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: We Learn From History That We Do Not Learn From History

Despite all the efforts of the most brilliant economists and analysts in the world to build models mimicking the methods of physics that follow their own self-contained logic, rules, and patterns to predict outcomes, when faced with failure, they dismiss it by claiming that “random shocks” had somehow disturbed equations and did not need to be explained since they are “nonrecurring aberrations.” War, pandemics, and politics are not abnormal historical events, only in economics.

However, many questions in economics can be approached more simply through history. In the most recent record, from 2010 to 2020, US CPI YoY averaged only 1.7%, below the Fed’s target (how much did that play a role in the recent late response from the Central Bank is anyone’s guess). However, in the history of the US, there are only a handful of times that the inflation picture could be described as stable.

Looking back at American history, we find six inflationary spiral events. The first occurred in the late 1700s just after the Revolutionary War; the second in 1813 after the War of 1812; the third in the 1860s during the Civil War; the fourth in the late 1910s after World War I; the fifth around and after World War II in the mid-1940s; and, the most current one in the 1970s associated with the Vietnam War. These periods were always followed by long periods of deflation. Evidence would point to politics, not economics, to explain inflationary spirals, and war looks like the common denominator. War in itself has many different impacts on inflation (as we discussed in this Macro Minute: The Reflexivity of Inflation and Conflict). Still, it is really the increase in money spent by the government, above what it collects in taxes, that makes inflation and negative real rates an attractive solution to the debt problem.

Looking back to the latest inflationary cycles of the 1970s, we find a few similarities and one significant difference. [1]

Similar to the present day, in 1975, the government balance sheet resembled conditions only tolerated during periods of war. And in the preceding years, just like recently, conservative governments that were supposed to be fiscally conservative were actually accelerating the deficit. In today’s world, for example, if interest rates rise above inflation, the Treasury’s interest expense goes up as debt rolls over, and the Fed reduces remittances to the Treasury. The Congressional Budget Office calculates that a 1% increase in real rates increases the annual deficit by $250 billion, about 1% of GDP, planting the seeds for an explosive debt dynamic.

In the 1970s, oil price inflation was a big problem, increasing to around 6 percent per month. More recently, on average, oil has been growing at 4.2 percent per month since January 2021. That includes the price corrections we saw in the last couple of months. The contribution to the CPI is still high at 47% YoY at current gasoline prices.

In the 1970s, real interest rates reached -4 percent. Today, we are living through the most extended period of negative real rates, currently sitting at -6 percent. That is before factoring in what can happen with nominal rates in a recessionary scare. We calculate real rates by subtracting the US Treasury 10-year yield by the current CPI YoY number. We believe this is a better indicator of real rates on Main Street than the real rates derived from the TIPS markets on Wall Street. This is the rate that alters the lives and actions of people who are not traders or advisors and who do not follow the FOMC decisions or read the Wall Street Journal. Different from the previous cycle, when the Fed was focused on impacting asset prices, to have an impact on goods and services prices, the central bank needs to focus on the decisions in the real economy and not in financial markets. 

“At 15 percent inflation, an investor lending $1 million at 10 percent ‘loses’ $50,000 a year. You cannot count on the lender being a complete idiot, sooner or later, he will stop lending at low-interest rates and invest the money himself in commodities or real estate.” – Senator William Proxmire. October 1979

Another interesting observation from looking at real interest rates is that every recession is proceeded by positive real rates. More importantly, real rates tend to turn negative to help the economy once a downturn starts. This brings us to the recessionary debate. Like today, in 1979, most economists, including the Fed, were forecasting a recession. They had been wrong for many months, and in September, data showed the economy was not tipping over; it was accelerating again. This was true even with a deceleration in housing and autos and the fear of recession. “A Gallup survey found that 62 percent of the public expected a recession sometime in 1979.”

In an inflationary economy, people behave differently. Inflation doesn’t slow people down. With inflation at 16 percent, borrowing at lower rates seemed like a good deal. Bank credit was expanding at an annual rate of 20 percent. Most consumers did not care about what the higher interest rates were, as long as the monthly payments could fit their incomes. This is not a foreign concept for Latin Americans.

