Debt ceiling solution arrives?

While acknowledging the considerable risk that this information becomes stale before the ink has dried, a resolution to the imminent U.S. debt ceiling now appears possible.

Democrat President Biden and Republican House Leader McCarthy have reached a tentative deal to suspend the debt ceiling for two years. This is in exchange for spending restraints over the next two years that would keep spending flat in fiscal 2024 before tolerating a small 1% increase in fiscal 2025. Certain programs would be exempted from this limit, including the defence budget, pension obligations and various healthcare programs.  Other smaller Republican-pleasing reforms include:

• incrementally stricter eligibility for some social services
• reduced funding for the Internal Revenue Service
• a commitment to faster environmental reviews for energy infrastructure projects.

A vote is expected on May 31. Some House Republicans have indicated they will oppose the deal (including one who called it a “turd sandwich”). They claim that McCarthy is violating promises that were made to them upon his election. But there are probably enough supportive Democrats to get the deal passed into law.

Should a deal not be achieved, the latest estimate from the Treasury Department is that the government will run out of money on June 5 rather than June 1. This buys a smidgen of extra time, ensuring that pensioners receive their monthly transfers scheduled for the first few days of June. It is possible that the actual debt ceiling could arrive even later than June 5: it depends on the exact amount of tax receipts over the next few weeks, with June 15 a particularly key date given the expectation of large corporate tax receipts pitted against massive government interest payments to be made that day. If the debt ceiling is again avoided on June 15, the government would probably have enough funds to reach August.

But, again, the most likely scenario at this point is that a deal is struck before the deadline. If a deal is not achieved, we assume that June 5 is the relevant deadline, and that the government prioritizes the payment of debt obligations so that a technical default is avoided, albeit at the cost of what amounts to a temporary government shutdown.

Banking stress today versus S&L crisis

Mid-sized American banks have struggled over the past three months, with two failures and one last-minute acquisition.

The main problem has been mark-to-market losses on the bond holdings of these financial institutions due to the abrupt increase in interest rates over the past year and a half. Mark-to-market is an accounting method that measures the value of assets based on their current price.

A secondary though arguably catalyzing force has been deposit outflows. This was originally just a trickle due to a combination of financially stressed clients (struggling tech and crypto firms that were obliged to tap their deposits to sustain their businesses) and the allure of superior returns available outside of chequing accounts. But as word spread of banking-sector troubles, deposit outflows became self-compounding and even fatal for certain mid-sized banks.

We continue to hypothesize that bank lending conditions will remain constrained for some time.

Fortunately, the overall deposit base of smaller U.S. financial institutions stabilized in May, with deposits now a hair higher on May 17 than in late April. This doesn’t guarantee that every small U.S. bank is fine – and some are still quite weak – but it is a promising sign for the viability of the sector as a whole.

Nevertheless, between sharply tighter lending standards – some of which predates this period of acute bank stress – and the newly constrained finances of the regional banks, lending is drying up (see next chart).

Sign of ebbing U.S. credit growth?

As of the week ending 05/17/23. Source: Federal Reserve, Macrobond, RBC GAM

We continue to hypothesize that bank lending conditions will remain constrained for some time. This may be acute in the near term if a recession transpires as we forecast, but a longer-lasting stinginess could persist among mid-sized U.S. banks for many years. Reasons include:

• Their deposit base has eroded.
• Their bond market losses will take several years to resolve.
• Greater political scrutiny should lead to more stringent regulations that structurally reduce the leverage (and thus lending capacity) of the sector.
• History suggests that banking sector problems usually take years to fully resolve, not just a handful of months.

The U.S. Savings and Loan crisis of the 1980s and early 1990s is not a perfect parallel and the problems today are considerably smaller, but it does provide some useful perspective.

Savings and Loans (S&Ls) were a form of U.S. financial institution largely oriented toward taking deposits and issuing residential mortgages. The sharp increase in inflation and thus interest rates in the late 1970s and early 1980s proved problematic for these institutions, much as higher interest rates are creating losses for banks today. But the problems for S&Ls were arguably much worse than those facing regional banks today, for several reasons.

1. S&Ls represented 15% of the U.S. banking market by the early 1980s and as much as half of the mortgage market, versus a much smaller 6.5% share of the banking market for regional banks (data from 2013).

