply shock. In addition, the aging of the population, not just in advanced economies but China too, means that the global working-age population is now growing at a much slower pace. (See Chart 16.) Most notably, the working-age population has peaked in both high-income countries and China, and it will contract over the next few decades. As a result, not only will the downward pressure from the earlier integration of workers on global wages wane, but the slowdown in global labour force growth could push wage inflation higher.

Chart 16: Working Age Population (%y/y)

Source: UN

2. Global productivity growth has slowed

China’s trend productivity growth has also slowed, bringing down the average for emerging markets and compounding the slowdown in productivity growth evident in most developed countries. (See Chart 17.) That will reduce the scope for China to further lower the costs of production for its goods. Given China’s size, there is only limited scope for production to move to other low-cost emerging market economies like Vietnam, Thailand or the Philippines.

The slowdowns in global labour force growth and productivity growth were well established before the pandemic hit and, since these are deep-seated structural forces, there is little reason to believe that the pandemic has changed the outlook. Nevertheless, these are still reasons – albeit not necessarily the key ones – why we think there is a risk of inflation rising more markedly in the longer term.

Chart 17: Labour Productivity (%y/y)

Sources: GGDC, Conference Board

3. Globalisation going into reverse

As far as globalisation is concerned, a lot has changed in the three years since we wrote our last Focus on the long-term inflation outlook. The pandemic may further accelerate the deglobalisation trend that has been building, with the global shortages of some key goods illustrating the dangers of over-reliance on imports and complex global supply chains.

We expect that US-China relations will continue to deteriorate over the next decade. The rift has been caused by China’s emergence as a geopolitical competitor to the US, rather than Donald Trump’s particular brand of populist protectionism. The first phase of the conflict focused on re-imposing tariffs on goods trade. But we expect the next phase to broaden out to include limits on market access and flows of technology (and possibly finance.) (See here and here.)

At the very least, the long trend of globalisation which, mainly via the dismantling of trade barriers, put downward pressure on goods prices for more than 20 years, is over. As a result of all those developments, the US inflation rate for core goods prices trended progressively lower during the 1990s and, since the early 2000s, goods prices have been falling outright for most of the time. (See Chart 18.)

Chart 18: CPI Core Goods Prices (%y/y)

Source: Refinitiv

The prospects for additional liberalisation of trade are close to zero – not just because of geopolitical shifts, but more simply because most barriers to free and unfettered global trade in goods had already been removed. That should remove one of the key sources of deflationary pressure over the last couple of decades, although we don’t expect it to develop into a renewed source of inflationary pressure.

4. Fed less focused on controlling inflation now

Back in 2017 we argued that “for a resurgence in inflation to take hold as it did in the 1970s, policymakers would need to be complicit again. It is not inconceivable that, with the high inflation experience of the 1970s and 1980s now 40 years ago, the lessons of that period could be forgotten, particularly after a decade when uncomfortably low inflation has become the biggest concern. Major central banks have already shown a willingness to engage in extraordinary policy action, including large-scale asset purchases.”

This is the key factor where developments since the pandemic began have significantly increased the risk of higher inflation in the future. The Fed embraced unconventional policy tools more fully during the pandemic – buying risk-free government securities at an unprecedented pace during the early stages and, for the first time, it bought corporate bonds and made direct loans to businesses and state governments. The result was a massive surge in money growth. (See Chart 19.)

Chart 19: Monetary Aggregates (%y/y)

Source: Refinitiv

To be clear, we think the Fed did the right thing. At a time when aggregate demand is depressed and underlying inflation is low, there are few downside risks of running a very accommodative monetary policy. Furthermore, since the economic damage was caused principally by the draconian lockdowns ordered by governments, it was only right that central banks and the fiscal authorities did everything within their powers to support affected households and businesses. If policymakers hadn’t been so aggressive, the contraction in real activity and the surge in unemployment would have been much worse.

The surge in the monetary aggregates will also not necessarily lead to higher inflation, particularly not at a time when the economy is still operating well below capacity. Finally, in a world where the Fed has the power to pay interest on excess reserves, the size of those reserves and the broader monetary base are largely irrelevant. The Fed can still control aggregate demand and money demand simply by raising its policy rate.