“Lenders were still surprised at how many families were willing to take on home mortgages at 13 percent or even higher. ‘ Perhaps it is not so hard to understand,’ Volcker said, ‘when you realize that the prices of houses have been going up at 15 percent or more.’” – 1979

Today, bank credit is growing at +12% for consumers and +8% for Commercial and Industrial clients. We’ve been following bank’s earnings calls very closely and we find that all the major banks see strong balance sheets, very low forward-looking default rates, and expect credit to grow in the mid-teens for the next few quarters. This past week, American Express reported that overall cardholder spending rose 30% from a year earlier.

Even when the Fed was finally able to create the presumed remedy, a prolonged recession that endured for 15 months with unemployment rising to 9.1 percent and industrial production shrinking to roughly 15 percent, as soon as the economy recovered, inflation came roaring back, rising even higher than before even with employment never getting close to its natural rate. With the supply of commodities constrained, even a short-term decrease in demand does not fix the inflation problem; it only postpones it to the following part of the cycle when policies revert to accommodative.

Lastly, the Fed genuinely did not know how much interest rates would have to rise to break inflation. If record levels of rates were not fixing the problem, how high would rates need to go to do it? Nor did it have the political capital to do what was necessary. Volcker acknowledges, ‘We could have just tightened, but I probably would have had trouble getting policy as much tighter as it needed to be. I could have lived with a more orthodox tightening, but I saw some value in just changing the parameters of the way we did things. (…) it would serve as a veil that cloaked the tough decisions.’”

“There is a wide concern about the Fed’s resolve in adhering to this policy in the face of an election year and the increasing likelihood of a recession. If strong words and actions are not followed by results, then holders of dollar-denominated financial assets in the US and abroad will conclude that the recent changes are no more significant than the statements and policy changes of prior years which did not reduce inflation. When rhetoric sufficed several years ago, tangible proof is now required of the Fed’s intentions.” – Federal Advisory Council 1979

The similarities are striking.

The main difference between the 1970s and today lies in the credibility that central banks around the world collected during a period of global deflationary forces that made them look like they could bend prices to their will and achieve their dual objective effortlessly, giving rise to the mantra “Don’t fight the Fed!” On July 14th, 2022, Governor Waller said, “The response of financial markets to the FOMC’s policy actions and communications indicate to me that the Committee retains the credibility and the public confidence that is needed to make monetary policy effective. (….) lenders and borrowers are still doing business at these rates, which indicates that they believe the FOMC’s policy intentions are credible, as broadly reflected in the interest rate paths in the Summary of Economic Projections (SEP).” Today, markets price the Fed’s projections to perfection.

What does history tell us about that? StoneX’s Vincent Deluard shows us that using post-war data from the World Bank of more than 350 events when inflation spiked above 7%, only 1.4% of the time, inflation slows to less than 3% in each of the next five years. Markets are pricing 1 in 70 odds as if it were 100 percent certain.

“Acting hastily is essential to [a trader’s] profitability. If today’s quickest-to-the-keyboard move makes little sense according to some notion of ‘fundamentals,’ who cares? Overshooting is a feature, not a bug.” – Alan S. Blinder, July 2022.

“Traders must and do therefore respond literally instantly to all news to which they think other traders might respond. Whether the news is considered economically significant or even true is immaterial.” – Albert Wojnilower, Chief Economist at First Boston 1964-1986

This confidence also has an impact on the USD. With the expectation that inflation will converge to 2% in the next 18 months, interest rate differentials make the currency attractive. That, in turn, keeps inflation in the US in check. The DXY Dollar index is more than 17 percent up YoY, while the US CPI is 9.1 percent. Being conservative, we can assume a short-run currency passthrough in the US at about 25 percent.[2] This means that if the US Dollar was flat year-over-year, inflation should be a whopping +13%! This blind faith in central banks is what is keeping everything together. But history also tells us that after a long deflationary cycle and the build-up in credibility, what comes next is the drawing down of goodwill until there is nothing left.