2. By virtue of being funded primarily with highly liquid deposits and then lending the money out via extremely long 30-year mortgages, S&Ls suffered from a massive duration mismatch. Banks definitionally have some form of duration mismatch – a central purpose of banks is maturity transformation – but rarely that large. Maturity transformation is a practice where banks accept traditionally short-term deposits and use these to make long-term loans.

3. Interest rates rose far more in the late 1970s and early 1980s than they have over the past 18 months. S&Ls had large numbers of mortgages on their books that were – in the context of a higher interest rate environment – issued at far too low a mortgage rate, rendering them loss-making. The mismatch is not nearly as great today.

4. S&Ls were initially legally prevented from offering more than a 5.5% deposit rate. This proved woefully inadequate to retain depositors as interest rates increased. While some depositors today are being lured away from banks by higher returns available in money market funds, banks today at least possess the legal ability to pay competitive rates.

5. By the early 1980s more than two-thirds of S&Ls were technically insolvent, having exhausted all of their capital. The vast majority were also unprofitable, providing no prospect of recapitalization. In comparison, while estimates vary considerably today, the Federal Deposit Insurance Corporation (FDIC) reports that U.S. banks of all sizes collectively share around US$620 billion in mark-to-market bond losses today. This is a big number, but only 2.7% of bank assets. The most pessimistic estimate is that around half of U.S. banks would fall below the minimal capital threshold set by regulators if these losses were fully reflected on balance sheets. But there is an enormous difference between falling below a capital threshold and becoming insolvent, and as bonds mature the sum of mark-to-market losses should shrink.

6. Regulators responded the wrong way to the initial problems with S&Ls. Instead of tightening regulations to address the sector’s problems, they de-regulated. Some of this made sense, such as removing the limit on the deposit rate, but much did not. The end result was that S&Ls were suddenly unconstrained in their activity, playing with what amounted to government money since they were already insolvent (the government would be on the hook for any failure via deposit insurance). Many undertook high-risk investments in real estate projects and elsewhere in an effort to become solvent again.

The timing for this was particularly poor as the S&L industry was centered in Texas, which then suffered an oil bust and a real estate collapse in the 1980s. There was also a great deal of incompetence and outright fraud on the part of S&L executives. Many of the big bets thus failed, increasing S&L losses. Even the regulatory incentives were misaligned, as often state regulators were easing restrictions, whereas federal money paid for the deposit insurance if things went wrong.

Today, in contrast, the talk primarily centres around imposing tighter regulations on regional banks in the future. That said, the decision to provide blanket deposit insurance for the two failed banks represents a parallel to the easier regulations that ultimately plagued S&Ls in the 1980s. Similarly, one could criticize the decision to provide unlimited liquidity via the Federal Reserve for banks today as it papers over liquidity and solvency problems.

We should not anticipate a rolling multi-decade banking crisis as profound as that experienced by S&Ls.

Eventually, after more than a decade of brewing problems, proper reforms were instituted for S&Ls in 1989. A new regulator was put in charge and a Resolution Trust Corporation (RTC) was created that resolved troubled S&Ls. A total of 747 S&Ls were shuttered via this process, out of a peak count of 4,000. This ultimately cost taxpayers US$124 billion, with the work completed in 1995. The regulatory divide between S&Ls has since shrunk to the point that they are little different than ordinary banks today.

Admittedly, there are some clear parallels between then and now. In both cases, financial institutions were caught offside by rising interest rates and arguably too much risk-taking, and have suffered significant capital losses. Some banks today may be technically insolvent, just like most of the S&Ls. Regulators then and now have eased rules in a way designed to keep insufficiently capitalized financial institutions afloat, despite the moral hazard this creates.

But, again, there are critical differences:

• Today’s regional banks are a much smaller fraction of the financial sector.
• The duration mismatch is smaller.
• Interest rates have not risen as aggressively this time.
• A far smaller fraction of banks are outright insolvent today.
• The regulatory environment was not as flawed going into the period of stress.
• The policy changes made in the heat of the crisis didn’t distort incentives as badly as in the 1980s.

Thus, we should not anticipate a rolling multi-decade banking crisis as profound as that experienced by S&Ls. But the idea that these things take several years to resolve and that tighter regulations are appropriate as soon as the initial dust has settled are useful takeaways.

Murky economic data persists

The economic data remains murky. Some metrics, such as the number of global commercial flights continue to rise happily, now substantially exceeding the pre-pandemic peak (see next chart). More broadly, the demand for discretionary services is strong as people make up for lost time.

