Odds and Ends - Market Update - 10/17/20
4Q20 has a lot of fireworks ahead – do you have a contingency plan? Reflecting this weekend and over the past couple of weeks, on my many clients and friends in Hotel CRE that are ‘hurting’ right now – and – I can’t help but think about the government’s role in this. As a politically center person; and, mostly apolitical; I have GOP and Dem friends, most of us do. While not to stereotype, probably roughly ½ of the hotels in this country are owned by AAHOA members – first and second generation immigrants to the United States – and many maybe most Democratic Party supporters. Consider this though, in California, given Governor Gavin Newsom’s draconian lockdown and control policies, the suicide rate and suicide attempts increased in greater proportion than the number of estimates lives saved from COVID. Yes, you read that correctly – if this was 1780 – Gavin Newsom would be HANGED alive – literally. In the 1800s he would have faced trial. Newsom is directly responsible for the loss of not only lives through bad policy, but also generational hard earned wealth of many people in Hotel CRE – people like clients of mine: Dr. Barry Lall, the Hansjis, the Thakkars and Desais (VA) and Bharat Patel (IN) – the insurers should be footing this bill via business interruption. A nation that allows people like Newsom to direct and control policy at the expense of lives and livelihoods is not the America of our founding fathers, prima facie. If I were President, I would try Gavin Newsom for Manslaughter – give one reason why not? .. and onto the macro and some good data to reflect upon here..
On the Macro Front, in the good news camp, ECRI’s US Weekly Leading Index (WLIW) printed another gain, at 3.6, marking its positive print in a row. In the not so great news, in addition to lower readings in GER inflation (deflating) and sentiment, in the U.S., despite a nice retail sales (lagging indicator) print on Friday, Capacity Utilization and Industrial Production knocked 1.61% off the NYFed’s 4Q20e Nowcast, from 4.82% to 3.55% (inclusive of +.17 from prior week data adjustments).
This past week was a stock-picker’s market. We (me) are watching weakness in Banks ($BKX, $XLF), Airlines ($XAL) and Transportation – as if there is follow-through, that may portend more weakness during this quarter; until vaccine prospects are more clear (encouraging news from Pfizer ($PFE) to end the week on presenting data in late-November) and Congress gets it’s act together on stimulus (they should piecemeal it at this point and not have Pelosi hold the entire country hostage).
Operating Leverage May Take Hold in 2021 (Moody’s, 10/15/20)
Unrivaled support from the Federal Reserve and fiscal stimulus have benefited corporate credit quality. Of related special importance has been the drop by the U.S. unemployment rate from April 2020’s 80-year high of 14.7% to September’s 7.9% and the climb by the percent of manufacturing capacity in use from April 2020’s record low 59.9% to September’s 70.2%.
Operating profits benefit considerably when rates of resource utilization rise from today’s atypically low levels. This phenomenon is referred to as “operating leverage,” or a situation where the percent increase by corporate earnings is a multiple of the accompanying percent increase by business sales. Operating leverage also implies a widening of profit margins, where the latter ordinarily enhances a company’s financial flexibility.
Operating leverage is implicit to 2021’s outlook for S&P 500 corporate earnings that has earnings per share expanding at least three-times as fast as revenues. According to October 9’s FactSet consensus, a recovery by the revenues of S&P 500’s member companies from 2020’s prospective 2.6% drop to 2021’s 8.0% advance is expected to prompt a dramatic change in direction for earnings per share from 2020’s expected 17.7% contraction to a projected 25.5% surge for 2021.
By contrast, the Blue-Chip consensus forecast of early October 2020 projected only a partial rebound by pretax profits from current production, or core profits, from 2020’s anticipated 11.4% contraction to a 5.6% increase for 2021. However, the latter may be too small if the Blue-Chip consensus projections for industrial production prove correct. Early October’s consensus expects the annual percent change for U.S. industrial production to recover from 2020’s 7.5% contraction—the deepest since 2009’s 11.5% plunge—to a 4.0% increase for 2021.