“We’ve lost that euphoria that we had fifteen years ago, that we knew all the answers to managing the economy.” – Volcker 1989

[1] A good friend of the firm and fellow investor, knowing of our quest to understand history, pointed out to us that the team at MacroStrategy research was studying a book written in 1989 by William Greider called “The Secrets of the Temple” about the Fed’s fight against inflation under Volcker to help them with a similar pursuit. This book has been invaluable in our understanding of the period, and all quotes in this letter are from the book. https://www.amazon.com/Secrets-Temple-Federal-Reserve-Country/dp/0671675567/ 

[2] Campa, Jose Manuel, and Linda S. Goldberg. “Exchange rate pass-through into import prices.” Review of Economics and Statistics 87.4 (2005): 679-690. (https://www.nber.org/system/files/working_papers/w8934/w8934.pdf)

Exchange Rate Pass-Through and Monetary Policy, Governor Frederic S. Mishkin, at the Norges Bank Conference on Monetary Policy, Oslo, Norway. March 07, 2008 (https://www.federalreserve.gov/newsevents/speech/mishkin20080307a.htm)

Takhtamanova, Yelena F. “Understanding changes in exchange rate pass-through.” Journal of Macroeconomics 32.4 (2010): 1118-1130. (https://www.frbsf.org/economic-research/wp-content/uploads/sites/4/wp08-13bk.pdf) 

Macro Minute: The Ides of June

When looking at the return of assets for the first half of the year, we find that US bonds posted their worst first half-year performance for over 100 years, while the S&P 500 declined 20.6% year-to-date, recording the worst first half of the year since 1970 and its 4th worst start on record. More broadly, the MSCI All Country World was down 20.9% for the period. Institutional investors are having one of their worst performance periods on record with the trusted 60/40 portfolio declining by 17% YTD, making it the second-worst start since the 1900s.

The month of June was marked by a sharp repricing of recession fears along with a VaR shock that led to risk reduction and high correlation across markets, providing very few opportunities for hedges and diversification. During the month, 10-year treasuries increased +16bps, with the difference between the 10 and 30-year bonds flattening by 3bps. The 2-year bond yield increased by almost +40bps for the month and jumped +54bps in two trading sessions, the largest move since 2008 when it moved +55bps. The S&P and Nasdaq were down -8.4% and -8.7%, while Energy and Materials sold off -18% and -15%, respectively. Commodities were down across the board, ranging from -10% to -40% in agriculture commodities, and -22% to -57% in industrial metals. In other words, you could not make money in June by being long.

Looking at the long-term we believe that commodity and commodity-related equities exposed to the green energy transition have an exceptional demand backdrop that arises from decarbonization initiatives that will only increase going forward while also possessing major supply challenges. As an example, the average EV consumes five to six times more copper than a combustion engine vehicle. Conservative estimates of EV production put copper demand, just from this source, increasing 20-25% over the next two decades. This does not even account for the increasing demand for copper arising from other electrification needs like batteries and cables.

This is happening against the backdrop of virtually no production increases and very low inventory levels. Mining companies learned from their mistakes in the previous CAPEX cycle of the early 2000s, and along with the more recent price declines and volatility, board members will not be in a rush to invest in capacity. Rather, they will prefer dividends and share buybacks. 

On a March 1st podcast interview with Eric Mandelblatt, he says “(…) three of the largest copper mines in the world were developed over 100 years ago. There’s been only one of the 10 largest copper mines in the world that’s been developed this century since (2001). So, you have this situation where supply in the near term is highly inelastic.”

Reflecting on the recent commodities drawdown, we put too much weight on the probability that markets would realize early that the Fed won’t be able to run the level of positive real yields required to bring down inflation to its target. Rather, when looking at prices, it appears markets are pricing the Fed outlook to perfection. We are now accounting for that and expecting that the crucible moment will occur after the Fed reaches their expectation of terminal rate, or just above, and inflation is still above target. At that moment, the Fed will either have to prove credible, or the market should then realize that the Fed will let inflation run above target for longer. It is worth pointing out that history is not on the Fed’s side. Vicent Deluard from StoneX shows that, historically, central banks only manage to bring inflation down to 3% in each of the next 5 years, following a spike above 7%, in less than 1.4% of the time. What we see today is a market that blindly believes in the Fed and prices that 1.4% probability scenario with full certainty, while completely dismissing the other scenarios. 

Also, we did not expect the market to aggressively price in a deflationary bust scenario so rapidly after a higher-than-expected inflation print and still extremely negative real rates. We expected that during this secular bull market in commodities, we would see some ups and downs in prices, but the speed and magnitude of these moves only compare to 2008, which was a massively deflationary bust period. We assign a very small probability of that scenario (for a detailed analysis on this, please see our recently published annual report), and we believe that if a recession is around the corner, it would be an inflationary bust instead. 