The number of global commercial flights rises happily

As of 05/25/2023. Includes commercial passenger flights, cargo flights, charter flights, and some business jet flights. Source: Flightrader24 AB, RBC GAM

Conversely, U.S. business confidence continues to collapse. It is now worse than in the most pessimistic moment of the pandemic (see next chart).

U.S. business confidence sags

As of April 2023. Shaded area represents recession. Source: The Conference Board, National Federation of Independent Business, Macrobond, RBC GAM

The blurry data even extends to contradictory perspectives on the same subject. U.S. Q2 real Gross Domestic Product (GDP) is tracking a decent 1.9% annualized gain according to the Atlanta Fed’s GDPNow index, but just +0.4% according to the Blue Chip Consensus, and a dismal -0.3% according to the St. Louis Fed’s GDP nowcast. This is quite a range of possible outcomes for a quarter that is already two-thirds complete.

Germany, meanwhile, reported an outright 0.3% (non-annualized) real GDP decline in the first quarter of 2023. The country has now strung together two consecutive quarters of contraction – a loose proxy for a recession (see next chart). Strangely, though, the central bank’s weekly activity index has been trending higher for several quarters (see subsequent chart).

Germany enters a technical recession

As of Q1 2023. Source: Statistisches Bundesamt, Macrobond, RBC GAM

Deutsche Bundesbank Weekly Activity Index continues to trend higher

As of the week ending 05/21/2023. Index estimates the trend-adjusted growth rate of economic activity by comparing the average over the past 13 weeks to the average of the preceding 13 weeks. Source: Central Bank of Germany (Deutsche Bundesbank), Macrobond, RBC GAM

The bottom line is that you can still take your pick of economic indicators if you want to build a constructive or a destructive narrative. It is a strange and highly uncertain time.

Productivity woes emerge

Another curious and largely unresolved economic matter is the recent sharp dip in productivity. Whether one looks at the U.S. or Canada, the level of productivity is outright lower than it was a few years ago (see next chart). It is highly unusual to see an outright productivity decline that lasts for several years. Did businesses abandon their latest computers and dig older models out of storage? Did the workforce spontaneously forget a year’s worth of education?

Labour productivity in U.S. and Canada now slipping

U.S. as of Q1 2023. Canada as of Q4 2022. Productivity measured as business sector/total economy real output per hour of all persons. Trendline based on growth rate from 2002 to 2019. Shaded area represents U.S. recession. Source: Macrobond, RBC GAM

The reality for productivity is considerably more nuanced. What actually happened is that productivity surged early in the pandemic to arguably unsustainable heights and has since slowly fluttered back toward to Earth. The initial surge was at least partly compositional in nature: a lot of lower skilled roles were eliminated, meaning that average productivity leapt without any one worker necessarily becoming more productive. By extension, as those lower skilled roles have returned in recent years, the average level of productivity has been dragged back down.

We continue to believe that productivity growth could be somewhat faster in the coming decades relative to the anemic rate of the past decade, reflecting our enthusiasm for a range of new technologies.

But that doesn’t explain the entire story. The level of U.S. productivity is now slightly below the trend growth rate that one would have anticipated without the pandemic. Canadian productivity is well short of that level (refer to the upward sloping dashed lines in the prior chart).

What might explain this?

One provocative possibility is that working from home diminishes productivity. The literature on this is not yet clear. Theory argues that working from home should deliver productivity gains in the form of saved commuting time and less workplace distraction. But it might alternately dampen productivity due to inferior technology at home, the temptation to work less, and less efficient communication with colleagues.

Perhaps working from home delivers an initial spurt of productivity but then a gradual drag as corporate culture erodes over time and as there are fewer interactions that spark the new ideas that are crucial to a business’ medium- and long-run success. The fact that businesses are clamoring for their workers to return to the office despite the high cost of leasing officer towers suggests that businesses believe there has indeed been some productivity loss through this channel.

Alternately, one might observe that recent hires usually aren’t as productive as longstanding employees. There is a learning curve with any new job, sometimes spanning years. We continue to believe that productivity growth could be somewhat faster in the coming decades relative to the anemic rate of the past decade, reflecting our enthusiasm for a range of new technologies.

Beneath the surface, and predating the pandemic, productivity growth has hardly been great shakes for a long period of time. Canada, in particular, has lagged badly and so the fact that it has again underperformed hints at a structural influence. Among a variety of potential explanations, the Canadian public policy of recent years has not been focused on productivity growth.

