Macro Minute: SVB takeover and Credit Suisse Acquisition- Signs of a looming Minsky Moment?

In recent weeks, concerns over bank solvency have resurfaced in the markets, with Silicon Valley Bank (SVB) being taken over by regulators and UBS acquiring Credit Suisse for $3.2 billion. The takeover of SVB marks the second-largest bank failure in U.S. history and the most significant since the 2008 financial crisis.

Why it matters: While some attribute the recent events to executive mismanagement, we contend that they are symptomatic of a broader systemic issue. Rather than isolated incidents, they may represent the opening act of a potential future collateral crisis. What’s particularly worrisome is that if the collateral in question is so-called “risk-free” government bonds, we could be entering uncharted territory, leaving us with little recourse to address the issue. The short-term impact of recent events could be a significant curbing of credit to the economy, while the long-term impacts could pose an even more significant challenge to the financial system than the Great Financial Crisis.

The short-term effects of the most recent banking stress will be to reduce credit in the economy.

  • Regional banks are likely to curb credit for three primary reasons:
    • First, increasing regulation and the risk of executive prosecutions for mismanagement will create a risk-averse environment, leading these banks to reduce lending.
    • Second, the newly formed mechanism that gives these banks direct access to the Fed will increase the relative attractiveness of securities that can be posted as collateral, thereby decreasing the attractiveness of loans.
    • Lastly, the drop in deposits in these banks will create a hole that must be filled through wholesale funding provided by money market funds. While access to funding won’t be a problem, the rates charged for this funding will be higher than those on deposits, increasing costs for these banks, thereby pressuring their margins, reducing their profits, and potentially disincentivizing them from making new loans.
  • The recent events will also impact large banks, causing them to take a risk-averse stance for two main reasons.
    • First, the FDIC is not backed by taxpayer money but by the banks, particularly the well-run large banks. Apart from the direct cost of insuring all deposits of SVB, the possibility of new regulations requiring the FDIC to insure a larger portion of deposits will drive the profitability of these banks lower, reducing their willingness to lend.
    • Second, with regional banks now having direct access to the Fed, there is a new price-insensitive buyer of Treasuries that could pressure long-term rates lower while the Fed keeps the short-end of the curve higher through its hiking cycle, making curves flatter and reducing bank profits and their willingness to lend. These factors could lead to tightening credit and financial conditions, which could have significant implications for economic growth and stability.

The long-term effects of the debasement of US Treasuries could be catastrophic to the global financial system.

  • While it’s true that critics have pointed to mismanagement as a contributing factor in the sudden collapse of SVB, this only scratches the surface of a more significant issue at play. Banks were led to invest heavily in risk-free (US Treasuries) and quasi-risk-free (government-guaranteed) assets due to financial repression and regulation. “Hold to maturity” accounts further fuelled this trend by increasing the allure of such assets, leading to a dangerous level of complacency among banks. Therefore, while SVB may have been the first to fall, the underlying problem is much more pervasive and requires a more comprehensive solution. The possibility cannot be dismissed that we may be witnessing the initial signs of what could be the most significant Minsky Moment in history.

Named after economist Hyman Minsky, a “Minsky Moment” refers to a sudden collapse of asset prices following a long period of growth. Minsky believed that during periods of economic stability and growth, investors and lenders become increasingly complacent and take on more and more risk. This leads to a build-up of financial fragility, with borrowers becoming increasingly over-leveraged and lenders becoming more lax in their lending standards.

Eventually, the economy reaches a tipping point where borrowers can no longer meet their obligations, and the value of assets used as collateral declines rapidly. This triggers a panic among lenders and investors, who try to liquidate their assets, causing a further reduction in asset prices and triggering a financial crisis.

Minsky moments have been observed in many historical financial crises, such as the 2008 global financial crisis, the dot-com bubble burst in the early 2000s, and the savings and loan crisis in the 1980s.