Of all readily available macro indicators, none explains annual percent changes by core profits better than the annual percentage point changes of the manufacturing and industrial rates of capacity utilization. Here, faster growth rates for manufacturing and industrial output imply bigger percentage point increases for the rates of capacity utilization. The year-over-year percent change of core pretax profits’ moving yearlong average shows its strongest correlation of 0.70 with the manufacturing capacity utilization rate’s yearly percentage point change followed by its 0.69 correlation with the yearly percentage point change for the industrial capacity utilization rate. Further behind are core pretax profits’ correlations of 0.45 with nominal GDP’s yearly percent change, 0.44 with real GDP’s yearly percent change, and -0.34 with the unemployment rate’s yearly percentage-point change.
As inferred from the historical record, the latest Blue-Chip consensus projection of a 4.0% annual increase by 2021’s industrial production portends an increase by the rate of industrial capacity utilization that favors a 15% midpoint for 2021’s increase by core profits, wherein core nonfinancial-corporate profits advance by a livelier 18%. So big of an increase by profitability can only improve corporate credit quality
The Rotten Year (Zandi, Moody’s, 10/15/20)
Twenty-twenty has been a rotten year, and any chance of it ending on a high note has faded. The COVID-19 pandemic continues to rage. Another wave of infections is a serious threat as the weather turns colder and social distancing and mask wearing are at best uneven across the country.
Rancor over the presidential election is mounting, and it is disconcerting that there is even a question that there will be a graceful transition of power if President Trump loses the election. Adding to the tumult, lawmakers appear unlikely to come to terms soon on another fiscal rescue package to shore up the fragile economy, even though the president and House Democrats both have endorsed trillion-dollar plans. It is tough to see how the economy can gain any traction until next year when, we hope, the pandemic and election are in the history books.
On the pandemic, there is a clear relationship between the healthcare crisis it has created and the economic crisis. More infections results in a weaker economy. With the pandemic entering its eighth month, there is now enough data to show this econometrically using a panel regression relating confirmed infections per capita to the unemployment rate across states, Puerto Rico, and the District of Columbia since the pandemic hit in March. The best fitting regression lags infections per capita by one month, so that for example an increase in infections in August results in an increase in unemployment in September. A panel regression simply captures this relationship across states and over time, increasing the number of datapoints (260) available for the regression. The regression results are definitive (and available upon request), with confirmed infections explaining 85% of the variation in unemployment rates across states since March.
Based on this regression and other cross-country analysis, we estimate that a sustained 10,000 increase in daily confirmed infections results in an approximately 0.5-percentage point increase in the unemployment rate, all else being equal. For example, if daily infections double from 40,000, the average over the past couple of months, to say 80,000, about where infections peaked in the summer, then the unemployment rate would ultimately increase by 2 percentage points—putting it back near double digits several weeks, the European economy is faltering again. Europe had meaningfully brought down infections with its stringent lockdowns early on in the pandemic, and its economy had begun a strong comeback in the summer, but that recovery looks to be flagging again with the reintensifying pandemic. The same negative dynamic also appears to have begun in the United States. As people start to move indoors with the colder weather, infections in recent weeks are up in much of the northern part of the country, including in the Northeast, which like Europe had gotten its infections down with stringent lockdowns. College reopenings haven’t helped; many college towns have quickly become hot spots. Our baseline economic outlook assumes that the worst of the pandemic is behind us, but this is clearly a tenuous assumption.
The process of electing a president has historically been neither here nor there when it comes to the economy. Even the highly contentious Bush versus Gore election in 2000 had no meaningful impact, save perhaps for a few bad days in the stock market, which was already struggling with the dot-com stock bust. This time may be different. The nation feels like it could boil over if Trump versus Biden is a close election and one side or the other believes that the election is being stolen.