Somewhat stubbornly, and flying in the face of the recent trend, we continue to believe that productivity growth could be somewhat faster in the coming decades relative to the anemic rate of the past decade. This is for a range of issues, but most importantly reflects our enthusiasm for a range of new technologies, especially in the health and artificial intelligence spaces. Some experts believe the generative AI technologies that have recently dominated headlines could provide not just an upward nudge to productivity, but a giant boost on par with the invention of the internet itself.

Recession musings evolve

We continue to anticipate a recession for much of the developed world in the latter half of 2023. What follows are three small thoughts pertaining to that subject.

1. As previously discussed, Germany has arguably just descended into recession given the recently announced decline in the country’s output for a second-straight quarter.

2. While it is widely appreciated that standard measures of the yield curve (such as the 2-year—10-year curve and the 3-month—10-year curve) are inverted, one can argue that during a period of high inflation those metrics are flawed as recession predictors. After all, a nominal bond yield implicitly includes both a growth and an inflation premium.  An inverted yield curve is normally taken to mean that investors expect a weaker economic performance in the future relative to the present. But it could alternately just mean that inflation is expected to be lower in the future than the present.

For this reason, it is useful to examine the inflation-adjusted yield curve (see next chart). This is also inverted, and to an extent that is usually associated with a recession. So the message is ultimately similar. But it is notable that the degree of the inversion in the real curve is less extreme, and thus less definitive in its conclusion than if one solely used the nominal yield curve.

Real yield curve inverts as recession risks rise

As of 05/18/2023. Shaded area represents recession. Source: Bloomberg, Macrobond, RBC GAM

3. Of relevance for the North American economy, CN railway’s CEO recently said that their current volumes are consistent with a mild recession, and that they are budgeting for an outright decline in industrial production. Railways have an unusually broad perspective on the industrial side of the economy.

Inflation still set to decline

Inflation prints for April were broadly disappointing, revealing inflation that descended only grudgingly during the month, if at all. However, there were still signs of important progress. Critically, the breadth of high inflation is now narrowing importantly. A significantly smaller fraction of the consumer price basket is rising at 10% per year or more relative to earlier months.

Furthermore, whereas real-time inflation indicators were stagnating in late April and early May, a downward trend has now reasserted itself as of mid to late May (see next chart). Another real-time inflation measure from Trueflation is also falling once again. Don’t give up on the declining inflation story yet.

U.S. Daily PriceStats Inflation Index falls again

PriceStats Inflation Index as of 05/23/2023. CPI as of April 2023. Source: State Street Global Markets Research, RBC GAM

The Bank of Canada recently published a lovely chart that argues food inflation should decelerate markedly in the coming quarters based on what they can see happening to the cost structure up the food chain (see next chart).

Easing pressures on input costs suggest food price inflation could decline further

As of February 2023. Food costs are estimated as a weighted average of the costs incurred as part of the food supply chain, including imports, production, transportation & labour. Source: Statistics Canada, Haver Analytics, Bank of Canada, RBC GAM

Attributing higher inflation

On a very different inflation note, a clever slicing of economic data allows for the attribution of the inflation experienced in the U.S. since the onset of the pandemic. The three main components of cost include labor costs, corporate profits, and “other” (see rightmost bar on the next chart).

Cumulative contribution to the change in U.S. unit prices

As of Q4 2022. Unit price is calculated for non-financial corporations. Source: U.S. Bureau of Economics Analysis (BEA), Macrobond, RBC GAM

The conclusion is that rising labour costs explain 51% of the cumulative increase in prices, higher corporate profits explain 34%, with assorted other factors explaining the rest. This is initially somewhat surprising, for a few reasons.

For one, at first glance the analysis looks like it is neglecting exogenous forces like higher commodity prices. But keep in mind that while higher commodity prices are a cost to one company, they are a source of revenue to another, neutralizing one another. And if one traverses all the way to the point at which a company is physically extracting natural resources out of the ground, the cost of extraction doesn’t change when commodity prices go up. Thus, higher commodity prices should boost the profitability of the first company and have something like a neutral effect the rest of the way down the supply chain. In other words, higher commodity prices should initially show up as larger corporate profits.

Second, it is unintuitive that more than half of the increase in prices is pinned on labour costs. Aren’t real wages actually down over the past few years? If workers are poorer than before, how can they be responsible for the inflation? Furthermore, wages didn’t lead inflation higher – they have just followed it upwards. How can wages then be to blame?