  • It’s worth noting that the estimated losses on securities only represent a fraction of the total unrealized losses that banks have experienced due to the rise in interest rates. Loans, much like securities, also experience a decrease in value when interest rates increase. With a total of $17.5 trillion in loans and securities as of December 2022 and an average duration of 3.9 years, the total unrealized losses on bank credit amounted to approximately $1.7 trillion (calculated as $17.5 trillion x 3.9 x 2.5%). This figure is only slightly less than the total bank equity capital of $2.1 trillion in 2022, indicating that the losses resulting from the interest rate increase are on par with the entire equity of the banking system.[1]
  • While it’s crucial to consider losses incurred on assets, even if they are unrealized, the most significant issue is the impact those losses have on a bank’s ability to refinance its outstanding debt. Financial institutions rely heavily on short-term, wholesale dollar funding through collateral considered “safe.” However, a lack of such “safe” collateral can lead to a catastrophic failure of the financial system. Therefore, while the losses on assets are important, the loss of confidence in the ability to refinance their outstanding debt poses the most significant risk to the banking system.

A prime example of the risks associated with a lack of “safe” collateral can be seen in the lead-up to the 2008 Global Financial Crisis. JP Morgan provided cash to Lehman Brothers to conduct its daily business. However, as Lehman’s collapse loomed, JP Morgan began to question the value of the collateral that Lehman had pledged, suspecting that it was worth less than initially claimed. As a result, JP Morgan required Lehman to pledge more collateral as a condition for continuing its operations. This scramble for the most “pristine” collateral highlights the dangers of a shortage of “safe” collateral, which can ultimately lead to a decrease in USD funding worldwide as chains of wholesale dollar transactions begin to unravel.[2]

  • What could cause the debt of a sovereign country with a free-floating exchange rate and holds reserve currency status to become risky? Inflation. The Fiscal Theory of the Price Level suggests that higher inflation may be forthcoming, which could lead to the “risk-free” status of government bonds being called into question. If this were to occur, it could potentially leave us in uncharted territory with few tools at our disposal to address the issue. Therefore, inflation poses a significant risk to the perceived safety of government bonds and could have far-reaching consequences for the broader financial system.

The majority of U.S. government debt is issued with a nominal face value in U.S. dollars, meaning its real value or purchasing power is determined by dividing the value of outstanding nominal debt by the price level. The intertemporal government budget constraint stipulates that the real value of debt is linked to the real value of future surpluses. When considering the intertemporal government budget constraint as an equality, changes on the right-hand side (such as increases or decreases in future fiscal variables) correspond to changes on the left-hand side (i.e., real debt). Since the nominal value of outstanding debt is predetermined, the price level is the variable that adjusts to reflect changes in fiscal variables on the right-hand side of the constraint.[3]

Our own analysis also shows[4] that we do not anticipate the favorable inflation outcomes of the past three decades to continue over the next few decades. Apart from a stronger fiscal position, the favorable inflation environment of the past three decades was partly due to the increased global productive capacity resulting from the dissolution of the Soviet Union and the inclusion of China and India in the global trading market. Additionally, financial markets had become more interconnected, resulting in the greater deployment of the world’s savings toward cross-border investment financing. These factors contributed to lower inflation expectations and reduced risk premiums.

[1] Source: Itamar Drechsler, Alexi Savov, and Philipp Schnabl, “Why do banks invest in MBS?,” New York University Stern School of Business, March 13, 2023.

[2] Modern Macro by Zachary Cameron.

[3] Lubik, Thomas A. (September 2022) “Analyzing Fiscal Policy Matters More Than Ever: The Fiscal Theory of the Price Level and Inflation” Federal Reserve Bank of Richmond Economic Brief, No. 22-39.

[4] Macro Minute: What If (Dec 2022), Macro Minute: A Tale of Two FOMCs (Nov 2022), Catalysts Into Year-End (Oct 2022), Macro Minute: We Learn From History that We Do Not Learn From History (July 2022), Norbury Partners 2022 Annual Report (June 2022), Macro Minute: Flat as a Pancake (Mar 2022), Macro Minute: The Reflexivity of Inflation & Conflict (Mar 2022), Macro Minute: Speak Harshly and Carry a Small Stick (Feb 2022), Special Report: A Changing Paradigm (Aug 2021) – for any of these reports, please contact [email protected] or visit our website.

Macro Minute: What If?

The consensus today is that the global economy, led by developed countries, is heading into recession in the next few quarters. The debate ranges between hard or soft landing. Bloomberg’s recession probability forecast stands at 65% today. To add to this bleak outlook, we have Fannie Mae and Visa, companies with real economy visibility, forecasting an 85% chance of recession. 

Some of this doom and gloom is based on the past relationship between surveys and hard data. Soft data points to the worst economic environment in half a century, only comparable to the Great Financial Crisis.