Fueling this political heat are the nation’s extraordinarily polarized electorate as well as Trump’s active effort to undermine confidence in the election and his strong suggestion that he would not leave office if he thinks he was cheated. The arrest of more than a dozen men plotting to kidnap the Democratic governor of Michigan ostensibly for the way she has handled the state’s pandemic, which is very different from the president's, crystallizes the concern that the election could ignite violence.
This would be much less of a concern if the winner wins handily, making it indefensible to question the outcome. Current polling suggests this may happen. Vice President Biden is up by as much as 10 percentage points nationwide and 5 to 7 points in key swing states. However, our election model suggests the results will be much closer. Assuming that turnout by Republicans and Democrats isn’t lopsided to one party or the other, which seems likely, then our modeling shows Biden winning the electoral college with 279 votes. Of course, 270 votes are needed to win. As an aside, our modeling identifies Pennsylvania as the most critical swing state, and turnout in the Philadelphia area as especially critical to who will win the state. This would seem to give Biden a further edge, since he lives just a few miles south of downtown Philly. Nonetheless, if the election outcome is as close as our modeling suggests, the election seems certain to be contested by Trump. That would make the next couple of months stomach churning while the mess gets straightened out. There is nothing but downside to the economy in such a disquieting scenario.
Given how hard the pandemic has been on the economy, it may be hard to fathom how our models, which include a range of both political and economic variables, show such a close race. Key is the stability of Trump’s approval rating. According to the Gallup poll, which we use in our modeling given its long history, the president’s popularity has been consistently the lowest among modern day presidents, but it has also been astonishingly stable—regardless of what Trump says or does his base of supporters remains loyal. It is the change in the president’s approval rating leading up the election, and not the level of that rating, that matters in our modeling of voter behavior.
The narrow election results are also due to the stock market’s recovery. How stock prices fare is increasingly important to driving voting decisions, since the large baby boom generation is focused on its retirement and stock portfolios, and perhaps because Trump has used the market as a barometer of how the economy and he are performing. Stocks have recovered from their slide when the pandemic hit, and while this is because of a range of factors, not least of which is the Federal Reserve’s aggressive action, Trump benefits. There is thus considerable irony in the stock market’s recent strength as investors have begun to discount the prospects for a Biden victory and even a Democratic sweep. Investors appear cheered by Biden’s focus on providing massive fiscal support to the economy and the implications for stronger corporate earnings. This, despite their discomfort with Biden’s proposal to scale back Trump’s tax cuts for corporations and well-to-do households
The most serious blow to the economy as the year ends will be the failure of Trump and Congress to come to terms on more help to those hit hardest by the pandemic. We have long assumed that lawmakers would agree on legislation providing $1.5 trillion in additional fiscal support. This makes economic sense. Without such support the already-fragile recovery threatens to come undone. It also makes political sense given the approaching election and the need for lawmakers to demonstrate that they have voters’ backs. Yet a deal has not come to fruition, and we are now assuming in our baseline outlook (35% probability) that the $1.5 trillion package won’t become law until February, after the next president is inaugurated. Our baseline also assumes that Biden will become president, and the House remains Democratic and the Senate Republican. Real GDP and job growth come almost to a standstill in the fourth quarter in the baseline, but the economy avoids a double-dip recession as the fiscal support provided early next year revives
Of course, there are a number of alternative scenarios for how the election will play out and what this means for the fiscal package we expect early next year. The next most likely scenario is a Democratic sweep (30% probability), which would give President Biden the political window to go big, something closer to a $3 trillion fiscal package. This would be consistent with the fiscal policy proposals he has put forward during the campaign, and the HEROES Act legislation passed by the Democratic House several months ago. This would include funds to bridge the economy to the other side of the pandemic, but also money to ramp up government investments in infrastructure, education, healthcare, housing, and other social policies to help the economy return more quickly to full employment. There are two other election outcome scenarios, including the status quo of a Trump victory with a split Congress (25% probability) and a Republican sweep (10%). These scenarios wouldn’t preclude another fiscal rescue package, but it would likely be much smaller, something similar to the less than $1 trillion package Senate Republicans have been holding out for in the recent negotiations.