Nevertheless, more activist monetary policy could eventually lead to a bout of higher inflation over the longer-term, not because central bankers don’t have the tools to rein in demand, but because they will be persuaded not to use them by pressure from politicians and society more broadly.

The modern-day Fed is at risk of repeating the “anguish of central banking”, as originally described by ex-Fed Chair Arthur Burns in his infamous speech in the late 1970s. Burns’ argument was that he and other central bankers had the tools to control inflation in the 1960s and 1970s, but chose not to do so because “the Fed was itself caught up in the philosophic and political currents that were transforming American life and culture”. As Burns learned to his cost in the 1970s, the less focus the modern-day Fed puts on controlling inflation, the bigger the risk is that inflation will eventually get out of control.

The institutional slide has already begun, with the Fed recently adopting a flexible average inflation target, which explicitly allows for a period of above 2% inflation, to recover some of the ground lost during the past decade of persistent undershooting of the target. Admittedly, the Fed only intends to aim for inflation “moderately” above target for “some time” and it could begin to normalise interest rates when inflation is merely “on track” to climb above 2%. That suggests Fed officials envisage allowing inflation to run perhaps at no more than 2.5% for a year or two and could begin to rein in some of the policy accommodation even before that.

But the Fed also made other changes following its strategy review. From now on it will interpret its maximum employment mandate as “a broad-based and inclusive goal” and will focus on “shortfalls” rather than “deviations” in employment from its maximum level. The latter means that, as long as inflation is contained, the Fed won’t begin to raise interest rates until officials believe the economy is back to full employment.

These changes in isolation do not represent a full-blown return to the negligent monetary policy of the 1970s, but they are the first step down that path. A return to high inflation is not yet inevitable, but the strategy changes at the Fed, partly triggered by the pandemic, have undoubtedly raised the odds of it happening over the next decade.

5. High public debt burdens

Historically, periods of high inflation in the US usually coincided with periods of high government debt, or at least periods when the debt burden had been rapidly rising. The surge in debt caused by the Civil War in the mid-19th Century and the two World Wars in the 20th Century all triggered rapid rises in inflation, which subsequently reduced the real value of that debt to more normal levels. (See Chart 20.)

Chart 20: CPI Inflation (%)

Sources: Refinitiv, NBER

As Chart 21 shows, the surge in inflation during the 1970s wasn’t pre-dated by a significant rise in the public debt burden, in part because inflation took off so quickly that the debt burden never had a chance to increase.

Chart 21: Federal Net Debt (As % of GDP)

Sources: Refinitiv, NBER

With the fiscal stimulus required to tackle the pandemic likely to push the net debt burden well above 100% of GDP, for only the second time in history, there must be a risk of history repeating itself. As with the potential institutional slide at the Fed, however, nothing is inevitable. Everything depends on the choices that policymakers make in the next decade.

There is not a direct causation between higher public debt and high inflation. Coming out of the Second World War, rather than the debt it was the removal of war-time price and wage controls that triggered the initial rise in inflation. But, faced with the prospect of an unsustainable debt burden, governments have a strong incentive to put pressure on independent central banks – or even take direct control – to ensure that interest rates are kept below the rate of inflation, so the real debt burden can be reduced. This financial repression eventually leads to inflation, as interest rates are kept too low to rein in private sector demand.

We have already seen the Fed buying vast amounts of government debt in the secondary market and working together with the Treasury to set up its 13(3) lending facilities. And advocates of so-called Modern Monetary Theory (MMT) are pushing for the Fed to go further and finance large fiscal deficits directly using newly printed currency. With the aging of the population likely to put even more pressure on the public purse over the next couple of decades, at a time when the debt burden is already near unprecedented levels, it is not hard to imagine a future where the Fed and its price stability mandate become subservient to the Treasury and its need to finance higher welfare spending.

Can inflation rise in a flat Phillips Curve world?

But can the US experience a bout of rising inflation in a world where the Phillips Curve appears to be essentially flat and where uncomfortably low inflation is likely to persist in other advanced economies, such as the euro-zone and Japan?

We think it can. First, the breakdown in the Phillips curve relationship over the past couple of decades was principally due to the surge in the size of the global labour force and globalization. Because of those deflationary effects on goods prices, inflation never rose even when the unemployment rate was low. But stripping out goods prices, core services inflation still has a significant (inverse) relationship with the output gap. (See Chart 22.)