Financial markets are also forecasting an imminent recession when looking at the shape of the yield curve. The spread between 2-year and 10-year US Treasuries is the lowest since the high inflation period of the 1970s.

If we could point to only one data point to explain such dreary levels of survey responses and market pricing, it would be the speed and magnitude of the change in short-term nominal rates. The Federal Reserve hiked 425 basis points in nine months. This represents the fastest and largest rate-hiking cycle since the 1970s. The market and economists alike are saying that the current level of interest rates is incompatible with the economy’s structure. Markets believe that this level of rates will invariably cause the economy to contract, inflation to go back to 2% in the short- and long-term, and the Fed to start cutting rates in the second half of 2023.

The conclusion is valid if we accept the assumption that the trends of the 1985-2019 decades are still in effect, and that what we have seen over the past two years was just the effect of transitory impacts of Covid measures. 

Having said that, markets are already broadly pricing these assumptions with a reasonably high confidence level. As investors, we must ask ourselves, ‘what if?’ What if there is a deeper reason for the past two years’ economic dynamics? What if we are not living through (only) transitory effects? Then, looking at nominal rates to predict a recession and a turning point for inflation would be misguided. And if so, the US treasury market would have to reprice materially in 2023, causing a structural shift in the global economy and financial markets.

Inflation in the period from 1985 to 2019 averaged 2.6%. This is when we saw the third wave of globalization, increased working-age population, plentiful fossil fuel energy, and the unquestioned Pax Americana. With inflation at such low levels, one would be excused if all its conclusions were based on nominal rates assumptions. But inflation is only low sometimes. From 1950 to 1985, as well as from 2019 to today, inflation averaged 4.5%. When inflation is higher, nominal measures become less important and it is essential to look at real interest rates. Here, we use the Fed Funds Rate deflated by YoY CPI.

Real interest rates tell a very different story. We have seen a sharp increase in real rates since the beginning of 2022, but that move started from a historically low level. Today, real rates are still extremely negative, even after 425 basis points of hikes from the Fed in 2022. The conclusions we draw from looking at this measure are very different from those based on the nominal rate. We see a monetary stance that is not restrictive and, therefore, supportive of growth. With that, we also see the probability of recession at very low levels in the next few quarters, and little reason for the Fed to start cutting rates in the second half of 2023 (let alone the 125 basis points of cuts the market is pricing in between 2H23 and 2H24 – see graph below). This measure helps explain why the labor market is so strong, something that keeps confounding central bankers and analysts alike. It also helps explain why surveys are so pessimistic. In periods of inflation, people tend to have a very pessimistic view of the economy, even when real growth is positive. 

We must then ask ourselves. What if real rates are more important for the economy than nominal rates? What if the structural trends of less globalization, a decrease in the working-age population, scarce fossil fuel energy, and a multipolar world materially increase R*? What if the recent weakness in inflation numbers is just a transitory effect as part of a long-term structural inflationary period? What if growth surprises to the upside in 2023, even with the Fed keeping rates above 5%? What if?

Macro Minute: A Tale of Two FOMCs

he FOMC’s November meeting might have been one of the most important meetings in a long time. 

At 2:00 PM EST, we saw a statement that was believed to be dovish by most and confirmed by market moves. It said that the FOMC expects that “ongoing increases in the target rate will be appropriate,” which even the most dovish observers would agree, but added that “in determining the pace of future increases in the target range, the Committee will take into account the cumulative tightening of monetary policy, the lags with which monetary policy affects economic activity and inflation, and economic and financial developments.”

Thirty minutes later at the press conference, Chair Powell struck a hawkish tone, focusing on the least dovish parts of the statement and provided more hawkish commentary, leading the markets to react accordingly.

He mentioned that rates would be higher for longer: “The incoming data since our last meeting suggest the terminal rate of Fed Funds will be higher than previously expected, and we will stay the course until the job is done.” There is no pause in sight: “It’s very premature to think about a pause in our interest rate hiking cycle.” And lastly, he would rather do too much than too little: “Prudent risk management suggests the risks of doing too little are much higher than doing too much. If we were to over-tighten, we could use our tools later on to support the economy. Instead, if we did too little, we would risk inflation getting entrenched and that’s a much greater risk for our mandate.”

In sum, we saw an intentional dovish shift in the language of the statement, followed by a much more hawkish message at the press conference. We have two main takeaways.