With the supply of commodities constrained, even a short-term decrease in demand would not fix the problem of inflation, it would only postpone it to the following part of the cycle when policies revert to accommodative to shore up demand. We know from history that during inflationary busts, commodities have two-thirds of their upward move after a recession begins. 

2022 Annual Report

We believe that 2021 marked the beginning of a secular bull market in commodities and commodity-related equities, with the usual peaks and troughs along the way. A decade of underinvestment by producers and refiners of natural resources coupled with burgeoning excess demand for those resources driven by a myriad of global initiatives including electrification, food security, and energy independence has shifted the long-term supply-demand outlook into deficit for many commodities.

Going back to the 1970s, a period where many investors are looking for clues given the recent run-up in inflation globally, commodity prices and related equities enjoyed a bull market that only ended in 1980 with the collapse of a commodity bubble. In the early 1980s, oil prices began to drop, and at the same time the Federal Reserve was credibly moving to “break the back of inflation”. Since then, we lived through a constant cycle of disinflationary forces that ended in the mid-2010s.

Almost all crises since the 1980s were balance sheet crises, and therefore deflationary. The Japan bust, the Asian crisis, the sub-prime crash, and the Euro crisis were all balance sheet and banking crises. Those crises were deeply deflationary in an already deflationary environment. As balance sheets were negatively impacted, borrowers constrained consumption and investment to pay down debt, while at the same time banks constrained lending, which in turn negatively impacted the price of assets used to collateralize said debt, restricting banks’ ability to lend in a never-ending vicious cycle… until governments stepped in.

Global demographics served as a tailwind for labor and led to an increase in savings that got recycled into US Treasuries – colloquially known as the global savings glut. With the fall of the Berlin Wall in 1990 and China being admitted to the WTO in 2001, globalization went into overdrive as companies could tap into a global labor force, resulting in even more disinflation.

Equities and commodities have swapped market leadership in cycles averaging 18 years in length for over a century. Over time these cycles have become shorter with technological advancements, but they are still fairly consistent, predictable, and long. Commodity price bubbles tend to bust after military or economic conflicts due to the well-known “peace dividend” which drives lower commodity and input costs, better profit margins, higher equity multiples, and more leverage brought on by lower rates and a low-inflation environment. Conversely, when equity bubbles deflate, inflation resurges. Large amounts of debt that were accumulated during the expansionary phase must be reduced using a combination of inflation and defaults. When that happens, easy monetary policy follows, and military or economic conflict once again occurs, perpetuating the long-term cycle.

In the subsequent sections of this report, we examine in detail three probabilistic scenarios for what the medium- to long-term outlook in markets may be, but a summary of the analysis is as follows.

(1) Sustained growth and higher prices via re-leveraging of consumers, a renewed corporate investment cycle, and the build-up of inventories in a more inelastic supply environment. This view is anchored in the belief that the world is transitioning from slack to generally tight commodities supply. (P = 55%)

(2) Rising conflict, disruptions, and nonlinear upside price movements leading to a prolonged period of stagflation. Major wars (or other exogenous shocks like pandemics) produce high inflation, and even minor wars can interrupt trade. Conflict and inflation are intrinsically linked, especially coming out of a period of extreme money supply growth. (P = 35%)

(3) Continued price disinflation or deflation, western-dominated status quo, resumption of the technology capital expenditure boom, and prolonged strength for US equity index returns. In this scenario, the belief is that the Fed will not be as aggressive in hiking rates this cycle given the unsustainable divergence between rising debt as a percentage of US GDP and the falling nominal GDP growth derived from that debt. (P = 10%)

The above scenario analysis and applied probabilities shape our forward-looking market views and positioning. Throughout this report, we provide the economic data and analysis in support of these ideas, and a summation of the key points can be found below:

  • We believe that for at least the next few years, we are entering a new environment for inflation with consistently higher price levels. Some indicators to watch for are surging real estate prices, high money supply growth, large fiscal deficits, strong commodity prices, increasing geopolitical instability, and stretched valuations for the US dollar.
  • In that period, we also expect strong nominal GDP growth, while the outlook for real GDP growth is more uncertain given the rising risk of conflict or a central bank miscalculation.
  • Rising inflation will lead to Fed rate hikes, but the governors may have little choice but to return to an accommodative stance due to the “hangover” of past financial excesses and, potentially, war. Government roll-over rate risk is very large, with approximately two-thirds of United States federal debt maturing in the next four years.
  • We expect strong global commodity demand and commodity prices to be a central theme as well. Poor profits have discouraged investment by commodity producers since the mid-2000s, and the growth of sustainability concerns has exasperated the underinvestment.
  • With supply already in short store, producers have moved from the last decade of short duration investment – restock, destock, capex binge, balance sheet distress, capital raise, boom, and bust – to longer duration, disciplined capex cycles.
  • As a result of these dynamics, we expect commodities to outpace the S&P 500 over the next few years. Commodities become a defensive asset in commodity-driven recessions.
  • We see single-digit compound returns for the S&P over the period. The first part of the period will see negative returns and the later part low positive returns, as the focus shifts from multiple compression and falling earnings to cheaper valuations. Free cash flow generation will be key in both phases.
  • Inflation could be made significantly worse if increasing geopolitical instability leads to wars.

Given this outlook, we have built positions across the commodity complex, the core ones being in industrial metals – namely copper, aluminium, and cobalt – emission allowances, and grains. Alongside those positions, we have further built upon the theme through equity allocations to global energy refiners investing in renewable fuels (e.g., sustainable aviation fuel, renewable diesel), miners and refiners of industrial metals who lead in low carbon intensity production, and industrial companies exposed to the renewable revolution with large market share and pricing power. We expect the supply-demand fundamentals facing commodity markets to persist for many years, pressuring prices and having a negative impact on prevailing market sentiment, with a particular emphasis on long-duration assets. We will be hedging a portion of our equities exposure by betting against indices we find to be richly valued given our probability-weighted scenarios. We expect interest rates, especially in the developed world, to make higher highs and higher lows over the next few years, a quasi-mirror image of the lower highs and lower lows of the deflationary past few decades. Lastly, we believe that the currencies of commodity-exporting nations will benefit greatly from this scenario.

To arrive at these views, we have done extensive research that involves proprietary information and third-party data. If you are interested in a full copy of the report, please contact ir@norburypartners.com.

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: The Russian Gambit

Last week, the European Union announced plans to ban Russian crude oil over the next six months and refined fuels by the end of the year as part of the sixth round of sanctions. Over the weekend, the proposed ban on its vessels transporting Russian oil to third-party countries was dropped, but the EU will retain a plan to prohibit insuring those shipments. Bloomberg reports that about 95% of the world’s tanker liability cover is arranged through a London-based organization that must heed European law. Without such insurance, Russia and its customers would have to find alternatives for risks, including oil spills and mishaps at sea, that can quickly run into multi-billion-dollar claims. (For more on the commodity trading business and how insurance impacts commodity flows, listen to “Javier Blas Explains How Commodity Trading Shops Really Work” on the Odd Lots Podcast).

No alt text provided for this image

The pain to European consumers is clear. The region last year got 27% of its oil imports and approximately 40% of gas from Russia (paying $104 billion for supplies of fossil fuels). Economists estimate a full embargo on Russian oil and gas reduces the area activity from 4% (Barclays) to 2% (Bundesbank), while some analysts have argued that the impact would be lower than that. Germany’s vice-chancellor Robert Habeck said his nation has already cut its reliance enough to make at least a full oil embargo manageable with the share of Russian crude in German imports falling to about 12% since the invasion.

For Russia, an oil embargo would limit the inflow of foreign currency and make difficult spending cuts necessary. Russia’s Finance Ministry expects its GDP to shrink as much as 12% this year, on par with the turmoil seen in the early 1990s, when the Soviet Union ultimately dissolved. On Monday, Russia said it expects its oil production to rise in May, and that it is seeing new buyers for its crude, including in Asia. But how much of this is true and how much of it is just posturing?

No alt text provided for this image

Changing oil shipping routes from Europe to Asia is not as trivial as Russia would lead us to believe. Different vessels are required to efficiently transport oil on different sea routes. When done properly, transport of crude oil by tankers is second only to pipelines in terms of efficiency, with the average cost of transport at $0.02 to $0.03 per gallon. When transporting oil, there are three main types of vessels: Very Large Crude Carriers / Ultra Large Crude Carriers, Suezmax, and Aframax vessels. There are about 800 VLCCs/ULCCs in the world used for long-haul routes and they carry around 2 million barrels each. There are about 700 Suezmax vessels capable of passing through the Suez Canal in a laden condition, and they can carry around 1 million barrels on long-haul routes. Lastly, there are about 600 Aframax carriers in the world, known as “go-fast boats,” moving about 600,000 barrels on short-haul routes.