Unfortunately, there is a lot of difficult script to be written between now and when the next president and Congress sign another fiscal package into law. It would be prudent to buckle in.
Banks – Cross Region: Pandemic induced credit losses set to rise (Moody’s, 10/15/20)
The coronavirus pandemic has pushed many countries into recession, leading to a deterioration in credit conditions that will trigger a significant increase in bank loan losses. Across the largest banks in Australia (Aaa stable), the United Kingdom (UK, Aa2 negative) and the United States (US, Aaa stable), growth in non performing loans (NPLs) has so far been limited, reflecting bank forbearance and government support measures. While provisions for expected credit losses rose materially during H1 2020 due to more severe macro economic assumptions, NPL growth will accelerate as bank and government support expires, followed by loan write-offs.
Economic growth and support measures will determine loan losses
We expect the UK’s real GDP to contract by 10.1% in 2020 as a result of the coronavirus pandemic, more than the US (-5.7%) and Australia (-5.3%). Weaker economic growth and higher unemployment will lead to a significant increase in bank loan losses. However, the growth trajectory of these losses is highly uncertain, given the unprecedented monetary and fiscal stimulus launched in response to the economic downturn, as well as forbearance measures provided by banks to borrowers facing repayment difficulties. In all three countries, monetary and fiscal support and the relaxation of “lockdown” restrictions designed to curb the spread of the virus has led to a pickup in economic activity since H1. This is reflected in a recovery in demand-side indicators such as growth in retail sales and mortgage applications. However, activity remains subdued, and there is significant downside risk of a resurgence in the virus, leading to the reimposition of localized or nationwide lockdowns1.
The true extent of asset quality deterioration will therefore only become clear as time passes, and as the benefits of policy and bank support wear off. We do not expect to see significant NPL increases until the fourth quarter of 2020 at the earliest. However, we believe that the recategorization during H1 2020 of some loans from Stage 1 to Stage 2 under the IFRS 9 accounting standard, and the sharp increase in “criticized” exposures at US banks, both of which indicate a deterioration in loan quality, are early indications of what is to come. Further migration of exposures into Stage 2 and 3 for IFRS 9 reporters, even if macro economic assumptions remain stable, will necessitate further increases in provisions. We expect the increases to be less than the amounts already booked in H1 2020. If the economic downturn is more severe or long-lasting than banks currently forecast, for example because of further restrictions, current provisions, as well as the levels banks have guided towards for year end 2020, will be insufficient. On the other hand, if the economic recovery and consequent increase in credit demand surprise to the upside, banks may be in a position to release provisions in 2021 and 2022.
Most pandemic-related credit losses will come from loans to coronavirus-exposed corporate sectors (Exhibit 1). These are primarily industries that rely on face to face interaction such as retail, hospitality and associated commercial real estate (CRE) lending, as well as transportation, aviation and tourism. Social distancing and other restrictions will continue to hold back revenues in these sectors until the pandemic subsides, or until an effective vaccine becomes available.
The CRE sector faces a potential drop in demand for office space as the coronavirus prompts more employees shift to working from home, at least on a part-time basis. However, given the long-term nature of most commercial office leases, we think most banks will have time to adjust to this trend, and are unlikely to suffer significantly higher loan losses in this sector over the next year or two.
Unsecured personal loans and credit cards, which are vulnerable to rising unemployment, are a further source of pandemic-related risk to banks, albeit contained so far by government support measures and bank forbearance (Exhibit 2). Higher exposure to at-risk corporate sectors or to unsecured personal lending will drive both higher provisioning and higher eventual loan write-offs. While residential mortgages are less exposed, they are potentially vulnerable in economies such as Australia where household indebtedness is already high, as this makes borrowers less resilient against rising joblessness
UK banks will post biggest rise in NPLs
We expect UK banks to report a bigger increase in NPLs than their US or Australian peers. This is because we anticipate a sharper economic contraction in the UK than in the US or Australia. UK banks have a relatively high volume of loans showing signs of deterioration and higher exposure to unsecured lending than Australian lenders. Some UK banks also have higher exposure to at-risk corporates than their US counterparts.