First, we might be seeing the first signs of a fracture happening within the FOMC. That is exactly what happened in the 1970s and it was the main reason that led Volcker to shift to a monetarist approach of targeting monetary aggregates, even though he was not a monetarist. Volcker was an incredible central banker not just because of his technical expertise, but also because he was a savvy politician. He understood that he could not bring all members of the FOMC along to raise rates as much as was necessary to curb inflation. In changing the way the Fed did monetary policy, he saw a way to unburden the FOMC members from this responsibility. He understood that it would otherwise be politically impossible to keep raising rates.

Secondly, Powell changed the shape of the distribution of potential rates outcomes with his comment on “prudent risk management.” If the FOMC follows Powell’s lead, we could see rates going higher for longer, but only at much smaller increments. That will be the compromise. With eight meetings in 2023, we are talking about a potential hawkish rate increase of 200-250bps for the full year (compared with 425-450 in 2022). On the other hand, if they find themselves to have overtightened, they will have roughly 500bps or more to cut, depending on when that happens. The risk of maintaining a paying rates position at the short end of the curve, which was probably one of the best risk-adjusted trades of the year, materially increased.

In A Tale of Two Cities, Dickens opens the book with a sentence that has become famous: “It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity (…)” This meeting might have marked the end of the golden age of monetary policy where consensus was the norm and developed markets’ central banks did not have to struggle with their dual mandate or politics.

Macro Minute: The Changing Colors of Politics

On the artist’s color wheel, red and green are considered complementary colors, diametrically opposed from one another but known to harmonize when used together. However, for at least a decade, the biggest political proponent of green energy in America has been the “blue” Democratic Party.

The administration’s most recent spending bill, The Inflation Reduction Act of 2022, has been heralded as a huge leap forward for renewable energy in the United States by Democrats, but was opposed by every Republican in the House and Senate. A closer look at where renewable infrastructure is being built, thereby creating jobs and increasing investment, demonstrates that while on Capitol Hill, the reds may be diametrically opposed to green legislation, red and green may actually be quite complementary. We believe that green investment will have meaningful repercussions come election season for years to come.

In our 2021 Annual Report, we discussed how our most probable scenario for achieving net-zero by 2050 would require expansive transmission and generation infrastructure to be built in the American heartland, primarily in traditionally Republican states. In turn we suggested that the development of said infrastructure would result in significant job creation and local investment, that would lead to one of two outcomes – more bipartisan support for investment in green infrastructure as Republicans acted in the interests of their constituents or a change in voting patterns by those being positively impacted by investment and new jobs.

A 2014 study by the University of Maryland found that a $1 spent on infrastructure investment added as much as $3 to US GDP[1] and suggested that the effect could be even larger in a recession. Historically, state and local governments have borne the majority of costs for spending on infrastructure – since 1956, they have been responsible for approximately 75 percent of spending on infrastructure. In that time frame, federal infrastructure spending has increasingly become a smaller percentage of the overall budget.

When the federal government does spend, it is typically through capital investment for new projects or modernization. The nonprofit, nonpartisan Tax Foundation estimates $116 billion of new energy and climate spending, excluding tax credits, from the newly passed legislation.[1] Including leverage available through components of the bill like the Energy Infrastructure Reinvestment Financing program, which provides $5 billion to finance up to $250 billion in projects for energy infrastructure, including repurposing or replacing energy infrastructure, takes new spending to more than $300 billion over 10 years. The last Congressional Budget Office estimate for federal government infrastructure spending was approximately $98 billion per year, meaning the bill would increase spending by around 30% annually, excluding tax credits that will encourage more private investment. Why is this important? Using percent changes in GDP, inflation, and the S&P 500 as barometers for economic conditions, Lewis-Beck and Martini[2] demonstrated the existence of a map from real economic conditions, to voter perceptions, to vote choice. Put simply, voters’ evaluation of the economy is real, and they punish or reward the incumbent candidate based on these conditions.

Bloomberg recently ran an article titled ‘Red America Should Love Green Energy Spending’, showing where a bulk of renewable infrastructure is being built. There are 435 congressional districts in America. 357 have planned or operating solar plants, with 70% of the power capacity found in republican districts. 134 have planned or operating wind plants, with 87% of the capacity found in red districts. Lastly, 192 have planned or operating battery storage facilities, with 58% of the capacity in right-leaning districts. Of the top-10 districts with planned or operating renewable infrastructure, nine are currently Republican-held seats, and within that group, 86% of total capacity is found in Republican districts.