Zoltan Pozsar of Credit Suisse estimates that roughly 1.3 million barrels of oil get shipped from Primorsk and Ust Luga to Europe on Aframax carriers and these journeys take a week or two to complete. Russia does not have pipelines to Asia so the only way to sell its product to new buyers is by using sea routes. However, it is uneconomical to transport crude on long-haul voyages on Aframax carriers and they would need more VLCCs/ULCCs to make that happen. Because Russian ports are not deep enough to dock VLCCs/ULCCs, they would first need to use Aframax vessels to get to a port to then transfer the crude to larger vessels. This transfer itself can takes weeks. After the transfer, the larger vessel would then take about 70 days to get to Asia, unload, and take a similar amount of time to return. This compares with just a couple of weeks when shipping to Europe. This would cause a sharp slowdown in Russia’s economic activity. The world would also need an extra 80 VLCCs/ULCCs to accommodate that change which represents 10% of the current global capacity of those vessels. Additionally, this only accounts for the re-routing of one product, oil, but Russia exports every major commodity.

The increasing competition for selling oil in Asia would have an impact on one of the Middle East’s biggest consumers. It is then no surprise that Saudi Arabia cut oil prices for Asia buyers over the weekend. This will not make Russia’s situation any easier. China benefits when Russia becomes weaker and more isolated and hence more dependent on Chinese goodwill. Let’s not forget that during Russia’s invasion of Crimea in 2014, China got Russia to agree to build (and possibly pay for) a dedicated pipeline at lower prices than it sells to Germany, even when the cost of gas from the new field is higher.

The cost of banning Russian oil might be large for Europe, but it can be even larger for Russia. This should increase the impact of sanctions and diminish the possibility of a unilateral cut in supply in the near future. 

Macro Minute: Inequality and Economic Growth

On a macroeconomic level, inequality can affect economic growth, productivity, and political stability, which in turn has direct implications for corporate profitability.

As with most social science endeavors, there is a healthy debate about the precise impact of inequality on growth. For starters, does it hinder or accelerate economic growth? Economic theory shows that with higher income and wealth come higher savings rates and, therefore, a higher level of investment and gross domestic product. In this case, if marginal productivity is higher for capital than for labor, one can see that more inequality would create higher economic growth. Also, in economies with underdeveloped credit markets, significant investments can only be made if wealth is accumulated. In the presence of imperfect capital markets and indivisibility of investments, an economy with higher levels of inequality may be able to introduce new industries, technologies, and markets and ultimately grow faster than the same economy with lower levels of inequality. Finally, there is an argument that reducing inequality reduces incentives to accumulate wealth through labor, entrepreneurship, and innovation, having, therefore, a negative impact on long-term growth.

However, a healthy debate should not be confused with a balanced debate. The bulk of the literature supports the theory that inequality hurts economic growth. Studies that found a positive relationship between inequality and growth were focused on the short term, and studies that found a negative relationship were focused on the long term. In other words, inequality can produce a small contribution to growth in the short term but will substantially adversely affect growth in the long term. Empirical results have increasingly supported the arguments for impaired economic growth and a negative impact on productivity in the face of rising inequality.

Poor people without access to credit markets often defer health care treatments, cannot procure housing or transportation and lack the means to further their education. This results in missed opportunities and diminished productivity and growth potential. The same applies to poor parents with multiple children compared to wealthy parents with few children. The inability of low-income families to invest in their children’s education, and the inability of poor workers to invest in developing job skills because of either financial constraints or time constraints while working multiple jobs, can result in a reduction in overall skills and knowledge of the potential employee pool. That can have a drag on growth, productivity, and business performance. Finally, there is the problem of the indivisibility of consumption. In the absence of developed credit markets, expensive items can only be acquired by accumulating wealth. If the economy became more equitable, part of the population initially excluded from the acquisition of these goods would enter the market, encouraging the creation of new domestic industries.