However, UK banks will not necessarily report the biggest rise in credit losses, as this will depend in each system on how many loans continue to perform after restructuring. Loans subject to restructuring including payment holiday extensions, which have lower loss given default, will incur lower provisioning requirements than loans which have defaulted, which will be written off.
Based on bank disclosures, we calculate that UK banks' exposure to at-risk sectors ranges between 35% and 184% of their tangible common equity, to some extent reflecting inconsistent sector disclosures. This compares with averages of 107% and 148% for their US and Australian peers respectively. 3.
The UK banks' provisioning charges rose materially to 81 basis points of gross loans during the first half of 2020 from 15 basis points six months earlier. The increase is more moderate than their US peers’, but substantially greater than that of the four Australian lenders (Exhibit 3).
Barclays reported the highest cost of risk among the large UK banks in H1 2020 (105 basis points), reflecting its higher exposure to small business, corporate and credit card lending. HSBC’s was lower at 66 basis points, reflecting its strong presence in Asia, where the economic downturn is likely to be less severe.
US banks' loan losses will depend on employment trends
The largest US banks’ exposure (funded and unfunded) to at-risk corporate and commercial real estate sectors is relatively contained, averaging 107% of tangible common equity based on available disclosures. This compares with an average of 120% across all twelve banks covered in this report.
We also believe that in many cases these exposures are well collateralized, particularly in the retail and oil & gas sectors, which have been under stress for some time, and where asset-based lending arrangements are common. However, the US banks' sizeable credit card operations give them greater exposure to unsecured consumer credit than their Australian peers and most UK lenders. High unemployment levels will lead to sharply higher losses on those portfolios in 2021.
During the first half of 2020, US banks more than doubled their allowance for loan losses, reflecting significantly higher provisions as well as the adoption of CECL on January 1. Aggregate cost of risk (provisions as a % of total loans) across the four banks increased from 28 basis points in the second half of 2019 to 147 basis points in the first half of 2020.
Much of the increase related to credit cards, reflecting the US banks' expectation that card losses will go up. At 30 June 2020, the allowance for losses on credit cards accounted for 63% of the total allowance at Citi, 55% at JPM, 48% at BAC, and 18% at WFC. This ranking reflects the weight of credit cards within each bank's respective loan book, with Citi the highest at 22%, JPM at 15%, BAC at 9%, and WFC at 4%.
Notwithstanding the increase in card related provisions, many US credit card borrowers reduced their debt and increased their savings during the first half of 2020, as in Australia and the UK. Credit card borrowers also benefited from loan deferral programs, during which time missed payments did not count as delinquencies. However, we expect reported credit card delinquencies to rise significantly over the next two quarters, as most of the deferral programs have now ended.
Since credit cards are typically charged off at 180 days past due, banks will likely start to incur significantly higher credit card chargeoffs by spring of 2021. Unemployed customers who have depleted their savings are most at risk, although additional government stimulus programs could moderate delinquencies from this source.
However, since US banks have improved their underwriting quality compared with the global financial crisis of 2007-08, we currently expect system-wide credit card charge-offs to peak at around 7.0% to 8.0% in 2021. This is well below the post-crisis peak reached a decade ago.
Several banks have reported that the proportion of their credit card customers who are unemployed is several percentage points below the national jobless rate. This is because the largest increases in unemployment in H1 2020 were among lower-wage workers, who are often not eligible for bank credit cards. This further supports our expectation that US bank credit card losses will likely be more manageable than they were a decade ago.
Sustained high unemployment in the absence of additional government stimulus is also likely to push up losses on US consumer auto loans. However, large US banks' have relatively moderate exposure to this sector, where non-bank lenders play a significant role. At the start of the downturn, US bank auto loan underwriting standards had tightened, but were still relatively loose by historical standards.