Chart 22: Core CPI Inflation & Output Gap

Source: Refinitiv

The domestic Phillips Curve relationship still exists, it is just that it has been masked by stronger global disinflationary forces over the past couple of decades. But the crucial point is that, as we’ve discussed, we expect those disinflationary forces to dissipate over the next few years. If, rather than just seeing an end to further globalization, there is actual deglobalization, with trade rules being tightened, then that could even put upward pressure on prices.

We also see no reason why inflationary trends can’t diverge across countries, not least because Japan has been the only country caught in deflation for the past decade and plenty of developing economies have experienced bouts of much higher inflation than the average, largely because of policy errors.

It is true that most countries experienced the same deflation during the Great Depression of the 1930s and the same high inflation during the 1970s and early 1980s. But the former owed a lot to the gold standard, with countries reflating once they devalued their currencies, and the latter in part due to the spike in global oil prices.

Nevertheless, some of the factors we have outlined above – such as slower labour force growth and the potential for globalisation to go into reverse – would be common shocks that hit all countries. But the biggest risk of higher inflation that we see is a further institutional shift at the Fed, with the central bank effectively abandoning inflation targeting to pursue social policy goals, or at least support efforts by Congress to improve social conditions, by leaving monetary policy on a semi-permanent accommodative setting. Some other major central banks might follow the Fed’s lead, but the ECB, which is built on the foundation of the stridently anti-inflationary Bundesbank, is a lot less likely to abandon its principles and the northern European countries have always been more committed to disciplined fiscal policy too. That is a key reason why we think the inflation outlooks for the US and the euro-zone can diverge.

Deflation risks have not disappeared entirely

In our 2017 Focus we also identified three factors that could act to lower inflation, possibly even to the extent that deflation became a risk:-

  • The lower neutral interest rate means that the zero nominal bound will be a recurring problem, making it difficult for the Fed to pull the economy out of future downturns.
  • Japan’s experience demonstrates that, if inflation expectations fall too far, deflation is possible even with fiat currencies.
  • Technological advances, with hi-tech capital replacing labour, could drive prices lower.

It is also worth looking at those factors again, to see how the pandemic may have altered the long-term inflation risks.

1. Low neutral rate will remain a problem

Although several factors unique to the pandemic mean that the recession the US economy is just emerging from will not be the one that tips it into deflation, the structural decline in the equilibrium (aka natural) interest rate means that the Fed could still struggle to provide enough monetary stimulus to tackle future recessions. The shift to a flexible average inflation target was, after all, designed to partly alleviate the problem of the Fed repeatedly hitting the near-zero bound for nominal interest rates.

In the near-term, the Fed’s policy impotence could prevent a full and timely economic recovery, particularly if Congress fails to provide additional fiscal stimulus. But the Fed’s inability to drive the real interest rate far enough below zero, to provide enough support to the real economy, is principally because inflation has remained so low. As we’ve noted, however, low inflation was partly due to structural factors – like global labour force growth and globalization – which the Fed had no control over. If we are right that the structural factors will drive a rebound in inflation over the next decade, then the higher inflation will, by extension, mean the Fed can achieve even more negative real rates when needed.

2. US learns from Japan’s mistakes

Despite having a public debt burden that dwarfs the US, Japan has already succumbed to a mild deflation and the euro-zone is getting dangerously close to suffering the same fate.

As we noted in our earlier Focus, however, Japanese policymakers were partly responsible for the slide into deflation when the bubbles in equities and real estate burst in the late 1980s. The Bank of Japan was slow to reduce nominal interest rates – the policy rate didn’t fall to near-zero until 2001 and the Bank then pre-emptively raised rates again in 2006. Quantitative easing was not tried until 2001 and even then the Bank of Japan bought mostly short-term bills that already had near-zero yields. Despite Japan’s high public debt burden, fiscal policy has not always remained loose either, with a couple of notable hikes in the sales tax introduced over the past decade. (See this Focus from our Japan macro service for more detail – here.)