Half of the answer to this mystery is that inflation-adjusted, productivity-adjusted wages in the U.S. are actually higher than they were before the pandemic (see next chart). True, they fell across much of 2020, 2021 and 2022, but only after spiking upwards in early 2020.  Surprisingly, workers appear to have been fully compensated for the extra inflation.

Inflation-adjusted U.S. wages are still in line with pre-pandemic trend

As of March 2023. Source: U.S. Bureau of Labor Statistics, Macrobond, RBC GAM

The other half of the answer is that while wages didn’t create the original spurt of inflation, wages have nevertheless risen more quickly than normal in response to it, adding to the cost structure of products sold to customers. The attribution effort isn’t about laying blame or determining causality. Still, it illustrates that we must keep a close eye not just on profit margins but also on wages in gauging where inflation goes from here.

U.S. infrastructure boom?

The U.S. Inflation Reduction Act, signed into law in August 2022, introduced a wide range of subsidies intended to encourage various forms of green-oriented infrastructure investment.

In the subsequent nine months, quite a garden industry has formed trying to estimate just how much of an infrastructure boom may result. The figures vary radically. The Congressional Budget Office estimates that the Inflation Reduction Act will actually suck a net US$306 billion out of the economy due to some tax hikes that more than outweigh the extra green investment. In exceedingly sharp contrast, some private-sector analysts have mused about as much as a $3 trillion boost to infrastructure.

The potential effect of the Inflation Reduction Act on the overall U.S. economy is significant, but not as radically game-changing as the $3 trillion figure initially suggests. The impact on U.S. carbon emissions, on the other hand, could be quite significant. Not everything is about GDP.

What can possibly explain a range of estimates so large that they can’t even agree on the sign in front of the number? There are two main answers.

1. The Congressional Budget Office (CBO) concedes that, beneath the surface, there will be approximately $369 billion of additional government money for green infrastructure coming from the Inflation Reduction Act. So there isn’t actually a disagreement over the sign if one focuses purely on the infrastructure aspect.

2. There is a lot of uncertainty over how big the infrastructure stimulus will actually be. The government hasn’t committed a fixed pot of money. Instead, it has promised to subsidize certain types of infrastructure. If no one takes the government up on the offer, the stimulus is $0. Conversely, some analysts think the uptake could be up to three times larger than the standard government estimate. That triples the effective stimulus.

Furthermore, the private sector will have to spend money to qualify for these credits. It is reasonable to think that every dollar of government spending will require about $1.50 of private spending. So $369 billion of government spending as per the CBO becomes $922 billion. And if the uptake is actually three times larger than the CBO imagines, the tally becomes $2.767 trillion.

But that’s cheating, because some of the private sector money would have been spent anyway, either on green projects or more polluting infrastructure. If you figure that half of the private-sector portion would have been spent anyway and that the economic multiplier of this extra infrastructure spending (public plus private combined) is around 1.0, the true figure is $1.94 trillion. The 1.0 multiplier is low for infrastructure stimulus, but keep in mind that some of the infrastructure being built doesn’t enhance productivity as much as normal since it is replacing perfectly functional if polluting power plants and vehicles.

However, we must do three more things:

1. Let us not forget about the economic drag from the tax hikes. Adjusting for that, we are left with a $1.26 trillion boost.

2. The money is expected to be deployed over a decade. That’s an additional $126 billion for the economy per year.

3. Let us not forget that this $126 billion represents the high end of the estimates. If you redo all the math with the CBO’s original infrastructure number, you end up with a small net drag of $3 billion per year.

Thus, the Inflation Reduction Act is set to add between 0.0% and 0.5% to the level of U.S. economic output each year for the next 10 years. Mathematically, that means up to 0.5 percentage point of extra growth in the first year, followed by nine years of unaltered growth, and then a tenth year in which the economy grows by 0.5 percentage point less than normal as the stimulus vanishes.

In other words, the potential effect of the Inflation Reduction Ac on the overall U.S. economy is significant, but not as radically game-changing as the $3 trillion figure initially suggests. The impact on U.S. carbon emissions, on the other hand, could be quite significant. Not everything is about GDP.

Finally, one should not neglect the fact that the U.S. infrastructure program is now motivating a host of other countries to introduce their own green subsidy programs. In turn, the net global economic boost in dollar terms may be considerably larger than the U.S.-only boost.