Societies with higher levels of inequality also tend to have higher levels of crime, keeping a larger share of the labor force from productive activities and decreasing potential growth. Also, with the increase in social instability, trust and social cohesion can erode, leading to conflict, political crises, and the resulting retraction of investments. One of the potential consequences is the rise of nationalism and the splintering of support for globalization. This scenario has played out in economies around the world over generations. Ray Dalio, the founder of Bridgewater Associates, the world’s largest hedge fund, has warned of the corrosive effects of inequality on faith in capitalism and, therefore, the stability of the institutions and markets in which business operates.*

*This excerpt appeared in the article “Business Risks Stemming from Socio-Economic Inequality” by Todd Cort, Stephen Park, and Decio Nascimento at the Columbia Law School Blue Sky Blog published on April 15, 2022. The article is based on the paper “Disclosure of Corporate Risk from Socio-Economic Inequality” by the same authors published on March 18, 2022. You can find the article here, and the paper here.

Macro Minute: A Phoenix From the Ashes

While we normally use this section to discuss macroeconomic concepts that help frame our top-down asset allocation views, or to present bottom-up macro asset analysis, today we are highlighting a company that stands out in a world of higher commodity prices.

At present, there is much discussion on how legacy companies, particularly in the commodity space, must change how they do business to adapt to a sustainability-focused world. There is currently a plethora of negative attention on fossil fuel companies as CEOs go before Congress, but today we’re focused on a company that is seven years removed from an activist campaign and fresh off an upgrade from high yield to investment grade last December.

Alcoa is the fifth largest aluminum “pure player” globally, and the largest upstream producer in the Western world. Its operations are geographically diverse, with an integrated aluminum production business from bauxite and alumina procurement and processing, to smelting.

Aluminum is a critical input to virtually every aspect of the forthcoming renewable energy complex and is often referred to as solid electricity given the sheer amount of power (and carbon intensity) required to process the metal.

By our estimates, global aluminum balances are expected to persist in deficit for at least the next 2-3 years, which is already being reflected on exchanges; LME aluminum inventories are running 40% below their five-year average. 

75% of Alcoa’s smelting capacity is powered by renewable energy, which positions the company as one of the least emitting players in a highly energy-intensive sector.

Alcoa is naturally leveraged to aluminum prices and has been consistently delivering on cash flow generation due to a disciplined capital allocation strategy focused on rationalizing its asset base and deleveraging. The company has also been quick to target green growth initiatives with an emphasis on low-carbon aluminum products, helping to differentiate the company from its peers. Some analysts believe that one such peer, Rusal, is expected to see aluminum production go to zero in 2022 due to Russian sanctions precluding the import of bauxite and alumina, which should have a positive impact on Alcoa’s market share.

In our view, Alcoa is a good example of a legacy company from an old-fashioned industry that has increasingly gained investor interest, not only due to the quality of the company’s management but also from a positive industry backdrop (e.g., green transition, supply imbalance).

Macro Minute: Flat as a Pancake

The flattening of interest rate curves is nothing new. In the US, swaps markets are pricing that in one year, the difference between the 2-year rate and the 10-year rate (the preferred reference for curve shape) will be -39bps, meaning the 2-year rate is 39bps higher than the 10-year. That compares to a difference of +140bps almost exactly one year ago, when the 10-year rate was materially higher than the 2-year.

Reasons abound, from the perception of more hawkish Fed policy given elevated inflation concerns, to lower pricing of terminal and neutral rate expectations as the Fed pulls forward the timing of the hikes.

What is more curious is how flat the very long end of the swap curve is right now in the US and Europe. In the United States, the difference between the 10-year rate and the 30-year rate sits at around -14bps, with the one-year forward at an eye-watering -21bps. In Europe, things are even more extreme, with the difference between the 10-year and 30-year rates at -17bps, and the one-year forward at -34bps. That part of the interest rates curve is not typically used to express a view on the path of interest rates like the 2y10ys, and assuming that the time-value of money is positive (something we’ll be hearing more about in this inflationary period), usually trades in positive territory with the 30-year rate above the 10-year rate.

Just how extreme the levels we are seeing now is the focus of this Macro Minute.

Let’s first look at the United States. In the past 30 years, the difference between 10-year and 30-year rates (10y30y for short) has been on average +40bps, staying most of the time within 1 standard deviation above or below the mean. As of today, the spread now sits 2 standard deviations below the mean. And how often does this rate differential stay at or below 2 standard deviations? Less than 0.2% of the time. In 30 years of data, the most consecutive days that it has ever stayed below that level is 5. Furthermore, in this period it has never touched 2.5 standard deviations below the mean (but it came incredibly close in 2008).