US policymakers, both monetary and fiscal, have been much more pro-active in trying to boost demand in the real economy, in part because lessons have been learned from Japan’s experience. The same cannot necessarily be said of the euro-zone, where the ECB still has a deep-seated institutional loathing of inflation that it inherited from the Bundesbank, and fiscal policymakers are constrained by rules and limits on deficits and debts.

Finally, we aren’t convinced that population ageing has directly contributed to Japan’s deflation problem. All else equal, an ageing population should reduce aggregate supply relative to demand, boosting price pressures. Retirees drop out of the labour force, but continue to spend by running down their accumulated savings.

Even if it was a factor, Japan’s demographics are relatively unique in their severity. (See Chart 23.) Dependency ratios in both Europe and the US are set to rise over the next couple of decades, but the projected increases are far smaller than what has already occurred in Japan, let alone what is still to come. Within another decade or so, it is China that will replicate the Japanese experience.

Chart 23: Dependency Ratio (%)

Source: Refinitiv

All things considered, we think the risk that the US will follow Japan into deflation is, perhaps counter-intuitively, lower now post-pandemic than it was before the coronavirus struck.

3. Technology drives only muted deflation

Some commentators have suggested that new technologies will drive a renewed wave of “good” deflation, with productivity gains lowering costs. The migration of sales from bricks-and-mortar stores to online platforms is claimed to be another potential source of deflationary pressure – in part because it makes it easier for consumers to compare prices and find the lowest. We are sceptical of both arguments.

Chart 24: CPI IT Hardware & Services (%y/y)

Source: Refinitiv

First, the deflation rate for IT prices in the CPI is weaker now that it has ever been since the personal computer boom began in the early 1990s. (See Chart 24.) At its peak in the late 1990s, the rate of deflation for IT hardware and services CPI hit 25% per year, as rapid developments in technology drove remarkable increases in hardware quality. But manufacturers have struggled to generate the same rapid gains in capabilities in recent years, in part because physical limits mean Moore’s Law – the doubling of processor power every two years – no longer holds. Prices are barely falling at all now.

Admittedly, rapid technological improvements have not just been restricted to personal computers. More recently, the mobile wave has driven rapid improvements in the quality of cell phones, while the quality of audio and video hardware (particularly televisions) have improved at a breakneck pace over the past couple of decades. (See Chart 25.)

Chart 25: CPI Selected Goods (%y/y)

Source: Refinitiv

Where this argument falls apart, however, is that when looking at the overall basket of household spending, people spend only a tiny fraction of their monthly budgets on hardware these days. That is, in part, precisely because the prices of those technology goods have fallen so rapidly. These days consumers spend a lot more of their incomes on digital services, and the prices of those services are rising – in some cases quite rapidly.

The CPI inflation rates for both video & audio services and wireless telephone services are currently above 4%. (See Chart 26.) Wireless telephone services prices did fall very sharply in 2017, when providers introduced unlimited data plans. But that was a one-off never-to-be-repeated step change in quality.

Chart 26: CPI Selected Services (%y/y)

Source: Refinitiv

As Table 1 shows, although a weighted average of digital goods prices in the CPI are falling at an annual rate of 5%, even when combined those goods account for less than 1% of the index. In contrast, the prices of digital services are rising at a weighted average of 4% and those services have a combined weight in the overall CPI of 4.2%.

Table 1: Digital Components of CPI

CPI Weight

Inflation Rate (Oct ’20)


Video & Audio













Video & Audio



Wireless Telephone



Internet Services









Goods + Services



Source: BLS

As a result, although all digital goods and services now account for around 5% of the overall CPI, the overall impact is far from deflationary, with the inflation rate for a weighted average of all those digital components currently running at 2.3%, putting it above the headline inflation rate.

Eventually other goods may enjoy a wave of rapid technological progress – with the transition to self-driving electric vehicles probably the big development that is coming in the next decade. But even with motor vehicles, progress will probably end up being more gradual that the tech evangelists often claim. Tesla still hasn’t managed to produce a low-cost mass-selling all-electric vehicle, and neither have more-traditional manufacturers – largely because battery technology has not progressed sufficiently quickly. Mass-produced self-driving technology seems to have been two years away for the past decade.