So why might Republicans who are overwhelmingly benefiting from job creation and investment in green infrastructure be against such legislation? First, some of the capacity listed is planned, and has yet to filter through into the local economies they represent. Second, there are elements of both NIMBY-ism and extreme partisanship throughout the country on both sides that lead people to immediately dismiss ideas from “opposing” parties. But most obvious to us is that Republicans also overwhelmingly represent areas with the most emissions. 80% of the top-100 emitting districts are represented in Congress by Republicans, including eight of the top-10.

n the 2020 election cycle, fossil fuel companies spent $63.6 million lobbying Republicans compared to $12.3 million for Democrats, and since 1990 the industry has spent approximately 4.3 times the amount lobbying for Republicans than Democrats. In other words, support for green investment will ultimately come at a cost for the party. However, a myriad of studies have demonstrated that infrastructure investment boosts productivity over time and the literature shows that this will ultimately have an impact on voter preferences. Voter preferences fundamentally drive political rhetoric, so as green infrastructure investment becomes more pervasive, particularly in red states, we expect an increasing impact of renewable energy development on elections. 

[1] Werling and Horst. “Catching Up: Greater Focus Needed to Achieve a More Competitive Infrastructure.”


[3] Lewis-Beck C, Martini NF. Economic perceptions and voting behavior in US presidential elections. Research & Politics. October 2020. doi:10.1177/2053168020972811

Macro Minute: GDP Deep Dive

Last week, just before the end of the month, we got the Second Quarter Advance GDP Estimate from the US Bureau of Economic Affairs (BEA). The quarter-over-quarter annualized number for real GDP printed a disappointing -0.9%, compared to a median expectation of +0.4%, but still better than the 1Q number of -1.6%. 

GDP releases are very important events for markets. Companies use them to help make investment decisions, hiring plans, and forecast sales growth. Investment managers use them to refine their trading strategies. The White House and Federal Reserve both use GDP as a barometer for the effect of their policy choices. 

These numbers are especially important for turning points in the economy. For some (but not the National Bureau of Economic Relations – the US agency responsible for classifying recessions), two consecutive quarters of negative real GDP growth is defined as a recession. If we took the early GDP releases at face value, this would imply that we are in a recession today, dating back to the first quarter. For all the above reasons, it is worth digging into how the BEA derives this number and how reliable the early releases are.

One of the tasks of the BEA is to calculate US GDP, measured as the total price tag in dollars of all goods and services made in the country for a given period. It is the sum value of all cars, new homes, lawnmowers, electric transformers, golf clubs, soybeans, barbeque grills, medical fees, computers, haircuts, hot dogs, and anything else sold in the US or exported during the period. When calculating current (or nominal-dollar) GDP, the agency adds the value of all goods and services in current dollars. But this herculean task does not end there, because what matters for most people is the real growth in the economy. And so, after tallying up everything in current dollars, the agency has to then make adjustments to try and come up with an estimate of the value of what was actually produced in the economy (e.g., ex-inflation). 

Imagine an economy that only produces two things, potato chips and mobile phones. Suppose that the economy is selling $1.1 million of goods this year, an improvement of 10% compared to the $1 million from last year. That $1.1 million number represents the nominal GDP for the economy this year. But that number does not tell us how much of that 10% increase is due to more goods being sold and how much derives from price increases. 

If last year there were 50,000 bags of chips sold for $10 and 500 mobile phones for $1,000, and this year there were 55,000 bags of chips and 550 mobile phones sold for the same price as last year, the economy had real growth of 10% and zero percent inflation. 

Alternatively, if this year the economy sold the same number of chips and mobile phones as last year but did so at a price of $11 and $1,100, respectively, the economy had zero real growth and 10% inflation. 

However, things are not so simple, for the methodology is designed not only to remove price inflation but also to adjust for the quality of the goods being sold. Let’s assume that this year the economy sold 55,000 bags of chips for $10, and 550 phones for $1,000 (the same as the first example). But in this example, the bags of chips sold this year only contain 40 chips versus the 50 chips in each sold last year, and the mobile phones sold this year have better computational power and an extra camera versus last year’s. In this case, the agency would have to account for those changes by calculating a positive price increase for the potato chips and a negative one for the mobile phones, even though the number consumers saw on the price tag did not change. Now imagine that the BEA must do this not just for all the goods sold in the US economy, but also for every service provided, and to deliver an advance estimate one month after the end of a quarter. 