When plotting the divergence of the rates differential to its trend we can see how this data is distributed. From the charts below we see that the data fits the bimodal distribution better than the normal distribution. However, both distributions overestimate the tails when compared to the data, meaning we can assume that they will yield conservative estimates for the probabilities of very small or very large numbers. Using the probability density functions to estimate the probability of the 10y30 moving below current levels, we get a probability of 1.4% when using the normal distribution and 0.20% when using the bimodal.

When looking at Europe, we recognize that the 10y30y’s moves are more extreme than in the United States. For the past 30 years, the 10y30y has averaged +42bps. Like in the US, the spread most often lives between 1 standard deviation above or below the mean, but it spends more time below 2 standard deviations than in the US, at about 3% of the period. Today, we find ourselves sitting nearly 3 standard deviations below the mean. How often does this rate differential stay at or below 3 standard deviations? About 0.4% of the time. In 30 years of data, the greatest number of consecutive days it has ever stayed below that level was 12. Additionally in this period, it has only touched 4 standard deviations below the mean once in 2008, before retracing toward the mean on the following day.

In Europe, we also find that the historical data better fits the bimodal distribution than the normal distribution. Here, the normal distribution underestimates the left-tail, while the bimodal distribution underestimates both tails, so we should take the results with a grain of salt. However, using probability density functions to estimate the probability of the 10y30ys moving below current levels, we get a probability of 0.23% by using the normal distribution and 0.24% from the bimodal.

We fundamentally believe that we are in a new normal of permanently higher inflation rates around the world (albeit less than today once supply chain disruptions ease) and with that, the return of higher term premiums. Combining that with the statistical analysis above makes us believe these markets are largely dislocated.

Macro Minute: The Reflexivity of Inflation and Conflict

In The Changing World Order, Dalio makes the point that history shows us that empires follow a predictable cycle of rising and decline that he termed Big Cycle. The cycle starts at “The Rise” phase when there is strong leadership, education, character, civility, and work ethic development. Innovation is also enhanced by being open to the best thinking in the world. This combination increases productivity, competitiveness, and income. At “The Top”, the country moves from “growing the pie” to “splitting the pie”. While the country enjoys higher standards of living, people get used to doing well and enjoy more leisure time in detriment of hard work. Financial gains come unevenly, and the elites influence the political system to their advantage creating wealth, value, and political gaps. Borrowing grows to make up for the loss in productivity, weakening its financial health. “The Decline” happens when debt is large and there is an economic downturn leaving the country two choices: default or printing new money. Typically, countries opt for new money. The faster the printing occurs, the faster the deflationary gaps close and worrying about inflation begins.

It is clear that major wars are greatly inflationary, and even minor conflicts can interrupt trade, causing disruptions and increasing prices. But what we’re finding out is, inflation actually precedes periods of war or economic warfare.

It should come as no surprise then that the global increase in general prices along with a strong price increase of commodities is leading the world to a new era of conflict. Ukraine and Russia are not just material in the global trade of wheat but now also make up close to 20% of the world corn trade. A prolonged conflict can lead to issues this spring when planting starts causing significant productions shortfalls when harvest comes later this year, causing prices to go higher, increasing the risk of an escalation and conflict. This self-reinforcing process we see between inflation and conflict has a name, reflexivity. [1]

[1]The theory of reflexivity in financial markets was proposed by George Soros. Considered by most as one of the best macro investors in history, delivering an average annualized net return of 33% from 1970 to 2020. His work around the theory of reflexivity first appeared in his 1987 book “The Alchemy of Finance”. Soros’ theory of reflexivity is based around human fallibility—human beings can be wrong in their beliefs about the world and therefore act based on misguided knowledge. Heavily influenced by Karl Popper’s account of the scientific method (Popper was Soros’ tutor at the LSE), Soros argues that when the subject of study involves thinking participants, the scientific method must be changed to account for that fact. When thinking participants try to understand a situation, the independent variable is the situation. And when participants try to make an impact on the situation, the independent variable is participants’ views. Reflexivity occurs when the participant’s view of the world influences the events in the world, and these same events will influence the participants’ view of the world. When reflexivity is in play, it can cause boom/bust processes—self-reinforcement that eventually becomes self-defeating. For a detailed description of reflexivity and the construction of social reality, read “When Functions Collide” at https://www.norburypartners.com/nascimento-decio-when-functions-collide/