The BEA now produces a comprehensive set of satellite accounts for the digital economy, which estimate that it accounted for 9% of the overall economy in 2018 (the latest year we have data available for). The share is growing rapidly, with real value added of the digital economy averaging 6.8% per year between 2006 and 2018, whereas the rest of the economy only grew by an average of 1.7% per year. But, mirroring the CPI weights, nearly half of the digital economy is priced digital services, with hardware accounting for only 10%. The reminder is made up of B2B e-commerce (17%), B2C e-commerce (8%), software (13%) and cloud services (4%).

Looking at how price inflation for those various components has been evolving, the rate of digital hardware price deflation has slowed over the past decade. (See Chart 27.) The prices of priced digital services are declining, but only at a very muted rate. Finally, despite the claims that the rise of Amazon and other online sellers will drive a new wave of deflation, e-commerce prices have increased in nearly every year over the past decade.

Chart 27: Digital Economy Price Indexes (%)

Source: BEA

Admittedly, the rate of price inflation for the total digital economy has been running slightly below zero in recent years, but it does not appear to be a strong source of deflationary pressure, with prices basically unchanged in 2018. (See Chart 28.)

Chart 28: Price Indexes (%y/y)

Source: BEA

We looked at whether the expansion of B2C e-commerce was responsible for the weakness in CPI inflation in more detail here. In short, there is no evidence of any “Amazon effect”, which fits with what the BEA’s digital economy accounts are telling us. The biggest issue is that the prices of goods and services now sold predominately online – such as books, newspapers, clothing and air fares – are not falling any faster than they were a decade ago and, in many cases, those prices are rising. (See Chart 29.)

Chart 29: CPI Books & Newspapers (%y/y)

Source: Refinitiv

The prices of air fares and clothing have fallen precipitously this year specifically, but that was largely due to the pandemic, which stopped nearly all air travel and devastated clothing sales while bricks-and-mortar stores were closed. (See Chart 30.)

Chart 30: CPI Clothing & Air Fares (%y/y)

Source: Refinitiv

New productivity “miracle” remains elusive

The original IT revolution in the 1990s – with the introduction of Windows PCs and the arrival of the internet – was a strong source of productivity gains not just because of the rapid improvements in the quality of the computers themselves. The adoption of computers and the internet by businesses in other sectors also generated significant efficiency gains economy wide. But that productivity “miracle” proved to be a short-lived one off and subsequent waves of technological progress – like the mobile revolution – have failed to boost productivity in the same way. Productivity growth had begun to fade even by the mid-2000s, slowed further following the financial crisis, and has remained muted for the past decade. (See Chart 31.)

Chart 31: Non-Farm Productivity (%y/y)

Source: Refinitiv

We remain optimistic that developments like AI and driverless vehicle technology will eventually power a new wave of faster productivity growth at some point in the future. But it may not happen for at least another five to ten years, if not longer.

Furthermore, even when it does happen, there is no guarantee that stronger economy-wide productivity growth will translate into downward pressure on prices. As Chart 32 shows, at the aggregate level, there is no obvious relationship between productivity growth and core inflation.

Chart 32: PCE Core Prices & Productivity

(%y/y, 3yr Ave )

Source: Refinitiv

That is because higher productivity boosts the demand for labour, pushing real wages higher. Efficiency gains can also lead to higher corporate profits rather than lower consumer prices.

As a result, we don’t expect technological progress to generate any significant deflationary pressure over the next decade. Prices will continue to fall rapidly for digital hardware, but that accounts for an insignificantly small share of overall spending in the economy. Digital services account for a much bigger share and those prices are rising.


The pandemic has increased the odds that the US will eventually experience a period of high inflation, principally because we expect the Fed to be less committed to ensuring price stability in the future. The higher public debt burden, slower global labour force growth and the possibility that globalisation will be partly reversed are additional reasons to expect inflation to rise over the longer-term. Any deflationary pressure from technology is likely to be muted, while several unique factors linked to the pandemic mean that the risk of a near-term slide into deflation is low.

This higher inflation isn’t likely to develop within the next year or two, not when the unemployment rate is still elevated, and the Fed has only taken its first baby steps toward abandoning inflation targeting. But over the next decade we anticipate a more telling shift in the Fed’s priorities, with inflation edging gradually higher to between 3% and 4%.

Paul Ashworth, Chief US Economist,