Which brings us to the question, how reliable are early GDP estimates? The answer is… it depends. Each revision incorporates more and better data and is believed to be a better estimate of the true value of GDP. For example, comprehensive data accounts for only 25.5% of advance estimates and 36.8% of second estimates, but it accounts for 96.7% of what we can call “final” estimates[1].

To assess the reliability of the GDP estimates we can look at revision patterns to understand if there is a bias in these revisions and how large they can be. To assess bias, we calculate Mean Revision (MR) where components tend to be offsetting and a large positive or negative number would indicate bias. To understand how large revisions can be, we calculate the Mean Absolute Revision (MAR) and the standard deviations, which are both complementary measures of the distribution for the revisions around their mean. We calculate these revision metrics for the Advance release that comes out one month after the end of a quarter, comparing with both, the Second releases (two months after the end of a quarter) and what we here call the “final” estimates (also called, comprehensive revisions, which are released approximately five years after the advance release).

What we find is that inflation has a meaningful impact on reliability. More specifically, it creates a pronounced bias for advance releases in underestimating real GDP growth. This makes intuitive sense. The task of calculating real GDP becomes even more challenging during inflationary environments. Looking at the numbers, we find that in periods of low inflation [3,4], bias is virtually inexistent with MRs for Second and Final at +0.10% and -0.01%, respectively. While during periods when US CPI is above 7%, MRs are +0.40% and +0.80%, respectively. That means that, on average, in high-inflation environments, Advance GDP numbers are underestimated materially. It is also important to note that MARs and standard deviations are essentially unchanged from one environment to another. This means that the size of revisions is similar in both circumstances. 

To clarify the point, let’s look at last week’s 2Q 2022 GDP Advance release of -0.9%. We can say that the second estimate will be between -1.5% and +0.4%, while the final estimate will be between -2.6% and 2.4%, with 90 percent confidence. This distinction between inflationary and non-inflationary environments is important because if we used the low-inflation scenario numbers, we would say that the second estimate would be between -1.9% and +0.2%, while the final estimate would be between    -3.6% and +1.7%, with 90 percent confidence. [5]

One way to increase the reliability of activity numbers is to look at the average of GDP and GDI. In theory, GDP and GDI should be equal, but in practice, GDP and GDI differ because they are constructed using different sources of information – both are imperfect in different ways. If both GDP and GDI are interpreted as the sums of unobserved, true economic activity and measurement errors, it is possible to infer that the weighted average series of the two is a more reliable measure of activity than either GDP or GDI alone, assuming some of the measurement errors are averaged out.

In short, calculating GDP is a mammoth undertaking, early estimates of real GDP tend to underestimate growth in inflationary environments, and you are better off taking a holistic view of the economy when data is as volatile as it is today. 

P.S. We talked a lot about real GDP, but we should not neglect nominal GDP. Historically, S&P earnings growth tended to stay in line with nominal GDP. And that is how corporate sales, revenues, and profits are recorded. In the second quarter of 2022, nominal GDP in the US was approximately +7.9% QoQ annualized.  

P.P.S. For a depiction of how and when GDP revisions and their vintages are made and maintained by the BEA, please see below.

[1]  Comprehensive revisions are performed every five years and include major updates to classifications and definitions for the entire GDP time series – for more information, please see the endnote

[2] Holdren, Alyssa – Gross Domestic Product and Gross Domestic Income – Revisions and Source Data (June 2014)

[3] Fixler, Francisco, Kanal – The Revisions to Gross Domestic Product, Gross Domestic Income, and Their Major Components (June 2021)

[4] Using 1996-2018 period used in above paper, when US CPI inflation averaged 2.2%

[5] Revisions follow a normal distribution and therefore we can calculate the combined probability that the true value of real GDP growth in the 1Q and 2Q was below zero, i.e., two consecutive quarters of negative GDP growth. P (2Q < 0% | 1Q < 0%) = 36%.

Macro Minute: Flip or Flop

With so much talk about a recession lately, it is hard not to look for clues in housing numbers. This past week, we had numbers for US housing starts and building permits. While homes only directly account for roughly 5% of GDP, related goods and services can account for nearly 20%. Aside from 2001, the US has never gone through a recession when housing is doing well. Conversely, the US has never emerged from a recession without the help of housing (2009 being the exception with a rebound while housing was stagnant). Fort these reasons, it comes as no surprise that so much attention is given to the release of housing data.