System-wide, we expect auto loan charge-offs to peak in 2021 at around double their 2019 levels. However, given the secured nature of such loans, auto loan loss rates of around 1.5% to 1.75% will be considerably lower than for credit cards.
US residential mortgage underwriting standards entering the downturn were solid at around the 30-year historical average. Currently, we forecast only a modest mid-single-digit decline in home prices after years of solid appreciation. Therefore, we expect residential mortgage asset quality to deteriorate far less than credit cards and auto loans. US banks' residential mortgage loan performance should be far more comparable to the 0.25% to 0.30% average annual charge-offs experienced during the early 1990s downturn than to the higher levels experienced after the 2007-08 financial crisis.
For the third quarter of 2020 provisions for loan losses were substantially smaller at the largest US banks, with JPM releasing a small amount of reserves, BAC and Citi adding modestly to theirs, and Wells Fargo leaving its reserves largely unchanged. A modest improvement in macroeconomic conditions led the banks to conclude that no significant further additions to reserves were necessary.
However, given the high level of uncertainty around the trajectory of the virus, the likelihood of additional government stimulus, and future macroeconomic conditions, managements also concluded there was not sufficient justification for reserve releases beyond those driven by a decline in loan balances.
NPL growth will accelerate from current modest levels
The rise in NPLs across the three systems has been moderate so far. This is because repayment moratoria and other forms of bank forbearance, together with job subsidy schemes and other government stimulus measures, have masked the underlying level of stress among borrowers (Exhibit 6). We expect nonperforming loan ratios to rise significantly over the next 12-months as these extraordinary support schemes expire or are wound back.
Increase in criticized loans provides a future indicator of NPL trends at US banks
US banks reported a slightly steeper increase in NPLs during the first half of 2020 than UK banks or Australian lenders. The average NPL ratio (excluding performing restructured loans) increased by 22 basis points to 0.92% of gross loans, driven primarily by higher commercial/corporate NPLs. At the same time, the four largest US banks reported a sharper rise in “criticized” commercial loans during the first half of 2021 (Exhibit 10). These include loans that are internally categorized as special mention, substandard or doubtful due to an elevated level of risk, and which may have a high probability of default or total loss. The sharp rise in criticized loans reflects the economic downturn in 1H20, and suggests a potential for higher NPLs in the future. Criticized loans are based on US regulatory definitions, and the interpretation of those definitions can vary from bank to bank.
On the consumer side, most US banks initially granted forbearance to any borrower claiming to be under financial stress as a result of the pandemic. As the economy and employment levels have improved since bottoming in April, forbearance levels have declined. At mid-year, US banks had granted forbearance relief to customers accounting for around 5% to 10% of the total loan portfolio. While loans are in temporary forbearance, they are not being reported as delinquent. As a result, reported delinquency levels actually declined in the second quarter. Lenders have reported that many borrowers continue to make at least some payments while under forbearance, indicating that some borrowers initially opted for forbearance out of caution rather than a pressing need to preserve liquidity, and most banks are reporting that a significant majority of the loans exiting deferral are current. Only after the forbearance period expires will banks report loans which borrowers have stopped repaying as delinquent.
On the commercial side, loan forbearance programs were generally more limited than on residential mortgage and other consumer loans. An exception to this was within banks’ CRE portfolios, where a significant proportion of loans to hotels and retail CRE borrowers were modified, deferred or both. In addition, loans to auto dealers and other commercial distribution businesses were deferred at a high rate.
We believe US banks’ wide-ranging forbearance and other deferral initiatives, combined with significant government support, have delayed the deterioration in asset quality. However, as the second half of 2020 unfolds and these programs abate, we expect reported delinquencies and NPLs to rise rapidly. This trend was apparent in the recently announced third quarter results which showed a modest rise in early-stage consumer delinquencies and NPLs. We expect further increases in NPLs going into next year