Housing starts record how much new residential construction occurred in the preceding month, while building permits track the issuance of construction permits. The number for both releases is reported in number of units, with the latest number for housing starts and building permits disappointing the Bloomberg median survey at 1.549 million and 1.695 million, respectively. But how disappointing are these numbers, if at all?

First, let’s look at housing starts. The month-over-month number came in at -14.4%, and comparing the latest release with the same time last year, the number of starts contracted by -3.5%; however, these numbers are very volatile and prone to significant revisions. When looking at the rate of change of the 12-month moving average in May versus the previous month, we encounter only a -0.28% contraction, and when comparing the average with the same period last year, we find a growth of +9.5%. Building permits decreased by -7% MoM and increased +0.2% compared to last year. Using the same 12-month moving average to smooth volatility, the rates of change from the previous month and last year are +0.2% and +6.4%, respectively. We can see some deceleration, but we are still at very healthy levels compared to the past.

One thing to keep in mind is that looking at housing starts and building permit numbers only gives us an idea of the real component of the economy. But for prices and company earnings, it is Nominal GDP that matters. Therefore, we have constructed a nominal index for housing starts and building permits using the S&P CoreLogic Case-Shiller U.S. National Home Price NSA Index. When looking at that number, we see some deceleration, but in aggregate both starts and permits are still running at very high growth rates.

We cannot draw firm conclusions from a single piece of evidence, but what a closer look at housing starts and building permits shows is that the probability of recession may not be as high as one perceives from reading headlines.

Macro Minute: Speak Harshly and Carry a Small Stick

In the early 1900s, President Theodore Roosevelt was known for the aphorism “Speak Softly and Carry a Big Stick.” The idea behind this adage is that it is the availability of raw power, not the use of it, that makes for effective diplomacy. In the case of Roosevelt, that policy worked very well. His two terms in office had been almost completely without conflict. “He has managed, without so much as firing one American pistol, to elevate his country to the giddy heights of world power.” – Literary Digest, December 22, 1906

Possibly empowered by the policy success almost 100 years earlier, in the early 2000s, the Fed decided to embark on a policy of forward guidance. Forward guidance is the act of communicating to the public the future course of monetary policy, namely, the path of interest rates. Said guidance, which along with the control of short-term interest rates and quantitative easing, had the aim of controlling the interest rate curve without so much as “firing one American pistol.” In hindsight, it worked. Forward guidance not only kept interest rates low through the expectations channel, but also helped dampen interest rate volatility (and along with that, equity and fx volatility).

However, the deflationary forces that provided the Fed with the availability of raw power began to dissipate around 2015. Raw power allows the Fed to introduce QE and rate hikes without needing to backtrack. This was ultimately eliminated once Covid incited a degree of globally coordinated fiscal and monetary policy never seen before. Since then, the Fed has flipped Roosevelt’s policy on its head. Today, the Fed is using FOMC press conferences and governors’ speeches to speak harshly on inflation. But when the time to act comes, we believe that the Fed does not have the same firepower as before. On one hand, it cannot materially tighten financial conditions without causing an enormous problem for the refinancing of historically high levels of debt (both corporate and government). On the other hand, with inflation rampant, it cannot continue to serve as a backstop to financial markets.

Going forward we believe that the Fed will speak harshly of inflation, but will consistently be behind the curve, possibly un-anchoring inflation expectations and the long end of the interest rates curve.

Macro Minute: US Labor Participation

This week, we will once again touch briefly on labor force participation and attempt to make sense of the US Employment Situation Report from Friday.

US labor force participation has been the subject of much discussion lately. Beginning in the 1960s when more women entered the workforce, it has steadily risen, moving from 59.1% to 66.9% by the year 2000. Since then, it has drifted lower and settled near 63% pre-Covid. A drop of almost 4% on the labor participation rate is equivalent to around 10 million jobs. At first glance this seems negative, but we find that most of this was due to strong levels of enrollment in post-secondary education among those aged 16 to 24. This trend began in the late 1980s, and accelerated into the 2000s, hence a deluge of social science majors and a dearth of truck drivers.

Turning to today, let’s analyze some of the most common arguments for explaining the slow recovery of the labor force participation rate.

(1) Self-employment is keeping labor participation low – One way to try to test for that, is to track the difference between the household and the establishment employment data. The household employment figure captures the self-employed, farm workers and domestic help, something the BLS payrolls survey doesn’t do. Here what we find is that household employment suffered more than payrolls during 2020, and still hasn’t recovered to pre-covid levels.

2) Women have been kept out of the labor force because of childcare – There is some indication this may be true. We saw nearly the same number of exits from the labor force for men and women in 2020 (3.9mm & 4.2mm in April ’20, respectively). Those aged 25-34 were the second most affected at the time, accounting for more than 1mm women exiting the labor force. By September 2021, there were still 550k less women aged 25-34 in the labor force than in January 2020, the largest discrepancy across all age brackets. With schools reopening, that number was cut in almost half to 283k in November.

(3) Retirement is keeping people out of the labor force – It is hard to see that clearly in the data. The age group 55 and over (55-64 & 65 and over), suffered the least in both genders and have the least amount of people out of the labor force (when compared to January 2020 levels). Today there is 100k more men 65 and over in the labor force than at the peak in January 2020.

Macro Minute: Fiscal Cliff vs. Excess Savings

Looking ahead to 2022, much has been written about the pending fiscal cliff and its impact on Real GDP Growth. As the impact of fiscal stimulus dissipates and the federal government mulls tax increases, analysts expect fiscal impulse to shift from positive to negative next year.

Figure 1: Effect of Fiscal Policy on Real GDP Growth (3Q CMA) [Source: Goldman Sachs]

In our estimation, given the levels of excess personal savings reached in the past 20 months, we believe there is enough pent-up savings to compensate for the forthcoming negative fiscal impact on GDP. Using seasonally adjusted personal income minus personal consumption expenditures as a proxy for personal savings, we find that from April 2020 through September 2021, Americans generated over $2.8 trillion in excess savings, amounting to approximately 12% of GDP. That compares with approximately 4% of fiscal drag projected for 2022.

Macro Minute: What’s going on in the US Labor Market?

With job openings, participation rate, and unemployment central to the current discourse on markets, the topic of this month’s memo is the United States labor force.  

Focusing on the four largest sectors (which add up to more than 60% of payrolls and job openings in the US economy), we can see that wage inflation is a pattern that predates the onset of covid. In other words, wage inflation is not simply a result of covid supply shocks, it is based on fundamentals in the economy, and therefore it is not transitory. 

1 – Trade, Transportation & Utilities (19% of total Payrolls, 18% of total Job Openings)  In 2018, demand for work (job openings) started to grow much faster than supply (using payrolls as a proxy). As a result, average hourly earnings growth for this sector has surged from an average of 2.25% percent in 2018 to over 4% today (and 3.35% pre-covid). 

2 – Education & Health Services (16% of total Payrolls, 18% of total Job Openings) Hereto the story is very similar, but it started even earlier. In 2014, demand for work accelerated faster than supply of workers, driving an increase in earnings from 1.5% to 3.4% today (and 2.5% pre-covid).

3 – Professional & Business Services (14% of total Payrolls, 18% of total Job Openings) In Professional & Business Services, we saw two waves. The first in 2014 and the second in 2018, causing an increase in earnings from 1.5% to 2.3% in the first wave, and from 2.3% to 4% today (and 3.3% pre-covid).

4 – Leisure & Hospitality (10% of total Payrolls, 14% of total Job Openings) Leisure & Hospitality is the only sector in the top 4 that has gone through two opposite cycles since 2014. The first was with the demand for work growing faster than supply starting in 2014, increasing earnings growth from 1% to 4% in 2017. The second cycle took place starting in 2018, with labor supply growing faster than demand, and earnings growth falling to 3.5%. Today we are back at 4% growth, last seen entering 2018. It is worth noting that today, the demand for work in this sector is at historical highs while supply is back near the levels of 2010.

Labor Supply Shock

The last point has to do with the temporary labor supply shock that happened due to covid. Comparing jobless claims numbers between states that ended extra unemployment benefits before the September 6th deadline and those that adhered to the target, we see that the states that finished earlier have a much more accelerated and consistent contraction in claims across latter weeks. With this in mind, we expect some of this labor supply shock to normalize as we get farther from the deadline. However, when we look at the pre-covid trend, we believe that this will not be enough to avoid wage inflation.